ArticlePDF Available


There are two major proposals regarding how to derive the VOS word order in the Mayan family. One is a right-specifier analysis, according to which specifiers of lexical categories are located to the right of the heads and the subject occupies a right-specifier. The other is a predicate fronting analysis, in which vP is preposed across the subject. Comparing two Mayan languages, Chol and Kaqchikel, this paper argues that Kaqchikel reaches VOS via a right-specifier route rather than a predicate fronting route, and suggests a possibility of extending the right-specifier analysis to Chol VOS sentences.* 1. Introduction Languages differ in the order in which the subject (S), the object (O) and the verb (V) are aligned. For example, in declarative sentences with a nominal subject and object, the unmarked or "basic" word order is SVO in English and SOV in Japanese, with the subject preceding the object. However, many Mayan languages exhibit the basic VOS word order with the subject following the object. There are two major proposals regarding how to derive the VOS order in the Mayan family. One is the right-specifier analysis by Aissen (1992), according to which specifiers of lexical categories are located to the right of the heads and the subject occupies a right-specifier. The other is the predicate fronting analysis by Coon (2010), in which vP is preposed across the subject. Comparing two Mayan languages, Chol and Kaqchikel, we argue that the right-specifier analysis is more suitable
言語研究(Gengo Kenkyu156: 25–452019 doi: 10.11435/gengo.156.0_25
Two Routes to the Mayan VOS:
From the View of Kaqchikel
K O K S
Kanazawa Gakuin University Kwansei Gakuin University
N Y M K
Miyagi Gakuin Women’s University Tohoku University
Abstract: ere are two major proposals regarding how to derive the VOS word
order in the Mayan family. One is a right-specier analysis, according to which
speciers of lexical categories are located to the right of the heads and the sub-
ject occupies a right-specier. e other is a predicate fronting analysis, in which
vP is preposed across the subject. Comparing two Mayan languages, Chol and
Kaqchikel, this paper argues that Kaqchikel reaches VOS via a right-specier
route rather than a predicate fronting route, and suggests a possibility of extend-
ing the right-specier analysis to Chol VOS sentences.*
Key words: Chol, Kaqchikel, right-specier, predicate fronting
1. Introduction
Languages dier in the order in which the subject (S), the object (O) and the
verb (V) are aligned. For example, in declarative sentences with a nominal subject
and object, the unmarked or “basic” word order is SVO in English and SOV in
Japanese, with the subject preceding the object. However, many Mayan languages
exhibit the basic VOS word order with the subject following the object. ere are
two major proposals regarding how to derive the VOS order in the Mayan family.
One is the right-specier analysis by Aissen (1992), according to which speciers
of lexical categories are located to the right of the heads and the subject occupies
a right-specier. e other is the predicate fronting analysis by Coon (2010),
in which vP is preposed across the subject. Comparing two Mayan languages,
Chol and Kaqchikel, we argue that the right-specier analysis is more suitable
* We are immensely grateful to Lolmay Pedro García Matzar, Juan Esteban Ajsivinac
Sián, and Filiberto Patal Majzul for providing us with the Kaqchikel data reported in this
paper. We are equally indebted to Yoshiho Yasugi for helping us create and analyze relevant
Kaqchikel sentences and to Kensuke Takita for sharing with us his idea on linearization.
We would also like to thank the audiences at the Keio-Nanzan Workshop on Minimalist
Syntax ( June 8th, 2019) and two anonymous reviewers for their invaluable comments and
suggestions. All remaining errors are of course our own. is work was supported in part by
JSPS KAKENHI Grant Numbers 15H02603, 19H05589 (PI: Masatoshi Koizumi).
26 K O, K S, N Y,  M K
for Kaqchikel, and then suggest the possibility of extending the right-specier
analysis to Chol VOS sentences with a few parametric dierences between these
2. Kaqchikel and Mayan languages
e Mayan family is comprised of about 30 languages primarily spoken in
Guatemala, Mexico, and Belize (for a recent overview of Mayan linguistics, see
Coon (2016) and the other articles in the same issue of Language and Linguistics
Compass). Kaqchikel is a language of the K’ichean branch of the Mayan family.
(1) Possible Classications of Mayan Languages
a. Huastecan: Wastek, Chicomuceltec [extinct]
b. Yukatekan: Yukatek, Lakantun; Itza’, Mopan
c. Greater Tseltalan:
i Cholan: Ch’orti’, Cholti [extinct], Chontal, Chol
ii Tseltalan: Tzotzil, Tseltal
d. Greater Q’anjob’alan:
i Chujean: Tojolabal, Chuj
ii Q’anjob’alan: Mocho (Motocintlec); Jakaltek, Akatek, Q’uanjob’al
e. K’ichean-Mamean (or Eastern Mayan):
i Mamean: Ixil, Awakatek; Mam, Teco
ii K’ichean: Sipakapense, Sakapultek, Tz’utujil, Kaqchikel, K’ichee’,
Poqomam, Poqomchi’, Uspantek, Q’eqchi’
(Adapted from Campbell and Kaufman 1985: 188)
Like other Mayan languages, Kaqchikel is a head-marking and morphologi-
cally ergative language in which subjects and objects are not overtly case-marked
for grammatical relations. Instead, grammatical relations are obligatorily marked
on predicates with two sets of person/number morphemes, traditionally called Set
A and Set B in Mayan linguistics. Set A corresponds to ergative (transitive sub-
jects) and genitive (possessors) marking, and Set B to absolutive (transitive objects
and intransitive subjects) marking. e order of the morphemes is [Aspect-B-Verb
stem] for intransitive verbs and [Aspect-B-A-Verb stem] for transitive verbs.¹
e word order of most Mayan languages is predicate-initial in pragmati-
cally neutral contexts (England 1991, Aissen 1992). According to García Matzar
and Rodríguez Guaján (1997), among others, the basic word order of Kaqchikel
is VOS, as exemplied in (2a), with neither the subject nor the object topicalized
or focused (Rodríguez Guaján 1994: 200, García Matzar and Rodríguez Guaján
1997: 333, Tichoc Cumes et al. 2000: 195, Ajsivinac Sian et al. 2004: 162. For
psycho/neuro-linguistic evidence of VOS being syntactically basic in Kaqchikel,
see Koizumi et al. 2014, Yasunaga et al. 2015, Koizumi and Kim 2016, Yano et al.
¹ Unless otherwise noted, the description of Kaqchikel grammar in this paper is based
on our eldwork with three native consultants, Lolmay Pedro Oscar García Mátzar (Chi-
maltenango), Juan Esteban Ajsivinac Sian (Patzicía), and Filiberto Patal Majzul (Patzún).
Two Routes to the Mayan VOS: From the View of Kaqchikel 27
2017, among others). VSO is also grammatically allowed, as shown in (2b):²
(2) a. X-Ø-u-chöy ri chäj ri ajanel. [VOS]
CP-B3sg-A3sg-cut DET pine.tree DET carpenter
b. X-Ø-u-chöy ri ajanel ri chäj. [VSO]
CP-B3sg-A3sg-cut DET carpenter DET pine.tree
“e carpenter cut the pine tree.”
It is possible to topicalize the subject by moving it in front of the verb, as
exemplied in (3a) (García Matzar and Rodríguez Guaján 1997: 334). Similarly,
the object may be fronted as a topic, as shown in (3b) (García Matzar and
Rodríguez Guaján 1997: 335):³
(3) a. Ri ajanel x-Ø-u-chöy ri chäj. [SVO]
DET carpenter CP-B3sg-A3sg-cut DET pine.tree
“e carpenter cut the pine tree.”
b. Ri chäj x-Ø-u-chöy ri ajanel. [OVS]
DET pine.tree CP-B3sg-A3sg-cut DET carpenter
3. Agreement and hierarchical structure in Kaqchikel
As the starting point of our discussion of Kaqchikel syntax, we adopt Imanishi’s
(2014) proposal about agreement and hierarchical structure in the language, which
is schematically shown in (4) and (5) below. (In the discussion in this section,
constituent orders are irrelevant and arbitrarily represented in syntactic diagrams.
We discuss linear ordering in the next section.) At the point of a derivation in
(4), Voice undergoes agreement with the subject (SUB), which will eventually be
reected as an ergative (Set A) agreement morpheme in the verbal complex.
² e following abbreviations are used in glosses: 1: rst person, 3: third person, A: set A
ergative, B: set B absolutive, CL: classier, CP: completive, DET: determiner, IP: incomple-
tive, pl: plural, sg: singular, PRFV: perfective, TV: transitive verb sux, AF: agent focus.
³ When a transitive subject undergoes some type of movement to the pre-verbal position,
such as wh-movement and focus movement, ergative agreement does not appear on the
verb and a special morpheme suxes to the stem, a construction commonly termed Agent
Focus (AF). We are not concerned with AF in this paper. All transitive sentences with the
SVO word order discussed in this paper have canonical transitive agreement. For AF, see,
among others, Preminger (2014) and Watanabe (2017).
Imanishi (2014) assumes that vP dominates VoiceP, whereas Coon (2010) assumes that
VoiceP dominates vP. is dierence does not have crucial bearing on the main points of
this paper, and we remain agnostic between the two possibilities. For the sake of exposition,
we assume in this paper that VoiceP dominates vP (for arguments for this structure, see
Pylkkänen 2002 and Harley 2013, among others).
More specically, Imanishi (2014) proposes that ergative Case is a type of Case which
is assigned by a phase head to the highest Case-less DP within the Spell-Out domain of a
phase. In (4), therefore, ergative Case is assigned to the otherwise Case-less subject by the
phase head, Voice. On the other hand, Imanishi (2014) assumes that absolutive Case is as-
28 K O, K S, N Y,  M K
(4) Ergative agreement
e subject then raises to Spec,TP to satisfy an EPP feature on T. is movement
makes it possible for T to agree with the object, resulting in the absolutive agree-
ment (Set B) morpheme. us, the subject movement feeds the absolutive agree-
ment between T and the object (Imanishi 2014: 60).
(5) Absolutive agreement
e structure in (5) captures the following two basic characteristics of
Kaqchikel transitive sentences. First, the absolutive agreement morpheme (Set B)
occurs outside the ergative agreement morpheme (Set A) in the verbal complex
([Aspect-B-A-Verb stem]). e Mirror Principle proposed by Baker (1985) (i.e.,
morpheme order should mirror syntactic structure) suggests that the functional
head responsible for the absolutive agreement should be structurally higher than
the functional head implicated in the ergative agreement, as is the case in (5).
Second, the subject is structurally higher than the object, in the sense that the
former c-commands the latter. is point can be shown by well-known syntactic
diagnostics. For example, anaphors (e.g., herself, himself, themselves, each other) must
be c-commanded by their antecedents in a local domain (i.e., anaphors must be
locally bound), a condition known as Binding Condition A (cf. Chomsky 1981).
us, (6a), in which the anaphor each other in a verb phrase (VP) is bound by Mary
and John, is grammatical, whereas (6b) is ungrammatical, because each other is not
c-commanded by its potential antecedent Mary and John.
(6) a. [Mary and John [VP saw each other]].
b. *[Each other [VP saw Mary and John]].
signed by T via Agree (Chomsky 2000).
Two Routes to the Mayan VOS: From the View of Kaqchikel 29
Now consider the following Kaqchikel sentences:
(7) a. X-Ø-ki-tz’ët (jub’ey chik) k-i’ a Lolmay chuqa’
CP-B3sg-A3pl-see (once again) each.other CL Lolmay and
a Xwan.
CL Juan
“Lolmay and Juan saw each other (again).”
b. * X-e’-ru-tz’ët (jub’ey chik) a Lolmay chuqa’ a Xwan
CP-B3pl-A3sg-see (once again) CL Lolmay and CL Juan
In (7a), the subject a Lolmay chuqa’ a Xwan “Lolmay and Juan” is cross-referenced
with the third person plural agreement morpheme of Set A, ki. e object is the
anaphor k-i’each other,” associated with the third person singular agreement
morpheme of Set B, Ø, which is phonetically null. In (7b), on the other hand, the
subject is the anaphor k-i’, which triggers the third person singular Set A agree-
ment ru, and the object is the plural DP a Lolmay chuqa’ a Xwan, which is cross-
referenced with the third person plural Set B agreement e’. e contrast in gram-
maticality between (7a) and (7b) indicates that the ergative subject c-commands
the absolutive object in Kaqchikel (see also Henderson 2012 for similar data).
Another syntactic test that can be used to detect a c-command rela-
tion between nominal arguments is Binding Condition C, which states that
R-expressions (e.g., Mary, the book that John wrote) must be free (i.e., R-expressions
must not be c-commanded by noun phrases coreferential with them) (cf. Chomsky
1981). e rst three examples in (8) are grammatical under the interpretations in
which John is coreferential with his or him. In contrast, (8d) is ungrammatical if
John and He are coreferential. is is because only in (8d) is John c-commanded by
the pronoun coreferential with it.
(8) a. John1 is looking for his1 wife.
b. His1 wife is looking for John1.
c. John1s wife is looking for him1.
d. * He1 is looking for John1s wife.
Turning back to Kaqchikel, consider the sentences in (9), which are parallel to the
English counterparts in (8). In (9c) and (9d), pro indicates a phonetically null pro-
noun in the object and subject position, respectively.
(9) a. N-Ø-u-kanoj [ri r1-ixjayil] a Xwan1.
IC-B3sg-A3sg-seek DET A3sg-wife CL Juan
“Juan1 is looking for his1 wife.”
b. N-Ø-u-kanoj a Xwan1 [ri r1-ixjayil].
IC-B3sg-A3sg-seek CL Juan DET A3sg-wife
“His1 wife is looking for Juan1.”
30 K O, K S, N Y,  M K
c. N-Ø-u-kanoj pro1 [ri r-ixjayil a Xwan1].
IC-B3sg-A3sg-seek DET A3sg-wife CL Juan
“Juan1’s wife is looking for him1.”
d. *N-Ø-u-kanoj [ri r-ixjayil a Xwan1] pro1.
IC-B3sg-B3sg-seek DET A3sg-wife CL Juan
Lit. “He1 is looking for Juan1’s wife.”
e grammaticality of (9c) suggests that the object does not c-command the
subject; the ungrammaticality of (9d) indicates that the subject c-commands the
object. Taken together, the paradigm in (9) shows again that in Kaqchikel transi-
tive sentences, the syntactic position of the ergative subject c-commands that of
the absolutive object, but not vice versa.
4. A right-specier analysis
In this section, we show that a right-specier analysis, such as that by Aissen
(1992), is readily applicable to Kaqchikel with minimal modication.
4.1. Right-specier analysis of Tzotzil (Aissen 1992, 1996)
To account for VOS word orders in Mayan languages in general and those in
Tzotzil (Tseltalan branch) in particular, Aissen (1992) proposes parameterizing
the order of speciers with respect to their heads as follows: e specier of a
functional category X precedes X, whereas the specier of a lexical category X
follows X. Assuming that the base position of the external argument (subject) is
Spec,VP, the VOS order is obtained when both the subject and object remain in
situ, as shown in (10). According to Aissen (1992, 1996), if the subject undergoes
movement to the specier of a functional category outside VP, the SVO order is
obtained, because the speciers of functional categories are all located to the left.
(10) [CP [IP [VP [V’
Data from weak crossover eects make the same point. It has been known since Postal
(1971) that an object wh-operator cannot cross over a bound pronoun contained in a struc-
turally higher subject DP. For instance, (ia), which involves no crossover conguration, can
be understood as a question asking the identity of the person x such that x respected x’s
mother, whereas (ib), in which the object wh-operator crosses over the subject DP contain-
ing the pronoun his, does not allow such a bound variable interpretation.
(i) a. Who1 t1 respected his1 mother?
b. W h o1 did his*1/2 mother respect t1?
e corresponding Kaqchikel examples in (ii) show the exact same pattern, which suggests
that the subject is structurally higher than the object in the language.
(ii) a. Achike1 x-Ø-kamela-n ru1-te’ t1?
who CP-B3sg-respect-AF A3sg-mother
b. Achike1 x-Ø-u-kamelaj t1 ru*1/2-te’?
who CP-B3sg-A3sg-respect A3sg-mother
Two Routes to the Mayan VOS: From the View of Kaqchikel 31
4.2. Extension to Kaqchikel
We now consider how various word orders in Kaqchikel can be derived if we apply
to this language a right-specier analysis similar to that by Aissen (1992, 1996).
Following Imanishi (2014), we have assumed a slightly more elaborated hierar-
chical structure than does Aissen (1992). Specically, the subject overtly raises
to Spec,TP in our analysis. To make this analysis compatible with Aissen’s right-
specier analysis, we propose that in Kaqchikel, in addition to lexical categories,
all categories up to TP have speciers to the right, correctly resulting in the VOS
order, as shown in (11). Following Imanishi (2014), we also assume that V raises
to C via v, Voice, and T.
(11) Kaqchikel VOS
If the object undergoes a right-ward scrambling to TP across the subject, the
VSO order is obtained, as shown in (12).
We stipulate that speciers are located to the right, except for CP. We suspect that this is
because the CP domain is closely related to discourse/information structure, and it might be
the case that UG requires elements related to discourse/information structure to appear to
the left. is is consistent with the fact that rightward wh-movement is virtually unattested
in spoken languages (cf. Richards 2010, 2016).
In Kaqchikel embedded clauses, a complementizer such as chin “that” appears as an inde-
pendent lexical item, as shown below.
(i) X-Ø-in-rayij [chin x-Ø-tzaq ri achin].
CP-B3sg-A1Sg-desire COMP CP-B3sg-fall DET man
‘I wanted the man to fall.’ (Clemens 2013)
Note that this fact does not preclude V-to-C movement in Kaqchikel once a split CP
structure (cf. Rizzi 1997) is adopted: the complementizer is located in the topmost category,
Force, which species clause types, and V moves to a lower category, presumably Topic or
32 K O, K S, N Y,  M K
(12) Kaqchikel VSO
e SVO word order is derived from a structure similar to (11) by moving the
subject to Spec,CP, which, following Aissen (1992), we assume is located to the
left, as shown in (13).
(13) Kaqchikel SVO
Similarly, we obtain the OVS order by topicalizing the object to Spec,CP, as
shown in (14).
In Kaqchikel, the SVO order is possible in matrix clauses and certain subordinate clauses.
Again, we assume a split CP structure for the left periphery.
Two Routes to the Mayan VOS: From the View of Kaqchikel 33
(14) Kaqchikel OVS
e right-ward scrambling in (12), subject topicalization in (13), and object
topicalization in (14) are all A-bar movements.¹ us, for example, they do not
alter A-binding relations between the subject and the object, as detected in the
anaphor binding test.
5. A predicate fronting analysis
In this section, we argue that Coon’s (2010) predicate fronting analysis of Chol
cannot easily be extended to Kaqchikel because of crucial grammatical dierences
between the two languages.
5.1. Predicate fronting analysis of Chol (Coon 2010)
Coon (2010) proposes an alternative account of the VOS order in Chol (Cholan
branch), which is called a predicate fronting analysis. According to this analysis,
all speciers precede their heads a la Kayne (1994), and the subject remains in situ
in Spec,VoiceP. e maximal predicate projection, vP, fronts to Spec,TP, giving
rise to the VOS order, as schematically shown in (15). Coon (2010) argues that
T in Chol has strong agreement features requiring the verb to move overtly to T;
V cannot move to T because head movement is generally absent in this language.
erefore, the entire predicate phrase vP must front as a last resort.
¹ Unlike the Spec,TP subject position, which is considered to have A-properties, the
scrambled object in (12) is located in a TP adjoined position, resulting in showing A-bar
properties. is is similar to the case where sentence-internal scrambling adjoining to TP
can be A-bar movement in Japanese, as shown in the following example (no condition C
violation is triggered).
(i) Zibun-zishin1-o Taro1-ga t1 hihanshita.
himself-Acc Taro-Nom criticized
“Taro criticized himself.”
34 K O, K S, N Y,  M K
(15) Chol VOS
Evidence for this analysis stems from dierent restrictions on objects in VOS
and VSO sentences. In Chol, the object must be a bare (determiner-less) noun
phrase (NP) in VOS sentences, and full determiner phrases (DPs) with an overt
determiner (D) are prohibited from occupying the VOS object position, as shown
in (16).
(16) VOS
a. Tyi i-kuch-u [NP si’ ] aj-Maria.
PRFV A3-carry-TV wood DET-Maria
“Maria carried wood.”
b. * Tyi i-kuch-u [DP jiñi si’ ] aj-Maria.
PRFV A3-carry-TV DET wood DET-Maria
“Maria carried the wood.” (Coon 2010: 355)
In contrast, the object must be a DP in VSO sentences, as shown in (17).
(17) VSO
a. Tyi i-kuch-u aj-Maria [DP jiñi si’ ].
PRFV A3-carry-TV DET-Maria DET wood
“Maria carried the wood.”
b. * Tyi i-kuch-u aj-Mari a [NP si’ ].
PRFV A3-carry-TV DET-Maria wood
“Maria carried wood.” (Coon 2010: 355)
ese restrictions on objects are readily accounted for by assuming that a bare NP
object must remain in situ within VP, as in (15), whereas a full DP object must
undergo an object shift to Spec,AbsP before the remnant vP preposing, as shown
in (18), yielding the VSO order.
Two Routes to the Mayan VOS: From the View of Kaqchikel 35
(18) Chol VSO
Further empirical support for Coon’s analysis pertains to adjunct positions.
e predicate fronting analysis makes two predictions. First, the verb and a bare
NP object are adjacent to each other within VP, such that nothing can intervene
between them. Second, the verb and a full DP object are separated by several func-
tional projections so that some adverbial expressions may intervene between them.
ese predictions are well borne out, as illustrated in (19), in which the subjects are
all phonetically empty (i.e., pro).¹¹
(19) a. Tyi k-wuts’-u abi [DP ili pisil ].
PRFV A1-wash-TV yesterday DET clothes
b. * Tyi k-wuts’-u abi [NP pisil ].
PRFV A1-wash-TV yesterday clothes
c. Tyi k-wuts’-u [NP pisil ] abi.
PRFV A1-wash-TV clothes yesterday
“I washed (the) clothes yesterday.” (Coon 2010: 367)
More concretely, if the adverb abi “yesterday” is left-adjoined to VoiceP, the pattern
in (19) is as expected, as shown in (20) and (21).¹²
¹¹ Since no examples with overt subjects are provided in Coon (2010), there remains a pos-
sibility that the sentence in (19a) has the underlying V-S-Adv-O order and the V-Adv-S-O
order is actually ungrammatical. If that is the case, the V-Adv-S-O order will no longer be
problematic for the right-specier analysis of Chol without assuming V-to-C movement
(see discussion in Section 5.3).
¹² Although the time adverb abi “yesterday” is adjoined to VoiceP in (20) following the
analysis by Coon (2010), we posit in our analysis of Chol and Kaqchikel that time adverbs
are adjoined to TP/T’, as their interpretations are related to tense, not to voice.
36 K O, K S, N Y,  M K
(20) Bare NP object
(21) Full DP object
5.2. Extension to Kaqchikel
In attempting to extend Coon’s predicate fronting analysis of Chol to Kaqchikel,
however, we face at least three problems. First, if vP raises to Spec,TP in
Kaqchikel parallel to the structure of Chol in (20), and the object in the fronted
VP must be a bare NP, we expect that the VOS object in Kaqchikel must likewise
be a bare NP. However, this expectation does not materialize. As shown in (22),
in Kaqchikel, unlike in Chol, both a bare NP object and a full DP object (either
denite or indenite) are acceptable in the VOS order. e object must be denite
in VSO, as shown in (23).
(22) a. X-Ø-u-qüm raxya’ ri ajanel.
CP-B3sg-A3sg-sip cold.water the carpenter
“e carpenter sipped cold water.”
b. X-Ø-u-ch’äj ri ch’ich’ ri a Xwan.
CP-B3sg-A3sg-wash DET car DET CL Juan
“Juan washed the car.”
c. X-Ø-u-ch’äj jun ch’ich’ ri a Xwan.
CP-B3sg-A3sg-wash a car DET CL Juan
“Juan washed a car.”
Two Routes to the Mayan VOS: From the View of Kaqchikel 37
(23) a. X-Ø-u-ch’äj ri a Xwan ri ch’ich’.
CP-B3sg-A3sg-wash DET CL Juan DET car
“Juan washed the car.”
b. * X-Ø-u-ch’äj ri a Xwan jun ch’ich’.
CP-B3sg-A3sg-wash DET CL Juan a car
“Juan washed a car.”
Second, as we saw in (19b), the predicate fronting analysis accounts for the
fact that adjuncts cannot intervene between the verb and the object in the VOS
order in Chol. If Kaqchikel had the same structure as Chol, adverbs could not
occur between the verb and its object in the VOS order in Kaqchikel, either, con-
trary to fact. In Kaqchikel, unlike in Chol, a time adverb such as iwir “yesterday”
may freely occur between the verb and the object, irrespective of its deniteness, as
shown in (24).
(24) a. V Adv Oindef S
X-Ø-u-ch’äj iwir jun ch’ich’ ri a Xwan.
CP-B3sg-A3sg-wash yesterday a car DET CL Juan
“Juan washed a car yesterday.”
b. V Adv Odef S
X-Ø-u-ch’äj iwir ri ch’ich’ ri a Xwan.
CP-B3sg-A3sg-wash yesterday DET car DET CL Juan
c. V O Adv S
X-Ø-u-ch’äj ri ch’ich’ iwir ri a Xwan.
CP-B3sg-A3sg-wash DET car yesterday DET CL Juan
d. V O S Adv
X-Ø-u-ch’äj ri ch’ich’ ri a Xwan iwir.
CP-B3sg-A3sg-wash DET car DET CL Juan yesterday
“Juan washed the car yesterday.”
e nal problem with the application of the predicate fronting analysis of
Chol to Kaqchikel is that it cannot account for the absolutive agreement. Chol is
a so-called low absolutive language, in which absolutive agreement of a transitive
object is licensed within a predicate phrase. e morpheme order in the verbal
complex is [Aspect-A-Verb stem-B] in this language. We might assume that v
enters into an absolutive agreement relation with the object. In contrast, Kaqchikel
is a high absolutive language, in which the functional head responsible for the
absolutive agreement with the object is structurally higher than the base position
of the transitive subject, as reected in the morpheme order [Aspect-B-A-Verb
stem]. As mentioned in Section 3, following Imanishi (2014), we assume that T
enters into an absolutive agreement relation with the object. If the subject stays
in situ, and the object together with the verb raises to Spec,TP parallel to (16)
in Kaqchikel, then the object should not be able to agree with T even before the
predicate fronting because of a defective intervener, i.e., the subject.
To summarize, it is entirely unreasonable and dicult to apply predicate-
38 K O, K S, N Y,  M K
fronting to account for the VOS word order in Kaqchikel because of the three
empirical dierences between Chol and Kaqchikel, as described.
On the other hand, the right-specier analysis of Kaqchikel outlined in the
previous section can readily account for problematic cases regarding the predi-
cate fronting analysis. First, the fact that the VOS object may either be denite
or indenite follows if we assume that a subject obligatorily moves to Spec,TP,
resulting in the VOS order, even if a denite object undergoes object shift to
Spec,VoiceP.¹³ Second, if an adverb such as iwir “yesterday” is adjoined to TP/T’,
as schematically shown in (25), then its distribution shown in (24) is as expected.
e [V Complex] in (25) stands for the verbal complex resulting from V raising to
C through v to T.
(25) Kaqchikel VOS
Finally, T can agree with the object before it raises to C because the subject does
not intervene between them on this account.¹
¹³ We assume that denite objects obligatorily undergo object shift to Spec,VoiceP in both
Chol and Kaqchikel, following the Mapping Hypothesis by Diesing (1992).
¹ As an anonymous reviewer points out, it would be interesting to see how other types of
adverbs, such as manner adverbs, behave in VOS sentences. Given that manner adverbs ap-
pear low in the structure (adjoined to vP or VP, for example), it seems dicult to obtain the
“V OBJdef ADVmanner SUB”/“V OBJ SUB ADVmanner order. (Whether these orders are pos-
sible or not is up to the availability of adverb scrambling in the language.) Imanishi (2014,
to appear) reports some facts concerning manner and time adverbs in Kaqchikel, although
they are used within nominalized clauses (usually with null pronouns) and it is unclear how
they behave in VOS sentences (see also Henderson and Coon 2018, who discuss various
adverbs in relation to AF in Kaqchikel).
Two Routes to the Mayan VOS: From the View of Kaqchikel 39
5.3. Right-specier analysis of Chol
To derive the VOS in Chol, Coon (2010) posits that (i) speciers uniformly occur
to the left, (ii) vP is obligatorily fronted to Spec,TP, (iii) V does not move (no head
movement in Chol), and (iv) the external argument stays in situ in Spec,VoiceP (T
does not have an EPP feature). On the other hand, on our account of Kaqchikel,
(i) speciers occur to the right except for CP, (ii) there is no vP fronting, (iii) V
raises to C, and (iv) the external argument moves to Spec,TP. ese dierences
between the two languages are summarized in (26).
(26) Parametric dierences between Chol and Kaqchikel (to be revised)
Chol (Coon 2010) Kaqchikel
Directionality of speciers Left Right (except for cp)
vP fronting to Spec,TP Present Absent
V raising to C Absent Present
Subject raising to Spec,TP Absent Present
We will show below that we can dispense with some of the parametric dierences
given in (26), and that it is possible to reduce the parametric dierences between
Chol and Kaqchikel into two.
Exploring the possibility that there is no predicate fronting even in the
derivation of Chol VOS sentences, we suggest that VOS sentences in Chol and
Kaqchikel have the structures represented in (27a) and (27b), respectively.
(27) a. Chol VOS b. Kaqchikel VOS
e only dierence between (27a) and (27b) is that the subject moves to Spec,TP
in Kaqchikel, but not in Chol (the subject remains in Spec,VoiceP). is dierence
is necessary to explain the fact that denite objects invariably result in the VSO
order in Chol: a full denite DP undergoes object shift to the periphery of the
phase (i.e., VoiceP), skipping over a subject, as shown in (28). A comparable object
shift in Kaqchikel, on the other hand, does not alter word order because the sub-
ject is located in Spec,TP.
40 K O, K S, N Y,  M K
(28) Chol VSO
Given the structure in (27a), an immediate question arises as to why noth-
ing can intervene between V and bare NPs in Chol. We assume that bare NPs
in Chol must undergo (pseudo-)incorporation to be licensed. is means that
(pseudo-)incorporation is prohibited in cases where some element (such as a
subject and adverb) intervenes between V and NPs. erefore, we can explain
the fact that V and bare NPs must be adjacent in Chol, without invoking vP-
fronting.¹ In contrast, such (pseudo-)incorporation is not necessary for indenite
objects to be licensed in Kaqchikel, presumably because it has the indenite article
jun. e table in (29) summarizes the parametric dierences between Chol and
¹ We thank an anonymous reviewer for suggesting the possibility that the presence/absence
of (pseudo-)incorporation could be a locus of parametric dierences between Chol and
Kaqchikel. As the reviewer suggests, this process is similar to “incorporation antipassive,” a
phenomenon discussed by Dayley (1981), Coon (2013), among others.
(i) a. Transitive
Tyi k-wuts’-u pisil.
PRFV A1-wash-TV clothes
“I washed clothes.
b. Incorporation antipassive
Tyi k-cha‘l-e wuts’ pisil.
PRFV A1-do-DTV wash clothes
“I washed clothes. (Coon 2013: 76)
Contrary to transitive objects, objects in incorporation antipassive may not be a full DP –
for example, determiners cannot appear with antipassive objects. We suspect that a similar
process is happening in transitive sentences with bare objects in Chol, thus prohibiting
them from being used with determiners and other elements (such as adverbs).
¹ A recent paper by Clemens and Coon (2018) proposes yet another theory of deriving
verb-initial word order in Mayan languages. ey argue that, all else being equal, the basic
word order of Mayan languages is invariably VSO (which is derived by a sequence of head
movement), and VOS is obtained by one of the following strategies (Clemens and Coon
2018: 238):
Two Routes to the Mayan VOS: From the View of Kaqchikel 41
(29) Parametric dierences between Chol and Kaqchikel (Revised)
Chol Kaqchikel
Subject raising to Spec,TP Absent Present¹
(Peudo-)incorporation of bare NP objects Present Absent
6. Conclusion¹
ere are multiple syntactic routes to the VOS order (Chung 2017). Dierent
VOS languages may have dierent syntactic structures. ere are two major pro-
posals regarding how Mayan VOS word order is grammatically obtained. We
proposed in this paper that Kaqchikel, and possibly Chol as well, derive the VOS
order through a right-specier route, rather than a predicate fronting route.
Before concluding the paper, we would like to briey comment on the
linearization of speciers. What is clear from the discussion above is that we need
to depart from Kayne’s (1994) LCA-based approach to word order, which assumes
that precedence relations are determined by dominance relations: since subjects in
the VOS order in Kaqchikel are structurally higher than objects, the LCA predicts
that subjects precede objects, contrary to the fact.
Recently, Takita (to appear) proposes the hypothesis that labeling is required
for linearization. More specically, departing from Chomsky (2013, 2015), he
argues that labeling does not contribute to semantic interpretation, but is required
solely for linearization in the sense that only labeled SOs (=syntactic objects) can
have the relative linear order of their members determined. For example, he con-
siders how the following structures in (30) are labeled.
(30) a. {X, Y} [head-head]
b. {X, YP} [head-phrase]
c. {XP, YP} [phrase-phrase]
(i) a. subject in high right-side topic position
b. heavy-NP shift of phonologically heavy subjects
c. prosodic reordering of bare NP objects
Of relevance to Kaqchikel VOS is (ia), which assumes that the VOS order is derived by
moving a subject to a right-side topic position, as illustrated in (ii).
(ii) [TopicP [CP V-Complex [TP tS O] ] S ]
is proposal, however, is incompatible with the experimental data reported in Koizumi et
al.’s (2014) study. Based on the results obtained from the sentence-plausibility judgment
task, Koizumi et al. (2014) report that VOS sentences induce less processing load than
VSO/SVO sentences for Kaqchikel speakers, suggesting that VOS is syntactically simpler
than the other two. If VSO is the basic word order and VOS is derived by movement to the
higher right-side functional projection, it is not clear why VOS has a processing advantage
over VSO.
¹ Erlewine (2016) also argues that ergative subjects in Kaqchikel move to Spec,TP, based
on the facts that ergative subjects in Kaqchikel trigger AF in constructions involving A-bar
dependencies (see also footnote 3).
42 K O, K S, N Y,  M K
(30a) represents the case where a head is merged with another head. e SO is
problematic for labeling because X and Y are symmetric and its label cannot be
determined. Takita (to appear) argues that (30a) can be pronounced if one of the
heads lacks its phonological realization due to its lexical property or movement,
thus making linearization of X and Y irrelevant. In (30b), there is an asymmetry
between the head X and the phrase XP, so minimal search takes the head X as the
label for (30b). Furthermore, he assumes the following linearization rules in (31),
which regulate the process of mapping a set to an ordered pair.
(31) a. Head-initial linearization rule (e.g., English): {X X, YP} <X, YP>
b. Head-nal linearization rule (e.g., Japanese): {X X, YP} <YP, X>
e SO in (30c), again, is symmetric and hence cannot be labeled. Takita (to
appear) argues that one way to avoid the problem is to make one of the set mem-
bers phonologically null (by movement, for example). Another way to pronounce
(30c), which is relevant to our discussion, is to label it by feature sharing (cf.
Chomsky 2013) and linearize the structure with the linearization rule in (32),
which states that XP with a valued feature precedes YP with an unvalued feature.
(32) Linearization rule for SOs labeled as <F, F>:
{<F, F> XPF[val], YPF[unval]} <XP, YP>
In eect, this rule yields the left specier order, provided that φ-features on DPs
are valued and their counterparts (on T) are unvalued.
We suggest that the linearization rule in (32) is not universal but subject
to parametric variation, as in the case in (31). More concretely, we assume that
Kaqchikel and Chol have the linearization rule in (33).
(33) Linearization rule for SOs labeled as <F, F> (Kaqchikel and Chol)
{<F, F> XPF[val], YPF[unval]} <YP, XP>
is linearization rule ensures that in Kaqchikel and Chol, the subject with a val-
ued feature, which enters into an Agree relation with Voice, appears to the right
of VoiceP, resulting in the OS order.¹,¹ Whether or not the parametric variation
between (32) and (33) is related to other properties of grammar (such as ergativity
¹ Actually, the order of the SO {TP,DP} in Kaqchikel cannot be determined by (33) be-
cause we assume, following Imanishi (2014), that subject movement to Spec,TP does not
involve φ-feature valuation (purely EPP-driven movement): hence, no feature sharing is
happening between TP and DP. We stipulate that, in cases where Agree is not involved in
movement, the order of the original position of the moved element (for example, VoiceP
precedes DP) must be preserved in the moved position. is stipulation might be relevant
to the fact that leftward scrambling, another candidate of non-Agree-driven movement, is
observed only in head-nal languages, which has the “object before verb” order in the origi-
nal position (cf. Saito and Fukui, 1998; see also Takita, to appear, for relevant discussion).
19 e order of adverbs is relatively exible in that they can right- or left-adjoin to
TP/T’. We suspect that this is because adverbs do not enter into an Agree relation with T,
thus being out of the scope of the linearization rule in (33).
Two Routes to the Mayan VOS: From the View of Kaqchikel 43
and head/dependent marking) is also an interesting research question, which we
leave for future research.
Aissen, Judith L. (1992) Topic and focus in Mayan. Language 68: 43–80.
Aissen, Judith L. (1996) Pied-piping, abstract agreement, and functional projections in
Tzotzil. Natural Language and Linguistic eory 14: 447–491.
Ajsivinac Sian, Juan Esteban, Lolmay Pedro Oscar García Mátzar, Martín Chacach Cut-
zal, and Ixkusamil EstelaAlonzo Guaján (2004) Gramática descriptive del idioma maya
Kaqchikel: Rutzijoxik rucholik ri Kaqchikel ch’ab’äl. Guatemala City: Academia de las
Lenguas Mayas de Guatemala, Comunidad Lingüistica Kaqchikel.
Baker, Mark (1985) e Mirror Principle and morphosyntactic explanation. Linguistic
Inquiry 16: 373–416.
Campbell, Lyle and Terrence Kaufman (1985) Mayan linguistics: Where are we now?
Annual Review of Anthropology 14: 187–198.
Chomsky, Noam (1981) Lectures on government and binding. Dordrecht: Foris.
Chomsky, Noam (2000) Minimalist inquiries: e Framework. In: Roger Martin et al.
(eds.) Step by step: Essays on minimalist syntax in honor of Howard Lasnik, 89–155. Cam-
bridge, Mass.: MIT Press.
Chomsky, Noam (2013) Problems of Projections. Lingua 130: 33–49.
Chomsky, Noam (2015) Problems of projection: Extensions. In: Elisa Di Domenica et
al. (eds.) Structures, strategies and beyond - Studies in honour of Adriana Belletti, 3–16.
Amsterdam/Philadelphia: John Benjamins.
Chung, Sandra (2017) VOS languages: Some of their properties. In: Martin Everaert and
Henk C. Van Riemsdijk (eds.) e Wiley Blackwell companion to syntax, Second Edition,
4787–4832. Malden, MA: Wiley-Blackwell.
Clemens, Lauren (2013) Kaqchikel SVO: V2 in a V1 language. In: Michael Kenstowicz
(ed.) Studies in Kaqchikel grammar (MIT working papers on endangered and less familiar
languages 8), 1–24. Cambridge, MA: MIT Press.
Clemens, Lauren and Jessica Coon (2018) Deriving verb-initial word order in Mayan. Lan-
guage 94: 237–280.
Coon, Jessica (2010) VOS as predicate fronting in Chol. Lingua 120: 354–378.
Coon, Jessica (2013) Aspects of split ergativity. Oxford: Oxford University Press.
Coon, Jessica (2016) Mayan morphosyntax. Language and Linguistics Compass 10: 515–550.
Dayley, Jon (1981) Voice and ergativity in Mayan languages. Journal of Mayan Linguistics 2:
Diesing, Molly (1992) Indenites. Cambridge, Mass.: MIT Press.
England, Nora C. (1991) Changes in basic word order in Mayan languages. International
Journal of American Linguistics 57: 446–486.
Erlewine, Michael Yoshitaka (2016) Anti-locality and optimality in Kaqchikel Agent
Focus. Natural Language and Linguistic eory 34: 429–479.
García Matzar, Lolmay Pedro and Pakal B’alam José Obispo Rodríguez Guaján (1997)
Rukemik ri Kaqchikel chi’: Gramática Kaqchikel. Guatemala City: Cholsamaj.
Harley, Heidi (2013) External arguments and the Mirror Principle: On the distinctness of
Voice and v. Lingua 125: 34–57.
Henderson, Robert (2012) A tour of Kaqchikel ergativity. Handout, McGill Ergativity Lab.
Henderson, Robert and Jessica Coon (2018) Adverbs and variability in Kaqchikel Agent
44 K O, K S, N Y,  M K
Focus. Natural Language and Linguistic eory 36: 149–173.
Imanishi, Yusuke (2014) Default ergative. Doctoral dissertation, MIT.
Imanishi, Yusuke (to appear) Parameterizing split ergativity in Mayan. Natural Language
and Linguistic eory.
Kayne, Richard S. (1994) e antisymmetry of syntax. Cambridge, MA: MIT Press.
Koizumi, Masatoshi and Jungho Kim (2016) Greater left inferior frontal activation for
SVO than VOS during sentence comprehension in Kaqchikel. Frontiers in Psychology 7:
1541. doi: 10.3389/fpsyg.2016.01541.
Koizumi, Masatoshi, Yoshiho Yasugi, Katsuo Tamaoka, Sachiko Kiyama, Jungho Kim, Juan
Esteban Ajsivinac Sian, and Lolmay Pedro Oscar García Mátzar (2014) On the (non-)
universality of the preference for subject-object word order in sentence comprehension:
A sentence processing study in Kaqchikel Maya. Language 90: 722–736.
Postal, Paul (1971) Cross-over phenomena. New York: Halt, Rinehart and Winston.
Preminger, Omer (2014) Agreement and its failures. Cambridge, MA: MIT Press.
Pylkkänen, Liina (2002) Introducing arguments. Doctoral Dissertation. MIT.
Richards, Norvin (2010) Uttering trees. Cambridge, Mass.: MIT Press.
Richards, Norvin (2016) Contiguity theory. Cambridge, Mass.: MIT Press.
Rizzi, Luigi (1997) e ne structure of the left periphery. In: Liliane Haegeman (ed.) Ele-
ments of grammar, 281–337. Dordrecht: Springer.
Rodríguez Guaján, J. O. (1994) Rutz’ib’axik ri Kaqchikel: Manual de redacción Kaqchikel.
Guatemala, Guatemala: Editorial Cholsamaj.
Saito, Mamoru and Naoki Fukui (1998) Order in phrase structure and movement. Linguis-
tic Inquiry 29: 439–474.
Takita, Kensuke (to appear) Labeling for linearization. e Linguistic Review.
Tichoc Cumes, Rosalino, Juan Esteban Ajsivinac Sian, Lolmay Pedro Oscar García, Ixchel
Carmelina Espantzay, Martín Chacach Cutzal, and María Estela Alonso Guaján
(2000) Runuk’ul pa rub’eyal rutz’ib’axik ri Kaqchikel ch’ab’ äl: Gramática normativa del
idioma Maya Kaqchikel. Chimaltenango: Comunidad Lingüistica Kaqchikel de la Aca-
demia de las Lenguas Mayas de Guatemala.
Watanabe, Akira (2017) e division of labor between syntax and morphology in the
Kichean agent-focus construction. Morphology 27: 685–720.
Yano, Masataka, Daichi Yasunaga, and Masatoshi Koizumi (2017) Event-related brain
indices of gap-lling processing in Kaqchikel. In: Samuel R. Harris (ed.) Event-Related
Potential (ERP): Methods, outcomes, research insights, 89–122. Waltham, MA: NOVA
Yasunaga, Daichi, Masataka Yano, Yoshiho Yasugi, and Masatoshi Koizumi (2015) Is the
subject-before-object preference universal? An event-related potential study in the
Kaqchikel Mayan language. Language, Cognition, and Neuroscience 30: 1209–1229.
Author’s contact information: [Received 31 December 2018;
Faculty of Letters Accepted 6 September 2019]
Kanazawa Gakuin University
10 Sue-machi, Kanazawa, Ishikawa, 920-1392, Japan
Two Routes to the Mayan VOS: From the View of Kaqchikel 45
要 旨】
マヤ語 VOS語順への2つの道
大滝 宏一 杉崎 鉱司 遊佐 典昭 小泉 政利
金沢学院大学 関西学院大学 宮城学院女子大学 東北大学
 マヤ諸語における VOS 語順の派生に関して,これまで主に二つの分析が提案されてい
少なくともカクチケル語の VOS 語順に関しては「右方指定部分析」の方が妥当であること
を示す。また,「右方指定部分析」をチョル語の VOS 語順にも拡張する可能性に関しても議
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
The purpose of this paper is to explain the variation of Case alignment in the accusative side of the ergative split of Kaqchikel, Chol and Q’anjob’al (Mayan). In particular, I will address contrastive alignments found in their accusative side. In the accusative side of Kaqchikel, the intransitive subject and the transitive subject alike are cross-referenced by the absolutive morpheme (also known as the set B marker in Mayan linguistics). On the other hand, the object of a transitive verb is cross-referenced by the ergative morpheme (or the set A marker). In the accusative side of Chol and Q’anjob’al, by contrast, both the intransitive subject and the transitive subject are cross-referenced by the set A marker, while the set B marker cross-references the transitive object. This contrast is unexpected, given that these languages have a (nearly) identical biclausal structure for their accusative side, as I will claim building on Laka (2006) and Coon (2010a, 2013a): the aspectual predicate forms a biclausal structure with a nominalized clause. I will argue that the contrastive alignments found in Kaqchikel, Chol and Q’anjob’al follow from a parametric difference regarding the nominalization involved in the accusative side of these languages. It will be proposed that the Restriction on Nominalization (RON) holds for Kaqchikel, whereas it does not apply to Chol and Q’anjob’al: the nominalized verb must lack a syntactically projected external argument. The RON will be developed, based on a similar observation made for nominalizations in Greek and some Indo-European languages among others (Alexiadou 2001). As will be demonstrated, the presence or absence of the RON and the type of alignment patterns in the accusative side of the ergative split are causally connected.
Full-text available
This paper demonstrates that the person-number hierarchy effects in the Agent-Focus construction in Kichean languages of the Mayan family such as Kaqchikel, K’iche’, and Tz’utujil can be attributed to the general mechanism governing morphological realization of person and number features, once an appropriate characterization of relevant agreement markers is given. In other words, this account only relies on the minimal theoretical machinery needed for a description of individual languages, unlike previous analyses. The binary nature of person and number features plays a crucial role in the proposed account. It is also suggested that the impossibility of pairing a non-3rd person subject with a non-3rd person object in the Agent-Focus construction is due to the [+participant]-targeting application of the Obligatory Contour Principle. Furthermore, a novel pan-Mayan characterization of Agent-Focus is proposed to capture variation concerning the alternation between Agent-Focus and the plain transitive marking in the context of the Obligatory Contour Principle violation. The new characterization of AF also enables us to capture the rather exceptional form of the Agent-Focus construction in Yucatec. The overall framework is thus shown to have validity that extends beyond the Kichean group.
Full-text available
In many languages with flexible word order, transitive sentences in which the subject precedes the object have been reported to have a processing advantage during sentence comprehension compared with those in which the subject follows the object. This observation brings up the question of why this subject-before-object (SO) order should be preferred in sentence comprehension, together with the related empirical question of whether this preference is universal across all human languages. In the present ERP study, we address these two issues by examining the word order preference in Kaqchikel, a Mayan language spoken in Guatemala, in which the verb-object-subject (VOS) order is the syntactically basic word order. In the experiment, native speakers of Kaqchikel were auditorily presented four types of sentences (VOS, VSO, SVO, and OVS), followed by a picture that either matched or mismatched an event described in a preceding sentence, while their EEGs were recorded. The result of the ERP experiment showed that VSO elicited a larger positive component, called a P600 effect, in the comparison to the canonical word order, VOS in the third region (i.e., O of VSO versus S of VOS), in which the filler-gap dependency was supposed to be established in VSO sentences. Furthermore, SVO also exhibited a P600 effect compared to VOS in the third region, reflecting an increased syntactic processing cost. These results indicate that the syntactically basic word order, VOS, requires a lower amount of cognitive resources to process than other possible word orders in Kaqchikel. Based on these results, we argue that the SO preference in sentence comprehension reported in previous studies may not reflect a universal aspect of human languages; rather, processing preference may be language-specific to some extent, reflecting syntactic differences in individual languages.
Full-text available
In many languages with ergative morphology, transitive subjects (i.e. ergatives) are unable to undergo A’-extraction. This extraction asymmetry is a common hallmark of “syntactic ergativity,” and is found in a range of typologically diverse languages (see e.g. Deal 2016; Polinsky 2017, and works cited there). In Kaqchikel, the A’-extraction of transitive subjects requires a special verb form, known in Mayanist literature as Agent Focus (AF). In a recent paper, Erlewine (2016) argues that the restriction on A’-extracting transitive subjects in Kaqchikel is the result of an Anti-Locality effect: transitive subjects are not permitted to extract because they are too close to C⁰. This analysis relies crucially on Erlewine’s proposal that transitive subjects undergo movement to Spec,IP while intransitive subjects remain low. For Erlewine, this derives the fact that transitive (ergative) subjects, but not intransitive (absolutive) subjects are subject to extraction restrictions. Furthermore, it makes the strong prediction that phrasal material intervening between IP and CP should obviate the need for AF in clauses with subject extraction. In this paper, we argue against the Anti-Locality analysis of ergative A’-extraction restrictions along two lines. First, we raise concerns with the proposal that transitive, but not intransitive subjects, move to Spec,IP. Our second, and main goal, is to show that there is variation in whether AF is observed in configurations with intervening phrasal material, with a primary focus on intervening adverbs. We propose an alternative account for the variation in whether AF is observed in the presence of adverbs and discuss consequences for accounts of ergative extraction asymmetries more generally.
Full-text available
Cortical activations during the processing of Kaqchikel transitive sentences with canonical and non-canonical word orders were investigated using functional magnetic resonance imaging. Kaqchikel is an endangered Mayan language spoken in Guatemala. The word order in this language is relatively flexible. We observed higher cortical activations in the left inferior frontal gyrus for sentences with the subject-verb-object (SVO) word order, as compared to sentences with the verb-object-subject (VOS) word order, suggesting that Kaqchikel sentences are easier to process when they have the VOS order than when they have the SVO order. This supports the traditional analysis of Mayan word order: the syntactically simplest word order of transitive sentences in Mayan languages, including Kaqchikel, is VOS. More importantly, the results revealed that the subject-before-object word order preference in sentence comprehension, previously observed in other languages, might not reflect a universal aspect of human languages. Rather, processing preference may be language-specific to some extent, reflecting syntactic differences in individual languages.
Full-text available
The processing load of sentences with three different word orders (VOS, VSO, and SVO) in Kaqchikel Maya was investigated using a sentence-plausibility judgment task. The results showed that VOS sentences were processed faster than VSO and SVO sentences. This supports the traditional analysis in Mayan linguistics that the syntactically determined basic word order is VOS in Kaqchikel, as in many other Mayan languages. More importantly, the result revealed that the preference for subject-object word order in sentence comprehension observed in previous studies may not be universal; rather, the processing load in sentence comprehension is greatly affected by the syntactic nature of individual languages.*
Individual languages in the Mayan family display either rigid VSO or alternating VOS/VSO word orders (England 1991). In this article we review problems with previous accounts of Mayan word order and argue that verb-initial (V1) order is consistently derived by head movement of the verb to a position above the subject and below Infl⁰, which accounts for uniformity in verb-stem formation across the family. After an in-depth examination of the factors that have been reported to determine postverbal argument order, we present three distinct paths to VOS: (i) postsyntactic reordering of NP objects (following Clemens 2014, 2017), (ii) right-side subject topicalization (Can Pixabaj 2004, Curiel 2007), and (iii) heavy-NP shift (Larsen 1988). This account makes testable predictions in the domains of word order and prosodic constituency and has implications for the derivation of verb-initial order crosslinguistically.
In this book, Omer Preminger investigates how the obligatory nature of predicate–argument agreement is enforced by the grammar. Preminger argues that an empirically adequate theory of predicate–argument agreement requires recourse to an operation, whose obligatoriness is a grammatical primitive not reducible to representational properties, but whose successful culmination is not enforced by the grammar. Preminger's argument counters contemporary approaches that find the obligatoriness of predicate–argument agreement enforced through representational means. The most prominent of these is Chomsky's ‘interpretability’–based proposal, in which the obligatoriness of predicate–argument agreement is enforced through derivational time bombs. Preminger presents an empirical argument against contemporary approaches that seek to derive the obligatory nature of predicate–argument agreement exclusively from derivational time bombs. He offers instead an alternative account based on the notion of obligatory operations better suited to the facts. The crucial data involves utterances that inescapably involve attempted–but–failed agreement and are nonetheless fully grammatical. Preminger combines a detailed empirical investigation of agreement phenomena in the Kichean (Mayan) languages, Zulu (Bantu), Basque, Icelandic, and French with an extensive and rigorous theoretical exploration of the far–reaching consequences of these data. The result is a novel proposal that has profound implications for the formalism that the theory of grammar uses to derive obligatory processes and properties.