Natsuko Tsujimura

20 Syntax and argument structure

1 Introduction
Contrasted with adjuncts, arguments serve as obligatory and integral elements of
predicates. Their analysis has raised a number of typological and theoretical issues
in the linguistics literature, encompassing morphology, semantics, and syntax as
well as their interface areas. Although nouns, verbs, adjectives, and pre- and post-
positions can all take arguments, the literature on a predicate’s arguments and argu-
ment structure has focused primarily on arguments of verbs. As important as they
are, arguments of a predicate are not always straightforwardly identified in a given
language since the degree of obligatoriness of linguistic expressions varies across
languages. Particularly in languages like Japanese that allow an extensive use of
contextually recoverable zero pronouns (or noun ellipsis), a predicate’s arguments
are semantically present for interpretations but are not necessarily realized syntacti-
cally as overt noun phrases. In these languages, the identification of a predicate’s
arguments cannot rely solely on the obligatory presence of noun phrases in a surface
string of words. This challenge, to some degree, applies to English, which allows no,
or at least extremely limited use of, zero pronouns. Examples like I’ve already eaten
(lunch), Have you been drinking (alcoholic beverages) again?, and even This tiger
kills (people) when he is hungry illustrate the point. The definition of a predicate’s
arguments for languages like Japanese may well rely more on semantic measures,
such as Comrie’s (1993: 907): “Thus our overall definition of argument would be:
a phrase that is either obligatory given the choice of predicate, or whose meaning
is a function of that of the predicate, or whose behavior is parallel to argument so
defined” (emphasis added).
While a variety of assumptions have been made in organizing arguments of
predicates, what seems to be of general agreement is that argument structure specifies
the number and semantic types (or thematic roles) of the arguments with which a
verb has a strong association, syntactically or semantically, given Comrie’s definition
above. In some approaches, furthermore, structured ordering or hierarchical organi-
zation among the arguments has been claimed to explain various syntactic behavior.
(Bresnan and Kanerva 1989; Grimshaw 1990) The argument structures provided in
(1), indicated within parentheses, display some of the standard representations of
intransitive verbs (both unergative and unaccusative types), transitive verbs, and
ditransitive verbs, as well as ‘psych verbs’ that pattern similar to transitive verbs in
the number of arguments but differ from them in their semantic types.

DOI 10.1515/9781614516613-021
752 Natsuko Tsujimura

(1) a. intransitive – unergative: warau ‘laugh’ (agent)

Akanboo ga hazimete waratta.
<agent>
baby NOM for.the.first.time laughed
‘A baby laughed for the first time.’

b. intransitive – unaccusative: kowareru ‘break’ (theme)

Hurui tokei ga kowareta.
<theme>
old clock NOM broke
‘An old clock broke.’

c. transitive: kaku ‘write’ (agent, theme)

Gakusei ga ii ronbun o kaita.
<agent> <theme>
student NOM good article ACC wrote
‘A student wrote a good article.’

d. ditransitive: ageru ‘give’ (agent, goal, theme)

Hahaoya ga musume ni seetaa o ageta.
<agent> <goal> <theme>
mother NOM daughter DAT sweater ACC gave
‘A mother gave her daughter a sweater.’

e. “psych verb”: odoroku ‘be surprised’ (experiencer, theme/stimulus)

Ryokoosya ga natu no atusa ni odoroita.
<experiencer> <theme>
travelers NOM summer GEN heat at surprised
‘Travelers were surprised at the heat of the summer.’

The organization of argument structure, as exemplified in (1), is further motivated by

the crosslinguistic generalization that a verb’s arguments are systematically mapped
onto grammatical functions in syntactic configurations, and that the mapping is
mediated by linking rules. A linking rule connects each member of argument structure
and a syntactic position with regularity. In languages of the nominative-accusative
case pattern, typically agent is projected onto the subject position, theme onto the
object position, and so on; and they generally have their morphological correspon-
dence of Case. The systematic nature of the linking between thematic roles and
grammatical functions effectively accounts for the typological differences between
the nominative-type and the ergative-type across the world’s languages. It further
captures the differences within a single language between the two types of intransi-
tive sentences – one with unergative verbs and the other with unaccusative verbs –
while explaining the similar syntactic behavior of the sole argument of unaccusative
sentences and the object argument of transitive sentences. The generalization over
Syntax and argument structure 753

the systematic linking patterns has been dealt with in various syntactic theories –
ranging from the Universal Alignment Hypothesis in Relational Grammar (Perlmutter
and Postal 1984), the Uniformity of Theta Assignment Hypothesis in Government
and Binding Theory (Baker 1988), to the Lexical Mapping Principles in Lexical Func-
tional Grammar (Bresnan and Kanerva 1989) – although the specific mechanism of
linking rules may vary.
In describing the relationship between semantic representations of arguments
and their syntactic realization, thematic roles alone are often considered to be insuf-
ficient to account for the complexity of the relationship. Instead, it is discussed in
terms of a more elaborate structure that represents a verb’s lexical meaning through
lexical decomposition. In such cases, argument structure is understood to be linked
to a lexically decomposed structure that is known under terms like lexical conceptual
structure, lexical semantic structure, and event structure. (Hale and Keyser 1986a,
1986b, 1988; Jackendoff 1983, 1987, 1990; Kageyama 1996; Rappaport and Levin 1988)
Examples of lexical conceptual structures of the verbs in (1) are given in (2).

(2) a. conceptual structure: [x ACT]

|
argument structure: (agent)

b. conceptual structure: [BECOME [x <state>]]

|
argument structure: (theme)

c. conceptual structure: [[x ACT] CAUSE [BECOME [y <state>]]]

argument structure: (agent, theme)

d. conceptual structure: [[x ACT] CAUSE [BECOME [y HAVE z]]]

argument structure: (agent, goal, theme)

e. conceptual structure: [BECOME [x BE AT [PSYCH-STATE ABOUT y]]]1

argument structure: (experiencer, theme)

A number of issues have been addressed that directly and indirectly pertain to argu-
ment structure from descriptive, typological, and theoretical points of view (Alsina
2006; Bresnan and Kanerva 1989; Butt 2006; Comrie 1993; Duguine, Huidobro, and
Madariaga 2010; Goldberg 1995; Grimshaw 1990; Harley 2010; Levin and Rappaport
Hovav 1998, 2005; Sadler and Spencer 1998); and Japanese has contributed to the

1 I follow Kageyama’s (2007) notation of psych verbs.

754 Natsuko Tsujimura

exploration as is demonstrated by many published works. Touching on only a few of

those issues here obviously does not do justice to the breadth and depth of this in-
triguing topic that a wide gamut of investigations have brought to light particularly
over the past thirty-five years. Instead, the goal of this chapter is rather modest. In
section 2 I will briefly discuss past and ongoing studies by focusing on two areas of
argument structure research in which Japanese data have had significant bearing
and have contributed to our general understanding of argument structure and its
relation to syntax. The first of these deals with the process of Argument Transfer.
The phenomenon has been observed with the light verb construction in Japanese,
and a representative analysis outlined here supports a hierarchical organization
of argument structure. The second case involves different types of double object
verbs in Japanese that exhibit non-uniform patterns of syntactic realization of their
arguments. This line of research suggests that various patterns of linking between
argument structure and syntax, as observed within Japanese and in their cross-
linguistic differences, can be explained by lexical semantic properties of the predicate
as well as morphosyntax of the arguments. In section 3 we will turn our attention to
theoretical approaches to argument structure, and examine to what extent Japanese
data can serve as a testing ground for them. To this end I will discuss projectionist
and constructionist approaches as two prevalent ways of viewing argument structure.
They differ in answering the question of where in the grammar argument structure
should be captured. I consider potential advantages and challenges these contrastive
views posit vis-à-vis the Japanese data drawn from language acquisition and mimetic
verbs.

2 Past and ongoing contribution

To illustrate the relevance that Japanese data have brought to the research on argu-
ment structure and related matters, this section singles out two previous and ongoing
investigations: (i) the process of Argument Transfer and hierarchical organization of
argument structure; and (ii) double object verbs and an analogy to dative alternation.

2.1 Argument Transfer and hierarchical organization of

argument structure
In their work on the light verb construction, Grimshaw and Mester (1988) propose
a process they termed Argument Transfer that operates over two sets of argument
structures. This work spurred a series of subsequent investigations on argument
structure within the context of the light verb construction and beyond. (Miyagawa
1989; Tsujimura 1990; Uchida and Nakayama 1993) The term “light verb” is not
unique to Japanese, as is evidenced by discussions on the topic that precede or
Syntax and argument structure 755

coincide with Grimshaw and Mester’s analysis of the Japanese construction (e.g.
Jespersen 1954; Cattell 1984; Kearns 1998; to name a few). For instance, English
examples of light verbs appear in phrases like make a decision, take a break, give a
hug, do a trick, and have a drink, where the verbs make, take, give, do, and have are
considered “light” in their semantic contribution while the accompanying nouns
carry weight in interpreting what the phrases mean. In Japanese, the verb suru,
which is most customarily translated as ‘do’ in isolation, can form complex predicates
with lexemes of varying strata: (i) torihiki-suru (‘business’ [Yamato compound]-suru)
‘trade’, (ii) kikoku-suru (‘return to the home country’ [Sino-Japanese compound]-suru)
‘return to the home country’, (iii) deeto-suru (‘dating’ [English loanword]-suru) ‘(to)
date’, and (iv) tin-suru (‘sound of bell’ [mimetic]-suru) ‘(to) microwave’. These complex
predicates are similar to the English examples above in that their interpretations
primarily rely on the meaning of the lexeme to which suru is suffixed. Furthermore,
at least in the first three types of Japanese complex predicates with suru, the accom-
panying nouns with Accusative Case on them may be morphologically separated
from suru and serve as the verb’s complement. (3a) and (3b) are minimally different
in this respect without any significant difference in meaning.

(3) a. Musume ga dookyuusei to deeto-siteiru.

my.daughter NOM classmate with dating-suru
‘My daughter has been dating with her classmate.’

b. Musume ga dookyuusei to deeto o siteiru.

my.daughter NOM classmate with dating ACC suru

The difference between (3a) and (3b) does not merely mean that (3a) is a noun-
incorporation alternant of (3b). An interesting set of properties that is relevant to
argument structure is associated with the pattern in (3b), and it is more notably
demonstrated when more than one non-agent argument is involved (i.e. parallel to
a ditransitive predicate). (4a) presents such a case.

(4) a. Sityoo ga yuusyoosya ni hanataba o zootei-sita.

mayor NOM winner DAT bouquet ACC gift-suru.PST
‘The mayor gave the winner a bouquet of flowers.’

b. *Sityoo ga yuusyoosya ni hanataba o zootei o sita.

mayor NOM winner DAT bouquet ACC gift ACC suru.PST

b’. (Sono) hanataba wa sityoo ga yuusyoosya ni zootei o sita.

(that) bouquet TOP mayor NOM winner DAT gift ACC suru.PAST

c. Sityoo ga yuusyoosya ni hanataba no zootei o sita.

mayor NOM winner DAT bouquet GEN gift ACC suru.PST
756 Natsuko Tsujimura

d. *Yuusyoosya ni hanataba o sityoo no zootei o sita.

winner DAT bouquet ACC mayor GEN gift ACC suru.PST

e. *Sityoo ga yuusyoosya e no hanataba no zootei o sita.

mayor NOM winner to GEN bouquet GEN gift ACC suru.PST

f. *Sityoo ga hanataba o yuusyoosya e no zootei o sita.

mayor NOM bouquet ACC winner to GEN gift ACC suru.PST

Unlike the complex predicate in (4a), the verb in (4b) is the free-standing suru, but
the three event participants – sityoo ‘mayor’, yuusyoosya ‘winner’, and hanataba ‘a
bouquet of flowers’ – are entities required by the Sino-Japanese noun, zootei ‘gift’.
That is, the mayor is the agent, the winner is the recipient, and a bouquet of flowers
is the theme of the event of giving. In contrast, suru does not have any semantic or
thematic contribution, hence “light”. Grimshaw and Mester’s (1988) proposal attempts
to explain this difference between the predicative noun and the light verb by claiming
that the noun has an argument structure that consists of three arguments: agent, goal,
and theme. In contrast, the light verb suru has an empty argument structure, and the
arguments belonging to the noun can be moved to the light verb’s empty argument
structure by the process of Argument Transfer. The light verb, however, has the ability
to assign Accusative Case on its own, ensuring the case assignment on the predicative
noun, as in (3b) and (4b). The interaction of the two sets of argument structure in-
volved in (4b) mediated by Argument Transfer is illustrated in (5), following Grimshaw
and Mester’s notation.

(5) a. zootei (agent, goal, theme)

b. suru ( ) <acc>
c. zootei ( ) + suru (agent, goal, theme) <acc>

Once all the members of zootei’s argument structure are transferred to the argument
structure of suru, they are projected onto the syntax and each of the arguments
receives case assignment as verbal arguments: the agent argument, sityoo ‘mayor’
with Nominative, the goal argument, yuusyoosya ‘winner’ with Dative, and the theme
argument hanataba ‘bouquet’ with Accusative. At the same time, the light verb
assigns Accusative Case to the noun zootei ‘gift’. The ungrammatical status of (4b),
however, is attributed to an independent reason, i.e. the Double O Constraint
(Harada 1973; Poser 1981), which blocks two occurrences of Accusative Case –o
within a single clause. The effect of the Double O Constraint is shown by the contrast
between (4b) and (4b’). In (4b’) the theme argument, (sono) hanataba ‘(that) bouquet’
Syntax and argument structure 757

is topicalized with the Topic marker, –wa. Since there are no longer two NPs that are
both marked with –o, the sentence is acceptable (Sells 1989).2
Argument Transfer does not require that the predicative noun’s argument struc-
ture be emptied, with all the arguments transferred to the argument structure of the
light verb. (4c), for example, demonstrates a partial transfer in which the agent and
goal arguments, but not the theme argument, are transferred. The theme argument
remains in the argument structure of zootei ‘gift’, and it also remains within the
noun in its syntactic projection, as is reflected by the Genitive Case –no on hanataba
‘bouquet’. The schematic representations are given in (6).

(6) a. zootei (agent, goal, theme)

b. suru ( ) <acc>

c. zootei (theme) + suru (agent, goal) <acc>

Argument Transfer, however, imposes constraints on how many and which argu-
ments can and should be transferred. The agent argument is always required to
be transferred, as is illustrated in the ungrammatical (4d), where the agent, sityoo
‘mayor’, remains in the argument structure of zootei ‘gift’ and its syntactic projection
is marked with the Genitive Case –no. Furthermore, at least one more argument in
addition to the agent role must be transferred. (4e) is ungrammatical because only
the agent argument participates in Argument Transfer, resulting in the goal and
theme arguments being left in the argument structure of the predicative noun and
marked with the Genitive Case.3 (4b’) and (4c), on the other hand, follow the con-
straint: in (4b’) all of the noun’s arguments are completely transferred to the argu-
ment structure of the light verb; and in (4c) the agent and goal arguments stay in
the noun’s argument structure while the theme argument is transferred to suru’s
argument structure.
Finally, the ungrammaticality of (4f) addresses the issue of prominence among
the members of argument structure. (4f) follows the constraint that the agent and
one additional argument (theme) be transferred, but nevertheless the process yields

2 Sells (1989) demonstrates in an example similar to (4b’) that in addition to the topicalization of the
theme argument, placing the theme argument in a position that is not adjacent to the predicative
noun would also derive an acceptable sentence even though it would end up with two NPs with –
o. The example in (i) – a slightly modified sentence of (4b’) – indeed seems to be more acceptable
although it still violates the Double O Constraint.
(i) Hanataba o sityoo ga yuusyoosya ni zootei o sita.
bouquet ACC mayor NOM winner DAT gift ACC suru.PST
3 In (4e) the goal argument, yuusyoosya ‘winner’, is expected to appear as yuusyoosya-ni-no ‘winner-
DAT-GEN’, but the juxtaposition of –ni–no is independently excluded on morphological grounds.
Whenever the sequence of –ni–no is expected, it is invariably replaced by –e–no.
758 Natsuko Tsujimura

an ungrammatical outcome. The situation in (4f) contrasts with that in (4c) regarding
the condition on the number of arguments to be transferred. The contrast between (4c)
and (4f) led Grimshaw and Mester to argue that argument structure is hierarchically
organized, and Argument Transfer takes place in the order of prominence among
thematic roles. The hierarchical structure they propose is (Agent/Source (Goal
(Theme))), where the outermost – Agent/Source – is the highest in prominence, and
Argument Transfer is applied in the order of prominence.4 (4c) follows this order. On
the other hand, in (4f) the least prominent theme argument is transferred without
the goal argument also being transferred. Since goal is higher in prominence than
theme, the Argument Transfer in this sentence violates the hierarchical order among
the arguments.
The discussion of Argument Transfer in the light verb construction sheds impor-
tant light on research regarding the general architecture of argument structure as it
is relevant to the way in which a predicate’s arguments are projected onto syntax.
The arguments that belong to the predicative nouns in the light verb construc-
tion can have various syntactic realizations in a non-arbitrary manner, and the
systematic pattern reflects the degree of prominence among the arguments. The
hierarchical organization of arguments that encodes relative prominence among
them and the formal machinery of Argument Transfer together provide a mechanism
that accounts for which syntactic patterns of argument realization are possible and
which are not. The data drawn from the light verb construction account for the
specific phenomenon internal to Japanese. At the same time, they further confirm
that the hierarchically organized argument structure, rather than a linearly ordered
list of thematic roles, makes it possible for arguments to be projected onto syntax in
a systematic and predictable fashion.5

2.2 Double object verbs and analogy of the dative alternation

Double object verbs have crosslinguistically ignited a great deal of active discussion
on the way in which their arguments are realized in syntax. Double object verbs in
English, for one, are the subject of a rich body of literature from a variety of different
theoretical perspectives particularly in relation to the dative alternation (or dative
shift) as in (7).

4 In addition to the Argument Transfer phenomenon in the light verb construction in Japanese,
Grimshaw (1990) further supports the hierarchical organization of argument structure by drawing
on data from English compounds. The contrast between gift-giving to children and *child-giving of
gifts is claimed to follow the same hierarchy in that the theme is internal to the compound while
the goal is outside in gift-giving to children; and the reverse pattern in *child-giving of gifts leads to
an ungrammatical status.
5 A discussion of hierarchical organization of argument structure is also found in Ito (2007) in rela-
tion to ditransitive verbs in Japanese.
Syntax and argument structure 759

(7) a. John gave Mary an expensive box of chocolates on her birthday.

b. John gave an expensive box of chocolates to Mary on her birthday.

The verb give has agent, goal, and theme arguments, and the last two can appear
either in the double object construction in (7a) or in the PP dative construction as
in (7b). Some view the two constructions as being independently generated, bearing
a unique semantic property specific to each (e.g. Oehrle 1976), while others adopt a
derivational view in which one construction is derived from the other (e.g. Chomsky
1955; Bowers 1981; Dryer 1987).6 While Japanese apparently lacks a dative alterna-
tion of the same sort, straightforward translation equivalents of English double object
verbs like give seem to correspond to the two variants in (8a) and (8b), which appear
to be merely different in the order of the two NPs.

(8) a. Roozin ga kodomo ni omotya o ageta.

old man NOM child DAT toy ACC gave
‘An old man gave a child a toy.’

b. Roozin ga omotya o kodomo ni ageta.

old man NOM toy ACC child DAT gave

Hoji (1985), however, argues that (8a) – [subject-indirect object-direct object] – reflects
the basic word order, from which (8b) is derived by scrambling the direct object,
and that the indirect object (goal argument) and the direct object (theme argument)
are hierarchically asymmetrical in syntax. To this end he uses quantifier scope to
demonstrate the asymmetrical relations in (9a) and (9b).

(9) a. Roozin ga dareka ni dono omotya mo ageta.

old man NOM someone DAT every toy also gave
‘An old man gave someone every toy.’
some > every, *every > some

b. Roozin ga dono omotyai mo dareka ni ti ageta.

some > every, every > some

The quantifier scope data like those in (9) have contributed to the now standard
view that in the syntactic configuration, goal and theme arguments of double object
verbs are hierarchically ordered in such a way that the goal NP (indirect object)

6 Drawing on data that contain idioms, Larson (1988) demonstrates that some sentences in the
PP dative construction are not derived from the double object construction, while Bruening (2001),
Harley (1995, 2002), and Richards (2001) argue that some sentences in the double-object construc-
tion are not generated from the PP-dative construction.
760 Natsuko Tsujimura

asymmetrically c-commands the theme NP (direct object).7 (cf. Harada and Larson
2009)
A question has been asked more directly regarding whether two alternating
constructions parallel to those in (7) can be found in Japanese beyond the word
order variation, from both language specific and crosslinguistic perspectives (e.g.
Miyagawa 1994, 1997; Kishimoto 2001; Matsuoka 2003; Miyagawa and Tsujioka 2004;
Ito 2007; Miyagawa 2012). In considering this question, it is important to note that
not all the verbs that appear in the double object construction as in (7a) can allow
for the PP dative construction as in (7b), and vice versa, as has been extensively
surveyed in English (Levin 1993). A more fundamental question is to ask whether
Japanese equivalents of double object verbs employ a structural dichotomy similar
to the one in (7) that showcases an interesting array of characteristics concerning
the way in which argument structure is mapped onto syntax. To this end, detailed
examinations have been undertaken of the lexical semantic properties of individual
verbs and fine-grained semantic analyses of the “goal” argument that is accom-
panied by –ni in its syntactic realization. Due to space limitation, I will summarize
Kishimoto’s (2001) discussion, as his analysis provides arguably the most compre-
hensive investigation to illustrate how the lexical-semantic and syntactic properties
of a verb’s arguments interact in relation to double object verbs in Japanese.
Kishimoto (2001) shows that there are two patterns in which argument structure
of double object verbs can be mapped onto syntax. He further demonstrates that
the two patterns emerge from the difference in the core meaning of the verb and the
semantic nature of the arguments. The two patterns, taken from Kishimoto (2001: 51),
are schematized in (10) and (11): (10) can be analogized to the double object con-
struction and (11) to the PP dative construction in (7).

(10) lexical representation: [x causes y to possess z]

| | |
argument structure: causer recipient theme
| | |
grammatical category: subject indirect object direct object
| | |
case marking: nominative dative accusative

(11) lexical representation: [x causes y to move toward z]

|
argument structure: causer goal theme
| | |
grammatical category: subject to-dative direct object
| | |
case marking: nominative dative accusative

7 Subsequent discussion has been taken up regarding the syntactic structure and the nature of
scrambling that connects the direct object construction and the PP dative construction in Japanese.
See Miyagawa and Tsujioka (2004), Kishimoto (2008), and Tsujioka (2011) for details.
Syntax and argument structure 761

Verbs of change of possession such as ageru ‘give’, ataeru ‘give’, watasu “hand’, and
wariateru ‘assign’ take the pattern in (10), while verbs of transfer whose members
include okuru ‘send’, nageru ‘throw’, and hakobu ‘carry’ take the pattern in (11). Each
pattern of mapping is associated with a cluster of semantic, morphological, and
syntactic characteristics, distinguishing the ways in which arguments are realized
in syntax.8
First, the fundamental lexical difference between these two classes of verbs is
“change of possession” vs. “change of location” (for verbs of transfer) as is described
in the lexical representations of (10) and (11). This distinction leads to the more
precise identification of the argument that corresponds to NP-ni as “recipient”9 for
change of possession verbs and “goal” for change of location verbs, rather than
both uniformly subsumed under goal. This difference is reflected in the following
entailments.

(12) a. #John wa Mary ni hon o ataeta ga, Mary wa

John TOP Mary DAT book ACC give.PST but Mary TOP
mada moratteinai.
yet receive.NEG
‘John gave Mary a book, but Mary has not gotten it yet.’
(Kishimoto 2001:40)

b. John wa Mary ni tegami o okutta ga, Mary wa

John TOP Mary DAT letter ACC send.PST but Mary TOP
mada uketotteinai.
yet receive.NEG
‘John sent a letter to Mary, but Mary has not received it yet.’
(Kishimoto 2001:39)

Second, the change in location, which is the core lexical meaning of verbs of trans-
fer, suggests not only that the marking of –ni with the goal argument in (12b) is the
directional postposition, but also that it can be replaced by another directional post-
position like –e or –made. The substitution indeed derives the grammatical sentence
in (12b) (John wa Mary(no uti) e/made tegami o okutta ‘John sent Mary(’s home) a
letter’). However, the same outcome is not obtained in (12a) (*John wa Mary e/made

8 In addition to verbs of change of possession and verbs of transfer, Kishimoto further discusses a
third class of verbs, i.e. verbs of transaction such as uru ‘sell’, kasu ‘rent’, and harau ‘pay’. He dem-
onstrates that these verbs represent a “hybrid” type in that they denote both change of possession
and change of location, sharing the properties of change of possession verbs and those of transfer
verbs.
9 The recipient argument is further constrained such that it has to be animate.
762 Natsuko Tsujimura

hon o ataeta ‘John gave Mary a book’). This difference means that –ni in (12a) repre-
sents the structural Dative case rather than a directional postposition. Third, the two
non-subject arguments that are realized as Dative NP (Mary) and Accusative NP
(book) in the syntactic configuration of (12a) can both be passivized, while only the
argument that is projected as Accusative NP (letter) in (12b) can be passivized. This
contrast confirms that the two seemingly identical occurrences of –ni in (12) should
not be uniformly analyzed: it is the Dative case in (12a) and a postposition in (12b).
Furthermore, it suggests that in Japanese, verbs of change of possession take the
analogue of the double object construction on par with (7a), and verbs of transfer
are projected to a structure parallel to the PP dative construction, analogous to (7b).
The close examination of verbs such as those in (10–11) makes it clear that double
object verbs in Japanese cannot simply be put together as a uniform group. These
verbs demonstrate that the general linking rule mentioned in the introductory
section must be enriched far more than a single way of mapping thematic roles to
grammatical functions and to syntactic positions. Double object verbs in Japanese
are divided into at least two (and possibly more) types, and in order to elucidate
the specific nature of the linking, analyses must refer to a variety of factors. Included
in those factors are the contrast between the two lexical-semantic notions of change
of possession and change of location, fine-grained semantic characterizations of
broadly conceived goal arguments, and their morphological and syntactic repercus-
sions. All of these factors jointly play significant roles in systematically projecting
a predicate’s arguments onto syntax, as is demonstrated by double object verbs in
Japanese, English, and beyond.10 (Levin 2008, 2010; Rappaport Hovav and Levin
2008)

3 Projectionist vs. constructionist views of

argument structure
Much of the discussion on argument structure in the early generative literature was
built upon the core premise that lexical semantic properties of a verb play an impor-
tant role in the syntactic structure. Under this premise, it has been understood that
the relationship between the verb’s argument structure and the syntactic realization
of the verb’s arguments are constrained by principles like the Universal Theta Assign-
ment Hypothesis (Baker 1988). The emphasis on the relevance of verb meaning to

10 A series of discussions is found in Miyagawa and Tsujioka (2004), Kishimoto (2008), and Tsujioka
(2011), among others, in relation to the dative alternation counterpart in Japanese and its implications
to the debate over base-generation vs. scrambling. Drawing on a different set of ditransitive verbs,
Matsuoka (2003) discusses PASS-type verbs and SHOW-type verbs in Japanese and demonstrates
that they exhibit non-uniform semantic and syntactic properties regarding the dative argument.
Syntax and argument structure 763

the syntactic realization of arguments remains as an assumption underlying what

has come to be known as the lexicalist or projectionist view of argument structure.
(Rappaport Hovav and Levin 1998; Beavers 2010) In this approach, multiple syntactic
patterns in which a verb appears (e.g. the English verb sweep appearing in the intran-
sitive frame, in the transitive frame, in the caused change of location pattern with a
direct object and a PP complement, and in the resultative construction) are attributed
to multiple meanings of the verb itself, each of which receives a unique lexical repre-
sentation. The projectionist view is a predominant approach taken by researchers
who have examined various aspects of argument structure and its relation to syntax
in Japanese.
An alternative approach claims that multiple meanings do not emerge from the
lexical properties of the verb but that syntactic structures in which a verb and its
arguments appear take substantial responsibility for them.11 (Borer 2005a, b) This
view has been extensively elaborated on by those who argue that argument structure
should be considered a construction, which is taken as a form-meaning(-function)
pairing. (Goldberg 1995, 2006, 2013) According to the constructionists, specific syn-
tactic realizations of arguments are taken to be part of argument structure con-
structions. For instance, the double object construction in (7a) constitutes one of
such argument structure constructions that is represented by a pairing of the form,
{Subjx, V, Obj1y, Obj2z}, with the meaning of “actual or potential transfer,” more spe-
cifically “x intends to cause y to receive z” (Goldberg 1992, 2013).12
The projectionist and constructionist views have provided analytical tools by
which Japanese data are examined internal to the language and are interpreted for
their crosslinguistic and typological implications. There seem to be several areas of
investigations related to argument structure in Japanese that may offer a testing

11 To illustrate the relevance of syntactic structure, rather than the lexical properties of a verb, to
interpretation, Borer (2005a) claims that native speakers can assign flexible but reasonable interpre-
tations to seemingly uninterpretable sentences like (i) below such that the red under is the subject
that corresponds to the agent of an action and lunch is the direct object that refers to an entity that
is acted upon; and both are somehow relevant to the verb five even if the meaning of the verb is not
identified.
(i) The red under fived lunch.
There are experimental works that illustrate the same point, in support of the constructionist view
(e.g. Kaschak and Glenberg 2000; Kako 2006; Goldwater and Markman 2009).
12 We focus on the two approaches to argument structure, the projectionist view and the construc-
tionist view, but I might add another approach, the one from a functional point of view: Preferred
Argument Structure, as proposed by Du Bois (2003). Preferred Argument Structure seems to contribute
to our understanding of the architecture of argument structure, broadly conceived, by way of its rela-
tion to information structure. Preferred Argument Structure hypothesizes that “. . . in spontaneous
discourse, certain configurations of arguments are systematically preferred over other grammatically
possible alternatives” (Du Bois 2003:33). This functional approach has been adopted to analyze typo-
logical and language-specific patterns in which arguments are distributed or surface in discourse.
(Du Bois, Kumpf, and Ashby (eds.) 2003)
764 Natsuko Tsujimura

ground for these approaches. In this section I will focus on two such topics, which
may be interpreted to posit some challenges to each of the two views of argument
structure.

3.1 Acquisition research

One of the research areas to which the projectionist and constructionist approaches
to argument structure have significant implications is children’s acquisition of verbs.
The projectionist view is closely connected to the semantic bootstrapping hypothesis
while the constructionist approach underlies the syntactic bootstrapping hypothesis.
Semantic bootstrapping hypothesizes that the basic information children use in verb
learning is a verb’s semantic representation while syntactic bootstrapping claims
that such information comes from syntactic structure. Thus, the basic premises of
the semantic and syntactic bootstrapping hypotheses are respectively in line with
the projectionist and constructionist views of argument structure. A number of
experimental studies have been reported for English in support of each approach to
verb acquisition (e.g. Pinker 1987; Gropen et al. 1991; Fisher et al. 1994; Niagles
1990), and they are in turn used to argue for or against the two approaches to argu-
ment structure. The contrast between the projectionist-based semantic bootstrapping
and the constructionist-based syntactic bootstrapping has interesting implications
for Japanese.
An example of experiments that supports the semantic bootstrapping hypothesis
can be cited from Gropen et al. (1991). In English it has been noted that the verb pour
as in pour water in the glass and fill as in fill the glass with water have different
patterns in the syntactic realization of the verb’s argument: with pour the direct
object is “figure” while with fill, the direct object corresponds to “ground”. Gropen
et al. demonstrate that the two seemingly contrastive patterns of argument realiza-
tions in syntax that these verbs show can be explained by the uniform linking rule
in (13).

(13) Link the argument that is specified as “caused to change” in the main event of
a verb’s semantic representation to the grammatical object.
(Gropen et al. 1991: 159)

The argument identified as “caused to change” is an “affected” entity, and Gropen

et al. claim that the semantic notion of “affectedness” plays an important role in
determining the way in which the verb’s arguments are realized in syntax. In pour
water in the glass and fill the glass with water, water and the glass in the respective
phrases correspond to “caused to change”, and are singled out as the affected argu-
ment. By associating figure-object (water) and ground-object (the glass) with the
semantic notion of affectedness, two apparently different projection patterns are
Syntax and argument structure 765

accounted for by a unique linking rule. Experiments with adults and children that
Gropen et al. conducted using nonsense verbs report that (i) the goal of motion is
picked as the direct object when the goal undergoes change of state regardless of
whether the change is of shape, color, or fullness, and (ii) when change of state (as
an end result) and manner (as a means) jointly describe an event, the affected entity
is realized as the direct object. This result is consistent with the linking rule of (13)
and substantiates the view that syntactic realization of a verb’s arguments is guided
by its lexical properties. Note that this is consistent with the way in which the pro-
jectionists understand a verb’s arguments to be realized in syntax.
Fisher et al. (1994) test the validity of the syntactic bootstrapping hypothesis by
using a pairing of a transitive sentence and an intransitive sentence with nonsense
verbs. Under the construction-based acquisition view, it is predicted that sentences
in the transitive frame – e.g. The rabbit is zilking the duck – would be interpreted as
causative events while those in the intransitive frame – e.g. The duck is zilking –
would refer to non-causative events because the running assumption is that the
meaning of causation is encoded in the form of transitivity. Fisher et al. tested this
hypothesis with 3- and 4-year-old children. The subjects were shown several scenes
that are relevant to nonsense verbs, and the scenes were presented with intransitive
sentences – e.g. The rabbit is pilking – for one group of subjects and transitive
sentences – e.g. The rabbit is pilking the elephant – for another group. Both groups
then heard Look, pilking, followed by the question, What does pilking mean? The first
group of subjects interpreted pilk as similar to eat while the second group guessed
its meaning to be like feed. This suggests that the transitive frame is strongly correlated
with causative events while the intransitive syntactic pattern is tied to non-causative
events. The results of the experiment, thus, support the syntactic bootstrapping
hypothesis that syntax provides information crucial to the verb’s meaning. This
conclusion is further in line with the constructionist view that the surface form of
the verb’s arguments (and other constituents) is directly linked to the way in which
the verb is interpreted.
Acquisition data in Japanese shed an interesting light on the two bootstrapping
hypotheses and in turn on the two approaches to argument structure. (Tsujimura
2007) The construction-based syntactic bootstrapping assumes that the syntax of
the sentences that children hear from their caregivers, from which they deduce the
verbs’ meanings, provides complete information about the verbs’ arguments. This
assumption presents a challenge to languages like Japanese that are rich in noun
ellipsis. Based on his earlier work on child-directed caregivers’ sentences of the
transitive pattern [NP-ga NP-o V] (Rispoli 1991), Rispoli (1995) cites that out of 226
such transitive sentences, those with both arguments present in them along with
corresponding Case particles (i.e. the sentence type of (14i) below) amount to only
1%. A more comprehensive picture of the caregivers’ transitive sentences is given
in (14).
766 Natsuko Tsujimura

(14) sentence type frequency percentage of total

a. V 73 32
b. NP V 103 46
c. NP-o V 15 7
d. NP-ga V 8 4
e. NP-ga V (passive) 2 1
f. NP NP V 18 8
g. NP-ga NP 3 1
h. NP NP-o V 2 1
i. NP-ga NP-o V 2 1
(Rispoli 1995: 342)

The range of the syntactic patterns in which transitive verbs appear summarized in
(14) indicates that the two most frequent patterns in (14a–b) give the least amount of
formal information about the verb with respect to the number of its arguments. Note
also that these most ubiquitous patterns lack morphological Case marking on the
verb’s arguments, although the morphological information provides important cues
for relevant grammatical functions of the arguments. Interestingly, Rispoli notes that
children do not make errors in the number of NPs that are required by verbs despite
the syntactic indeterminacy of their input sentence patterns in (14), to which they are
exposed during the acquisition process. On the one hand, Japanese input sentences
with pervasive noun ellipsis and Case drop do not serve as a rich source for figuring
out the nature of argument structure and semantic relations among arguments. On
the other hand, despite the low degree of syntactic information, children seem to
understand the verb’s meaning correctly. These observations point to the conclusion
that Japanese children cannot resort to the structural properties for the acquisition
of verb meaning, and that the syntactic bootstrapping hypothesis and the construc-
tionist view of argument structure are problematic in explaining Japanese children’s
verb acquisition.
This conclusion, however, may not directly lend support to the semantic boot-
strapping hypothesis that shares with the projectionist view its premise on the
lexical properties of a verb as the guiding principle for the syntactic realization
of its arguments. Japanese appears to present a more intricate situation than what
experimental data in English demonstrate to argue for the semantic bootstrapping
hypothesis and the projectionist view. Recall that the projectionist approach to argu-
ment structure invokes the systematic mapping rule to account for the way in which
a verb’s arguments are projected onto syntax. An example illustrating such a linking
mechanism has been given in (10–11) for Japanese. As these two schemata show,
Case marking constitutes an important means to ensure that the thematic roles of
the verb’s arguments correspond to the appropriate grammatical functions. This is
because unlike English, the word order among arguments does not register gram-
matical functions, nor does every argument associated with the verb necessarily
Syntax and argument structure 767

appear overtly due to noun ellipsis. Researchers who have examined longitudinal
data of Japanese children’s verb acquisition often report children’s errors in Case
particles, but they explicitly and implicitly state that the errors do not immediately
suggest their lack of acquisition of the transitive vs. intransitive distinction of verbs.
(Clancy 1985; Rispoli 1987; Morikawa 1997) Examples that illustrate errors with Case
marking are given in (15–17); the correct Case is indicated in square brackets.

(15) Omizu *ga[o] ireta noni

water *NOM[ACC] put.in.PST but
‘she put in water, but. . .’
(Clancy 1985: 389)

(16) Akatyan *ga[o] turete iku.

baby *NOM[ACC] take.GER go
‘(We will) take the baby along.’
(Morikawa 1997: 56)

(17) Otootyan ga o *ga[o] kitta.

Dad NOM thong *NOM[ACC] cut.PST
‘Dad cut the thong.’
(Morikawa 1997: 57)

The verbs in these examples are transitive: in (15) and (16), the sole overt NPs, omizu
‘water’ and akatyan ‘baby’, are direct objects and are expected to be marked with
Accusative Case, –o, in adult language; in (17) otootyan ‘Dad’ is the subject and
correctly marked with Nominative Case, –ga, but the direct object o ‘thong’ is
incorrectly marked with Nominative Case, –ga. Given the expected linking of [agent-
subject-Nominative] and [theme-direct object-Accusative] for a transitive verb, the
type of errors demonstrated in (15–17) seems to indicate that the theme argument is
not properly projected onto syntax. That is, the linking problem that children have is
not with ALL of the verb’s arguments, but rather it appears to be centered on the
linking of theme, direct object, and the Accusative Case –o. If the semantic boot-
strapping hypothesis and the projectionist view of argument structure are at play,
children’s acquisition data would be only partly explained while remaining issues
would have yet to be accounted for. Japanese acquisition data of the sort discussed
above, thus, challenges both of the bootstrapping hypotheses and also both approaches
to argument structure.
Drawing on longitudinal data of two Japanese children, Rispoli (1987) observed
that the children rarely Case marked theme arguments of transitive verbs. Based on
this and also referring to Clancy’s example in (15), he suggests that these children
have not yet learned that direct objects are marked with –o as a morphological
768 Natsuko Tsujimura

reflection of syntactic projection because input data are not informative enough, as
(14) shows. Furthermore, Rispoli goes on to claim that the children nevertheless
know how to figure out the transitivity of verbs based not on morphosyntactic cues
of Case particles but on semantic and contextual information that includes animacy
of theme arguments and the planned nature of an action. For example, the two
children’s transitive sentences in his study are strongly tied to the presence of
inanimate theme arguments while their intransitive sentences tend to have their
sole arguments that refer to animates, animate surrogates, or vehicles. Of the total
of 40 transitive sentences that one child produced, 63% can be characterized by having
inanimate theme arguments in them. In contrast, of the total of 59 intransitive
sentences, 78% have the sole argument referring to an animate, animate surrogate,
or vehicle. Similarly, the other child produced 43 transitive sentences, of which 81%
had inanimate theme arguments; and out of his 85 intransitive sentences, 59% were
ones whose sole arguments were animate, animate surrogate, or vehicle. Rispoli’s
claim, supported also by Morikawa (1997), is that children are able to distinguish
between transitive and intransitive sentences and that they do so according to the
(inherent) semantic nature of the nouns, such as animacy, when they have not yet
acquired how to pair them up morphosyntactically with appropriate Case markers.
These discussions of acquisition data in Japanese make it clear that a comprehen-
sive account of Japanese children’s verb acquisition and issues related to argument
structure needs to take into consideration language-specific properties that may
interact directly or indirectly with lexical, morphological, and syntactic characteristics.
(Slobin 2001; Fausey, Long, and Boroditsky 2009) And, such an account may have to
incorporate additional (or alternative) factors that are needed to address language-
specific acquisition patterns. As illustrated earlier, research on children’s verb acqui-
sition, such as in English and many other languages, offers not only testing grounds
for acquisition theories like the bootstrapping hypotheses but also evaluative mea-
sures for the two approaches to argument structure under discussion. Japanese
acquisition data contribute to the current theoretical debate on argument structure
by adding a level of complexity to the range of phenomena being considered, result-
ing in a better understanding of the relation between argument structure and syntax.

3.2 Mimetic verbs

The projectionist approach to argument structure has been extensively adopted in
the large literature on argument realization and related issues as they pertain to the
Japanese data. One of the core assumptions that is prevalent is that a verb’s meaning
comprises two types, which Rappaport Hovav and Levin (1998) refer to as “struc-
tural” and “idiosyncratic”. They distinguish between the two meaning types as
follows: “The structural part of a verb’s meaning is that part which is relevant to
Syntax and argument structure 769

determining the semantic classes of verbs that are grammatically relevant, while the
idiosyncratic part of a verb’s meaning distinguishes that verb from other members of
the same class” (p. 106). The structural aspect is considered to be the properties over
which regularities and generalizations relevant to the architecture of argument
structure and to syntax are stated. Although approaches along these lines have
been effectively adopted for analyses of lexical verbs in Japanese, as I will discuss
below, these analyses may be challenged by verbs that are built around mimetic
lexemes.
Mimetics are ubiquitous in Japanese, and there has been an enormous amount
of literature that investigates them in virtually every sub-area of linguistics. (Kita
1997; Hamano 1998; Tamori and Schourup 1999; Nasu 2002; Akita 2009; to name
only a few) While investigations of the mimetic word class across languages have
often been restricted to language-specific descriptions, similarities to and differences
from other word strata in a given language have been increasingly re-evaluated in
recent years particularly with respect to their relevance to linguistic theories and
language typology. Focusing on the elusive nature in meaning and argument struc-
ture properties of mimetic verbs, Tsujimura (2005, 2007, 2010, 2014, 2017) claims that
mimetic verbs are different from non-mimetic verbs or prosaic verbs on several
grounds, and argues for the construction approach in analyzing mimetic verbs. I
will elaborate on the flexibility in meaning and in argument structure that mimetic
verbs exhibit, showing that the degree of flexibility that mimetic verbs demonstrate
is precisely the reason why they should be better approached from the construction
view.
What I refer to as mimetic verbs are verbs formed by combining mimetic bases
(or lexemes) with the light verb suru. There are countless such mimetic verbs.
Examples include tin-suru ‘(to) microwave’, sappari-suru ‘be refreshed’, bikkuri-suru
‘be surprised’, and burabura-suru ‘(to) stroll’. I have given what seems to be the most
typical and conventionalized “definitions” of these mimetic verbs. It is widely believed
that mimetic verbs have specific and fixed meanings assigned to them, and this asser-
tion is probably true to the extent that the “definitions” reflect the conventionalized
interpretations of given mimetic verbs. In what follows, however, I wish to demon-
strate that mimetic verbs are remarkably flexible in what they “mean” and in the
argument structure patterns in which they can appear, often far beyond what their
conventionalized uses allude to.
Let us take batabata-suru for our first example. Kakehi, Tamori, and Schourup
(1996: 46) give its “definition” as “to make a flapping or rattling noise”, although
the exact instantiations of a flapping or rattling noise can vary widely as long as
the image that is connected to such noise is somehow reflected in what is being
depicted, as (18–22) demonstrate. For our purpose, attention should be paid to the
range of argument structure patterns in these examples; relevant aspects of argu-
ments are annotated under each example.
770 Natsuko Tsujimura

(18) Ami ni haitta yatu [anago] wa[ga] . . . batabata-suru kara

net in enter.PST conger.eel TOP because
azi ga otimasu yo.
taste NOM decrease SFP
‘Conger eels in the net move in the bata-bata way, so they get less tasty.’
e intransitive frame
e subject-NOM: agent (=fish)
(Kakehi, Tamori, and Schourup 1996: 46)

(19) Mayakutyuudokusya ga kusuri ga kireruto

drug.addict NOM drug NOM run.out.when
batabata-suru no to onazi. . .
that as same
‘In the same way as a drug addict moves in the bata-bata way when
he runs out of the drug. . .’
e intransitive frame
e subject-NOM: agent (theme?) (=incognizant human being)
(Kakehi, Tamori, and Schourup 1996: 47)

(20) Naya no to wa[ga] . . . batabata-site, mimizawari da.

Barn NOM door TOP irritating COP
‘The barn door is moving noisily in the bata-bata way and is making
a rough (irritating) noise.’
e intransitive frame
e subject-NOM: theme (=door)
(Kakehi, Tamori, and Schourup 1996: 47)

(21) Musume wa teasi o ima made izyooni

daughter TOP arms.legs ACC now until more
batabata-suru yooni.narimasita.
has come to do come.to.PST
‘My daughter has been moving her arms and legs in the bata-bata way
more than ever.’
e transitive frame
e subject-NOM: agent (=human being)
e object-ACC: theme (=body parts)
Syntax and argument structure 771

(22) Rooka o batabata-suru.

hallway along
‘(He) moves along the hallway in the bata-bata way.’
e intransitive frame
e subject-NOM: agent (=human being)
e adjunct: path (=hallway)

(18–20) and (22) all take the intransitive frame with the subject marked with the
Nominative Case although the subject argument does not have to be overtly present
in syntax. The mimetic verbs in these sentences describe a manner of motion whose
image relates to a “flapping” or “rattling” noise in one way or another. The subject
arguments, however, are not uniform in their corresponding thematic roles. The
arguments that are realized as the subjects in (18) and (22) are agents. The same
may be said of (19) although the agent of the motion is not cognizant of the action.
The mimetic verb in these three sentences is arguably an intransitive verb of the
unergative type. (20) also takes the intransitive frame, but the sole argument realized
as the subject bears the theme role. The mimetic verb in (20) then is an unaccusative
verb. The unergative mimetic verb in (22), furthermore, is slightly different from
those in (18–19) in that the motion described in (22) assumes intended change of
location, as expressed by an adjunct referring to a path along which the motion
takes place. The mimetic verb in (21) appears in the transitive frame, describing a
caused motion. (18–22) show that a single mimetic verb can appear in the two types
of intransitive patterns – unergative and unaccusative – as well as in the transitive
frame. Batabata-suru is one of numerous mimetic verbs that are highly recognized
and widely used, and even with relatively conventionalized meaning as is demon-
strated in (18–22), the patterns in which arguments are realized in syntax take
different frames. Importantly, the extent to which the mimetic verb batabata-suru
flexibly appears in different argument realization patterns is not paralleled by prosaic
verbs. (Tsujimura 2017)
A similar observation can be made with mimetic verbs whose “meanings” are
often distant from their conventionalized ones and are innovatively assigned. This
is shown by a series of unconventional “meanings” for gatyagatya-suru in (23–26).
These examples and others akin to them have been found on the Internet and in
daily conversation. Kakehi, Tamori, and Schourup (1996: 353) do not independently
list the mimetic verb form of gatyagatya-suru but give the entry of the mimetic base
gatyagatya as “[t]he clattering or rattling sound made by relatively thin metallic or
ceramic objects knocking together repeatedly; the sound of an adding machine or
similar mechanical device”.

(23) Doanobu o gatyagatya-suru nante mottenohoka

door.knob ACC like out of question
‘It is out of the question to move a door knob in a gachagacha way.’
772 Natsuko Tsujimura

(24) XX-sensei to hutaride gatyagatya-siteitandesu kedo. . .

XX.prof. with two but
‘I was gachagacha-ing with Prof. XX, but. . .’

(25) Hai o otosu rebaa ga tuiteite, haiotosi no koto

ashes ACC drop lever NOM attached ash.dropping GEN that
o “gatyagatya-suru” tte ittemasita. Kodomo ni wa
ACC quotative is.called children to TOP
kekkoo gatyagatya-suru no ga omosirokattandesu.
very that NOM was.fun
‘It accompanies a lever with which you can let the ashes fall, and dropping the
ashes is called “doing gachagacha”. Doing gachagacha was fun for children.’

(26) Heya de gatyagatya no kapuseru o niyaniyasinagara

room in GEN capsule ACC grinning.while
aketeitara, haha ni “anata sono tosi de gatyagatya-suru
was.opening mother by you that age at
no” to akireraremasita.
Q QUOT was.amazed
‘With a smile on my face, I was opening the capsule of gacha-gacha in my
room, and then my mother was amazed, saying “you do gacha-gacha at
your age!”.’

In the set of examples above, only the instance of gatyagatya-suru in (23) appears in
the transitive frame, in which an unexpressed subject corresponds to the agent role
and the theme argument doanobu ‘door knob’ is realized as the object marked with
the Accusative Case –o. The mimetic verb in this example directly reflects Kakehi,
Tamori, and Schourup’s dictionary definition of the mimetic base: Somebody makes
a noise by repeatedly rattling the door knob, which is generally metallic. The same
mimetic verb in (24–26) takes the intransitive frame with the sole argument – agent –
interpreted as the subject (i.e. implicit in these examples). While these sentences
seem to be uniform in the argument realization pattern, the mimetic verb, which is
the common denominator among them, refers to distinct actions and states in each
case. (24) was uttered by my colleague in a daily conversation, describing a situation
in which she and her colleague, Professor XX, were taking care of odds-and-ends in
a somewhat disorganized manner and without following a planned course of action.
Some commonly used dictionaries such as Kōjien lists a disorganized state of affairs
as one of the descriptors of the mimetic base, gatyagatya. Such a description, how-
ever, is generally applied to inanimate objects, unlike (24) in which a somewhat
chaotic nature of the action by volitional and sentient human beings is referred to.
Syntax and argument structure 773

What the mimetic verb depicts in (25) is related to the conventionalized descrip-
tion given by Kakehi, Tamori, and Schourup but is extended innovatively to fit a
specific situation. As the example itself defines, the mimetic verb gatyagatya-suru
in (25) means to remove ashes from a coal stove by maneuvering the lever attached
to it. Given the conventional “definition”, we might expect the lever to be the theme,
or the ashes to be the theme while the lever to be the instrument. The uniqueness of
the meaning definition in (25) is that the potential arguments of the mimetic verb
based on our conventional knowledge of the mimetic verb are actually lexicalized
in such a way that they do not form independent memberships of its argument struc-
ture, except for the agent argument. In order to fully understand what the mimetic
verb refers to, including the membership of its argument structure, contextually rich
information needs to be invoked in these particular cases. Sensitivity to contextual
information and cultural knowledge that is relevant to the context is further illus-
trated by (26). The mimetic verb gatyagatya-suru in (26) can be considered a denominal
verb: It is connected to the mimetic base gatyagatya, which, as a noun, refers to a
type of small toy contained in transparent plastic capsules. These toys are generally
sold in vending machines for purchase, and they are removed from a vending
machine by maneuvering its handle, which makes a noise that is described as
gatyagatya. The mimetic verb in (26) means playing with these small toys called
gatyagatya or related activities that involve them. This example is similar to (25) in
that what would generally be conceived of as the theme argument is lexicalized in
the mimetic verb. The semantic contents of these two instances of the identical verb
are entirely different although the only thread that distantly connects the two is
the sound and image that the mimetic base gatyagatha invokes. These examples
are particularly instrumental in recognizing that the lexical content of a verb that is
relevant to the meaning is not always inherently determined but is governed by con-
texts that may call for knowledge of social background.13 Going back to the earlier
examples, the interpretations of the sentences in (18–22) that would correspond to
the projectionist’s structural meaning are determined not by the lexical semantic
properties that the mimetic verb batabata-suru inherently has (precisely because it
does not have any, beyond its idiosyncratic meaning) but by the constructions in
which batabata-suru appears. In (21), the transitive construction of the form {Subjx,
Objy, V} is linked to the interpretation of “x acts on y”, while in (22) the intransi-
tive motion construction of the form {Subjx, Obly, V} gives rise to the construal of
“x moves y”.
It seems that the situation illustrated by examples of mimetic verbs above raises
the fundamental question that Fillmore (1977a, b) discusses, namely, what Du Bois
(2003: 24) terms the question of “participant inclusion”: Which participant roles of
an event are selected to be regarded as arguments? Fillmore (1977a) explains that

13 This type of contextually determined semantic content is discussed in Tsujimura and Davis (2011)
for innovative verbs in Japanese, and applies to denominal verbs in general. (Clark and Clark 1979)
774 Natsuko Tsujimura

the four English verbs, buy, sell, spend, and cost all describe a commercial event, but
specific participants or a particular combination of participants are included in lin-
guistically relevant representations depending on which of them are in perspective.
For instance, when the seller and the goods are in perspective, sell is used, while
when the perspective of the goods and the money is taken, cost is used. This is why
Fillmore (1977b: 72) refers to meanings as being “relativized by scenes”. Contextually
determined semantic frames that include the information regarding which partici-
pant roles achieve the argument and co-argument status demonstrated in all the
examples of mimetic verbs above seem to cast a particular challenge to the projec-
tionist approach to argument structure. The lexical properties of mimetic verbs are
essentially different from those of prosaic verbs, and they provide the construction
view of argument structure (e.g. Tsujimura 2005, 2010, 2014, 2017) reasonable grounds
to serve as an alternative approach.
Kageyama (2007), however, argues against the construction treatment of mimetic
verbs and demonstrates that mimetic verbs and prosaic verbs should take a uniform
approach that follows the projectionist-based templatic analysis to explain a variety
of meanings that mimetic verbs can have. Following the lead of Rappaport Hovav
and Levin (1998), who present an architecture of lexical representation for multiple
meanings by using the mechanism called Template Augmentation, Kageyama claims
that a mimetic base and the light verb suru are each associated with independent
Lexical Conceptual Structures (LCSs), as in (27) and (28), and that the meaning of
a mimetic verb is generated by way of augmenting, or incorporating, the two sets
of LCSs, as in (29). An example of the process is illustrated by the derivation of
the mimetic verb, akuseku-suru ‘work hard’. The mimetic base akuseku ‘busily’ is
represented by the Type 1 LCS content in (27), and it is combined with the Type 1
LCS template in (28). The desired meaning representation is obtained with the argu-
ment projection pattern as an intransitive pattern of the unergative type, as in (30).

(27) Lexical Conceptual Structure (LCS) contents for mimetic base

Type 1: Manner α of action
Type 2: ON<Manner α> y
Type 3: x MOVE<Manner α> [Route]
Type 4: [EVENT BECOME [STATE x BE AT-[PSYCH.STATE<Manner α> ABOUT-z]]]
Type 5: [EVENT x’s BODY-PART MOVE<Manner α>]

Type 6: [EVENT y MOVE<Manner α>]

Type 7: [EVENT BECOME [y BE AT-[STATE<Manner α>]]]

(28) Lexical Conceptual Structure (LCS) templates for suru

Type 1: [EVENT x ACT]
Type 2: [EVENT x ACT]
Syntax and argument structure 775

Type 3: [EVENT x CONTROL[. . .]]

Type 4: [EVENT x EXPERIENCE[. . .]]
Type 5: [STATE x COGNIZE[EVENT. . .]]
Type 6: [STATE x COGNIZE[EVENT . . .]]
Type 7: [STATE x COGNIZE[EVENT . . .]]

(29) suru’s LCS template: [EVENT x ACT]

↑________ <Manner α>: LCS content of a mimetic
word

(30) a. LCS template of suru: [EVENT x ACT]

b. LCS content of akuseku: <Manner: BUSILY>
c. → Semantic incorporation akuseku-suru: [EVENT x ACT<BUSILY>]

The foundation of listing seven types of LCSs for mimetic bases and another seven
types of LCSs for suru results from his thorough survey of existing mimetic verbs and
their conventionalized meanings, as in (31).

(31) a. Group A (Agent/Experiencer subjects)

Type 1: [activity] (e.g. akuseku-suru ‘work hard’)
Type 2: [impact] (e.g. dondon-suru ‘bang (the wall)’)
Type 3: [manner-of-motion] (e.g. urouro-suru ‘wander aimlessly’)
Type 4: [psychological] (e.g. gakkari-suru ‘be surprised’)

b. Group B (Theme subjects)

Type 5: [physiological] (e.g. zukizuki-suru ‘throb (with pain)’)
Type 6: [physical perception] (e.g. guragura-suru ‘wobble’)
Type 7: [characterizing predication] (e.g. assari-suru ‘(taste) simple
and light’)

This is a very brief and somewhat abbreviated perusal of Kageyama’s projectionist-

based analysis, but it seems clear that the core of his argument is that the nature of
the semantic characterization and the representation does not motivate a mechanism
specific for mimetic verbs that is independent of the mechanism necessary for prosaic
verbs. Kageyama’s LCS-based analysis results from a meticulous examination of
so-called “meaning” of mimetic verbs as in (31), but it seems that central to his
characterization is a largely conventionalized or fossilized meaning of mimetic verbs
which corresponds to primary meanings listed in mimetic dictionaries and general
dictionaries. As the examples of mimetic verbs above demonstrate, there are mimetic
verbs (e.g. 18–23) whose meanings are well defined and are conventionalized, and
yet they can exhibit an innovative range of argument structure properties. There
776 Natsuko Tsujimura

are also mimetic verbs (e.g. 25–26) whose meaning definitions and structure of
core arguments may be understood by only a restricted part of the speech com-
munity but nevertheless achieve a remarkable degree of expressiveness without
hindering communication.
As long as the LCS templates and the mechanism of template augmentation are
used as descriptive tools, they may serve their purpose equally for prosaic verbs and
for mimetic verbs. However, the machinery of building verb meaning that is advanced
by some of the proponents of the projectionist view, such as Rappaport Hovav and
Levin (1998), has the intent of capturing much broader generalizations. As Rapparpot
Hovav and Levin argue, one such generalization is regular polysemy (Apresjan 1973):
A variety of syntactic patterns in which a verb can appear, as in (32), is observed
across the members of the same semantic class.

(32) a. Terry swept.

b. Terry swept the floor.
c. Terry swept the crumbs into the corner.
d. Terry swept the leaves off the sidewalk.
e. Terry swept the floor clean.
(Rappaport Hovav and Levin 1998: 97)

The verb sweep belongs to verbs of surface contact through motion, and this semantic
class includes wipe, scrape, dust, and rub as other members. The range of syntactic
patterns in (32) in which the verb sweep appears is available to other members of
the same verb class. As I have argued elsewhere (Tsujimura 2010, 2014, 2017), regular
polysemy is one of the major characteristics that significantly separates mimetic verbs
from prosaic verbs. The nature of polysemy that mimetic verbs exhibit, in fact, should
be better characterized as “irregular polysemy”. That is, the range of meaning exten-
sions and syntactic patterns of arguments that one member of a semantically (seem-
ingly) coherent class of mimetic verbs exhibit is not systematically available to other
members of the same class of mimetic verbs (Tsujimura 2014, 2017). Instead, multi-
ple “meanings” and argument structure properties that are associated with a
mimetic verb are specific to that particular mimetic verb. And there seem to be no
systematic patterns beyond individual mimetic verb that would lead to generaliza-
tions over the nature of the relationship between verb meaning and argument struc-
ture patterns. Thus, the type of regularities for which the projectionist approach is
motivated is lacking or meager in mimetic verbs. At the very least, the regularities
are not consistently observed with mimetic verbs to the same extent that prosaic
verbs demonstrate. On the one hand, the projectionist view has been influential in
analyzing prosaic verbs in Japanese. On the other, the degree of flexibility in argument
structure patterns of mimetic verbs and what could constitute their “meanings” may
Syntax and argument structure 777

shed interesting light on what argument structure really means for mimetic verbs.
Mimetic lexemes form an important member of the Japanese lexical strata, and verbs
that are built around them may well offer us a deeper insight into investigations of
argument structure in general.

4 Conclusion
The topic of argument structure appears in a wide range of linguistics literature, out
of which a number of important issues emerge with respect to its nature, both con-
tent and structure. The topic has been discussed both independently and in terms of
its relation to morphology and syntax. There have already been a number of interest-
ing observations made for Japanese from theoretical and comparative perspectives.
Obviously this chapter touches on only a very small part of this research topic.
Nevertheless, I hope to have shown through a brief sketch of argument structure
and related matters in Japanese that language specific characteristics, far beyond
translation equivalents, can contribute to the elucidation of some central factors
that need to be included in approaching general and broad issues revolving around
argument structure.

Acknowledgements
I would like to thank Shigeru Miyagawa and Stuart Davis for reading earlier versions
of this chapter and offering me helpful comments and suggestions. I would also like
to thank John Haig and Hideki Kishimoto whose editing has made the chapter more
readable. All remaining issues are my responsibility.

References
Akita, Kimi. 2009. A grammar of sound-symbolic words in Japanese: Theoretical approaches to
iconic and lexical properties of mimetics. Kobe: Kobe University dissertation.
Alsina, Alex. 2006. Argument structure. In Keith Brown (ed.), Encyclopedia of language and linguistics.
2nd edition, 461–468. Amsterdam: Elsevier.
Apresjan, Jurij D. 1973. Regular polysemy. Linguistics 142. 5–32.
Baker, Mark. 1988. Incorporation. Chicago: University of Chicago Press.
Beavers, John. 2010. The structure of lexical meaning: Why semantics really matters. Language 86.
821–864.
Borer, Hagit. 2005a. In name only: Structuring sense, volume I. Oxford: Oxford University Press.
Borer. Hagit. 2005b. The normal course of events: Structuring sense, volume II. Oxford: Oxford Uni-
versity Press.
778 Natsuko Tsujimura

Bowers, John. 1981. The theory of grammatical relations. Ithaca, NY: Cornell University Press.
Bresnan, Joan and Jonni Kanerva. 1989. Locative inversion in Chichewa: A case study of factorization
in grammar. Linguistic Inquiry 20. 1–50.
Bruening, Benjamin. 2001. QR obeys Superiority: frozen scope and ACD. Linguistic Inquiry 32. 233–
273.
Butt, Miriam. 2006. Theories of case. Cambridge: Cambridge University Press.
Cattell, Ray. 1984. Composite predicates in English. Sydney: Academic Press
Chomsky, Noam. 1955. The logical structure of linguistic theory. New York: Plenum.
Clancy, Patricia. 1985. The acquisition of Japanese. Hillsdale, NJ: Lawrence Erlbaum.
Clark, Eve and Herbert Clark. 1979. When nouns surface as verbs. Language 55. 767–811.
Comrie, Bernard. 1993. Argument structure. In Joachim Jacobs, Arnim von Stechow, Wolfgang
Sternefeld and Theo Vennemann (eds.), Syntax: An international handbook of contemporary
research, volume 1, 903–914. Berlin: Walter de Gruyter.
Dryer, Matthew. 1987. On primary objects, secondary objects, and antidative. Language 62. 808–
845.
Du Bois, John W. 2003. Argument structure: Grammar in use. In John W. Du Bois, Lorraine E. Kumpf
and William J. Ashby (eds.), Preferred argument structure: Grammar as architecture for function,
11–60. Amsterdam: John Benjamins.
Du Bois, John W., Lorraine E. Kumpf and William J. Ashby (eds). 2003. Preferred argument structure:
Grammar as architecture for function. Amsterdam: John Benjamins.
Duguine, Maia, Susana Huidobro and Nerea Madariaga (eds.). 2010. Argument structure and syntactic
relations. Amsterdam: John Benjamins.
Fausey, Caitlin M., Bria L. Long and Lera Boroditsky. 2009. The role of language in eye-witness
memory: Remembering who did it in English and Japanese. Proceedings of the 31st Annual
Meeting of the Cognitive Science Society, 2426–2431.
Fillmore, Charles. 1977a. Topics in lexical semantics. In Peter Cole (ed.), Current issues in linguistics
theory, 76–138. Bloomington: Indiana University Press.
Fillmore, Charles. 1977b. The case for case reopened. In Peter Cole (ed.), Syntax and semantics 8:
Grammatical relations, 59–81. New York: Academic Press.
Fisher, Cynthia, Geoffrey Hall, Susan Rakowitz and Lila Gleitman. 1994. When it is better to receive
than to give: Syntactic and conceptual constraints on vocabulary growth. Lingua 92. 333–375.
Goldberg, Adele. 1992. The inherent semantics of argument structure: The case of the English
ditransitive construction. Cognitive Linguistics 3. 37–74.
Goldberg, Adele. 1995. Constructions: A construction grammar approach to argument structure.
Chicago: University of Chicago Press.
Goldberg, Adele. 2006. Construction at work: The nature of generalization in language. Oxford:
Oxford University Press.
Goldberg, Adele. 2013. Argument structure constructions versus lexical rules or derivational verb
templates. Mind & Language 28. 435–465.
Goldwater, Micah B. and Arthur B. Markman. 2009. Constructional sources of implicit agents in
sentence comprehension. Cognitive Linguistics 20. 675–702.
Grimshaw, Jane. 1990. Argument structure. Cambridge, MA: MIT Press.
Grimshaw, Jane and Armin Mester. 1988. Light verbs and ɵ-marking. Linguistic Inquiry 19. 205–232.
Gropen, Jess, Steven Pinker, Michelle Hollander and Richard Goldberg. 1991. Affectedness and
direct objects: The role of lexical semantics in the acquisition of verb argument structure. In
Beth Levin and Steven Pinker (eds.), Lexical conceptual semantics, 153–195. Cambridge, MA:
Blackwell.
Hale, Ken and Samuel Jay Keyser. 1986a. Some transitivity alternations in English. Lexicon Project
Working Papers 7. Cambridge, MA: Center for Cognitive Science, MIT.
Syntax and argument structure 779

Hale, Ken and Samuel Jay Keyser. 1986b. A view from the middle. Lexicon Project Working Papers 10.
Cambridge, MA: Center for Cognitive Science, MIT.
Hale, Ken and Samuel Jay Keyser. 1988. Explaining and constraining the English middle. Lexicon
Project Working Papers 24. Cambridge, MA: Center for Cognitive Science, MIT.
Hamano, Shoko. 1998. The sound-symbolic system of Japanese. Stanford, CA: CSLI Publications.
Harada, Naomi and Richard Larson. 2009. Datives in Japanese. Proceedings of the 5th Workshop on
Altaic Formal Linguistics: MIT Working Papers in Linguistics 54, 3–17. Cambridge, MA: MITWPL.
Harada, S. I. 1973. Counter equi NP deletion. Annual Bulletin, Research Institute of Logopedics and
Phoniatrics 7. 113–147. Tokyo: University of Tokyo.Harley, Heidi. 1995. Subjects, events and
licensing. Cambridge; MA: MIT dissertation.
Harley, Heidi. 2002. Possession and the double-object construction. In Pierre Pica and Johan Rooryck
(eds.) Linguistic variation yearbook 2, 31–70. Amsterdam: John Benjamins.
Harley, Heidi. 2010. A minimalist approach to argument structure. In Cedric Boecks (ed.), The Oxford
handbook of linguistic minimalism, 427–448. Oxford: Oxford University Press.
Hoji, Hajime. 1985. Logical form constraints and configurational structures in Japanese. Seattle, WA:
University of Washington dissertation.
Ito, Atsushi. 2007. Argument structure of Japanese ditransitives. Nanzan Linguistics: Special Issue 3.
127–150.
Jackendoff, Ray. 1983. Semantics and cognition. Cambridge, MA: MIT Press.
Jackendoff, Ray. 1987. The status of thematic relations in linguistic theory. Linguistic Inquiry 18.
369–411.
Jackendoff, Ray. 1990. Semantic structure. Cambridge, MA: MIT Press.
Jespersen, Otto. 1954. A modern English grammar on historical principles. London: George Allen &
Unwin.
Kageyama, Taro. 1996. Dōshi imiron [Verbal semantics]. Tokyo: Kurosio Publishers.
Kageyama, Taro. 2007. Explorations in the conceptual semantics of mimetic verbs. In Bjarke Frellesvig,
Masayoshi Shibatani and John Smith (eds.), Current issues in the history and structure of
Japanese, 27–82. Tokyo: Kurosio Publishers.
Kakehi, Hisao, Ikuhiro Tamori and Lawrence Schourup. 1996. Dictionary of iconic expressions in
Japanese. Berlin: Mouton de Gruyter.
Kako, Edward. 2006. Thematic role properties of subjects and objects. Cognition 101. 1–42.
Kaschak, Michael P. and Arthur M. Glenberg. 2000. Constructing meaning: The role of affordances
and grammatical constructions in sentence comprehension. Journal of Memory and Language
43. 508–529.
Kearns, Kate. 1998. Extraction from make the claim constructions. Journal of Linguistics 34. 53–72.
Kishimoto, Hideki. 2001. The role of lexical meanings in argument encoding: Double object verbs in
Japanese. Gengo Kenkyu 120. 35–65.
Kishimoto, Hideki. 2008. Ditransitive idioms and argument structure. Journal of East Asian Linguistics
17. 141–179.
Kita, Sotaro. 1997. Two-dimensional semantic analysis of Japanese mimetics. Linguistics 35. 379–
415.
Larson, Richard. 1988. On the double object construction. Linguistic Inquiry 19. 335–391.
Levin, Beth. 1993. English verb classes and alternations: A preliminary investigation. Chicago: Uni-
versity of Chicago Press.
Levin, Beth. 2008. Dative verbs: A crosslinguistic perspective. Lingvisticæ Investigationes 31. 285–
312.
Levin, Beth. 2010. The semantic bases of Japanese and Korean ditransitives. Handout from the 20th
Japanese/Korean Linguistics Conference.
780 Natsuko Tsujimura

Levin, Beth and Malka Rappaport Hovav. 1998. Morphology and lexical semantics. In Andrew
Spencer and Arnold M. Zwicky (eds.), The handbook of morphology, 249–271. Oxford: Blackwell.
Levin, Beth and Malka Rappaport Hovav. 2005. Argument realization. Cambridge: Cambridge Univer-
sity Press.
Matsuoka, Mikinari. 2003. Two types of ditransitive constructions in Japanese. Journal of East Asian
Linguistics 12. 171–203.
Miyagawa, Shigeru. 1989. Light verbs and the ergative hypothesis. Linguistic Inquiry 20. 659–668.
Miyagawa, Shigeru. 1994. Scrambling as an obligatory movement. Proceedings of Nanzan University
International Symposium on Japanese Language Education and Japanese Language Studies,
81–92. Nagoya: Nanzan University.
Miyagawa, Shigeru. 1997. Against optional scrambling. Linguistic Inquiry 28. 1–26.
Miyagawa, Shigeru. 2012. Case, argument structure, and word order. London: Routledge.
Miyagawa, Shigeru and Takae Tsujioka. 2004. Argument structure and ditransitive verbs in Japanese.
Journal of East Asian Linguistics 13. 1–38.
Morikawa, Hiromi. 1997. Acquisition of case marking and argument structures in Japanese. Tokyo:
Kurosio Publishers.
Nasu, Akio. 2002. Nihongo onomatope no gokeisei to inritsu-kōzō [Word formation and prosodic
structure of Japanese onomatopoeia]. Ibaraki: University of Tsukuba dissertation.
Niagles, Letitia. 1990. Children use syntax to learn verb meaning. Journal of Child Language 17. 357–
374.
Oehrle, Richard. 1976. The grammatical status of the English dative alternation. Cambridge, MA: MIT
dissertation.
Perlmutter, David, and Paul Postal. 1984. The 1-advancement exclusiveness law. In David Perlmutter
and Carol Rosen (eds.), Studies in relational grammar 2, 81–125. Chicago: University of
Chicago Press.
Pinker, Steven. 1987. The bootstrapping problem in language acquisition. In Brian MacWhinny (ed.),
Mechanisms of language acquisition, 399–441. Hillsdale, NJ: Lawrence Erlbaum.
Poser, William. 1981. The “double-o constraint”: Evidence for a direct object relation in Japanese.
Unpublished ms., MIT.
Rappaport, Malka and Beth Levin. 1988. What to do with ɵ-roles. In Wendy Wilkins (ed.), Syntax and
semantics 21: Thematic relations, 147–165. New York: Academic Press.
Rappaport Hovav, Malka and Beth Levin. 1998. Building verb meanings. In Miriam Butt and Wilhelm
Geuder (eds.), The projection of arguments: Lexical and compositional factors, 97–134. Stanford,
CA: CSLI Publications.
Rappaport Hovav, Malka and Beth Levin. 2008. The English dative alternation: The case for verb
sensitivity. Journal of Linguistics 44. 129–167.
Richards, Norvin. 2001. An idiomatic argument for lexical decomposition. Linguistic Inquiry 32. 183–
192.
Rispoli, Matthew. 1987. The acquisition of transitive and intransitive action verb categories in
Japanese. First Language 7. 183–200.
Rispoli, Matthew. 1991. The acquisition of verb subcategorization in a functionalist framework. First
Language 11. 41–63.
Rispoli, Matthew. 1995. Missing argument and the acquisition of predicate meaning. In Michael
Tomasello and William E. Merriman (eds.), Beyond names for things: Young children’s acquisi-
tion of verbs, 331–352. Hillsdale, NJ: Lawrence Erlbaum.
Sadler, Louisa and Andrew Spencer. 1998. Morphology and argument structure. In Andrew Spencer
and Arnold M. Zwicky (eds.), The handbook of morphology, 206–236. Oxford: Blackwell.
Sells, Peter. 1989. More on light verbs and theta-marking. Unpublished ms., Stanford University.
Syntax and argument structure 781

Slobin, Dan I. 2001. Form-function relations: How do children find out what they are? In Melissa
Bowerman and Stephen C. Levinson (eds.), Language acquisition and conceptual development,
406–449. Cambridge: Cambridge University Press.
Tamori, Ikuhiro and Lawrence Schourup. 1999. Onomatope: Keitai to imi [Onomatopoeia: Form and
meaning]. Tokyo: Kurosio Publishers.
Tsujioka, Takae. 2011. Idioms, mixed marking in nominalization, and the base-generation hypothesis
for ditransitives in Japanese. Journal of East Asian Linguistics 20. 117–143.
Tsujimura, Natsuko. 1990. Ergativity of nouns and case assignment. Linguistic Inquiry 21. 277–287.
Tsujimura, Natsuko. 2005. A constructional approach to mimetic verbs. In Mirjam Fried and Hans C.
Boas (eds.), Grammatical construction: Back to the roots, 137–154. Amsterdam: John Benjamins.
Tsujiura, Natsuko. 2007. An introduction to Japanese linguistics, 2nd edition. Oxford: Blackwell.
Tsujimura, Natsuko. 2010. Onomatope dōshi no imi, kōkōzō no ichikōsatsu [On the meaning and
argument structure of mimetic verbs]. Proceedings of the Kansai Linguistic Society 29. 334–
343.
Tsujimura, Natsuko. 2014. Mimetic verbs and meaning. In Franz Rainer, Wolfgang U. Dressler, Hans
Christian Luschützky, and Francesco Gardani (eds.), Morphology and meaning, 303–314.
Amsterdam: John Benjamins.
Tsujimura, Natsuko. 2017. How flexible should the grammar of mimetics be? A view from Japanese
poetry. In Noriko Iwasaki, Peter Sells, and Kimi Akita (eds.), The grammar of Japanese mimetics:
Perspectives from structure, acquisition, and translation, 103–128. Oxford: Routledge.
Tsujimura, Natsuko and Stuart Davis. 2011. A construction approach to innovative verbs in Japanese.
Cognitive Linguistics 22. 797–823.
Uchida, Yoshiko and Mineharu Nakayama. 1993. Japanese verbal noun constructions. Linguistics 31.
623–666.