Resultative constructions with "implied-result" and "entailed-result" verbs in Thai and English: a contrastive study.
This study aims to make a contrastive investigation of the Thai and English transitive-based resultative constructions which consist of a causative predicate indicated by a transitive verb and a resultative predicate linguistically realized as a verb in Thai and an adjective in English. The resultative constructions which are the object of study in this article are those in which the causative predicate is manifested by two kinds of transitive verbs postulated in this article, namely, "'implied-result verbs" and "entailed-result verbs". This article examines one of the syntactic-semantic behavioral properties of both types of verb in both languages when they co-occur with resultative predicates. It is found in this article that English implied-result and entailed-result verbs are much more restricted in taking resultative predicates than the Thai counterparts. In other words, Thai resultative constructions are more productive than English ones. The productivity in the case of the former is attributed to the aspectual profile shift operating in serial verb constructions in Thai. Thai resultative constructions are arguably an instantiation of serial verb constructions. Resultative constructions in Thai allow both conventionalized as well as unconventionalized scenes to be expressed. In contrast, only conventionalized scenes can be expressed by resultative constructions in English.
There is indeed no single cross-linguistically accepted definition of the term "resultative," and different and conflicting ranges of examples are cited in the literature as "resultative." This article will therefore adopt a working semantic, theoretically neutral definition of the term "resultative". The term "resultative" can be semantically defined in a broad way as a term which refers to linguistic forms that express a state and a previous event. The resultative situation may be linguistically realized by many types of linguistic forms across languages, such as single verbs, compound verbs, two serialized predicates without any intervening linker and two predicates with an intervening linker. It is thus apparent that the resultative meaning may reside in a single verb and may also be expressed by a syntactic construction. Note that these types of resultative forms differ from one another in the degree of "lexicality" and "syntacticality" which they exhibit. The form which is the most lexical or the least syntactic is single verbs, which typically express a resulting state while implying a causing action, such as broken in The vase is broken, or mended in The shoes are mended in English. On the other hand, the form which is the least lexical or the most syntactic is two predicates with an intervening linker such as He beat a snake until it was dead in English.
The resultative forms which are intermediate between the two extremes are compound verbs and two serialized predicates without an intervening linker. The former type is more lexical than the latter. In the latter form, the causative predicate can be instantiated by either a transitive or an intransitive verb referred to in this article as transitive-based and intransitive-based resultative constructions, respectively. Resultative compound verbs are found to be prevalent in Chinese as in la-kai 'pull-open,' sha-si 'kill-die,' and da-sui 'strike-be in pieces' (Thompson 1973). Examples (1) and (2a)-(2c) below illustrate the intransitive-based and transitive-based resultative constructions without an intervening linker in
(1) nok naaw taay yuu khaanncck
bird feel cold die, dead be located at outside
'A bird froze to death outside.'
(2) a. khaw chok pwan lom
he punch friend collapse
'He punched his friend and he collapsed.'
b. khaw sak swa sa?aat
he wash shirt clean
'He washed his shirt clean.'
c. khaw khaa phuuraay taay
he kill criminal die, dead
'He killed the criminal (and the criminal died as a result)' (1)
There are also two types of resultative construction in English, namely, intransitive and transitive. Examples (3a)-(3e), in which the causative predicates are realized as intransitive verbs, illustrate intransitive-based resultative constructions in English. Examples (3a)-(3b) are different from (3c)-(3e) in that the causative predicates in the latter are followed by noun phrases whereas those in (3a)-(3b) are not. However, these noun phrases are not subcategorized arguments of the intransitive verbs functioning as the causative predicates.
(3) a. The river froze solid. (Napoli 1992: 66)
b. The clothes steamed dry. (Levin and Rappaport Hovav 1999: 206)
c. The boy cried himself sick. (Napoli 1992: 60)
d. The joggers ran their Nikes threadbare. (Carrier and Randall 1992: 173)
e. The kids laughed themselves into a frenzy. (Carrier and Randall 1992: 173)
Sentences (4a)-(4c) illustrate the transitive-based resultative construction in which the resultative predicates are realized as adjectives in (4a)-(4b) and as a prepositional phrase in (4c).
(4) a. The gardener watered the tulips flat. (Carrier and Randall 1992: 173)
b. I painted the car yellow. (Napoli 1992: 56)
c. The grocer ground the coffee beans (in)to a fine powder. (Carrier and Randall 1992: 173)
This study aims at contrastively investigating the transitive-based resultative constructions in Thai and English in which the causative predicate is manifested by two types of transitive verb which are postulated in this article, namely, "implied-result verbs" and "entailed-result verbs." Specifically, this article aims to identify constraints on co-occurrences between causative and resultative predicates in the transitive-based resultative constructions in the two languages. The similarities and differences in constraints on the co-occurrences between the two predicates between the two languages will be accounted for in semantic and functional terms.
In Section 2 of this article, the implied-result and entailed-result verbs will be characterized. In Section 3, findings regarding constraints on the co-occurrences between causative and resultative predicates in English and Thai will be presented and accounted for in semantic and functional terms. In Section 4, we review previous studies which accounted for co-occurrence constraints in different languages. We propose another perspective to account for such constraints in Section 5. Section 6 concludes the article.
2. Implied-result verbs and entailed-result verbs
The temas "implied-result" and "entailed-result" verbs which are postulated in this article are taken and modified from the terms "implied-fulfillment" and "attained-fulfillment" verbs postulated by Talmy (2000: 262-263) for reasons to be elaborated later. In setting forth the theory of realization, which refers to "an event of fulfillment or confirmation in realizing the agent's intention or goal in carrying out an action", Talmy postulates four verbal patterns, namely, (i) intrinsic-fulfillment verbs co-occurring with a further event satellite (2), (ii) moot-fulfillment verbs co-occurring with a fulfillment satellite, (iii) implied-fulfillment verbs co-occurring with a confirmation satellite, and (iv) attained-fulfillment verbs co-occurring with a pleonastic satellite. Talmy's four verbal patterns are largely based on English verbs. The main verbs in the four verbal patterns exhibit varying degrees of realization of the agent's intention. In the intrinsic-fulfillment verb pattern, the agent's goal in carrying out the action referred to by the verb is exactly fulfilled by the action itself. It does not extend beyond the action. An example of this type of verb is kicked. A "further-event satellite" can be added to ah intrinsic-fulfillment verb to denote a meaning that is extrinsic to the meaning referred to by the verb as in flat in I kicked the hubcapflat (Talmy 2000: 262). In the moot-fulfillment verb pattern, the agent further intends that the action lead to a particular result. However, based on the referential content of the verb, the fulfillment of the agent's intended result is left moot or questionable without a satellite. An example given by Talmy to illustrate this type of verbal pattern is The police hunted the fugitive down (Talmy 2000: 262). Since this study is concerned with those verbs which implicate and entail that certain resulting events will take place after the performance of the actions named by the verbs, only the other two types of verb postulated by Talmy are relevant to this study, namely, the implied-fulfillment verb and the attained-fulfillment verb. These two types of verbs are described below.
(a) Implied-fulfillment verb + confirmation satellite
The implied-fulfillment verb also consists of two main components, namely, (1) the agent's intended and executed action, and (2) the agent's further intention that this action lead to a particular desired result. However, the implied-fulfillment verb conveys the implicature that the agent's goal to bring about a certain result has been fulfilled. Since the agent's goal is merely an implicature, this reading is defeasible or cancelable by a disclaiming phrase as in I washed the shirt but it carne out dirty (Talmy 2000: 265). The addition of a satellite confirms what is otherwise only implied. In I washed the shirt clean, the satellite clean confirms that the implicature of the shirt's becoming clean has been fulfilled. Consequently, the implied-fulfillment verb accompanied by a satellite cannot co-occur with any disclaiming phrase as in * I washed the shirt clean but it carne out dirty.
(b) Attained-fulfillment verb (+ pleonastic satellite)
Like the other three types of verb mentioned above, the attained-fulfillment verb consists of two major components, i.e., (1) the agent's intended and executed action, and (2) the agent's further intention that this action lead to a particular desired result. However, it indicates the actual fulfillment of the agent's intention. The attained-fulfillment verb cannot be accompanied by a satellite to indicate the realization of the agent's intention; otherwise it would result in redundancy. An example of this type of verb given by Talmy is the transitive verb drown, which cannot be accompanied by the satellite dead indicating the agent's intention in executing the action of submerging an animate being in liquid. Thus, the sentence * I drowned him dead is unacceptable.
It is obvious that Talmy's verbal patterns in (a) and (b) above correspond to transitive-based resultative constructions examined in this study. The transitive verbs in Talmy's verbal patterns correspond to causative predicates and Talmy's satellite to resultative predicates in our terms. Thus, our study borrows many notions and analyses from his framework. However, it should be noted at the outset that our approach differs from his from one fundamental standpoint. We argue that the notion of the agent's intention is not crucial at all in the domain of realization as claimed by Talmy. In other words, the agent's intention is a part of the prototypical reading of the verb representing the causative predicate since the former is pragmatically associated with the latter in normal circumstances. We argue that the satellite or the resultative predicate in our terms does not indicate the fulfillment of the realization of an agent's intention as claimed by Talmy. It is argued that the result of the causing action, rather than the agent's intention, is a semantic property which is intrinsic in the semantics of the causing verb and that the degree of its intrinsicness varies from verb to verb. The fact that the English verbs classified by Talmy as "implied-fulfillment" and "attained-fulfillment" verbs can co-occur with the adverbs accidentally or unintentionally as shown in (5a)-(5d) proves that they are not always associated with an agent's goal that a certain result takes place on the part of an affected entity.
(5) a. I wiped the table clean unintentionally.
b. I kicked the hubcap flat unintentionally.
c. The police killed the criminal accidentally.
d. He drowned her unintentionally.
It is likely that the agent of these verbs sets a goal in carrying out an action but we argue that it is not a necessary part of the causative verb meanings. We can always create a context in which a sentence can be interpreted in such a way that the agent's goal concerning the affected entity is lacking.
We argue that what is at issue here is not the degree of the agent's intention inherent in a transitive verb. Rather, it is the degree of the likelihood that an event take place as a result of the causing action that is inherent in the causing verb. Such a likelihood is called implicature and entailment in this study. We postulate two types of verb which function as the causative predicate, namely, implied-result and entailed-result verbs. These two verb types are defined in terms of Vendler's event types (Vendler 1967), namely, activities, accomplishments, achievements and states. Vendler's event types which are relevant here are activities and accomplishments. Activities as defined by Vendler (1967) refer to processes going on in time in a homogeneous way, consisting of successive phases following one another in time, such as run, walk, swim, push a cart, drive a car. Activities are thus durative and unbounded, and can be schematically illustrated as a zigzag line in Figure 1.
[FIGURE 1 OMITTED]
Accomplishments refer to processes going on in time and proceeding to a terminal point, such as paint a picture, make a chair, deliver a sermon, draw a circle. Accomplishments are thus durative and bounded, and can be schematically shown in Figure 2. The zigzag line represents an activity whereas the vertical line represents a terminal point or the boundary of the activity.
[FIGURE 2 OMITTED]
Our implied-result verbs as defined in this study are a subtype of activity verbs. They express activities which have an implicature that a patient undergoes a change in state as a result of the agent's action. The result is not guaranteed to take place; it is only implied. Thus, implied-result verbs can be schematically shown by Figure 3 with the dots in the box representing the implicature.
[FIGURE 3 OMITTED]
Some examples of implied-result verbs are wash (clothes, glasses), wipe (a table), sweep (a floor), iron (a shirt), polish (a floor) in English, and laan (koeoew) 'wash (glass)', sak (phaa) 'wash (clothes)', thuu (phuwn) 'wipe (floor)', khat (rccnthaw) 'polish (shoes), shoeshine', kwaat (phwwn) 'sweep (floor)', taak (swa) 'expose to the sun (clothes)', riit (phaa) 'iron (clothes)' in Thai.
According to Vendler (1967), temporal adverbials are sensitive to the type of event being modified. For example, activity verbs can be modified by durative adverbials, such as for an hour, while accomplishment verbs cannot. In contrast, accomplishment verbs can be modified by frame adverbials, such as in an hour, while activity verbs cannot. Therefore, the (im)possibility to co-occur with certain temporal adverbials can be used as a criterion to determine the event type of a given verb. All of the examples of implied-result verbs given above can co-occur with a durative adverbial as shown below, which proves that implied-result verbs are activity verbs.
(6) a. John washed his clothes for one hour.
b. John wiped a table for one hour.
c. John swept the floor for one hour.
d. John ironed the shirt for one hour.
(7) a. somchaay laan kaeaew naan nwn chuamoon
Somchaay wash glass for one hour
'Somchaay washed the glasses for one hour.'
b. somchaay sak phaa naan mwn chuamoon
Somchaay wash clothes for one hour
'Somchaay washed the clothes for one hour.'
c. somchaay kwaat phuiutn naan nwn chuamoon
Somchaay sweep floor for one hour
'Somchaay swept the floor for one hour.'
d. somchaay riit phaa naan nwn chuamoon
Somchaay iron clothes for one hour
'Somchaay ironed the clothes for one hour.'
Our entailed-result verbs, on the other hand, are a subtype of accomplishment verbs. They express accomplishments which have an entailment that a patient undergoes a change in state as a result of the agent's action and that the state results after the terminal point has been reached. In an entailed-result verb, a resulting state is entailed. Entailed-result verbs are schematically shown in Figure 4. Notice that the schematic representation of entailed-result verbs is different from that of implied-result verbs in that the former is composed of a zigzag line denoting the activity phase, a vertical line denoting the boundary of the activity and a black box denoting the entailment of the accomplishment whereas the latter is composed of only a zigzag line representing the activity and a dotted box representing the implicature.
[FIGURE 4 OMITTED]
Some examples of entailed-result verbs are tear (paper), kill (a criminal), break (a twig), destroy (a building), drown (a child), strangle (a man) in English, and chiik (kradaat) 'tear (paper)', khaa (phuuraay) 'kill (criminal)', hak (kinmay) 'break (branch)', tat (phaa) 'cut (cloth)', thamlaay (?aakhaan) 'destroy (building)', pcck (plwak) 'peel (skin)', thccn (fan) 'pull out (teeth)' in Thai. The fact that ali instances of entailed-result verbs given here can co-occur with a frame adverbial as shown below proves that they are accomplishment verbs.
(8) a. John tore these pieces of paper in an hour.
b. John broke the twigs in an hour.
c. John destroyed the building in an hour.
d. The dentist pulled out his tooth in one hour.
(9) a. somchaay chiik kradaat phaaynay nwn chuamoon
Somchaay tear paper in one hour
'Somchaay tore the paper in one hour.'
b. somchaay hak kinmay phaaynay nwn chuamoon
Somchaay break twig in one hour
'Somchaay broke the twigs in one hour.'
c. somchaay thamlaay aakhaan phaaynay nwn chuamoon
Somchaay destroy building in one hour
'Somchaay destroyed the building in one hour.'
d. somchaay thccn fan khaw phaaynay nwn
Somchaay pull out tooth he in one
'Somchaay pulled out his tooth in one hour.'
It is noted that implied-result and entailed-result verbs can be defined semantically as well as syntactically. Implied-result verbs can be accompanied by both a confirming clause as in (10a)-(10c) and (11a)-(11c), and a disclaiming clause in biclausal constructions as in (12a)-(12c) and (13a)-(13c).
(10) a. John washed the shirt and it carne out clean.
b. John wiped the floor and it came out clean.
c. John ironed the shirt and it came out smooth.
(11) a. somchaay sak swa loe? swa kcc sa?aat
Somchaay wash shirt and shirt topic marker clean
'Somchaay washed the shirt and the shirt came out clean.'
b. somchaay thuu phwwm loe? phwwn kcc sa?aat
Somchaay wipe floor and floor topicmarker clean
'Somchaay wiped the floor and the floor came out clean.'
c. somchaay riit swa loe? swa kcc riap
Somchaay iron shirt and shirt topic marker smooth
'Somchaay ironed the shirt and the shirt came out smooth.'
(12) a. John washed the shirt but it did not come out clean.
b. John wiped the floor but it did not come out clean.
c. John ironed the shirt but it did not come out smooth.
(13) a. somchaay sak swa toeoe swa may sa?aat
Somchaay wash shirt but shirt not clean
'Somchaay washed the shirt but it did not come out clean.'
b. somchaay thuu phwwn toeoe phwwn may sa?aat
Somchaay wipe floor but floor not clean
'Somchaay wiped the floor but the floor did not come out
c. somchaay riit swa toeoe swa may riap
Somchaay iron shirt but shirt not smooth
'Somchaay ironed the shirt but the shirt did not come out smooth.'
In contrast, entailed-result verbs cannot be accompanied by a disclaiming clause in biclausal constructions as in (14a)-(14c) and (15 a)-(15c).
(14) a. * John killed some bugs but they did not die.
b. * John cut a piece of paper but it was not cut.
c. * John broke a twig but it was not broken.
(15) a. * somchaay khaa maloeoen toeoe maloeoen may taay
Somchaay kill bug but bug not die/dead
'Somchaay killed some bugs but they did not die.'
b. * somchaay tat kradaat toeoe kradaat may khaat
Somchaay cut paper but paper not be tom
'Somchaay cut a piece of paper but it was not cut.'
c. * somchaay hak kinmay toeoe kinmay may hak
Somchaay break twig but twig not be broken
'Somchaay broke a twig but it was not broken.'
Entailed-result verbs can take a confirming clause as in (16a)-(16c) and (17a)-(17c). However, they sound odd because of redundancy.
(16) a. ?John killed some bugs and they died.
b. ?John cut a piece of paper and it was cut.
c. ?John broke a twig and it broke.
(17) a. ?somchaay khda malcecen loe? maloeoen kcc taay
Somchaay kill bug and bug topic marker die/dead
'Somchaay killed some bugs and they died.'
b. ?somchaay tat kradaat loe? kradaat kcc khaat
Somchaay cut paper and paper topic marker be tom
'Somchaay cut a piece of paper and it was cut.'
c. ?somchaay hak kinmay le? kinmay kcc hak
Somchaay break twig and twig topic marker be broken
'Somchaay broke a twig and the twig was broken.'
As mentioned above, the notion of agent's intention is not relevant to the characterization of implied-result and entailed-result verbs. The prototypical situation expressed by an implied-result verb is one in which an agent volitionally performs an action with the intention that a certain resulting state will occur, and that it is likely to occur. Likewise, the prototypical situation expressed by an entailed-result verb is one in which an agent volitionally performs an action with the intention that a certain resulting state will occur, and that it does occur. However, some non-prototypical cases can exist as exemplified in (5a)-(5d). In the next section, we will present findings regarding constraints on the co-occurrences between such causative verbs and their resultative predicates in Thai and English.
3. Constraints on the co-occurrences between the causative and resultative predicates in Thai and English In this section, we will look into the syntactic-semantic behavioral properties of transitive verbs of implied-result and entailed-result types when they co-occur with resultative predicates in Thai and English. Specifically, we will study resultative sentences with implied-result and entailed-result verbs in the following aspects:
a) examine syntactic and semantic types of resultative predicates which can co-occur with implied-result and entailed-result verbs in each language,
b) identify constraints on the co-occurrences between causative and resultative predicates in each language,
c) account for collocational constraints in semantic and functional terms.
We will divide this section into three subsections. The first subsection deals with Thai and English resultative sentences with implied-result verbs whereas the second one deals with those with entailed-result verbs. We will postulate constraints on the co-occurrences between the two predicates and will account for them in the third subsection.
3.1. Thai and English resultative constructions with implied-result verbs
3.1.1. Thai resultative constructions with implied-result verbs. The resultative construction in Thai in which the causative predicate is instantiated as an implied-result verb is linguistically realized as a monoclausal transitive-based resultative sentence in which three components, namely, a transitive verb expressing a causing action, a direct object argument and a verb indicating a resulting state, are placed in juxtaposition without any intervening linker. The syntactic pattern of the transitive-based resultative construction in Thai is as follows:
NP1 VI NP2 V2 (subject) (causative predicate) (object) (resultative predicate)
There are many Thai verbs which can be semantically classified as belonging to the implied-result verb type, such as chet 'wipe', thuu 'wipe forcefully', laan 'wash, rinse', kwaat 'sweep', and sak 'wash (clothes).' As we saw above, these verbs are activity verbs which have the implicature that the activities lead to certain results and that the results take place. In the case of these five implied-result verbs in Thai, they convey the implicature that the objects that were affected by the agents' performance became clean as a result of the agents' activities. The addition of the stative verb sa?aat 'clean' after the direct object noun phrases confirms the occurrence of the implied results which are inherent in the transitive verbs. Such a stative verb is thus termed "confirmed-implicature" resultative predicate. Sentences (18a)-(18e) illustrate resultative constructions with confirmed-implicature resultative predicates.
(18) a. somchaay sak swa sa?aat
Somchaay wash shirt clean
'Somchaay washed his shirt clean.'
b. somchaay thuu phwwn sa?aat
Somchaay wipe floor clean
'Somchaay wiped the floor clean.'
c. somchaay riit swa riap
Somchaay iron shirt smooth
'Somchaay ironed his shirt smooth.'
d. somchaay laan rot sa?aat
Somchaay wash car clean
'Somchaay washed his car clean.'
e. somchaay kwaat phwwn sa?aat
Somchaay sweep room clean
'Somchaay swept the floor clean.'
It is noted that the resultative construction with the confirmed-implicature resultative predicate sounds more natural if the resultative predicate is accompanied by an intensifier can meaning "very" as in (19). The intensifier can creates the context of a high degree of the result rather than the realization in and of itself (3). This accounts for why the confirmed-implicature resultative predicate sounds natural in (19).
(19) somchaay sak swa sa?aat can
Somchaay wash shirt clean very
'Somchaay washed his shirt so clean.'
In addition to the confirmed-implicature resultative predicate, it is possible to add an "anti-implicature" resultative predicate to some implied-result verbs to cancel the implicature inherent in a causative verb and to indicate that the opposite resulting state takes place instead as in (20a)-(20e):
(20) a. somchaay sak swa sokkaprok
Somchaay wash shirt dirty
'Somchaay washed the shirt but it came out dirty.'
b. somchaay thuu phwwm sokkaprok
Somchaay wipe floor dirty
'Somchaay wiped the floor but it came out dirty.'
c. somchaay riit stua yap
Somchaay iron shirt wrinkled
'Somchaay ironed the shirt but it came out wrinkled.'
d. ?somchaay laan rot sokkaprok
Somchaay wash car dirty
'Somchaay washed the car but it came out dirty.'
e. ?somchaay kwaat phuium sokkaprok
Somchaay sweep room dirty
'Somchaay swept the floor but it came out dirty.'
As indicated above, sentences (20d)-(20e) are questionable whereas sentences (20a)-(20c) sound more acceptable. It is noted that the actions expressed by the verbs sak (swa) 'wash (clothes)', thuu (phwwn) 'wipe (floor)', riit (swa) 'iron (shirt)', which allow an anti-implicature resultative predicate, must be performed in an "appropriate" or "favorable" situation; otherwise, certain unexpected events might result. For example, if the agent in (20a) washed a white shirt with clothes in bright colors, the shirt could be stained and come out dirty. In (20b), if the agent did not wipe the floor in a neat manner or wiped it with a dirty piece of cloth, the floor would become dirty rather than clean. In (20c), if the agent ironed the shirt unskillfully, the shirt might come out wrinkled. The agent's skill and use of good instruments and materials to perform the actions obviously constitute the "appropriate" or "favorable" situations of the actions in (20a)-(20c) and are required in order for the expected results of the actions to obtain. In contrast, the agent's skill and use of good instruments and materials are less likely to play a crucial role in obtaining the expected results in the case of laan (rot) 'wash (car)' and kwaat (phwwn) 'sweep (floor)'. In other words, even if the actions expressed by these two verbs are performed in an unfavorable situation, such as by an unskillful agent or with abad instrument, it is likely that the expected results take place. This accounts for the oddness of (16d)--(16e), in which laan (rot) 'wash (car)' and kwaat (phwwn) 'sweep (floor)' co-occur with anti-implicature resultative predicates, which express unexpected resulting situations. It can be generalized that the implied-result Verbs which do not take anti-implicature resultative predicates express actions which do not require a favorable situation for an expected result to obtain.
It is also possible to negate the confirmed-implicature resultative predicate in Thai, as shown in (21a)-(21e). The negated resultative predicate functions as a disclaiming phrase and is termed "cancelled-implicature" resultative predicate in this article. The cancellation of the implicature inherent in the matrix verb results in vagueness in that it can be interpreted either as anti-implicature ("dirty" and "wrinkled" in [20a]-[20e] above) or non-implicature (moving toward the implied result, i.e., the state of cleanliness and smoothness in (20a)-(20e), but not reaching that state yet.)
(21) a. somchaay sak swa may sa?aat
Somchaay wash shirt not clean
'Somchaay washed his shirt but it did not come out clean.'
b. somchaay thuu phwwn may sa?aat
Somchaay wipe floor not clean
'Somchaay wiped the floor but it did not come out clean.'
c. somchaay riit swa may riap
Somchaay iron shirt not smooth
'Somchaay ironed his shirt but it did not come out smooth.'
d. somchaay laan rot may sa?aat
Somchaay wash car not clean
'Somchaay washed his car but it did not come out clean.'
e. somchaay kwaat phwwn may sa?aat
Somchaay sweep floor not clean
'Somchaay swept the floor but it did not come out clean.'
The last type of implicature which can be accompanied with implied-result verbs is the "other-event" (4) resultative predicate, which expresses a state that results from an action that does not lie on the axis leading to the implicated result.
(22) a. yaa sak swa khaat ha
do not wash shirt tom final particle
'Do not wash the shirt in such a way that it gets tom in the process.'
b. yaa laan koeoew toeoek na
do not wash glass broken final particle
'Do not wash the glass in such a way that it gets broken in the process.'
c. yaa riit swa may na
do not iron shirt burnt final particle
'Do not iron the shirt in such a way that it gets burnt in the process.'
In summary, the implied-result verbs in Thai which function as the causative predicate can be accompanied by various types of resultative predicate, namely, confirmed-implicature, anti-implicature, cancelled-implicature and other-event resultative predicates (5). It is noted that only some implied-result verbs can be accompanied by the anti-implicature resultative predicate.
3.1.2. English resultative constructions with implied-result verbs. In this section, we will examine what types of resultative predicate can accompany implied-result verbs which function as the causative predicate in English. The syntactic pattern of the resultative construction in English is basically the same as that in Thai. Our quick survey of some American English speakers revealed that implied-result verbs in English can be accompanied by the confirmed-implicature resultative predicate with varying degrees of acceptability as below.
(23) a. John waxed the floor glossy. (marginal)
b. John washed the glass clean. (marked)
c. John ironed the shirt smooth. (less marked than [23a] and [23b])
d. John swept the floor clean. (very common)
e. John wiped the table clean. (very common)
Sentences (23a)-(23e) are ordered from the lowest to the highest degree of acceptability. Sentence (23a) has the lowest degree of acceptability whereas (23e) has the highest degree (6). Sentence (23b) is marked in that it must occur in an appropriate context such as an emphatic one.
Implied-result verbs in English cannot take the anti-implicature resultative predicate as shown below. According to native speakers of English, the unacceptability of (24a)-(24e) arises from the fact that these sentences express inconceivable situations.
(24) a. * John waxed the floor dull.
b. * John washed the glass dirty.
c. * John ironed the shirt wrinkled.
d. * John swept the floor dirty.
e. * John wiped the table dirty.
Implied-result verbs in English cannot take the cancelled-implicature resultative predicate in the structural pattern parallel to that in (21a) (21e) in Thai, in which may 'not' appears immediately before the resultative predicate. The English sentences with the cancelled-implicature resultative predicate parallel to (21a)-(21e) in Thai would be ungrammatical as shown in (25a)-(25e) because English simply does not allow such a syntactic structure (7).
(25) a. * John waxed the floor not glossy.
b. * John washed the glass not clean.
c. * John ironed the shirt not smooth.
d. * John swept the floor not clean.
e. * John wiped the table not clean.
It is found that implied-result verbs in English cannot take the other-event resultative predicate either as shown below.
(26) a. * Do not wax the floor slippery.
b. * Do not wash the glass broken.
c. * Do not iron the shirt tom.
d. * Do not wipe the floor slippery.
To summarize, implied-result verbs in English exhibit more constraints in taking resultative predicates than those in Thai. The former can take only confirmed-implicature resultative predicates whereas the latter can take as many as four types. Furthermore, not every implied-result verb in the former can take confirmed-implicature resultative predicates as shown in (23a)-(23e). The fact that English implied-result verbs are much more heavily restricted in taking resultative predicates than the comparable Thai verbs is accounted for in Section 5.
3.2. Thai and English resultative constructions with entailed-result verbs
3.2.1. Thai resultative constructions with entailed-result verbs. The verbs in Thai which can be classified as entailed-result verbs include khaa 'kill,' chiik 'tear,' hak 'break,' tat 'cut,' pcck 'peel,' and thamlaay 'destroy.' These verbs which function as the causative predicate can take a "confirmed-entailment" resultative predicate to cortfirm that the entailed result inherent in the causing verbs take place as shown below.
(27) a. tamruat khaa phuuraay taay
police kill criminal die, dead
'The police killed the criminal (and he/she was dead).'
b. somchaay chiik phaa khaat
Somchaay tear cloth tom
'Somchaay tore the cloth (and it was tom).'
c. somchaay hak kinmay ?cck (8)
Somchaay break branch exit
'Somchaay broke the branch (and it was broken).'
d. somchaay tat phaa khaat
Somchaay cut cloth torn
'Somchaay cut the cloth (and the cloth was torn).'
e. somchaay pcck plwak som ?cck
Somchay peel peel (noun) orange exit
'Somchaay peeled the orange (and the peel came off).'
f. somchaay thamlaay kamphoeoen phan
Somchay destroy wall collapse
'Somchaay destroyed the wall (and it was destroyed).'
It is also possible to negate the confirmed results in (27a)-(27f). In other words, it is possible for entailed-result verbs in Thai to take "cancelled-entailment" resultative predicates as shown below.
(28) a. tamruat khaa phuuraay may taay
police kill criminal not die, dead
'The police tried to kill the criminal but he/she was not dead.'
b. somchaay chiik phaa may khaat
Somchaay tear cloth not tom
'Somchaay tried to tear the cloth but it was not tom.'
c. somchaay hak kinmay may ?cck
Somchaay break branch not exit
'Somchaay tried to break the branch but it was not broken.'
d. somchaay tat phaa may khaat
Somchaay cut cloth not tom
'Somchaay tried to cut the cloth but the cloth was not tom.'
e. somchaay pcck plwak som may ?cck
Somchay peel peel (noun) orange not exit
'Somchaay tried to peel the orange but the peel did not come off.'
f. somchaay thamlaay kamphoeoen may phan
Somchay destroy wall not collapse
'Somchaay tried to destroy the wall but it was not destroyed.'
The notion of entailment is traditionally defined as that which is not cancellable. The phenomenon in (28a)-(28f) in which the entailment of entailed-result verbs in Thai can be cancelled thus seems to contradict the traditional definition of the notion of entailment. We will account for the seemingly exceptional behavior of the entailment of entailed-result verbs in Thai in Section 5.
3.2.2. English resultative constructions with entailed-result verbs. English verbs which can be classified as entailed-result verbs include drown, kill, burn, cut, break, and destroy. These verbs are classified by Talmy (2000) as attained-fulfillment verbs, which indicate the actual fulfillment of the agent's intention. Syntactically, they cannot be accompanied by a satellite to indicate the realization of the agent's intention since all the semantic elements, including the agent's intention referred to by the verbs, are realized. Talmy (2000: 267) discusses two verbs of this type, namely, kill and drown. According to Talmy, the addition of a satellite to attained-fulfillment verbs to confirm the agent's intention is considered redundant and results in unacceptability as in *I killed him dead and *I drowned him dead. We have argued that the agent's intention is not relevant in characterizing this type of verb. Talmy is right in saying that attained-fulfillment verbs, which correspond to entailed-result verbs in this study, cannot take a satellite or resultative predicate in our terms as shown below.
(29) a. * I killed him dead.
b. * I drowned him dead.
c. * I cut the cloth tom.
Sentences (29a)-(29c) are unacceptable because of redundancy. Interestingly, it is found that the verb kill in English can take the confirmed-entailment resultative predicate only in the case of advertising insect-killing spray or computer bugs-killing software such as below.
(30) a. Raid. Kills Bugs Dead.
(A RAID commercial advertisement)
b. ZD Net: kill Y2K Bugs Dead.
(A computer program commercial)
It can be concluded that the verb kill in English cannot take a confirm-edentailment resultative predicate in normal cases. It can take one only as a fixed phrase in an emphatic context, such as in a commercial advertisement, in which the speaker or writer wants to emphasize the dead condition of an affected entity, which is the entailment inherent in the verb kill. The emphasis aims at convincing the hearer or reader of the effectiveness of an advertised product. Therefore, the emphatic context pragmatically licenses the appearance of the confirmed-entailment resultative predicate of kill.
In this section, we have examined the constraints on the co-occurrences between causative predicates with implied-result and entailed-result verbs in Thai and English on the one hand and resultative predicates in both languages on the other. It is found that the resultative constructions containing the two types of verbs in English exhibit more constraints than those in Thai. That is, the causative predicates with implied-result verbs in Thai can co-occur with confirmed-implicature, anti-implicature, cancelled-implicature and other-event resultative predicates, whereas those in English can co-occur only with confirmed-implicature with varying degrees of acceptability. As for causative predicates with entailedresult verbs, those in Thai can co-occur with confirmed-entailment and cancelled-entailment resultative predicates whereas those in English cannot co-occur with either one. There is only one exception, i.e., the case of the verb kill in English, which can co-occur with the confirme-dentailment in such emphatic contexts as advertisements. In the next section, we will examine how these constraints on co-occurrences between causative and resultative predicates have been accounted for in previous studies on various languages.
4. Previous studies on collocational eonstraints between causative and resultative predicates
In this section, we will review five pieces of work which examine co-occurrences between causative and resultative predicates both in monoclausal and biclausal structures, namely, Teng (1972), Tai and Chou (1975), Ikegami (1985), Talmy (2000) and Pederson (2008). These works examine co-occurrences between the two types of predicates in Chinese, Japanese, Tamil, German and English.
Teng (1972) is the first researcher to observe that while the English verb kill consists of two subevents, namely, a causing action and a resulting state, the corresponding verb sha in Mandarin Chinese has only the causing action although it may imply the resulting state of si 'dead.' Tai and Chou (1975) also compares and contrasts the verbs kill in English and sha in Mandarin Chinese and presents evidence and arguments to support Teng's claim. They claim that "As opposed to the verb compound sha si, sha can occur in the pattern of X sha Y Y not si 'X "kill" Y, Y not die.' This indicates that sha does not necessarily imply si." (Tai and Chou 1975: 48-49). Tai and Chou (1975) label English verbs which imply an attainment of the agent's goal in carrying out an action, such as kill, an "implicative action verb." They postulate a general principle governing some discrepancies between Chinese and English lexicalization which states that while English has implicative action verbs such as kill, find, and learn, (9) which imply an attainment of a certain goal, their correspondences in Chinese must be expressed by means of verb compounds in which the first element indicates the action and the second one indicates the attainment of the goal of the action which is the resulting state. They make a further claim that so far they have found no Chinese action verbs which imply the attainment of goal.
Ikegami (1985) also provides a contrastive investigation of a number of English and corresponding Japanese "goal-directed" action verbs defined as verbs which contain two subevents, i.e., an action and a goal, bearing a relation to each other in such a way that the former is directed to the latter. Some examples of this kind of verb include those meaning 'kill', 'burn', 'boil', 'drop', 'cheat', 'dry', 'float', 'melt', etc. According to Ikegami (1985), a goal-directed action may or may not achieve its goal. It is found in this work that there is a systematic contrast between English and Japanese verbs referring to goal-directed actions in terms of the implication of achievement or non-achievement of the goal. In other words, corresponding verbs in English and Japanese may differ in the emphasis they place on the action or the achievement phase. English verbs are found to focus on the achievement phase whereas Japanese verbs focus on the activity phase. He concludes that English verbs are goal-oriented whereas the Japanese ones are process-oriented. Ikegami also discusses the English and Japanese verbs with the meanings 'kill' and 'burn.' He points out that the English verb kill and the corresponding Japanese verb korosu are in the same semantic category since the achievement of the goal is implied in both verbs in the two languages. On the other hand, the English verb burn and the corresponding verb moyasu in Japanese are in different semantic categories since the achievement of the goal is implied by the English verb but not by the Japanese one. The English and Japanese sentences cited by Ikegami (1985: 273) to support this claim are as follows.
(31) * John killed Mary, but Mary didn't die.
(32) * John-wa Mary-o korosita keredomo, Mary-wa sinanakatta
John-Top Mary-Acc killed though Mary-Top didn't die
'* John killed Mary but Mary didn't die.'
(33) * I burned it, but it didn't burn.
(34) moyasita keredo, moennakatta burned though didn't burn
'* (Someone) burned (something), but it didn't burn.'
Talmy (2000: 269) accounts for the same phenomenon by postulating the notion of "lexicalized implicature," which refers to the kind of implicature which is defeasible and which is associated with a lexical item. Talmy claims that different verbs in a certain semantic field in a single language, such as choke, stab, strangle and drown, may have different degrees of strength of lexicalized implicature which might correlate in part with different degrees of strength of the agent's intention for a further result. Furthermore, the different degrees of strength of implicature tend to correlate also with the verbs' ability to take a satellite that confirms the fulfillment. The verb choke is located at one end of the cline of strength of implicature since it has no implicature of entailed result, at least for some speakers, whereas the verb drown is located at the other end since the occurrence of the resulting event of dying is not merely implied but asserted in the lexical semantics of the verb itself. The verbs stab and strangle are located between the two extremes with some degree of strength of implicature.
Pederson (2008) explores how different languages encode the realization of an event. The data in this study is drawn from Tamil, German and English. He claims that the core meanings expressed by typical translation equivalents or corresponding verbs in English and Tamil are the same. However, Tamil verbs are more flexible in that they can be used extensionally in ways prohibited by English. For example, although the verb for 'kill' in Tamil does mean 'kill' in its basic use, it can be used to refer to only the doing part without asserting the final realized state (dead). Pederson (2008) argues that the whole-for-part metonymy is a strategy which allows Tamil speakers to use a transitive verb such as the verb for 'break' to refer only to the first part(s) of the event. However, they cannot use this strategy when they use the converb construction, which entails realization, otherwise it would be contradictory. On the other hand, English has a number of words and constructions which explicitly deny realization, such as almost, nearly, shoot at the soldier. The availability of this strategy reduces the motivation to use event verbs in a whole-for-part metonymy. In addressing the question of learnability of German-speaking and Tamil-speaking children, this study argues that all of these children need not acquire different understandings of the semantics of the simple verbs. Rather, they acquire a language-sensitive appreciation of what is an allowable extended use of the verbs and what is not. It is claimed in this article that Tamil-speaking children learn that they can be flexible with their verbs in a way that is not allowed for German-speaking children.
In accounting for properties of entailed-result verbs across languages, all of the works reviewed above take the same position with regard to two issues. First, all works above, except Pederson (2008), claim that lexical semantic properties inherent in entailed-result verbs have a bearing on whether they can co-occur with resultative predicates or not. Such lexical semantic properties of entailed-result verbs can vary from language to language and from verb to verb in the same semantic field in a single language. Second, all works reviewed above assume that the agent performs an action expressed by an entailed-result verb with a certain goal in mind. The goal is a resulting state which happens to an affected entity after the agent has performed the action. If the resulting state takes place, it means the agent's goal has been attained. This article takes a different position with regard to the agent's intention as argued in Section 2 and also takes a different approach in analyzing the constraint in co- occurrences between causative and resultative predicates as argued in the next section.
5. An account of collocational constraints in terms of aspectual profile shift
5.1. Aspectual contour and aspectual profile
In this section, we will propose a new perspective of analyzing constraints on co-occurrences between causative and resultative predicates in Thai and English. We will account for such constraints in terms of aspectual profile shift, which is the notion postulated by Croft (in prep.). Before we discuss the mechanism of aspectual profile shift, it is imperative to understand the semantic representation of lexical aspect as set forth by Croft (in prep.).
Aspect is generally defined as the temporal structure of a situation. Croft argues that the notion of aspect involves an important second dimension, namely, its qualitative structure, which is defined as the qualitative states it possesses over time. The qualitative structure dimension represents only the relevant qualitative states and changes in a situation. The participants involved in a situation possess many different qualitative states at a given time. According to Croft, aspect must be understood as the interaction of two dimensions, namely, qualitative state and change on the one hand and time on the other. These two dimensions are represented geometrically in the two-dimensional diagram shown in Figure 5. The symbol A represents the dimension of qualitative states and changes in a situation where as the symbol t represents the dimension of time.
A situation is represented as a contour in the two-dimensional diagram above. This contour is called the "aspectual contour." The aspectual contour describes the course of a situation from its beginning to its end and represents a speaker's encyclopedic knowledge about the course of an event. The aspectual contour does not represent the aspectual meaning of a sentence containing a verb stem such as open and a tense-aspect construction. It is important to specify a particular phase in the situation as asserted by a sentence, such as I opened the door. The phase expressed by this sentence is the instantaneous change of state of the door from being not open to being open. This phase of the situation is called the "profile" of the sentence meaning. The aspectual contour and profile for I opened the door is shown in Figure 6.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
In the figure above, the three central lines represent the aspectual contour of the situation of opening the door, which indicates the course of a situation from its beginning to its end. The vertical line represents the profile of the sentence I opened the door, which corresponds to the meaning of the transitive verb stem in combination with the tense-aspect construction, which is past tense in this case. The symbol < P indicates a point in time preceding the present.
The combination of a verb stem and a tense-aspect construction gives rise to a "construal operation", which refers to a reconceptualization of the situation's aspectual structure. It may be a minor or major reconceptualization. A minor reconceptualization is simply a shift in what part of the aspectual contour is profiled. A major one is a more substantial restructuring of the aspectual contour.
According to Croft, the profile of the aspectual contour of a situation can be shifted if the verb stem indicating that situation appears in a "profile-changing construction." A good example of a profile-changing construction is the VERB-ing aspect construction, which normally serves as the progressive aspect marker in English. This progressive marker appears with activity verbs in normal circumstances, such as John is running, John is reading a book. However, stative verbs can occur in this construction in marked contexts. In that case, this construction serves to change the aspectual profile of the stative verbs from states to activities. Please look at the examples below.
(35) a. He is naughty.
b. He is being naughty.
The verb stem be naughty in sentence (35a) profiles a state. When this verb stem is put into the VERB-ing construction in sentence (35b), the original aspectual profile, which is a state, is reconceptualized or construed as an activity which is going on at the moment.
5.2. Aspectual profile shift in resultative constructions in Thai
In this section, we will account for the fact that the causative predicate with implied-result and entailed-result verbs can co-occur with many types of resultative predicate in terms of aspectual profile shift in Croft's sense. However, the profile-changing construction which applies to Thai resultative constructions is not the tense-aspect construction as in English. Rather, we argue that it is serial verb construction which changes the profile of the aspectual contour of a situation. A serial verb construction is generally defined as a sequence of verbs which act together as a single predicate and which are put in juxtaposition without any linker. Serial verb constructions are prevalent in West African languages, Chinese and Southeast Asian languages including Thai, creole languages and Oceanic languages. There are many types of relation between verbs in serial verb constructions. One of the typical uses of serial verb construction is to indicate events occurring in sequence without a noticeable time span in between. The resultative meaning in which an action is followed by a resulting state of an affected entity can be considered a sequence of events. If an entailed-result verb representing the causing event and an inchoative verb representing the resulting event occur in sequence in the serial verb construction, these two verbs will be coerced by the construction to express a sequence of causing and resulting events without a noticeable time span. One important coercion effect is the disappearance of the entailment in the original entailed-result verb. It is argued in this article that the crucial factor which allows the entailed-result and the implied-result verbs in Thai to take various types of resultative predicates is the reconstrual or reconceptualization of the original aspectual meaning of the verbs. The phenomenon under investigation is therefore accounted for in terms of Croft's notion of aspectual profile shift.
We argue that when an entailed-result verb in Thai, which expresses a sequence of subevents in itself, takes place in a serial verb construction, the accomplishment which characterizes the entailed-result verb is reconceptualized as an activity. The culmination phase of the situation denoted by the entailed-result verb is, in the serial verb construction, encoded as a separate achievement verb following the entailed-result verb in the case of a confirmed-entailment resultative predicate, or denied and encoded as the negative marker followed by the achievement verb in the case of a cancelled-entailment resultative predicate. In other words, the serial verb construction serves to shift the aspectual profile of the entailed-result verb from the accomplishment to the activity. For example, the verb khaa 'kill' in Thai is an entailed-result verb since it cannot take a disclaiming clause and can take a confirming clause in a biclausal construction as exemplified in examples (15a) and (17a) repeated here for convenience as examples (36a) and (36b).
(36) a. * somchaay khaa maloeoen toeoe maloeoen may taay
Somchaay kill bug but bug not die/dead
'Somchaay killed some bugs but they did not die.'
b. ?somchaay khaa maloeoen loe? Maloeoen kcc taay
Somchaay kill bug and bug topic marker die/dead
'Somchaay killed some bugs and they died.'
However, the aspectual profile of an entailed-result verb such as khaa 'kill,' which is originally an accomplishment with an inherent entailment can be reconceptualized as an activity if such a verb appears in a serial verb construction. In other words, the serial verb construction in Thai serves as a profile-changing construction which shifts the profile of an entailed-result verb from an accomplishment with an entailment to an activity. Thus, the entailment which is originally inherent in the verb stem of an entailed-result verb is lifted in the serial verb construction, and is linguistically realized as a separate achievement verb taay 'die/ dead' in the case of a confirmed-entailment resultative predicate as in (37a). In the case of a cancelled-entailment resultative predicate, the entailment is eliminated in the serial verb construction, and is linguistically negated by the negative morpheme may preceding the achievement verb taay 'die/dead' as in (37b).
(37) a. somchaay khaa maloeoen taay
Somchaay kill bug die, dead
'Somchaay killed some bugs.'
b. somchaay khaa maloeoen may taay
Somchaay kill bug not die, dead
'Somchaay tried to kill some bugs but they did not die.'
In the same vein, the aspectual profile of an implied-result verb with an inherent implicature occurring in a serial verb construction is reconceptualized as an activity without an implicature. This allows them to co-occur with different kinds of resultative predicate, namely, confirmed- implicature as in (18a)-(18e), anti-implicature as in (20a)-(20e), cancelled-implicature as in (21a)-(21e), and other-event resultative predicate as in (22a)-(22c).
5.3. An account of collocational constraints in English
To recapitulate, some English implied-result verbs functioning as the causative predicate in the resultative construction can take only a confirmed-implicature resultative predicate. They cannot take other kinds of resultative predicate as those in Thai. As for entailed-result verbs functioning as the causative predicate, they cannot take any resultative predicate. There is only one exception, namely, kill, which can take dead only in certain emphatic contexts, which advertising often tries to employ. As discussed in Section 5.2, the serial verb construction in Thai serves as a profile-changing construction. However, we observe that English has no construction which performs a function comparable to that of the serial verb construction in Thai, namely, shifting the aspectual profile of the aspectual contour of a situation from an accomplishment to an activity in the case of the entailed-result verb (10). Therefore, entailed-result verbs in English cannot take a resultative predicate to confirm the entailment because it becomes redundant. Nor can they take a cancelled-entailment resultative predicate because it contradicts the meaning in the entailed-result verbs. In the case of implied-result verbs, the implicature inherent in them can be confirmed by being realized as an adjective functioning as the resultative predicate. Such confirmation does not shift the aspectual profile of the implied-result verbs from the activity with an implicature to the activity without an implicature. If it did, it would be able to take other types of resultative predicate as in the Thai case. (11)
As pointed out in Section 3, resultative constructions with implied-result and entailed-result verbs in Thai are more productive than those in English. The productivity in the case of Thai resultative constructions is attributed to the aspectual profile shift operating in serial verb constructions in Thai. In other words, Thai resultative constructions are productive because serial verb constructions in Thai, of which Thai resultative constructions are an instantiation, allow both conventionalized as well as unconventionalized scenes to be expressed. By a conventionalized scene, we mean a sequence of action and a result caused by the action which is lexically implied and entailed by a transitive verb (12). In contrast, only conventionalized scenes can be expressed by resultative constructions in English.
We might wonder whether or not the new account of the productivity in resultative constructions in Thai in terms of aspectual profile shift is in conflict with the previous accounts of comparable phenomena in terms of varying degrees of strength of implicature inherent in transitive verbs in a semantic field in a single language and in corresponding verbs in different languages as pointed out in Section 4. The answer is no. The notions of aspectual profile shift and varying degrees of strength of implicature can work hand in hand. In analyzing linguistic phenomena, Langacker (1987) warns that an analyst should not fall into the "exclusionary fallacy," in which "one analysis, motivation, categorization, cause, function or explanation for a linguistic phenomenon necessarily precludes another" (Langacker 1987: 28). In our case, the adoption of the notion of aspectual profile shift in accounting for the productivity of resultative constructions in Thai does not preclude the notion of varying degrees of strength of implicature. Both notions can work hand in hand and help support each other. It is conceivable that if Thai verbs with an entailment such as khaa 'kill' repeatedly undergo aspectual profile shift, i.e., if they are used frequently in the serial verb construction, it is possible that the speaker and hearer accept this phenomenon in which the strength of lexical implicature gets weaker and weaker. From the language change perspective, it can be said that an entailment in this type of verb may become an implicature one day. This change is possible given the fact that there is no formal distinction marking aspectual contour difference between implied-result and entailed-result verbs in the construction under investigation in Thai. Furthermore, the two types of verbs resemble each other in taking resultative predicates which confirm and cancel their implicature and entailment, respectively. If a lexical semantic change in this direction occurs in the future, the ill-formed Thai sentence as exemplified in (38) may one day become acceptable as in the case of the verb sha 'kill' in Chinese.
(38) * somchaay khaa maloeoen toeoe maloeoen may taay
Somchaay kill bug but bug not die/dead
'Somchaay killed some bugs but they did not die.'
Chulalongkorn University Tohoku University
Received 18 January 2005
Revised version received
25 August 2006
Carrier, Jill & Janet Randall. 1992. The argument structure and syntactic structure of resultatives. Linguistic Inquiry 23(2). 173-234.
Chauncey, Chu C. 1973. Implied verbal aspects in English and Chinese. Paper presented at the Southeastern Conference on Linguistics (SECOL) IX, Chadottesville, VA.
Croft, William. In prep. Verbs: Aspect and argument structure. Oxford: Oxford University Press.
Ikegami, Yoshihiko. 1985. 'Activity'--'accomplishment'--'achievement'--A language that can't say 'I burned it, but it didn't burn' and one that can. In Adam Makkai and Alan K.
Melby (eds.), Linguistics and Philosophy. Essays in Honor of Rulon S. Wells, 265-304. Amsterdam: John Benjamins.
Langacker, Ronald W. 1987. Foundations of cognitive grammar, volume I." Theoretical prerequisites. Stanford, CA: Stanford University Press.
Levin, Beth & Malka Rappaport Hovav. 1999. Two structures of compositionally derived events. In Tanya Mathews & Devon Strolovitch (eds.), Proceedings of semantics and linguistic theory IX (SALT IX), 199-223. Ithaca, NY: CLC Publications, Cornell University.
Napoli, Donna Jo. 1992. Secondary resultative predicates in Italian. Journal of Linguistics 28. 53-90.
Pederson, Eric. 2008. Event realization in Tamil. In Melissa Bowerman & Penelope Brown (eds.), Crosslinguistic perspectives on argument structure: Implications for learnability, 331-356. New York: Lawrence Erlbaum.
Tai, James H.-Y. & Jane Yang Chou. 1975. On the equivalent of 'kill' in Mandarin Chinese. Journal of the Chinese Language Teachers Association 10(2). 48-52.
Talmy, Leonard. 2000. Toward a cognitive semantics, volume II. Typology and process in concept structuring. Cambridge, MA: MIT Press.
Teng, Shou-xin. 1972. A semantic study of the transitivity relations. University of California Berkeley dissertation.
Thepkanjana, Kingkarn. 2006. Properties of events expressed by serial verb constructions in Thai. Paper presented at the 11th Biennial Symposium: International Approaches to Complex Verb Construction, Rice University, Houston, TX.
Thepkanjana, Kingkarn & Satoshi Uehara. 2002. Grammaticalization of directional verbs into success markers in Thai. In Mary Andronis, Erin Debenport, Anne Pycha & Keiko Yoshimura (eds.), Proceedings of the 38th annual meeting of the Chicago Linguistic Society, 649-664. Chicago, IL: Chicago Linguistic Society.
Thompson, Sandra Annear. 1973. Resultative verb compounds in Mandarin Chinese: A case for lexical rules. Language 49(2). 361-379. Vendler, Zeno. 1967. Linguistics in philosophy. Ithaca, NY: Cornell University Press.
* This research was supported by a Basic Research Grant from the Thailand Research Fund (No. BRG 4780019) and Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science (No. 15520241). We are grateful to Masayoshi Shibatani and Susumu Kuno for their comments and suggestions provided at an early stage in the preparation of this article. Our thanks also go to two anonymous reviewers of Linguistics for comments and suggestions on an earlier draft of this article. We also thank Andy Klatt for providing English data and to Rob Troyer for textual improvements. Correspondence address: Kingkarn Thepkanjana, Department of Linguistics, Faculty of Arts, Chulalongkom University, Phayathai Road, Bangkok 10330, Thailand. E-mail: email@example.com.
(1.) The sentence translation for (2c), i.e., "He killed the criminal (and the criminal died as a result)" does not reflect the actual meaning of sentence (2c). There is no intervening time span between the events represented by the causative and resultative predicates in sentence (2c). However, this fact does not necessarily hold true for the English translation for (2c) above. Such English translation merely approximates the actual meaning of the Thai sentence (2c). The authors did not opt to translate (2c) as "He killed the criminal dead" because this sentence would sound odd in English
(2.) A satellite is "the grammatical category of any constituent other than a noun phrase or prepositional phrase complement that is in a sister relation to the verb root. It relates to the verb root as a dependent to a head" (Talmy 2000: 102).
(3.) This observation was made by one of the anonymous reviewers.
(4.) The term "other-event" is adopted from Talmy (2000: 277). A type of satellite in the domain of realization is termed by Talmy an "other-event satellite," which is defined as the state expressed by the satellite that does not lie on the conceptual axis leading to the verb's intended goal (Talmy 2000: 277).
(5.) The terms "confirmed-implicature" and "anti-implicature" postulated in this study are also modified from Talmy's terms for different kinds of satellite, i.e., confirmation satellite, and antifulfillment satellite, respectively.
(6.) The sentences in (23) are intended to show that degrees of acceptability for confirmed-implicature resultatives vary with the verb. One of the anonymous reviewers pointed out that degrees of acceptability seem to vary with the type of the complement as well as in John waxed the floor glossy (marginal) vs. John waxed the floor to a fine gloss (very common).
(7.) Sentences (21a)-(21e) would be syntactically possible with a morphological negation such as unclean. However, they would be semantically odd.
(8.) The verb hak in Thai belongs to the class of verbs which can be used either transitively or intransitively. The transitive hak means to break something whereas the intransitive hak can be interpreted as a process or a state. The former refers to a dynamic situation in which an entity enters into a state, which is a static situation. If such an "alternating" transitive verb occurs as the causative predicate in the resultative construction, one of the four directional verbs, i.e., khun 'ascend,' 109 'descend,' khaw 'enter,' and ?cck 'exit,' will be used as the resultative predicate instead of the homophonous form of the causative transitive verb to avoid repetition. (See Thepkanjana and Uehara  for details on this use of the four directional verbs in Thai.) The action denoted by this type of transitive verb results in the change of location of an affected entity. Any of the four directional verbs which functions as the resultative predicate therefore simultaneously denotes the resulting change of location of the affected entity and the direction of its motion. These four directional verbs can also appear as the resultative predicate in the resultative construction in which the transitive verb is not an alternating verb such as in (27e).
(9.) Tai and Chou's (1975) notion of "implicative action verbs" is based on Chauncey's (1973) observation that some English action verbs such as find and learn imply the attainment of a goal in carrying out an action. It is noted that the verbs find and learn are different from kill in that the agent's action is not linguistically encoded in the former whereas it is in the latter. Therefore, the former are not action verbs. On this basis, only the verb kill can be appropriately claimed to be an implicative action verb since it is made up of two subevents, namely, an agent's action and a resulting state.
(10.) It can be said that the exceptional case of kill taking dead in the English resultative construction which occurs in certain emphatic contexts is closest to the serial verb construction in Thai in question. It should be noted that the English resultative construction is not an aspectual profile-changing construction as the Thai serial verb construction, but can sometimes be an argument structure-changing construction. For example, the verb run is normally an intransitive verb but it can take a direct object in the resultative construction as in He ran his shoes threadbare.
(11.) One of the anonymous reviewers suspects that serial verb constructions in other languages are "quite commonly of this type" (i.e., Thai type), and we share his/her impression. However, we would like to leave the question of the universality of the aspectual profile shift of serial verb constructions for future study due to a lack of sufficient cross-linguistic data at hand.
(12.) The difference between conventionalized and unconventionalized scenes is a matter of degree and is, presumably, culturally dependent (Thepkanjana 2006).
|Printer friendly Cite/link Email Feedback|
|Author:||Thepkanjana, Kingkarn; Uehara, Satoshi|
|Publication:||Linguistics: an interdisciplinary journal of the language sciences|
|Date:||May 1, 2009|
|Previous Article:||Null subjects: a reanalysis of the data.|
|Next Article:||Feature-based argument mapping and animacy optimization in impersonal passives.|