Classification of Source-Domain and Target-Domain Concepts

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

bodily experience or socio-cultural practices to comprehend such metaphor. The sentence

‘The foundation of his theory is strong’ is an expression of the complex metaphor THEORIES

ARE BUILDINGS, which comprises the primary metaphors: STRUCTURES ARE PHYSICAL

STRUCTURES and PERSISTING IS REMAINING ERECT (Gibbs 2005: 79). There is no

straightforward connection between THEORIES and BUILDINGS. Yet the knowledge of the

experience that buildings remain erect due to their strong structure and foundation motivates

the metaphorical inference of THEORY.

In this section, nine kinds of metaphor are presented: body-part metaphor, causation

metaphor, conduit metaphor, container metaphor, entity metaphor, fictive-motion metaphor,

orientation metaphor, personification metaphor, and complex metaphor. The present study

analyzes the metaphoric expressions based on the definitions provided in this section. Except

for the body-part metaphor and the personification metaphor, the other kinds of metaphors

are produced from the current data.

3.4 Classification of Source-Domain and Target-Domain Concepts

The present study also attempts to investigate the sources and targets involved in the

metaphors. To examine the sources employed in the metaphoric expressions, the notion of

image schema is taken into account in the present study. Image schema is a recurring pattern

of our body experiences and has been studied as the embodied sources for metaphoric

extensions in past research. Hence, the source-domain concepts are categorized based on the

‧

image schemas. It is proposed that an image schema may relate to other image schemas

(Johnson 1987; Cienki 1997; Clausner & Croft 1999; Santibáñez 2002). Based on the

selective image schemas proposed by Johnson (see Table 3 in Chapter 2), Cienki (1997)

suggested some schemas are more general and represent cover terms for the more specific

schemas as shown in Table 4. For instance, OBJECT schema has several specified version of

schemas like PART-WHOLE, MASS-COUNT, etc.

Table 4. Image schemas proposed by Cienki (1997: 12)

More general More specific

PROCESS MATCHING MERGING CONTACT LINK

SPLITTING

PATH STRAIGHT SCALE CYCLE ITERATION

OBJECT PART-WHOLE CENTER-PERIPHERY SURFACE MASS-COUNT

COLLECTION

CONTAINER FULL-EMPTY SURFACE CENTER-PERIPHERY IN-OUT

FORCE ATTRACTION ENABLEMENT COMPLUSION

COUNTERFORCE BLOCKAGE RESTRAINT REMOVAL

Clausner and Croft (1999: 16-20) proposed that image schemas are a subtype of

domain called “image schematic domain” which has internal structure and supports profiles

of different concepts. They offered an inventory similar to Cienki’s list and suggested that the

image schemas are related. The SPACE schema is added as a more general image schema

which involves several specific image schemas like UP-DOWN, FRONT-BACK, and NEAR-FAR

in Clausner and Croft’s study. Santibáñez (2002) put emphasis on the OBJECT schema and its

dependent schemas. He argued that the OBJECT schema is a basic image schema that

comprises other more specific embodied patterns (e.g., PART-WHOLE,MASS-COUNT). The

‧

OBJECT schema in Santibáñez’s study includes the specific schemas (e.g., MATCHING,

MERGING, CONTACT, LINK, and SPLITTING) which are grouped as the sub-schema of the

PROCESS schema in Cienki’s study. The OBJECT schema is based on the experience that we

can manipulate objects which may change their property or relations with other objects. This

knowledge of OBJECT provides a guideline for the LINK schema, which consists of two or

more objects connected by some linking devices.

Although the grouping of image schemas is varied in different research projects, there

is a consensus that some image schemas are subordinate to the more general image schemas.

Following the framework of the above studies, the present study utilizes the more general

image schemas (PATH, OBJECT, CONTAINER, FORCE, and SPACE) to classify the source

-domain concepts. ⁵ The present study also identified BODY-PARTS, PERSON, and

FICTIVE-MOTION as the sources. BODY-PARTS and PERSON are based on the OBJECT image

schema. The two concepts are the elaborations of a basic object; hence, BODY-PART and

PERSON are separated from OBJECT in the present study. The concept of fictive-motion

relates to the SELF-MOTION image schema proposed by Mandler (1992); this image schema

involves a trajectory that moves on its own. Moreover, it should be noted that Johnson and

Lakoff did not claim that all metaphors require an image-schematically based source domain.

The complex metaphors are not directly associated with our sensory-motor experiences, and

5 The present study follows Santibáñez’s work to treat the image schemas like CONTACT, LINK, and SPLITTING (which were regarded as the specific schema of the PROCESS schema by Cienki) as the specific schema of the more general OBJECT schema.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

they may not be directly related to image-schematic patterns. In the current data, all the

complex metaphors are based on physical activities we practice in ordinary life (e.g., picking

an object and weighing an object); therefore, ACTIVITY is identified as a kind of source. To

generalize, 9 kinds of sources are identified: ACTIVITY, BODY-PART, CONTAINER,

FICTIVE-MOTION, FORCE, OBJECT, PATH, PERSON, and SPACE. Except for BODY-PART and

PERSON, the other kinds of sources can be found in the current data.

A great variety of notions can be realized via metaphors. A target domain is typically

concrete; in some cases, a target can be abstract. Among the 247 metaphoric expressions

examined in the present study, 197 (79.8%) of them involve abstract targets (e.g., STATE and

TIME) and 50 (20.2%) of them include concrete targets (e.g., ACTIVITY and COLOR). In the

current data, 61 kinds of targets were found; however, not all of the target-domain concepts

appear frequently. Regarding the issue about the targets realized in habitual metaphoric

expressions, the present study respectively identified 8 kinds of targets that have at least five

tokens in the current data: GROUP, MENTAL ACTIVITY, (physical) ACTIVITY, DEGREE,

SEQUENCE, SPEECH CONTENT, STATE, and TIME. The remaining 53 kinds of targets will be

generalized as ‘others’.