Transitivity alternation is a cross-linguistically recognized phenomenon in which a single verb undergoes a change in transitivity. It is part of diathesis alternations, a well-known verbal behavior which, by definition, involves shifts in the syntactic representations of arguments such as dative alternations and locative alternations (Levin 1993:2). While change in transitivity is crucially a syntactic phenomenon (i.e. change in valency of the verb) in English and Chinese, some languages like Japanese mark the difference in transitivity morphologically as well as syntactically (Teramura 1982:305). The present paper deals with the morphologically marked transitivity alternation in Japanese in comparison with the non-morphologically marked equivalent in English.

    The Japanese transitivity alternating verbs with overt morphological features (henceforth "morphological pairs") have drawn a great deal of attention of various researchers for centuries (cf. Shimada 1979). One main goal intended to be achieved by most of those researchers has been to classify the morphological pairs based on derivational patterns between alternants (Okutsu 1967). Most recently, Jacobsen (1992) conducted an extensive research, classifying 342 transitivity morphological pairs into sixteen categories according to suffix types and directionality of derivation. One of the conclusions drawn from his analysis of each class is that the derivational oppositions between morphological pairs are unproductive and, accordingly, each member of a pair needs to be memorized as a separate lexical item (1992:56).

    In this paper I will argue that such a lexicalist view is rather extreme. Instead, I will suggest that we focus more attention on the semantic relationships between morphological pairs. The primary objective of the paper is to conduct a semantic analysis of morphological pairs in Japanese, demonstrating that these pairs can be divided into two types. The first type involves the pairs whose semantic relationships are transparent, and the second involves the pairs whose semantic relationships are opaque. My argument is that the second type of morphological pairs are hardly said to be derivationally correlated in Modern Japanese due to their weaker semantic relationships. I will conclude that Japanese morphological pairs should be interpreted as being derived from a non-lexical operation when they maintain semantic correspondence.

    The organization of the paper is as follows. In section 2, I outline the general characteristics of the transitivity alternation. In section 3, I introduce two different--lexical and non-lexical--approaches to the English transitivity alternation. In section 4, I conduct a detailed morphological and syntactic analysis of Japanese morphological pairs. After suggesting a need for a comprehensive semantic analysis of Japanese morphological pairs in section 5, I point out a semantic discrepancies discernible in a number of morphological pairs in section 6. In section 7, I argue that semantic transparency between morphological alternants reflect a non-lexical derivational process in Japanese.

    One key feature of the transitivity alternation across languages is that the syntactic subject in the intransitive construction of the verb occurs as the syntactic object in the transitive construction of the verb, as illustrated below:

(1) a. Tom broke the vase.
b. The vase broke

The semantic characteristic of the transitive alternant in (1a), in contrast to the intransitive counterpart in (1b), is the addition of the abstract causative element, which is roughly represented as "cause to break-intransitive" (Levin 1993:2).¹

    The syntactic and semantic relationships between the alternants described above are also widely observed in the Japanese transitivity alternation:
 

(2) a. Tom-ga kabin-o wat-ta (< war-)
Tom-NOM vase-ACC break_vt-PAST

‘Tom broke the vase’

b. Kabin-ga ware-ta
vase-NOM break-PAST

‘The vase broke’

One additional feature of the Japanese transitivity alternation, however, is that a shift in transitivity is marked morphologically. That is, while transitive and intransitive alternants share a verbal stem, one or both of the alternants are overtly marked by suffix morphemes, resulting in distinct verbal forms as in war/ware in (2). Other morphological pairs point to this derivational characteristic of the Japanese transitivity alternation.
 

(3) TRANSITIVE     INTRANSTIVE     GLOSS

        or                     ore                            break
        sizume             sizum                         sink
        mage                magar                        bend
        kuzus               kuzure                       demolish/collapse

Given such an additional feature surrounding the morphological pairs, it is necessary to identify the implications, if any, that the derivational suffixes may have on the transitivity alternation in Japanese. Before addressing this issue, I will outline a frequently debated question about the transitivity alternation, that is; whether the alternation is a lexical process or a non-lexical process.

In generative theories of syntax, the distinction of transitivity between transitivity alternating verbs is represented in the lexicon. Under theoretical frameworks such as semantic selection and the projection principle, verbs participating in the transitivity alternation are provided with two lexical subentries, one for a transitive and the other for an intransitive (Keyser and Roeper 1984, Marantz 1984, Chomsky 1986, Spencer 1991, Manzini 1992, Schlesinger 1995, Cook 1996). With the projection principle, for instance, the alternating verb sink is provided with both transitive and intransitive lexical entries in the lexicon. In other words, the verb’s transitivity representations at the syntactic level as illustrated in (4) are interpreted as being projected from its lexical entries.
 

(4) a. The enemy sank the boat
b. The boat sank

    Under an analysis of lexical semantics, on the other hand, the transitivity alternation is viewed as a non-lexical process. Levin and Rappaport Hovav (1995) first postulate two lexical representations for transitivity alternating verbs, that is lexical semantic and lexical syntactic (or argument structure) representations. Under this analysis, the causative alternating verbs have a single causative lexical semantic representation, as illustrated below for the verb break (Levin and Rappaport Hovav 1995:83):
 

(5) break: [[x do-something] cause [y become BROKEN]]

Levin and Rappaport Hovav hold that the inchoative intransitive alternant which corresponds to the second part (‘y become BROKEN’) of the schematization above will obtain when the external cause x is not projected into argument structure, a process they refer to as detransitivization. The immediate cause of detransitivization has to do with the lexical binding of the external cause within the semantic lexical representation (Levin and Rappaport Hovav 1995:108). Levin and Rappaport Hovav do not clearly assert that detransitivation should occur in syntax. However, the dual levels of representations just described enable us to view the difference in transitivity as the result of derivational process, not as a mere reflection of inherent lexical entries.

    An approach adopted by Hale and Keyser (1993, 1997, 1998) to the issue of lexical versus non-lexical derivations with regards to transitivity alternation contrasts with the view held by generative grammarians described above. Under an argument structure approach, Hale and Keyser conduct a thorough analysis of denominal and de-adjectival verbs, maintaining that the process called conflation involved in the derivation of those verbs from adjectives and nouns, respectively, is a purely syntactic phenomenon. In their view, one characteristic of de-adjectival verbs like clear, distinct from denominal verbs like shelve, is that the former verbs undergo the transitivity alternation as shown below:
 

(6) a. They cleared the screen
b. The screen cleared

(7) a. They shelved the book
b. *The book shelved

Note that Hale and Keyser include a certain group of non-deadjectival lexical items in the group of de-adjectival verbs simply due to their ability to participate in the transitivity alternation. Thus, non-deadjectival alternating verbs such as break or melt are assumed to undergo the same syntactic process of conflation as de-adjectival verbs do (1998:84).

    In short, generative theorists assume in general that the transitivity of a verb should be specified in the lexicon, whereas the view shared by Levin and Rappaport Hovav and Hale and Keyser is that the transitivity alternation involves a non-lexical process. Hale and Keyser make clear their position that the alternation is strictly a syntactic phenomenon (Tyler 1999:114). I would argue for the non-lexicalist view of the English transitivity alternation in that it is syntactically represented in a strict sense and thus speakers of the language need to rely on syntactic configuration to determine whether a given alternant is transitive or intransitive. While the analysis just described might be proven to be viable in dealing with the English transitivity alternation, it remains to be seen whether the same approach is applicable to the Japanese transitivity alternation, which is morphologically, as well as syntactically, represented.

As mentioned earlier, Japanese morphological pairs are represented both morphologically and syntactically. In this section, I describe these two distinct aspects in more detail, particularly focusing on how they overlap with or/and complement each other in denoting the transitivity of each member of the morphological pairs.

4.1. Morphological Characteristics

Since the derivational patterns of the morphological pairs are distinguishable to a certain extent and are limited in number, one major objective of the literature on the Japanese morphological pairs has been to classify these pairs into distinct groups (cf. Okutsu 1967, Shimada 1979). In the most extensive classification in recent years, Jacobsen (1992) grouped a total of 341 morphological pairs into sixteen distinct classes.

    Given these classified morphological pairs, the immediate question that may arise is whether it is possible to associate particular suffix forms with transitivity; in other words, is it possible to predict whether each member of a given pair is transitive or intransitive merely by its suffix? If this is the case, then it follows that the significance of syntactic configuration will be enormously diminished in the domain of the morphological transitivity alternation in Japanese. This view would be further supported by the fact that the suffixes -se(ru), -os(u), and -as(u) always attach to verb stems to form transitive alternants (Jacobsen 1992:57).
 

(8) a. -se(ru):
abise "pour (over another)" (cf. abi "pour (over oneself)")
kise "put on (another’s) body" (cf. ki "put on (one’s own) body")
nekase "put to bed" (cf. ne "go to bed")

b. -os(u):
okos ‘wake up’ (cf. oki "get up_in")
modos "return_tr" (cf. modor "return_in")
yogos "soil" (cf. yogore "become dirty" )

c. -as(u):
hitas "soak _tr in" (cf. hitar "soak_in in")
mawas "turn_tr" (cf. mawar "turn_in")
kowas "break_tr" (cf. koware "break_in")

Table 1: Derivational patterns of Class I and Class 2 of Jacobsen (1992)
 
 

As evident in the above table, the derivational patterns of the two classes are diametrically opposed to each other. In Class I, the suffix pattern ø ~ e represents an intransitive-transitive relationship, whereas in Class II, the corresponding pattern represents a transitive-intransitive relationship.² In short, the dual functions of the suffix -e suggest that the transitivity of morphological pairs is not completely predictable from the form of a derivational suffix.³

That the syntactic configuration cannot be dispensed with in dealing with morphological pairs is further supported by the fact there are also transitive and intransitive alternating verbs in Japanese which do not mark the difference in transitivity morphologically at all. The difficulties of identifying the transitivity of alternants solely on the basis of suffix forms will be illustrated by the following examples:
 

(9) INTRANSITIVE         TRANSITIVE

    hirak ‘open’                 hirak ‘open’
    mas ‘increase’              mas ‘increase’
    syuuryoo-suru ‘end’     syuuryoo-suru ‘end’
    kaifuku-suru ‘recover’ kaifuku-suru ‘recover’
    kakudai-suru ‘expand’ kakudai-suru ‘expand’

As will be shown in the next section, the transitivity of these examples can only be denoted configurationally. In short, despite their distinct characteristics, Japanese morphological pairs cannot be fully accounted for solely from a morphological point of view. In what follows, we will see how transitive and intransitive alternants are syntactically represented in Japanese.

4.2. Syntactic Characteristics

Recall that one major syntactic characteristic of the transitivity alternation is that the object of a transitive predicate corresponds to the subject of an intransitive predicate (cf. section 2). It was demonstrated that in English, this configuration is denoted by the process of valency shift, as observed in (1a) and (1b) above, repeated as (10a) and (10b) below.⁴
 

(10) a. Tom broke the window
b. The window broke

The same configurational relationship also holds for Japanese morphological pairs, while the subject-object correlation is indicated by case-marking rather than word order. For instance, mado ‘window’ in (11a), which is followed by the accusative case marker -o, is marked by -ga in (11b), a nominative case marker.⁵
 

(11) a. Tom-ga mado-o wat-ta
Tom-NOM window-ACC break_tr-PAST

‘Tom broke the window’

b. Mado-ga ware-ta
window-NOM break_in-PAST

‘The window broke’

The syntactic relationship as demonstrated in (11), however, does not always obtain between transitive and intransitive constructions. For one thing, some -o marked objects occurring with transitive predicates fail to occur as subjects in the corresponding intransitive predicates.
 

(12) a. Dareka-ga watasi-no ude-o tukan-da
somebody-NOM I-GEN arm-ACC grab-PAST

‘Someone grabbed my arm’

b. *Watasi-no ude-ga dareka-ni tukamat-ta
I-GEN arm-NOM someone-to hold on to-PAST

‘My arm held on to somebody’)

b’. Dareka-ga watasi-no ude-ni tukamat-ta
somebody-NOM I-GEN arm-to hold on to-PAST

‘Somebody held on to my arm’                                                    (Teramura 1982:307)

While ude ‘arm’ is marked with -o when occurring with the transitive verb tukam ‘grab,’ the same noun is not marked with -ga but with the location marker -ni when occurring with the intransitive counterpart tukamar ‘hold on to.’ More importantly, as pointed out by Suga (1981) and Kanaya (2000), some intransitive members of morphological pairs can take the object case marker -o.
 

(13) a. Taroo-wa seki-o nankaimo kawat-ta (< kawar-; cf. kae ‘change_tr’)
Taro-TOP seat-ACC many times change_in-PAST

‘Taro changed seats many times’

b. Taroo-wa seki-o tat-ta (< tat; cf. tate ‘make stand’)
Taro-TOP seat-ACC stand up (and leave)-PAST

‘Taro stood up and left his seat"

c. Korede watakusi-no hanasi-o owari-masu (< owar-; cf. oe ‘end_tr’)
Now I-GEN talk-ACC end_in-POL

‘I end my talk now’

One problem arising from such anomalous configurations has to do with the question of unpredictability. In other words, merely by the forms of morphological suffixes there is no telling which morphological pairs will or will not follow the standard configurational pattern described above. In view of these facts, Shimada (1979) suggests that the configuration-based approach utilized for explicating the English transitivity alternation does not fully account for a range of problems posed by the Japanese transitivity alternation.

    To summarize, the morphological and syntactic approaches to morphological pairs outlined thus far reflect the complexities of the transitivity alternation in Japanese. In particular, the unproductive, as well as unpredictable, derivational morphology of this particular alternation led Jacobsen to conclude that "each member of a transitive/intransitive pair must therefore be learned as a separate lexical item" (1992:56).⁶ Furthermore, the idiosyncratic configurational behaviors involving morphological pairs as shown in (13) above seem to lend support to Jacobsen’s lexicalist view.

In the preceding section, we have observed that morphological oppositions, which characterize most of the Japanese transitivity alternating verbs, have encouraged a number of researchers to identify and classify such oppositions. As a consequence, in contrast to the non-lexical nature of the English transitivity alternation (see section 2 above), the Japanese transitivity alternation tends to be ascribed to a greater extent to the operation within the domain of the lexicon (cf. Jacobsen 1992).

    Notwithstanding a range of idiosyncratic relations between morphology and transitivity, I am reluctant to follow Jacobsen’s strong lexicalist view that all morphological pairs should be lexically listed. The rationale for my reservation is that there seems to be a line to be drawn across the entire classes of morphological pairs when these pairs are observed from a semantic point of view. Compare the two morphological pairs war/ware ‘break’ and ake/akas ‘dawn/spend (a night).’ As seen in section 4.2, the syntactic and semantic relationship between the first pair war/ware is fairly straightforward in spite of the presence of the unpredictable morphological suffix -e. By contrast, the pair ake/akas, whose morphological opposition and syntactic configuration are transparent (section 4.1 and 4.2), can hardly be considered to be correlated with each other because they have a weaker semantic link.
 

(14) a. Taroo-ga yo-o akas-ita
Taro-NOM night-ACC spend-PAST

‘Taro stayed up all night’

b. Yo-ga ake-ta
night-NOM dawn-PAST

‘The day dawned’

Thus, as an alternative view to Jacobsen’s, I would argue that morphological pairs which show semantic, as well as configurational, correspondence as illustrated by war/ware should not be considered "unproductive"?in Jacobsen’s terms?and, accordingly, need not be listed in the permanent lexicon. This non-lexical view, on the one hand, reflects the idea of diminishing the role of derivational morphology which has been strongly associated with Japanese morphological pairs. The same view, on the other hand, serves as shifting our attention to semantic relationships between transitive and intransitive alternants, an important area of research which, in my view, did not receive due attention in the past. In what follows, I will discuss semantic aspects of the transitivity alternation, illustrating the semantic discrepancies observed in several morphological pairs in Japanese.

The significance of semantic analysis of the transitivity alternation in Japanese has been noticed in the literature (Okutsu 1967, Nishio 1978, Hayatsu 1987, Jacobsen 1992, Kageyama 1996). Several studies based on lexical semantics, for instance, have demonstrated that the aforementioned suffix -e (see section 4.1 above) in Jacobsen’s Class I and II reflects different semantic features of verb stems to which it attaches. That is, the transitivizing suffix -e in Jacobsen’s Class I suggests that its verb stem involves a change caused by a volitional agent, whereas the intransitive suffix -e in Jacobsen’s Class II denote a self-induced change associated with the meaning of its verb stem (Jacobsen 1992, Kageyama 1996, Kanaya 2000). It is important to note that these studies attempt to explicate the morpho-semantic characteristics of derivational suffixes which presumably reflect lexical semantic characteristics of their verb stems, such as "volition" and "change of state."

    No extensive research, however, has ever been conducted on the semantic relationships between transitive and intransitive alternants of Japanese morphological pairs (cf. Teramura 1982:307, Hayatsu 1987:93, Morita 1994:160-1, Kageyama 1996:180-183). This strikes me as noteworthy considering the fact that there are many instances of morphological pairs which seem to show remote semantic relationships. In the list of morphological pairs given in Jacobsen’s (1992) appendix, for instance, it can be readily observed that in addition to the example given in (14) above, a number of other morphological pairs are semantically less related. Such semantic discrepancies are reflected in his translations:⁷
 

(15) INTRANSITIVE        TRANSITIIVE

    tar ‘suffice’                     tas ‘add, supplement’
    sutar ‘fall into disuse’     sute ‘throw away’
    hate ‘come to an end’     hatas ‘carry out’
    kom ‘become crowded’  kome ‘fill with’
    matagar ‘sit astride’         matag ‘straddle’

It is evident in his translations that he must have recognized semantic mismatches among a number of pairs. Nevertheless, Jacobsen makes no specific mention of this type of semantic discrepancy in his 1992 work.

Moreover, there are several other pairs in Jacobsen’s list which appear to have a weak semantic link while his translations indicate that they do. For instance, Jacobsen translates the pair mome/mom into ‘become wrinkled’ and ‘wrinkle,’ respectively, but I argue that the translations are inaccurate and that there is no apparent semantic connection between the pair, as illustrated below:
 

(16) a. Watasi-wa titioya-no kata-o mon-da
I-TOP father-GEN shoulder-ACC massage-PAST

‘I massaged my father’s shoulder’

b. Ano katei-wa itumo ironna koto-de mome-teiru
that family-TOP always various things-over have trouble-ASP

‘There is always trouble over many things in that family’

Interestingly, such semantically tenuous pairs tend to lack configurational relationships characteristic of the transitivity alternation (see section 4.2 above).
 

(17) a. Taroo-ga suupu-ni sio-o tasi-ta
Taro-NOM soup-to salt-ACC add-PAST
‘I added salt to the soup’

b. *Suupu-ni sio-ga tari-ta
soup-into salt-NOM suffice-PAST
(‘The salt in the soup sufficed’)

(18) a. Watasi-wa iranai hon-o sute-ta
I-TOP unnecessary book-ACC throw out-PAST
‘I threw out unnecessary books’

b. *Iranai hon-ga sutat-ta
unnecessary book-NOM fall into disuse-PAST
(‘Unnecessary books fell into disuse’)

(19) a. Taroo-wa sono yakusoku-o hatasi-ta
Taro-TOP the promise-ACC fulfill-PAST
‘Taro fulfilled his promise’

b. *Sono yakusoku-ga hate-ta
The promise-NOM come to an end-PAST
(‘The promise came to an end’)

(20) a. Watasi-wa kenzyuu-ni tama-o kome-ta
I-GEN gun-into bullet-ACC fill-PAST
‘I loaded the gun’

b. *Kenzyuu-ni tama-ga kon-da
gun-into bullet-NOM crowd-PAST
(‘The gun got loaded with bullets’)

b. *Watasi-no karada-ga matagat-ta
I-GEN body-NOM sit astride-PAST
(‘My body got sat astride’)

It is important to note that examples (17) — (21) do not imply that configurationally related morphological pairs are always semantically close-related: Even when the configurational correspondence obtains between transitive and intransitive alternants, some of the pairs simply lack a semantic correspondence.
 

(22) a. Kaigan-de natuyasumi-o sugosu
coast-on summer vacation-ACC spend
‘(I) spend my summer vacation on the coast’

b. Natuyasumi-ga sugiru
summer vacation-NOM pass
‘Summer vacation passes’ (Morita 1994:161)

(23) a. Taroo-wa migi asi-o itame-ta
Taro-TOP right leg-ACC hurt-PAST
‘Taro hurt his leg’

b. Taroo-no migi asi-ga itan-da
Taro-GEN right leg-NOM ached
‘Taro’s right leg ached’

Recall that the English transitivity alternation is essentially a meaning-preserving process (see section 3; cf. Haspelmath 1993:90). In general, the only semantic change involved in the process is the addition of the semantic primitive cause, if it is transitivization, to the inchoative intransitive (represented as become; see Guerssel et al. 1985) or the elimination of cause if it is a detransitivation process (Levin and Rappaport Hovav 1995).⁸ Rather, the English transitivity alternation is viewed strictly as a valency-changing process between a transitive and an intransitive. Specifically, the difference in transitivity is more likely to reflect the absence or presence of the argument Agent (Hale and Keyser 1986:2, Levin 1995:54).

In studies of verbs which participate in the causative alternation, Smith (1978) and Levin and Rapport Hovav (1995) note that a certain number of verbs of sound emission which appear to alternate in transitivity should not be identified with what they call "causative alternating" verbs like break and open.
 

(24) a. The baby burped
b. The nurse burped the baby (Smith 1978:107)

(25) a. The doorbell buzzed
b. The postman buzzed the doorbell (Levin and Rappaport Hovav 1995:115)

Levin and Rappaport Hovav consider these example "spurious" causative pairs, maintaining that they represent "two distinct verb meanings…that are not derivationally related" (1995:115). One reason for claiming that the causative pairs in (23) and (24) are spurious, according to Levin and Rappaport Hovav, is that the intransitive alternants will causativize "only for certain highly choices of objects for the transitive use" (1995:115).
 

(26) a. The doctor burped
b. *The nurse burped the doctor

(27) a. The bee buzzed
b. *The postman buzzed the bee

In the view of Levin and Rappaport Hovav, it is due to such idiosyncratic behaviors that spurious causative pairs like burp and buzz are assumed to have "two distinct semantic representations, not related by any productive rule" (1995:119). I would assume that the dual semantic representations suggest the possibility of listing causative and non-causative uses of some idiosyncratic causative pairs into two separate lexical entries.

    Following the contention by Levin and Rappaport Hovav, I would argue that the Japanese morphological pairs whose semantic relationships are understood to be weak are listed as separate lexical entries (cf. Aronoff 1976:19, Tyler 1999:8). Conversely, for those pairs such as ware/war ‘break’ which obey the standard configurational pattern and are semantically transparent, I would assume that only the verb stem (war-) is listed in the permanent lexicon in Japanese. On this view, the distinction between a transitive and intransitive is a corollary of morphological derivation and valency shift, both of which I consider to be part of a non-lexical process.⁹ In contrast, for semantically mismatched pairs as illustrated in section 6, I propose that they are listed as separate lexical items in the lexicon. In this view, not only idiosyncratic meanings but also transitivity status are listed as part of our lexical knowledge. The rationale behind setting up these two different levels of derivation is that it will help lessen the burden of our lexical memorization. As a consequence, the strong lexicalist view held by Jacobsen that all the morphological pairs are learned as separate lexical items will be reduced to the extent that only semantically tenuous pairs are listed separately in the lexicon.¹⁰

In this paper I have presented a semantic analysis of the transitivity alternation in Japanese. It was shown that a number of morphological pairs, while their valency shift patterns like that of the English transitivity alternation, show a varying degree of semantic discrepancies between transitive and intransitive alternants. Bearing in mind that such discrepancies may have resulted in part from semantic changes that have occurred to these pairs over time, I proposed that the morphological pairs whose semantic relationships are remote be listed as separate lexical items in the modern Japanese lexicon. This view is simply based on the idea that it is next to impossible for us to make a predication of the meaning of one member of a morphological pair idiosyncratically from the other member. Furthermore, for the morphological pairs which retain configurational and semantic coherence, I proposed to view them as having a verbal root in the lexicon. In this view, the distinction in transitivity result from a non-lexical (perhaps a syntactic) derivational process which might be comparable to the non-lexical processes?decausativiation under the analysis of Levin and Rappaport and conflation under the analysis of Hale and Keyser?applicable to the English transitivity alternation.

As a final remark, the analysis of the semantic mismatches seems to point to several other facts about the Japanese morphological pairs. First, there are hardly any such semantic discrepancies among the verbs in Class I. As noted earlier, this particular class of morphological pairs exhibit certain semantic features of verb stems such as "destruction" and "violence." The argument that change of state is the most significant semantic factor in determining whether a given verb will alternate in transitivity may be further supported by the semantic stability between the alternants as observed in Class I verbs. Secondly, Jacobsen’s Class XVI, which consist of the pairs with idiosyncratic morphological patterns, contains a relatively higher portion of semantically tenuous verbs. This may reflect the general view that morphological consistencies have the tendency to go hand in hand with semantic consistencies. The extreme case of this morpho-semantic relationship might be the verb pair sin/koros ‘die/kill,’ in which the two morphologically unrelated pair are understood to be completely separate lexical items despite the semantic affinity, that is "death."

*In the literal gloss, the following grammatical abbreviations are used:

ACC=accusative particle PAST=past tense marker

ASP=te-iru aspect marker POL=polite form

GEN=genitive TOP=topic particle

NOM=nominative particle
 
 

Aronoff, M. 1976. Word Formation in Generative Grammar. Cambridge, MA: MIT Press.
 
Chomsky, N. 1986. Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger.
 
Chomsky, N. 1995. The Minimalist Program. Cambridge, Mass.: The MIT Press.

Cook, V. C. and M. Newson. 1996. Chomsky's Universal Grammar: An Introduction. Oxford: Blackwell Publishers.

Hale, K. and J. Keyser.1986. Some Transitivity Alternations in English, MIT, Cambridge, Mass.
 
Hale, K. and J. Keyser.1987. A View from the Middle, MIT, Cambridge, Mass.
 
Hale, K. and J. Keyser. 1993. "On Argument Structure and the Lexical Expression of Syntactic Relations." In K. Hale and S. J. Keyser (eds.), The View from Building 20: Essays in Honor of Sylvain Bromberger, Cambridge, Massachusetts: The MIT Press, 53-109.
 
Hale, K. and J. Keyser. 1997. "On the Complex Nature of Simple Predicators." In A. Alsia, J. Bresnan and P. Sells (eds.), Complex Predicates, Stanford, California: CSLI, 29-65.
 
Hale, K. and J. Keyser. 1998. "The Basic Elements of Argument Structure." The MIT Working Papers in Linguistics 32: 73-118.

Haspelmath, M. 1993. "More on the Typology of Inchoative/causative Verb Alternations." In B. Comrie and M. Polinsky (eds.), Causatives and Transitives, Amsterdam: John Benjamins, 87-120.

Hayatsu, E. 1987. "Taioosuru tadoosi no aru zidoosi no imiteki-toogoteki tokutyoo (Characteristics of Intransitive verbs with corresponding transitive verbs)." Gengogaku Kenkyuu 6: 79-109.
 
Jacobsen, W. M. 1989. "Tadoosei to puroto-taipu ron (Transitivity and prototype)." In Kuno, S and M. Shibatani (eds.), Nihongo-gaku no shin-tenkai, Tokyo: Kurosio Publishers, [Reprinted in Suga & Hayatsu 1995a:166-178].
 
Jacobsen, W. M. 1992. The Transitive Structure of Events in Japanese. Tokyo: Kurosio Publishers.

Kageyama, T. 1996. Doosi imi ron (Verbal Semantics). Tokyo: Kurosio Shuppan.

Kanaya, T. 2000. "Doosi no zita-tairitu no keitairon-teki kanzen kaimei (Morphological analysis of Japanese transitivity alternation). ms., The Eighth Princeton Japanese Pedagogy workshop.
 
Keyser, J. and T. Roeper. 1984. "On the Middle and Ergative Constructions in English." Linguistic Inquiry 15: 381-416.

Kitagawa, C. and H. Fujii. 1999. "Transitivity Alternations in Japanese. Papers from the UPenn/MIT Roundtable on the Lexicon: 87-115.

Kuno, S. 1973. The Structure of the Japanese Languages. Cambridge, MA.: MIT Press.
 
Kuroda, S. Y. 1993. "Lexical and Productive Causatives in Japanese: An Examination of the Theory of Paradigmatic Structure." Journal of Japanese Linguistics 15: 1-81.
 
Lemmens, M. 1998. Lexical Perspectives on Transitivity and Ergativity. Amsterdam: John Benjamins Publishing Company.
 
Levin, B.1985. Lexical Semantics in Review: an Introduction, MIT: 1-62.

Levin, B. 1993. English Verb Class and Alternations: A Preliminary Investigation. Chicago: The University of Chicago Press.

Levin, B. 1995. "Approaches to Lexical Semantic Representation." In D.E. Walker et al. (eds.), Automating the Lexicon: Research and Practice in a Multilingual Environment, Oxford, Oxford University Press, 54-91

Levin, B. and M. R. Hovav. 1994. "A Preliminary Analysis of Causative Verbs in English." Lingua 92: 35-77.
 
Levin, B. and M. R. Hovav. 1995. Unaccusativity: At the Syntax-Lexical Semantics Interface. Cambridge, Mass.: The MIT Press.
 
Manzini, M. R. 1992. "The Projection Principle(s): A Reexamination." In I. M.I. Roca (ed.), Thematic Structure Its Role in Grammar, Berlin: Foris Publications, 271-291.
 
Marantz, A. 1984. On the Nature of Grammatical Relations. Massachusetts: The MIT Press.
 
Matsushita, D. 1923. "Doosi no zita-hisidoo no kenkyu (Study of transitivity/intransitivity and passivity/causativity of verbs)." Kokugakuin Zasshi 29-30: [Reprinted in Suga & Hayatsu 1995a:13-40].
 
Miyagawa, S. 1989. Syntax and Semantics 22: Structure and Case Marking in Japanese. New York: Academic Press.
 
Morita, Y. 1994. Doosi no imiron-teki bunpoo kenkyuu (The semantic and syntactic study of verbs). Tokyo: Meiji Shoin.
 
Nishio, T. 1954. "Doosi no hasei ni tuite (On the derivation of verbs)." Kokugogaku 17: [Reprinted in Suga & Hayatsu 1995a:41-57].
 
Nishio, T. 1978. "Zidoosi to tadoosi ni okeru imiyoohoo no taioo ni tuite (On a semantic relationship between intransitive and transitive verbs)." Kokugo to Kokubungaku 55: 173-186.
 
Nomura, T. 1982. "Zidoo tadoo ukemidoosi ni tuite (On intransitive, transitive, and passive verbs)." Nihongo Nihonbunka 11: [Reprinted in Suga & Hayatsu 1995a:137-150].
 
Okutsu, K. 1967. "Zidooka, tadooka, oyobi ryookyokuka tenkei (Patterns of intransitivization, transitivization, and bipolarization." Kokugogaku 70: 46-66.

Ootsuki, F. 1897. Koo nihon bunten (Japanese grammar). Tokyo: Benseisha.

Pinker, S. 1989. Learnability and Cognition: The Acquisition of Argument Structure.
Cambridge, Mass.: The MIT Press.
 
Schlesinger, I. M. 1995. Cognitive Space and Linguistic Case: Semantic and Syntactic Categories in English. Cambridge: Cambridge University Press.

Shibatani, M. 1990. The Languages of Japan. Cambridge: Cambridge University Press.
 
Smith, C. S. 1978. "Jespersen's 'Move and Change' Class and Causative Verbs in English." In H. A. Jazayery et al. (eds.), Linguistic and Literary Studies in Honor of Archibald A. Hill. Vol.2: Descriptive Linguistics, Mouton, The Hague, 101-9.
 
Spencer, A. 1991. Morphological Theory: An Introduction to Word Structure in Generative Grammar. Oxford: Blackwell.
 
Suga, K. 1981. "Zi-ta tigai: zidoosi to mokutekigo sosite zita no bunrui (Differences between transitives and intransitives: intransitive verbs and their objects, and the classification of transitives and intransitives)." In , Mabuchi Kazuo hakase taikan-kinen kokugaku ronzyuu, Tokyo: Taishukan Shoten, [Reprinted in Suga & Hayatsu 1995a:122-136].
 
Suga, K. and E. Hayatsu. 1995a. Dooshi no jita (Transitivity and intransitivity of verbs). Tokyo: Hitsuji Shoboo.

Suga, K. and E. Hayatsu. 1995b. "Doosi no zita o minaosu tameni (In order to re-examine the transitivity/intransitivity of verbs)." In Suga, K. and E. Hayatsu 1995a, 207-231.
 
Teramura, H. 1982. Nihongo no sintakusu to imi I (The syntax and semantics of Japanese Vol. I). Tokyo: Kurosio Publisher.
 
Tyler, L. J.1999. The Syntax and Semantics of Zero Verbs: A Minimalist Approach, Ph.D. thesis, University of Florida.
 
 

APPENDIX:

I. -e-/-ø-

INTRANSITIVE                             TRANSITIVE

hirakeru "open_in"                            hiraku "open_tr"

momeru "become wrinkled"             momu "wrinkle"

sireru "become known"                     siru "come to know"
 
 

II. -ø-/-e-

INTRANSITIVE                             TRANSITIVE

itamu "hurt_in"                                itameru "hurt_tr"

komu "become crowded"                 komeru "fill with"

mukau "face"                                    mukaeru "meet, welcome"

sou "go along with"                         soeru "add"

sukumu "crouch"                             sukumeru "duck (one’s head)"

tagau "differ"                                     tagaeru "break (one’s word)"

tigau "differ"                                     tigaeru "change"

tumu "become packed"                       tumeru "pack"

tutau "go along"                                 tutaeru "transmit"

yamu "stop_in"                                    yameru "stop_tr"

yasumu "rest_in"                                yasumeru "rest_tr"
 
 

III. -ar-/-e-

INTRANSITIVE                                 TRANSITIVE

azukaru "keep"                                     azukeru "entrust to"

kurumaru "become wrapped up in"     kurumeru "lump together with"

maziwaru "mingle with"                     mazieru "mix with"

osamaru "subside"                             osameru "pacify"

osowaru "learn"                                    osieru "teach"

sazukaru "receive"                                 sazukeru "grant"

sutaru "fall into disuse"                         suteru "throw away"

suwaru "site" sueru "set"

tazusawaru "participate in"                     tazusaeru "carry on one’s person"

ukaru "pass (an exam)"                         ukeru "take (an exam)"
 
 

IV -ar-/-ø-

INTRANSITIVE                                 TRANSITIVE

matagaru "sit astride"                         matagu "straddle"

tamawaru "be granted"                         tamau "grant"

tukamaru "be caught"                             tukamu "catch"
 
 

V. -r-/-s-

INTRANSITIVE                             TRANSITIVE

kitaru "come"                                     kitasu "bring about"

naru "become"                                     nasu "make"

noboru "rise"                                     nobosu "bring up, serve"

satoru "realize"                                 satosu "make realize"

taru "suffice"                                 tasu "add, supplement"
 
 

VI. -re-/-s-

INTRANSITIVE                             TRANSITIVE

kogareru "burn with passion for"     kogasu "scorch"

mureru "become steamed"                 musu "steam"
 
 

VII. -ri-/-s-

INTRANSITIVE                             TRANSITIVE

kariru "borrow"                             kasu "lend"

tariru "suffice"                                 tasu "add, supplement"
 
 

VIII. -ø-/-as-

INTRANSITIVE                             TRANSITIVE

aku "open_in"                                akasu "reveal"

au "go together"                             awasu "bring together"

hagemu "be diligent in"                 hagemasu "encourage"

huku "blow_in"                                hukasu "puff, smoke"

koru "become absorbed in"             korasu "concentrate_tr on"

meguru "come around"                 megurasu "turn around"

wazurau "be troubled"                     wazurawasu "trouble"
 
 

IX. -e-/-as-

INTRANSITIVE                         TRANSITIVE

akeru "dawn"                             akasu "spend (the night)"

bakeru "turn into"                      bakasu "bewitch"

hateru "come to an end"             hatasu "carry out"

hukeru "grow late"                     hukasu "stay up late at (night)"

kakeru "become lacking"             kakasu "miss (a meeting)"

kureru "(day, year) comes to an end" kurasu "pass (time)"
 
 
 
 

X. -i-/as-

INTRANSITIVE                         TRANSITIVE

dekiru "come into existence"     dekasu "bring about"

koriru "learn (from experience)" korasu "give (one) a lesson"

toziru "close_in"                            ozasu "close_tr"
 
 

XI. -i-/-os-

INTRANSITIVE                         TRANSITIVE

hiru "become dry"                     hosu "dry"

sugiru "go past"                         sugosu "pass (time)"
 
 

XII. -ø-/-se-

INTRANSITIVE                         TRANSITIVE

miru "see"                                 miseru "show"
 
 

XIII. -e-/-akas-

INTRANSITIVE                         TRANSITIVE

amaeru "act dependent (on)"         amayakasu "spoil"

hagureru "stray from"                     hagurakasu "put off, evade"

obieru "become frightened at" obiyakasu "frighten, threaten"
 
 

XIV. -or-/-e-

INTRANSITIVE                         TRANSITIVE

komoru "be fully present"             komeru "fill with"
 
 

XV. -are-/-e-

INTRANSITIVE                     TRANSITIVE

sutareru "fall into disuse"         suteru "throw away"
 
 

XVI. Miscellaneous affix pairs not fitting the above patterns

INTRANSITIVE                     TRANSITIVE

hosoru "become thin"             hosomeru "make narrow"

kakeru "run"                             karu "drive, spur"

kikoeru "become audible"         kiku "hear"

kudaru "go down"                     kudasaru "bestow"

nobiru "become extended"         noberu "extend"

tukiru "run out_in"                    tukusu "use up"

tumoru "become accumulated" tumu "accumulate"

useru "disappear"                        usinau "lose"