CHAPTER 2 LITERATURE REVIEW AND GRAMMAR SKETCH OF YAMI
2.5 F RAMEWORK
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
observed in (36a) in constructing a clause with an actor subject in Yami, and the affix m- is also found in actor subject clauses, as exemplified in (36a’). The affix ma- is observed to construct clauses with an actor or patient subject, as in (36b) and (36d), respectively. The lack of full-fledged verbs (a verb with a full set of voice variants), the mismatch problem, many-to-one affix (more than one affix serves the same function), and one-to-many affix (one affix serves more than one function) are rather confusing when one seeks to explore the language from the perspective of grammatical voice without consideration of other factors.
Hence, the present study aims to re-examine the function or functions of Yami verbal affixes through examination of the roots/stems these affixes attach to, the thematic role of clause subjects, the number of participants involved, and its case marking via different types of clauses including declaratives, negatives, negative imperatives, and interrogatives. In doing so, we hope to get a better understanding of the function or functions of these verbal affixes, their relation with clause subjects, and influence of verbal meaning on choice of these verbal affixation if possible.
As previous studies have shown, verbal meaning or verb classes do have some influences on the voice alternations and thematic role of participant(s) encoded by its argument(s) (Huang and Huang, 2007). When verbal meaning is an influencing factor, the study of Yami verbal affixes should take semantic criteria into consideration. The revised macro-role hierarchy proposed by Foley (2005) that is built based on Dowty’s (1991) proto-roles is employed to determine the thematic role of clause subjects in the present study as Dowty’s proto-roles are determined solely based on verbal semantics.
In the following section, the framework that is employed to re-examined Yami verbal affixed and the thematic role of clause subjects are discussed, and the terminologies that are used in the present study are defined.
2.5 Framework
The theory of thematic roles has been developed for decades and inventory of thematic roles is numerous, as different terminologies have been used to refer to similar notions. A unified theory of thematic roles is hardly found; hence, the choice has to be made among the numerous inventories of thematic roles when one aims to explore a language via a
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
theory of thematic roles. A thematic role of participants is determined by the meaning of the verb, and different classes of verbs require participants with different thematic roles, such as agent, experiencer, patient, theme, …etc. Certain thematic roles are found to only co-occur with certain classes of verb. For instance, verbs of emotion involve an experiencer, but never an actor. The thematic role list keeps expanding, as different classes of verbs require different roles. This leads to a problem of an unlimited list of thematic roles. Dowty (1991) has
recognized this problem and suggested a solution for it. He has proposed that thematic roles should not be treated as discrete roles, but as two cluster-concepts—agent and proto-patient—and these two role types can describe an argument selection efficiently. He has proposed a list of entailments that characterize these two cluster-concepts, given in (37) and (38).
(37) Contributing properties for the Agent Proto-Role:
a. volitional involvement in the event or state b. sentience (and/or perception)
c. causing an event or change of state in another participant d. movement (relative to the position of another participant) e. (exists independently of the event named by the verb)
(Dowty 1991: 572)
(38) Contributing properties for the Patient Proto-Role:
a. undergoes change of state b. incremental theme
c. causually affected by another participant
d. stationary relative to movement of another participant e. (does not exist independently of the event, or not at all)
(Dowty 1991: 572)
In addition to proto-role entailments, Dowty has proposed an argument selection principle for these two roles, provided in (39). Subject and object of a transitive clause are determined by the number of entailments an argument has. An argument with the greatest number of proto-agent entailments is realized as a clause subject, and an argument with the greatest number of proto-patient entailments is realized as a clause object. The entailments within Dowty’s framework are not ranked. The role determination is solely based on the
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
number of proto-role entailments an argument has. Note that Dowty’s argument selection principle is for nominative-accusative languages, not for ergative-absolutive languages.
(39)
ARGUMENT SELECTION PRINCIPLE: In predicates with grammatical subject and object, the argument for which the predicate entails the greatest number of Proto-Agent properties will be lexicalized as the subject of the predicate; the argument having the greatest number of Proto-Patient entailment will be lexicalized as the direct object.
(Dowty 1991: 576) Foley and Van Valin (1984) have proposed macro-roles—actor and undergoer to account for the problem with multitudinous semantic roles. The difference between Dowty’s proto-role and Foley and Van Valin’s macro-role is the assignment of the roles. In Dowty’s framework, the role determination is solely based on the number of entailments that a verb has imposed on an argument. In Foley and Van Valin’s (1984) framework, the role is read off from the semantic specification of the verb’s argument in a lexical decomposition of the verbal meaning (e.g. DO (x,…)).
For the purpose of assigning a macro-role to arguments of a verb, the semantic roles available to arguments are arranged along a hierarchy in Foley and Van Valin (1984). Within this hierarchy, the most agent-like role is at the top of the hierarchy and the most patient-like role is at the bottom, as in Table 13.
Table 13 Foley and Van Valin (1984) macro-role hierarchy Agent (volitional)
Effector (causer)
Locative Actor Undergoer
Theme (movement) Patient (change of state)
(Foley 2005: 390)
Foley (2005), who investigated the mapping problem of intransitive verbs across various Austronesian languages, has revised the macro-role hierarchy by inserting Dowty’s entailments beside the original semantic roles in parentheses in Table 13. Unlike Foley and Van Valin (1984), the proto-agent or proto-patient assignment is based on the number of entailments a participant has within Dowty’s theory of proto-role entailments. Foley (2005:
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
388) has pointed out that some of the entailments are weighted more heavily than the others on Dowty’s list, so he has suggested that these entailments should be ranked accordingly.
Foley takes the English verbs hit and break discussed in Fillmore (1970) to support his proposal on ranking Dowty’s proto-role entailments. The verbs hit and break share many semantic similarities, but their accessibility to the undergoer macro-role differs, as illustrated by the examples in (40). Sentences (a) and (b) are synonymous, whereas (c) and (d) are not.
The undergoer of the verb break, cane or vase, in sentence (c) and (d), must undergo a change in state, but the undergoer of the verb hit, cane or vase, in (a) and (b) is causally affected, does not undergo a change in state. Hence, Foley (2005) has proposed that a ranking among Dowty’s proto-role entailments can better describe the subtle differences between verbs of hitting and breaking.
(40)
a. Fred hit the cane against the vase.
b. Fred hit the vase with the cane.
c. Fred broke the cane against the vase.
d. Fred broke the vase with the cane.
(Foley 2005: 388)
Based on Dowty’s theory of entailments for a verb’s argument, Foley takes his and Van Valin’s (1984) earlier work of macro-role hierarchy and works out the equivalents between the two systems. He has come up with a revised version of macro-role hierarchy to examine intransitive verbs of various Austronesian languages (Foley 2005). The revised macro-role hierarchy is provided in Table 14. Both systems do not completely match up, but according to Foley (2005), this revised macro-role hierarchy works well in predicating and describing the different distributions of the unaccusative-unergative split across the four groups of Austronesian languages that have been surveyed.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 14 Revised macro-role hierarchy Actor: volitional performer
causing an event or change of state sentience
movement A U
stationary
causally affected incremental theme
Undergoer: undergoing a change in state
(Foley 2005: 391)
The macro-roles within Van Valin and Foley’s framework are read off from the semantic specifications of verb’s arguments in lexical decomposition of verb while Dowty’s proto-roles are determined by the verb’s meaning imposed on its arguments. For Foley and Van Valin (1984), the macro-roles are categories within a distinct module of grammar, but for Dowty proto-roles are not considered as part of the grammatical apparatus. Foley (2005) has adopted Dowty’s approach and the properties listed on the revised macro-role hierarchy are determined via the theory of entailments proposed by Dowty while the term actor and undergoer are used instead of agent and patient. Note that Ross (2015) has adapted Foley’s revised macro-role hierarchy and modified and reordered the properties while he works on reconstruction of Formosan verb classes. Ross’ (2015: 288) proto-role hierarchy is
reduplicated in Table 15 for comparison. The property of being in a state is included in patient-end of Ross’ hierarchy. The semantic properties of argument in Ross’ hierarchy have been adjusted and re-ordered to account for data in Formosan languages. The original ranking of the semantic properties of arguments is included in the square brackets and its order is given in numbers (1=volitional performer; 2=causing an event or a change of a state;
3=sentient; 4=moving; 5=stationary; 6=causally affected; 7=incremental theme;
8=undergoing a change in state).
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 15 Proto-role hierarchy, adapted from Foley (2005: 391)
(Ross 2015: 288) The present study will adopt Foley’s revised macro-role hierarchy and follow his terminologies—macro-roles (actor and undergoer). The incremental theme is omitted as an example of it was not found in our data. Following Ross (2015), the property of being in a state is added to the undergoer end of hierarchy as arguments in our data have exhibited this property. Our revised version of macro-role hierarchy is provided in Table 16. Note that this revised macro-role hierarchy is a modified version of Foley’s (2005) revised macro-role hierarchy.
Table 16 The modified version of Foley’s (2005) revised macro-role hierarchy Actor: volitional performer
causing an event or change of state sentience
movement A U
stationary
causally affected
Undergoer: undergoing a change in state or being in a state
The aim of the present study is to investigate the function or functions of Yami verbal affixes that show some relation with the thematic role of the clause subject. First, the clause containing these affixes will be investigated and its categories and functions are verified via different syntactic constructions including declarative, negative, negative imperative, and interrogative. The semantic property of the clause subject is examined based on Dowty’s and Foley’s concept of discrete roles and modified version of Foley’s (2005) revised macro-role hierarchy in the following chapter while the clauses containing different verbal affixes in Yami are re-examined.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
The Yami verbal affixes that are to be discussed in the present study include Ø, m-, mi-, om-/<om>, ma-, man-/mang-, -en, -an, i-, and ka- -an. Ho (1990), Rau and Dong (2006), and Y. Chang (2011) treat these affixes as focus or voice affixes and Shih (2013) considers them as primary affixes that alter valency of the verbs they are attached to. The major function of these verbal affixes is claimed to signal the thematic role of the clause subject.
Other functions of these affixes that have been mentioned or discussed in previous studies include signaling potentiality or participating in applicative construction. The aim of the present study is to show that these verbal affixes participate in different paradigms—
encoding voice, tense, aspect, or mood information at the same time, and they might be multi-functional, and are sensitive to lexical properties of roots/stems they attach to. The aim is achieved by investigating their conjugations, meanings, semantic properties of involved participants, and case marking of involved participants within different types of clauses including negative, negative imperative, and interrogative constructions.