Chapter 2
The Contemporary Theory: A Review
2.1. Views of metaphor: Classical vs. contemporary
1n classical theories, there are three. main views of metaphor: the comparison
view, the substitution view, and the interaction view. The comparison view can
be traced back to Aristotle, who regarded metaphors as implicit comparisons
between a metaphorical expression and a literal paraphrase based on underlying
analogy or similarity. The substitutiorì view, ofwhich the comparison view
is a special case according to Black (1962, 1993 [1979]), holds that a metaphor
is where a metaphorical expression is used in place of some equivalent literal
expression.' The interaction theory, proposed by Black (1962, 1993 [1979]),
states that metaphorical meaning is a result of an interaction between a metaphorical
expression, termed "focus," and its"surrounding literal 企ame" (1993
[1979]: 27). AlI these views share a common feature: they view metaphor ~ a
linguistic phenomenon, and assume a fundamental distinction between liter创
and figurative (or metaphorical in its broad sense) senses.s
As Lakoff (1994) poi附 out， a major difference between the contempor町
theory of metaphor and the classical ones is based on the old literal-figurative
distinction. Traditionally, the word ‘ literal' is defined in terms of "an idealized
and oversimplified model of language and thought" to include all of the following
four senses (Lakoff 1986b: 292):
Literal 1, or conventional literality: ordinary conventional language一-contrasting
with poetic language, exaggeration, approximation, embellishment,
excessive politeness, indirectness, and so on.
Literal 2, or subject matter literality: language ordinarily used to talk about
some domain ofsubject matter.
A REVIEW 11
Literal 3, or nonmetaphorical literality: directly meaningful language-nol
language that is understood, even partly, in terms of something else.
Literal 4, or truth-conditional literality: language capable of ‘ fitting the
world' (i.e. of referring to objectively existing objects or of being obj饵，
tively true or 也lse).
Going with the four-sense definition of ‘literal' is the following set of assumptions
that has been proved to be false (L也off1994: 43-44):
1. All everyday conventionallanguage is literal, and none is metaphorical.
2. All subject matter can be comprehended literally, without metaphor.
3. Only literallanguage can be contingently true or false.
4. All definitions given in the lexicon of a language are literal, not meta-
phorical.
5. The concepts used in the grammar of a language are alllit巳ral; none is
metaphorical.
The traditional definition of the word ‘ literal' is wrong, according to Lakoff
(1994), because a huge system of conventional, conceptual metaphors has been
discovered structuring our everyd:1y conceptual system and pervading our everyday
language. "The discovery of this enormous metaphor system has destroyed
the traditional literal-figurative distinction, because the term literal, as
used in defining the traditional distinction, carries with it all those false assumptions"
(p. 44). Assuming the literal-figurative distinction, the traditional
theory held that metaphor was mutually exclusive with the realm of ordinary
everyday language. Over the centuries, the classical theory of metaphor was
taken so much for granted that it came to be taken as 气iefinitional': "The word
‘metaphor' was defined 'as a novel or poetic linguistic expression where one or
more words for a concept are used outside of its normal conventional meaning
to express a similar concept" (p. 42).6
The contemporary theory of metaphor, as Lakoff (1993a: 244) argues, "is
revolutionary in many respects." Lakoff and Johnson (1980: 3) summarize the
contrast between the traditional and contempor缸y views of metaphor as fol-
lows:
Metaphor is for most people a device of the poetic imagination and the rhetorical
flourish-a matter of extraordinary rather than ordinary language. Moreover,
metaphor is typicalJy viewed as characteristic of language alone, a matter of
12 THE CONTEMPORARY THEORY OF METAPHOR
words rather than thought or action. For this reason, most people think they can
get along perfectly well without metaphor. We have found, on the con位ary， that
metaphor is pervasive in everyday life, not just in language but in thought and
action. Our ordinary conceptual system, in terms of which we bòth think and act,
is fundamentally metaphorical in nature.
In this way, Lakoff and Johnson have redefined the tenn metaphor. Since they
argue that human thought processes are largely metaphorical, and that the human
conceptual system is metaphorical1y structured and defined, metaphor in
their sense is no longer a way of expression, but also a way of conceptualization.
The tenn metaphor has come to mean "a cross-domain mapping in the
conceptual system" (Lakoff 1994: 43). Defined as such, metaphor is in effect
ubiquitous in everyday language and thought.
In the contempor缸y theory of metaphor, as Lakoff (1986b: 293) has suggested,
the tenn literal is restricted to the meaning of Literal 3; "the sense of
being directly meaningful, without the intervention of any mechanism of indirect
understanding such as metaphor or meton沪ny." With such a definition, it
can be said that, a1though many abstract concepts are metaphorical in character,
a significant part of our concepωal system consists of nonmetaphorical
concepts. For example, sentences such as ‘The bal100n went up' and <The cat is
on the mat' are not metaphorical. "But as soon as one gets away 企om concrete
physical experience and starts talking about abstractions or emotions, metaphorical
understanding is the nonn" (Lakoff 1994: 44).'
2.2. Cognitive linguistics and cognitive semantics
In the discipline of linguistics, the contemporary theory of metaphor is closely
associated with cognitive linguistics, which comprises cognitive grammar (e.g.
Langacker 1987, 1988a, 1991) and cognitive semantics (e.g. John~on 1987,
Lakoff 1987a, 1988, Sweetser 1990, Turner 1991). As a new school, cognitive
linguistics departs 企om the mainstream generative linguistics in comrnitments
and background assumptions (Lakoff 1989a, 1990, 1991). Gentrative linguists
make a distinction between competence and performance, ke叩ing their focus
ofstudy on competence, that is, on the intemal r，叩resentation ofrules 也at generate
grammatical sentences in the ideal speaker-hearer. G阳erativists see
metaphor as deviant and parasitic upon nonnal language, believing that it cannot
be studied in any reasonable or systematic way. Parmegiani (1988: 2) has
A REVIEW 13
noted that to generative linguists metaphor is but "a kind of semi-grammatical
phenomenon" violating semantic rules. It should therefore be brushed aside
into the domain of rhetoric, stylistics, or pragmatics. In short, in generative linguistics,
metaphor, viewed as a semantically-deviant phenomenon, is either
excluded from its study or relegated to the 企inges of attention. But, since
metapnor pervades human language and thought, any science of language or of
the mind which excludes any consideration of, or pays little attention to, metaphor
wil1 tum out to be far too narrow or of little lasting value (Danesi 1988b,
Mac Cormac 1985).
Cognitive linguistics, in contrast, "sees language as making use of conceptual
structure and general cognitive mechanisms" (Pütz 1992b: lii). The cognitive
paradigm holds a set of common views on language and cognition including
the following (Gibbs 1996a, Radden 1992, Rudzka-Ostyn 1993)扎It believes
that naturallanguage is a product ofthe human mind, based on the same
organizing principles that operate in other cognitive domains. As one domain
of human cognition, language is intimately linked with other cognitive domains
and as such mirrors the interplay of psychological, cultural, social, ecological,
and other factors. Language stru;ture depends on (and itself influences)
conceptualization, the latter being conditioned by our experience of
ourselves, the extemal wor1d, and our relation to that world. In other words,
language is not just a system consisting of arbitrary signs，阻d its structures are
related to and motivated by human conceptual knowledge, bodily experience,
and the communicative functions of discourse. Linguistic units are subject to
categorization which commonly gives rise to prototype-based networks and
critically involves metaphor and metonymy. Meanings of linguistic units are
based on embodied experience with and within the real world, and can be characterized
with respect to relevant knowledge structures such as those called
folk models,cultural models,or cognitive models.
As Fesmire (1994b: 150) has summarized, having departed 企om the mainstream
generative linguistics, cognitive linguistics "grapples with how human
beings actually make sense oftheir wor1d," and "dwells in the stream of human
experience rather than in a supposedly pure realm of form," thus "cultivating a
theory of the ecology of human understandi
14 THE CONTEMPORARY THEORY OF METAPHOR
models that underlie and mak:e possible our language activities. It hypothesizes
that our ‘higher' cognitive functions that are supposed to mak:e meaning and
reasoning possible are indeed continuous with and ins叩arable 企'om our sensorimotor
activities (Johnson 19自9a). It claims that knowledge is embodied: our
very conceptual system is grounded in and structured by various r回urring pattems
of our perceptual interactions, bodily orientations, and manipulations of
objects (Johnson 1993b). A central task of cognitive semantics, then, is to examine
the empirical evidence for embodied knowledge of this kind. To fulfil1
this task, cognitive semantics has worked out methods of analysis that mak:e it
possible to investigate the experiential grounding ofour conceptual system and
its irreducibly imaginative character. As Johnson (1989a: 112) claims, "What
is new in cognitive semantics is 也e way in which it has been able to be more
concrete and specific about the way in which structures of our perc叩tual interactions
work their way up into our understanding of more abstract conceptual
domains."
The contempor缸y theory of metaphor should be viewed as product of 伽is
endeavor of cognitive linguistics in general and of cognitive semantics in p缸，
ticular.
2.3. Conceptual and Iinguistic metaphors and metaphor systems
What mak:es the contempor:缸y theory of metaphor unique is the important distinction
that has been drawn between cODceptual metaphors or metaphorical
cODcepts on one hand, and Iinguistic metaphors or metaphorical expresSiODS
on the other hand (Lak:off and Johnson 1980). The former refers to those
abstract notions such as ARGUMENT IS WAR and LOVE IS A JOURNEY while the
latter is actual linguistic phrases that realize or instantiate those notions in one
way or another. Metaphor, according to 也is 也eory， is fundamentally ∞nceptual
rather than linguistic in nature. Metaphorical language, consisting of specific
linguistic expressions, is but a surface manifestation or realization of conceptual
metaphor. Conc叩tual metaphors are systematic mappings across conceptual
domains: one domain of experience, the sourçe domain, is mapped
onto another domain of experlence, the target domain. "In short, the locus of
metaphor is not in language at al1, but in the way we conc叩tualize one mental
domain in terms ofanother" (Lak:off 1994: 43).
A major discovery of Lak:off and Johnson is that people use metaphorical
expressions in a systematic way because metaphorical conc叩ts are systematic.
AREVIEW 15
For instance, under the metaphorical concept LOVE IS A JOURNEY, which is discussed
in detail in Lakoff (1986a, 1994), and Johnson (1993坊， there are these
metaphorical expressions which are highly conventionalized in daily English:
(1) LOVE IS A JOURNEY.
a. Look how far we've come.
b. It's been a long, bumpy road.
c. We can't tum back now.
d. We're at a crossroads.
e. We may have to go our separate ways.
f. We're spinning our wheels.
g. The relationship isn't going anywhere.
h. Our relationship is off the track
i. The marriage is on the rocks.
As is shown; conventional metaphorical expressions are govemed by concep明
tual metaphor in a syst巳matic way. Lakoff argues that conceptual metaphor reflects
"a general principle". that is "part of the conceptual system underlying
English" (1994: 45一46). 1 his principle for understanding the domain of love in
terms ofthe domain ofjoumeys can be stated as "a metaphorical scenario" (p.
46):
ηle lovers are travelers on a joumey together, with their common life goals seen
as destinations to be reached.ηle relationship is their vehicle, and it allows them
to pursue those common goals together. The relationship is seen as fulfilling its
purpose as long as it allows them to make progress toward their common goals.
The journey is noteasy. There are impediments, and there are places (crossroads)
where a decision has to be madl. about which direction to go in and whether to
'keep traveling together.
The metaphor here is thus a conceptual mapping 企om a source domain (journey)
to a target domain (love), with both ontological correspondences and
epistemic correspondences entailed by the mapping. The ontological coπ。
spondences are those in which the entities in the source domain are mapped
onto the entities in the target domain, while the epistemic co汀espondences 缸e
tho臼 in which knowledge of the sour臼 domain is mapped onto knowledge of
the target domain to form inference pattems. With the LOVE AS JOURNEY meta-
16 THE CONTEMPORARY THEORY OF METAPHOR
phor, for instancc, thc ontological correspond巳nces between the two domains
are as follows (from Johnson 1993b: 417):
(2) a. The lovers correspond to travelers.
b. The love relationship corresponds to th巳 vehicle.
c. The lovers' common goals correspond to their common destinations
on the joumey.
d. Difficulties in the relationshi~ correspond to impediments to travel.
Thc LOVE AS JOURNEY metaphor is thus a systematic mapping based on ontological
corrcspondences. It gives rise to a set of epistemic correspondences in
which knowledge of the source domain (joumeys) is mapped onto knowledge
of the target domain (love). "Consequently, the way we conceptualize, reason
about, and talk about our love relationship will be deterrnined by these, and
other, epistemic correspondences" (Johnson 1993b: 417).
In sum, each metaphorical mapping at thc conceptual level is a fixed set of
ontological corrcspondences between entities in the source domain and those in
the target domain. Once the fixed correspondenιes are activated, mappings can
project source domain inference pattems onto target inference pattems, resulting
in epistemic correspondences (Lakoff 1993). In such a way, conceptual
metaphors in our conceptual system forrn intricate systems. Lakoff (1994: 41-
42) believes that the study of systems of conventional conceptuaI metaphor is
"the most elaborate and conceptually radical branch of contempor町y conceptual
system research." Lakoff 皿d Johnson's (1980) methodology has demonstrated
that such studies can be accomplished by close examinations of linguistic
metaphors instantiating the underlying conceptual metaphors in our
conceptual systems. That is, one can gain an understanding of the nature of
human concepts by systematically studying linguistic expressions. With a new
definition of metaphor given, Lakoff and Johnson have also provided a new
methodology that makes it possible to study metaphor in a systematic way.
Although it is still under debate as to whether and to what extent this new research
methodology can actually reveal underlying conceptuaI systems (see,
e.g., Gibbs 1996a, 1996b, Gibbs and Colston 1995, Murphy 1996, 1997, Ortony
1988), there is no doubt that it has been the most productive and influential
approach to date.
In the above it is shown that metaphorical expressions are systematically
tied to a conceptual metaphor, with each of the forrner as a particular linguistic
instantiation or manifestation of the latter. That is, each concepω111 metaphor
AREVIEW 17
heads and govems a system of linguistic metaphors. The system of metaphor is
highly structured by its ontological and epistemic correspondences operating
across conceptual domains. The systematicity of metaphor, however, exists in
a larger scope than described above. Not only are metaphorical expressions
systematically govemed by a conceptual metaphor, but conceptual metaphors
may also be systematically related to each other to form a hierarchical structure.
"Metaphorical mappings do n,ot occur isolated 企om one another. They are
sometimes organized in hierarchical structures, in which ‘ lower' mappings in
the hierarchy inherit the structures of the ‘higher' mappings" (Lakoff 1994:
62). Lakoff calls this phenomenon ‘metaphor inheritance hierarchies.' Given
below is an example of such a hierarchy including the LOVE IS A JOURNEY
metaphor (adopted 企om Lakoff 1994: 62):
Levell:
Level2:
The Event Structure Metaphor
LIFE IS AJOURNEY
Leve13: LOVE IS AJOURNEY; ACAREER IS AJOURNEY
Here the two versions of metaphor at Level 3一-LOVE IS A JOURNEY and A
CA阻ER IS AJOURNEY-inherit the structure ofthe higher mapping at Level 2一
LIFE IS A JOURNEY-which is a more general metaphor containing the two
metaphors at Level 3 as its more specific manifestations. The LIFE IS A
JOURNEY metaphor may contain the following ontological correspondences or
metaphorical mappings (from Winter 1995: 235):
The LIFE IS AJOURNEY Metaphor
JOURNEY (Source)
traveler
point ofdeparture
initial conditions
baggage
obstacles
distance
distance covered
destination
termmat10n
--•-…+
一一-梦
一一~
~
一-→
-一-+
~
~
As Winter (1995: 235) points out:
LIFE (Target)
person
birth
personal endowments
personal problems
extemal difficulties
duration
accomplishments
life purpose
death
18 THE CONTEMPORARY THEORY OF METAPHOR
The ‘LIFE IS AJOURNEY' metaphor enables many different metaphorical expressions
and 阴阳ms of inference. Thus, we try to give our children an education so
they will get ‘a good start' in 1ife. If they act out, we hope that they are ‘just going
through a stage' and 让lat they will ‘ get over it.' As adults, we hope they won't
be ‘ burdened' (or ‘ saddled') with financial worries or ill health and, if they face
such difficulti邸， that they will be able to ‘overcome' them. We hope they wi11
have a ‘long 1ifespan' and that they wi11 ‘go far in life.' We know that, as mo阳Is，
they will ‘ go to their final resting place.'
Lakoff (1994) has cited the following English ex缸nples under the LIFE IS A
JOURNEY metaphor:
(3) LIFE IS AJOURNEY
a. He got a head start in life.
b. He's without direction in his life.
c. I'm where 1want to be in life.
d. I'm at a crossroads in my life.
e. He'll go places in life.
f. He's never let anyone get in his way.
g. He's gone through a lot in life.
The conceptual metaphor LIFE IS AJOU阳EY can thus summarize and account for
many English idiomatic expressions such as cited above.
Since love is an important aspect of life, the LOVE IS A JOURNEY metaphor,
therefore, inherits the structure of the LIFE IS A JOU刚EY metaphor. As Lakoff
(1994) points out, what is special about the LOVE IS AJOURNEY metaphor is that
there are two lovers, who are travelers, and that the love relationship is a vehicle,
while the rest of the mapping is a consequence of the LIFE IS A JOURNEY
metaphor. In a similar vein, a career is another important aspect of life. So the
CAREER IS A JOURNEY metaphor inherits the structure of the LIFE IS A JOURNEY
met呐。r just as the LOVE metaphor does. What is special about the CAREER
metaphor, however, is that a successful career is always a joumey UPWARD,
smce STATUS IS UP.
As shown above, the LIFE IS AJOURNEY metaphor is but a Level-2 metaphor,
on top of which at Level 1 is the Event Structure Metaphor. According to Lakoff
(1994), the Event Structure Metaphor has events as its target domain and
space as its source domain. Its general mapping in terms of ontological correspondences
goes as follows (Lakoff 1994: 62):
AREVIEW 19
(4) a. States are locations (bounded regions in space).
b. Changes are movements (into or out ofbounded regions).
c. Causes are forces.
d. Actions are self-propelled movements.
e. Purposes are destinations.
f. Means are paths to destinations.
g. Difficulties are impediments to motion.
h. Expected progress is a travel schedule; a schedule is a virtual traveler,
who reaches prearranged destinations at prearranged times.
i. Extemal events are large, moving objects.
j. Long-term, purposeful activities are joumeys.
It should be apparent that the LIFE IS A JOURNEY metaphor makes use of all the
structure ofthe Event Structure Metaphor, since events in a life are subcases of
events in general. Lakoff(1994: 62-63) describes the coherence ofinference in
the Event Structure Metaphor and the inheritance involved in the LIFE IS A
JOURNEY metaphor as follows:
In our culture, 1ife is assumed to be purposeful, that is, we are expected to have
goals in life. In the event structure metaphor, purposes are destinations and purposeful
action is selιpropelled motion toward a destination. A purposeful life is a
long-term, p山poseful activity, and hence a joumey. Goals in life are destinations
on the joumey. The actions one takes in life are self-propelled movements, and
the tota1ity of one's actions form a path one moves along. Choosing a means to
achieve a goal is choosing a path to a destination. Difficulties in 1ife are impediments
to motion. Exter.lal events are large moving objects that can impede motion
toward one's 1ife goals. One's expected progress through life is charted in
terms of a life schedule which is conceptua1ized as a virtual traveler that one is
expected to keep up with.
Considering the above three-level metaphor system and metaphorical expressions
that realize it linguistically, the ubiquity and systematicity of metaphor in
human language and thought should be obvious. Little wonder that the contempor缸y
theory of metaphor claims that metaphor is one of the essential elements
constituting and structurin京 human cognition.8
Lakoff (1994: 64) suggested that the hierarchical organization is a very
prominent feature of metaphor system, and that "the metaphors higher up in
the hierarchy tend to be more widespread than those mappings at lower levels."
20 THE CONTEMPORARY THEORY OF METAPHOR
He proposed the Event Structure Metaphor as his "candidate for a metaphorical
universal" (p. 88). In Chapter 5 below, 1 will make a detailed study of the
Event Structure Metaphor in Chinese. My study supports Lakoffs proposal
企om the perspective ofChinese.
Now, the fundamental distinction between the traditional and contemporary
theories of metaphor is obvious. The traditional approach studies metaphor as
individual linguistic expressions or rhetorical devices: what artistic or aesthetic
effects they have produced in a particular piece ofdiscourse, which is primarily
literary or poetic in nature. Just as Lakoff (1987d: vii-viii) points out, "traditional
theories of metaphor assume that metaphors occur one by one, that each
distinct metaphorical expression is individual1y created." In the contempor缸y
paradigm, on the other hand, metaphor is studied as systems of human conceptualization,
operating deep in human thought and cognition and, at the same
time, surfacing in everyday language in a systematic manner! on this view,
metaphor in poetηor in literature at large is but a special case of metaphor in
general, based on the same mechanisms (Gibbs 1994, Lakoff and Turner 1989,
Lakoff 1990, 1993a, 1994, Sweetser 1992, Tumer 1987, 1991). Section 2.6
wil\ retum to this claim.
2.4. Experiential basis of metaphor: The notion of embodiment
As Johnson (1987) and Lakoff (1987a) have argued, in the past the dominant
philosophical tradition in the West was objectivism. In his review of Lakoff
(1987吟， Langacker (1988b: 384) describes the predominant status ofobJectivism
in the Westem scholarly community as being "so pervasive as to be almost
invisible, and so fundamental as to be virtual1y immune to chal1enge." According
to the objectivist doctrine (Hampton 1989, Johnson 1987, 1989a, Lakoff1987a,
etc.), the world consists of mind-independent 0时ects which have
determinate properties and stand in definite relations to each other. The nature
ofthese objects is independent ofthe ways in which people experience and understand
them. The world, therefore, can be described objectively, independent
of any particular culture or observer's viewpoint. That is, there exists a God'seye
view of reality. Meaning, according to this view, is an abstract relation
between symbolic representations and objective reality. The symbols are 红bitr缸y
and meaningless in themselves, but supposedly given meaning by virtue
of their capacity to correspond to things, propert邸， and relations existing objectively
in the world out there. Meaning, defined as the relation between
AREVIEW 21
words and those things in the world to which they refer, is thus fundamentally
literal, holding a one-to-one or mirror-image relationship with the extemal
world. It follows that there can be no irreducibly figurative or metaphorical
concepts, because metaphorical projections cut across basic experiential domains,
and such cross-categorical projections are held to have no counte叩arts
in the real world, which supposedly has discrete and definite categorical
boundaries. The task for semantics, according to objectivism, is to describe the
way in which words and utterances correspond to the real world. Human understanding,
on the other hand, is distinguished 仕om meaning, which is held to
be objective and in no way dependent on any person's or community's understanding
of it. ldeal understanding proceeds by building an intemal representatÎon
that correct1y mirrors extemal reality. Reason is just the mechanical manipulation
of abstract symbols which are meaningful only via conventional correspondences
to things in the world. Correct reason merely mirrors the logic of
the extemal world.
The objectivist paradigm so described, as Lakoff points out (l987a: 157一
158), is "an idealization," brought from "our intellectual background into the
foreground." Langacker (1988b: 388) further points out:
the power of the objectivist world view does not depend on anybody accepting it
in its entirety. Rather, it works its influence through the pervasiveness and tacit
acceptance of numerous attitudes, working assumptions, and methodological
principles for which it can be recognized as the ultimate source. Even if, in its
pure form, the objectivist philosophy is universally rejected, it is nonetheless the
reference point with respect to which the actual world views of many scholars
can be measured-it stands as the archetype that gives these views their coherence,
shapes their research agenda in terms of both subject matter and approaches,
and determines whether an idea is adopted as a default-case assumption
or considered inherent1y suspect.
While criticizing 0均 ectivism， Lakoff(1987a) outlined an altemative called
experieJ1tialism or experiential realism. According to him, experientialism and
objectivism are two versions of ‘basic realism,' which is featured by a commitment
to the existence of a real world and stable knowledge of it, and by a
rejection ofthe view that theconception oftruth is merely ba~ed on intemal or
subjective coherence, yielding the view that ‘m灿ing goes. fHowever, experientialism
differs 企om objectivism in the definition of meanitìg. "Where objectivism
defines meaning independent1y of the nature and experience of thinking
22 THE CONTEMPORARY THEORY OF METAPHOR
beings, experiential realism characterizes meaning in terms of embodiment,
that is, in terms of our collective biological capacities and our physical and social
experiences as beings functioning in our environment" (Lakoff 1987a:
266-267). Here the key concept is emþodiment, a notion that has been most
forcefully articulated by Johnson (1 987, 1989a, 1989b, 1991, 1992, 1993b).
Meaning is based on experience, especially bodily experience. As Johnson
(1989a) 吨ues， the fact that every human being has a body bears directly on
the nature of meaning. To be human is to be embodied: "What we can experlence,
what it can mean to us, how we understand that experience, and how we
reason about it are all integrally tied up with our bodily being" (p. 109).ηlat
is, our bodily experience in and with the world sets out the contours of what is
meaningful to us and determines the ways ofour understanding.
111 a fundamental sense, human cognition is embodied. Our world, as ìt
means to us and as we understand it, is not something objectively given. Instead,
it is something ‘construed' by our embodied cognition. For this reason,
"It is ‘construals of the world' thatare properly regarded as the object of linguistic
semantics" (J. Taylor 1995: 4).10
The experientialist view of reason as being embodied in its context is
sUDlIDarized in Lakoff(19~7a). Again, it is argued, human reason is made possible
by the body. Instead ofùeing an instantiation oftranscendental reason, it
grows out of the nature of the organism. The contributing factors include the
organism's genetic inheritance, the nature of the environment it lives in, the
way it functions in that environment, the nature ofits social functioning, and so
forth.
In short, experientialism assigns a central role to bodily experience in
meaning, understanding, and reasoning. It holds that human knowledge arises
out of the interaction between the experiencing organism and the experienced
environment. The locus of 也at interaction is the human body; the human body
is the result ofsuch interaction. That is, "we have always existed only in and in
relation to our evolving environment. We are what we are at 也is instant, and
our world is what it is 刨出is instant, only because of our embodied interactions"
(Johnson 1991: 8). Therefore, it is necessary to put the body back into
the mind (Johnson 1987).
In line with experientialism in philosophy, the contempor:缸y 也e。可 of
metaphor maintains that human conc叩tual s
AREVIEW 23
bodied nature.)'hat is, m创呐。r is motiva时 by， and grounded in, our bodily
experience--how our bodies function in and interact with the world (e.g. Johnson
1987, 1989a, 1989b, 1 写91 ， 1992, 1993b, Fesmire 1994b, Lakoff 1987a,
1990, 1993a, 1993b, 1994). The basic idea is that "conceptual structure has
everything to do with one's body and with how one interacts as part of one's
physical environment" (Lakoff 1994: 42).
, The evidence supporting the claim that metaphor is constrained by human
bodily experience in the real world has been discovered in various target domains.
,But it stands out most prominently in the domain of_eIl1~t}oll.s~Numerous
studies have shown that human emotions are conceptualized metaphorically
in terms of bodily processes or activities (e.g. Emanatian 1995, Fesmire
1994a, King 1989, K，δvecses 1986, 1988, 1990a, 1990b, 1991, 1995a, 1995c,
LakoffandKδvecses 1987, Matsuki 1995, Sh归 1989， Yu 1995).
It seems that cognitive linguists have put more emphasis on the interactive
aspect of the grounding of meaning because they were criticized in the past for
having neglected the cultural and social aspects of human understanding and
reduced it to the biological or physiological only. Therefore, according to
Johnson (1~~2: 347), this .interaction is "at once biological, social, culturaJ,
economic, moral, and political. ... Thus, the way things can be meaningfully
understood by us depends, in large measure, on the kinds of bodies we have
and the ways we interact with our physical and social surroundings." App缸ently,
efforts have been made to make certain terminology more explicit to include
the sense of interaction between the body and the environment. For instance,
Fesmire (1994b) clarifies the term ‘embodied' as having the rich sense
ofan 龟J1，cultured， interactive body rather than just a physiological one.
To explicitly stress the importance of the interaction between the body and
the cultural and social environment in the grounding of metaphorical mappings,
cognitive linguistics is bound to expand its scope of investigation of
human cognition across linguistic and cultural boundaries.
2.5. Image schemas and the Invariance Principle
When metaphorical mappings are said to be not arbitrary, it means that they
are, in large measure, constrained by the so-called image schemas (or image
schemata). Johnson (1987), according to Lakoff (1987a: 271), "makes an
overwhelming case for the embodiment of certain kinesthetic image schemas."
As Johnson (1987: xiv) defines it, "An image schema is a recurring, d泸1部ηlC
24 THE CONTEMPORARY THEORY OF METAPHOR
pattem of our perceptual interactions and motor programs that gives coherence
and structure to our experience." Image schematic structures, which are central
in the organization ofmeaning and in the formation of inferences based on that
meaning, have two characteristics: they are nonpropositional and imaginative
in character. That is, they are preconceptual schematic structures that emerge
from our bodily experience and that are constant1y operating in our perceptual
interaction, bodily movement through space, and physical manipulation of objects.
Generated as "typical structures ofrecurring aspects ofhuman bodi1y experience,"
image schemas "play a crucial role in what we take as meaningful
and in how we reason" (p. xxxvii); they "make it possible for us to experience,
understand, and reason about our world" (p. 19).
As pointed out by Johnson (1987), image schemas are recurrent pa忧ems
that "emerge 企om our constant and usually unnoticed encounters with physical
containrnent" (p. 22), and therefore, they are relatively few in number, predominantly
visual, though not tied to any single perc叩tual modality. However,
image schemas are not concrete rich images: they "have a generality that raises
them a level above the specificity of particular rich images" (p. 24). As Johnson
summarizes, image schemas, existing at a level of abstraction, and operating
at a level of mental organization between propositional structures and concrete
images, "serve repeatedly as identifying pattems in an indefinitely large
number of experiences, perceptions, and image formation for objects or events
that are similarly structured in.the relevant ways" (p. 28).
From the viewpoint ofpsychology, Gibbs and Colston (1995) 缸gue that a
large body of research in psych()linguistics, cognitive psychology, and developmental
psychology, though not conducted in terms of cognitive linguistic
ideas on image schemas, actually'''can be interpreted as supporting the claim
that image schemas are indeed psychologically real and function in many aspècts
ofhow people process linguistic and nonlinguistic information" (p. 347).
Structural坊， image schemas are "extremely skeletal" (Tumer 1990: 250),
possessing a limited number of parts or components which stand in fixed relations
to one another. They inc1ude, for instance, CONTAINMENT, PA田， LINKS,
BALANCE, SCALARITY, OBJECTS, FORCES, and so on. Some image sche
AREVIEW 25
the structuring of more abstract conc叩ts， propositions, and pa饥ems of inference"
(Johnson 1992: 349). A good example is the PA四 or SOURCE-PATHGOAL
schema, which consists of three elements: a source point A, a terminal
point B, and a vector tracing a path between them. The three basic elements
stand in a definite relation, specified as a force vector moving 企om A to B.
This schema, as Johnson (1987) points out, is recurrently manifested in seemingly
different events, such as walking 企'om one place to another, giving
someone a present, and the melting of ice into water. All these even饵， varying
企om spatial movements to change of state of a substance, are structured by the
same image schema with the same basic parts and relations. As seen in these
ex缸nples， an image schema, which is "more general, abstract, and malleable" (
p. 28) than a concrete rich imagel can characterize "many similar, but different,
situations that manifest a recurring underlying structure" (p. 30). In this sense,
image schemas "operate as organizing structures of our experience and understanding
at the level of bodily perception and movement" (p. 20). Johnson
(1987: 29-30) suggests that there are two important respects in which image
schemas are dynamic: (a) they organize our experience in ways that we can
comprehend; and (b) they are flexible in that they can take on any number of
specific instantiations in varying contexts.
The pervasiveness of image schemas in our experience is well reflected in
our language. A well-known passage in Johnson (1987: 30-31) describing the
first few minutes of an ordinary day is a good ex缸nple ofhow the CONTAINER
schema is reflected in our language talking about daily experience. The
CONTAINER schema consists of a boundary distinguishing an interior from an
exterior in terms of in-out orientations. It structures not only our spatial experiences
but also, by metaphorical extension, our abstract experiences. And the
latter kind of structuring is especially interesting. The following sentences are
derived 企om Johnson's (1 987: 30-31) original passage:
(5) a. You wake out ofa deep sleep.
b. You gradually emerge out ofyour stupor.
c. You walk in a daze.
d. You brush your teeth in a hurry.
e. You might enter into a conversation.
They are instances of metaphorical projections of the CONTAINER schema in
our understanding of abstract states. Abstract states are interpreted as spatially
bounded entities or locations. The power of image schemas lies in the fact that
26 THE CONTEMPORARY THEORY OF METAPHOR
we can metaphorically extend them from the physical to the nonphysical so as
to structure and order our experience in abstract domains. In the following are
somemore ex缸nples 企om Johnson (1987: 34):
(6) a. Tell me your story again, but leave out the minor details. (STORY
EVENT AS CONTAINER)
b. 1give up, I'm getting out ofthe race. (RACE EVENT AS CONTAINER)
c. Whenever I'm in trouble, she always bails me out. (STATE AS
CONTAINER)
These examples further illustrate how a single image schema, as a recurring
organizing s肌ct町、 can help us understand and structure different kinds of
experiences and reason about them.
In Lakoff (1987纱， the image-schematic structure is treated as one of the
two pr民onceptual structures in our bodily experience that give rise to conceptual
structure, the other one being basic-level structure.12 According to Lakoff
(1 987a: 278), image schemas should have the following qualifications: they are
(a) pervasive in experience, (b) well-understood because it is pe凹asive， (c)
well-structured, (d) simply structured, and (e) emergent and well-demarcated.
The image-schematic structures are directly meaningful, whereas abstract conceptual
structures are indirectly meaningful, arising 企om image-schematic
structures by metaphorical projection 企om physical domain to abstract domain.
That is, abstract conc~tual structures "are understood because of their systematic
relationship to directly meaningful structures" (p. 268). In other words,
meaningfulness is embodied. In short, as Lakoff points out,."Image schemas
provide particularly important evidence for the claim that abstract reason is a
matter oftwo things: (a) reason based on bodily experience, and (b) metaphorical
projections from concrete to abstract domains"ω. 275). They structure our
experience preconceptually, and are mapped by metaphors into abstract domains,
with their basic logic preserved. In such a sense, metaphorical mappings
are not all arbitrary 企om any source domain to a target domain, but are very
often motivated by, and based on, structures inhering in everyday bodily expe-
nence.
To illustrate his point, Lakoff cites the example of the PURPOSES ARE
DESTINATIONS metaphor. The metaphor is motivated by "a structural correlation
in everyday experience" (p. 277): we go to a place, namely our destination,
to fulfill a purpose. Therefore, "there is an isomorphism between the
structural correlation and the metaphorical mapping" (p. 278), as follows:
AREVIEW
Source Domain Target Domain
movement to a destination 一一步 achievement of a purpose
Location A or starting point 一一→ the initial state
Location B or end point 一→ the desired state/purpose
motion along a path 一一步 action sequence
27
Here the metaphor is based on the SOURCE-PATH-GOAL schema, one ofthe most
common schemas that emerges from our constant bodily functioning. The
mapping seems natural because the connection between the source and t缸get
domains is motivated by experience. Thus, as Lakoff (1987a: 278) summarizes,
image schemas that structure our bodily experience preconceptually have
a basic logic, which is mapped onto abstract domains by metaphor motivated
by preconceptual structural correlations in experience. Therefore, abstract reason
has a bodily basis in our everyday physical functioning.
What Lakoff (1987a) claims is that those image schemas which structure
our experience of space also structure our concepts in abstract domains. For
instance, as he maintains, categories are understood in terms of CONTAINER
schemas, hierarchical structure in terms of PART-WHOLE and UP-DOWN schemas,
relational structure in terms ofUNK schemas, radial structure in categories
in terms of CENTER-PERIPHERY schemas, foreground-background structure in
terms of FRONT-BACK schemas, and so forth. Lakoff calls his claim 'The Spatialization
ofForm hypothesis' (p. 283):
Strict1y speaki吨， the Spatialization of Fonn hypothesis requires a metaphorical
mapping 仕om physical space into a ‘conceptual space.' Under this mapping, spatial
s衍ucture is mapped into conceptual s衍ucture. More specifically, image schemas
(which strucωre space) are mapped into the corresponding abstract configurations
(which structure concepts).ηle Spatialization of Fonn hypothesis thus
maintains that conceptual structure is understood in tenns of image schemas plus
a metaphorical mapping.
He then concludes that image sch巳mas play two roles: "They are concepts that
have directly-understood structures of their own, and they are used metaphorically
to structure other complex concepts" (p. 283).
The recognition ofthe pervasive function of image schemas as the basis for
metaphorical mappings led to the formulation of the lnvarianc巳 Hypothesis，
which was later revised and renamed as the Invariance Principle (Lakoff and
Turner 1989, Lakoff 1990, 1993a, 1994, Turnerl990, 1992, 1993).13 Proposed
28 THE CONTEMPORARY THEORY OF METAPHOR
as the general constraint on met叩horical mappings, it states: \;Metaphorical
mappings preserve the cognitive topology (that is，也.e image-schema structure)
of the source domain, in a way consistent with the inherent s位uc:ture ofthe target
domain" (Lakoff 1993a: 215).
An important co'rollary of the Invariance Principle, according to Lakoff
(1993a, 1994), is that image-schematic structure inherent in the target domain
cannot be violated, and 也at inherent target domain structure limits the possibilities
for mappings automatically. The phenomenon is called the target domain
overrides. In the metaphor ACTIONS ARE 四ANSFE邸， for jnstance，也tions
are conceptualized as 。同回ts transferred from an agent'to a pat~ent. But with
the target domain overrides，伽.e agent can ‘give' 也e patient ‘ a lcick' without
the latter's having it afterward. Turner (1990: 252) also points out 也at as a
general constraint on metaphor, the Invariance Principle is not inviolable, but
the violation is to be taken as "a carrier of significat邸，" which is found in
novel metaphors only.
According to Lakoff (1990, 1993a, 1994)，也e Invariance Pri盟.ciple raises
the possibility that a great many, if not all, abstract inferences 缸@邵阳ally
metaphorical versions of 叩atial inferences inherent in the structures of image
schemas. 白1at is, metaphors based on image schemas give rise to abstract reasoning,
and abstract reasoning is based on spatial reasoning via metaphorical
projections ofimage schemas. Lakoff(l990, 1993a, 1993b, 1994) has demonS位ated
in English that what have been called propositional structures or 曲，
stract concepts such as time, states, changes, causes, purposes, and categories
are really understood via metaphor in terms of spatial concepts structured by
image schemas. As he suggests, "These are conc叩ts that en伽 nonnally into
the grammars oflanguages, and ifthey are indeed metaphorital.in nature, then
metaphor becomes central to gratnm缸" (1993a: 212). Consequently, the Invariance
Principle has put the nature of abs位部t reason into I}ew perspective:
"what have been seen in 也e past as propositional inferences are really 恤lagebased
inferences" (Lakoff 1993a: 229). If the Invariance Principle is correct,
Lakoff (1 993a: 229) claims, it has a remarkable consequence: "Abstract reasoning
is a special case of image-based reasoning." "Image-based reasoning is
fundatnental and abstract reasoning is image-based reasoning under metaphorical
projections to abstract domains" (p. 229). If L
AREVIEW 29
sistent with what is genera11y known about human cognition. Part of such a
commitment is to characterize what abstract concepts are, how they can be understood,
and how abstract reason could have been acquired by human beings
The lnvariance Principle provides an answer to these questions, since it claims
that many abstract concepts arise 企om metaphorical mappings of spatial concepts
and that abstract reason arises via metaphorical mapping of image schemas
which prese凹es the inferential structure of those spatial concepts. As Lakoff
(1990) points out, abstract reasoning is that aspect of human beings that
has traditiona11y been taken as separating man from the lower animals, but that
activity of human brain has evolved 企om certain mechanisms for the perception
ofspatial relations that appear to be present in lower animals.
At the present stage, however, the lnvariance Principle is still, as Lakoff
\1990: 72) put it quite a few years ago, "a扭n emp抖iri巾ca刽1 hypot白he町s剖is，" and ‘"‘'it沁s
estatus is anything but c1ea矶r巳飞咄s剑ince a precise f岛or口rmηmu叫lla甜tio∞n would req伊ui忱r陀
knωowledg伊e of the ft缸创11山11 inventory of imag伊e-啊hema挝S." Moreover, it is still not
clear as to how strong this principle is: whether or not a11 abstract inferential
structure is image-schematic; and whether or not image-schematic structure is
only one of a number of aspects of generic-level structure (Lakoff 1990). The
answer to these questions ca11s for thorough investigation within and across
linguistic and cultural boundaries.
2.6. Conventional and novel metaphors
ln their More than Cool Reason, Lakoff and Tumer's (1989) central claim is
也at metaphor in poetry is not an essentia11y different phenomenon 企om metaphor
in ordinary language; poetic metaphor basica11y uses the same cognitive
mechanisms as everyday metaphor; and what makes poetic metaphor look dif二
ferent, however, is its extension, elaboration, and combination of those mechanisms
in ways that go beyond the ordinary. They argue, therefore, that creative
literary metaphor genera11y depends on conventional metaphor in generation
and interpretation. According to them, there are three basic mechanisms for
interpreting linguistic expressions as novel metaphors: extensions of conventional
metaphors, generic-level metaphors, and image metaphors. Poetic rnetaphor
uses a11 ofthem superimposed on one another.
A good ex缸nple ofnovel extension ofa conventional metaphor, as cited by
Lakoff(1993纱， is a line of a song lyric:
30 THE CONTEMPORARY THEORY OF METAPHOR
(7) We're d.riving in the fast lane on the 企eeway oflove.
Although what is regarded as novel is not absolute and is relative synchromcally
as well as diachronically, the rnetaphorical expression in this line, with its
rnodem flavor, is novel in the sense that it is not conventionalized in the everyday
language of English. However, it is only an extension, narnely a novel instantiation,
of the conventional conceptual rnetaphor LOVE IS A JOURN盯. The
understanding of the forrner is a consequen::e of the preexisting rnetaphorical
correspondences of the latter. As Lakoff argues, a conventional rnetaphor, as a
fixed pattem of ontological correspondences across conceptual dornains, defines
an open-ended class of potential correspondences across inference pattems.
"When activated, a rnapping may apply to a novel source dornain knowledge
structure and characterize a corresponding t缸get dornain knowledge
structure" (p. 210). At the lexicallevel, lexical iterns conventional in the source
domain may or may not be conventional in the target domain, depending on
whether or not the ontological coηespondences between two domains are activated.
白lat is,
each source dornain lexical item may or may not make use of the static mapping
pa阳m. lfit does, it has an extended lexicalized sense in the target domain, where
that sense is characterized by the mapping. lf not, the source domain lexical item
will not have a conventional sense in the target domain, but may still be actively
mapped in the case ofnovel metaphor. (p. 211)
Those lexical items that have g创ned a conventional sense in the target domain
are referred to as polysemies，也at is, words with related senses. The lexical
Iterns ‘企'eeway' and ‘ fast lane' in (7) above are not conventionally used in the
dornain of love, but they comprise a novel extension of the conventiona1 metaphor
LOVE IS AJOURNEY.
As Lakoff (1993a) has cited, exarnples of novel extension of conventional
rnetaphor are also found in the following lines ofpoems:
(8) a. In the rniddle oflife's road
1found myself in a dalk wood.
(Dante: Divine Comedy)
b. Two roads diverged in a wood, and 1一
1took the one less traveled by,
And that has rnade all the difference.
A REVIEW 31
(Robe口 Frost: Stopping by Woods on a Snowy Evening)
Here, (8a) is an instance of novel extension and combination of two conventional
metaphors: LIFE IS A JOU阳EY and KNOWING IS SEEING. (8b) extends and
elaborates the first of these two. In Sweetser's words (1 992: 707), "Jiterary
metaphors are creative uses of precisely those metaphors that shape our everyday
language and thought," and this is because "literary language is not ‘阻，
other language' from everyday language"; in effect they "coexist as aspects of
the varied language .use of a single community" (p. 706). As Sweetser has suggested,
being a powerful 红tistic structure in literature, metaphor is not merely
a literary tool,.but also generally basic to cognitive and linguistic structures.
Therefore, "we can't understand literary metaphor outside of its context in language
structure" (p. 708).14
Another type of metaphors that are common in poetry is called image
metaph,ors (Lakoff1987c, 1990, 1993a, Lakoffand Turner 1989, Gibbs 1994a,
Turner 1991), those that function to niap one mental image onto another (e.g.
‘My wife .., whose waist is an hourglass'). Image metaphors are ‘one-shot'
metaphors, mapping only one image (e.g. an hourglass) onto another image
(e.g. a woman's waist) by virtue of their commonshape. For this reason, it is
very unlike a conventional conceptual metaphor (e.g. LIFE IS A JOURNEY) that
maps rich inference pattems across conceptual domains. In other words, an image
metaphor maps the knowledge of one image onto another image whereas a
conceptual metaphor maps the knowledge of a conceptual domain onto another
conc叩tual domain. While the former is realized in a particular linguistic 饵'
pression, the latter is manifested in a system of linguistic expressions. 飞iWhat is
worthy of note, however, is that image-mapping metaphors work in'just the
same way as other metaphors, by mapping the structure of one dòmain onto the
structure of another. What is special about them is that their domains are conventional
mental images. But conventional mental images are also structured
by image schemas, which are preserved by image metaphors. That is to say,
both conceptual metaphors and image metaphors are unified under the general
Invariance Principle, which states that metaphors preserve the image-schematic
structure ofthe source domain in a way that is consistent with the inherent image-schematic
structure ofthe target domain.
According to Lakoff and Tumer (1989), novel metaphors, often f
32 THE CONTEMPORARY THEORY OF METAPHOR
instantiation of the fonner at a more specific level. The fonner only entails the
understanding of events in tenns of actions perfonned by agents, but leaves
unspecified its details, which are to be fumished by specific-level metaphors.
Lakoffand Tumer (1989) have found, for instance, that the so-called personification
in literature is summarized by the generic-level metaphor EVENTS ARE
ACTIONS. According to Lakoff (1 993吟， generic-level metaphors preserve generic-level
schemas, which may be image-schematic structures. If that is the
case, then the lnvariance Principle is operating here too, constraining the generic-level
metaphors.
Lakoff (1993a) argues that our ev町yday metaphor system is constantly active,
and is used maximally in interpreting novel metaphorical uses of language.
This is because literary language sh缸es much with ordinary language,
and artistic usage with everyday usage. As Sweetser (1992: 722) ∞ncludes，
"There are thus strong arguments in favor of approaching artistic metaphor together
with everyday metaphor, even via everyday metaphor."
More general1y, "ev町yday language and literary language are not s叩缸'ate
domains," and "discoveries about one bear on the other" (Lakoff 1987d: vii).
2.7. Summary of findings of the contemporary theory
The summ缸y of research findings provided here is based on Lakoff (1993a,
1994). Having resulted from the studies conducted under 伽 eontemporary
theory of metaphor in the past decade and a half, it addresses three. aspeçts of
metaphor: its nature, its structure, and its properties. In terms of its nature,
metaphor is fund缸nentally conc叩tual， not linguistic. Metaphoricallanguage is
merely a surface manifestation of ∞nceptual metaphor. Much of our ∞nceptual
system is metaphorical, although a significant part of it is nonmetaphorical.
Metaphorical und笛standing is built up on nonmetaphorical undmtanding.
As a basic cognitive structure, metaphor allows us to under宫tand a relatively
abstract or inherently unstructured subject matter in t町ms ofa more ωncrete，
or at least a more highly structured subject matter. In fact, m皿y subject matt町s，企om
the most mundane to the most abstruse scientific theori饵， can only
be comprehended via metaphor. In short, metaphor is the m锺in mechm世sm
through which we comprehend abstract concepts and perfonn abstract reason-
mg.
Structurally, metaphors are mappings across conceptual domains. Such
mappings are asymmetric in that they are one-directional, involving projections
AREVIEW 33
fr'Om a source domain to a target domain. They are partial in that only part of
the structure of the source domain is projected to the target d'Omain. Each
metaph'Orical mapping is a fixed set of ontological coηespondences between
entities in the source domain and entities in the target domain. When those
fixed 'Ont'Ol'Ogical corresp'Ondences are activated, mappings can project source
domain inference pa忧erns onto target domain inference patterns. Metaphorical
mappings are not arbitrary, but grounded in the body and in everyday experic:nce
and knowledge. A conceptual system contains thousands of conventional
metaph'Orical mappings, which form a highly structured metaphorical subsystem
of the c'Onc叩tual system. Metaphorical mappings, with both conceptual
and image mappings, all obey 也e Invariance Principle: The image-schematic
structure 'Ofthe source domain is pr'Ojected onto the target domain in a way that
is consistent with inherent target domain structure.
In its current sense of the contempor缸y theory, metaphor has the following
important Qrgp~rties， among others. First, the system of conventional conceptual
metaphor is mostly unconscious, automatic, and is used c'Onstantly, with
no noticeable effort, just like our linguistic system and the rest 'Of our conc叩·
切al system.η世s metaphor system plays a major role in both grammar and
lexicon of a language. It is central to 'Our understanding of experience and to
the way we act on that understanding. Second, novel or poetic metaphor is, for
the most part, an extension of our everyday conventional system of metaphorical
thought, constrained by the same principle as the conventional metaphor.
!hi.rd, metaphorical mappings v町 in universality; some seem to be w盯的al，
others are widespread, and some seem t'O be culture-specific.
2.8. Criticisms of the contemporary theory
Ever since its birth about a decade and a half ago, the contemporary theory of
metaph'Or has grown increasingly influential in the interdisciplinary field of
metaphor studies. It has also received some criticisms (e.g. Holland 1982, Mac
Cormac 1985, Wierzbicka 1986, Ortop.y 1988, Jackendoff and Aaron 1991,
Quinn 1991, Alverson 1991, Indurkhya 1992, Kennedy and Vervaeke 1993,
Cacciari and Glucksberg 1994, Murphy 1996, 1997) 企om different. theoretical
and disciplinary perspectives along its way of development. In this section 1
examine a few criticisms.
It seems that some criticisms have stemmed 企om different theoretical
views regarding what counts as a metaphor. The focus of study 'Of the contem-
34 THE CONTEMPORARY THEORY OF METAPHOR
porary theory is on (1) conc叩tual metaphors in human conceptual system and
(2) conventional metaphors in everyday language, which were largely ignored
by more traditional theories of metaphor. In the contemporary theory, it is believed
that metaphor is primarily conceptual in nature and that conventional
metaphors at the linguistic level yield important clues to human thought and
cognition. The rejection ofthese two beliefs wil1 surely lead to the denial ofthe
central claim of the contemporary theory that human conceptual system is fundamentally
metaphorically structured.
This central claim has been challeng时， for instance, by Wierzbicka (1986).
Citing the conceptual metaphor LOVE IS AJOuru也Y， Wierzbicka argued that the
journey metaphors are not applicable to the entire range of the use of the tenn
‘love. ' Instead, their use is limited to the kind of love between, say, erotic p缸b
ners while not applicable to the love between a mother and a child. ''This
means that ‘joumey' is not in any way included in the semantic invariant ofthe
concept ‘ love'" (p. 291). Wierzbicka (1986: 292) proposed the following
"REAL" definition of love which is 企ee ofmetaphor:
(9) X loves (person) Y. =
when X thinks ofY, X feels good feelings towards Y
X feels that he wants to be with Y
X feels that he wants to cause good things to happen to Y
As this definition suggests, people can have a clear ∞ncept of love without
having experienced journeys. This is because, Wierzbicka 缸gued， mental experiences
are given to us more directly than physical ones, and our inner world is
more familiar and more accessible to us than the extemal world. Therefore, we
know and understand love without the mediation of the journey metaphors,
which are used only "for talking about love" (p. 297).
It should be admitted that Wierzbicka was right when she said that the
LOVE IS A JOURNEY metaphor is not applicable to the entire range of the use of
the tenn love, and that she provided a nonmetaphorical definition of love in (9)
above. However, it should also be noted 也at this nonmetaphorical definition of
love is only one definition, and that it_does not represent all there is of human
cognition or conceptualization of love. In his reply to Kennedy and Vervaeke
(1993), who denied the claimed constitutive role ofmetaphor in human cognition,
Johnson (l993b: 418) emphasized that "the structure of a concept is not
an all-or-nothing matter," and that "it is not the case that conceptual structure
either preexists in a finished and fixed realm of its own, or else that it is all
A REVIEW 35
radically constructed." Instead, as with the concept ‘ love,' most basic concepts
are defined by multiple conceptual metaphors that are sometimes mutually inconsistent
(though not incoherent). As acknowledged by the Invariance Principle,
there are preexisting conceptual struct旧e in both the source and target domains,
but conceptual metaphor will be at least partially constitutive of our
cognition of the target domain, by virtue of additional structure carried over
企om the ‘logic' ofthe source domain. Johnson further pointed out that "cognitive
semantics never c1aims that image-schematic structures and metaphors are
all there is to cognition," and rather, it only claims that "this is a crucial, insuf二
日ciently explored, and undervalued dimension ofmeaning" (p. 421).
So, it is obvious that the LOVE IS A JOURNEY metaphor cannot replace the
nonmetaphorical definition of love in (9) provided by Wierzbicka; it nevertheless
enriches human understanding of, and reasoning about, the concept of
love, in a particular aspect. It carries p缸t of the structure of the source domain
(journey) over into the t缸get domain (love), which has its own preexisting
s伎ucture. It is worth reminding that the contemporary theory of metaphor is
~\lperior in one aspect: it has reached a higher level of generality by having
discovered the hierarchical structure of metaphor, as discussed previously in
2.3. In this structure, the conceptual metaphor LOVE IS A JOURNEY, which govems
a system I)f metaphorical expressions, is merely a subsystem of higher
metaphorical system LIFE IS A JOURNEY, which, in tum, is but a subsystem of
the even higher system, the Event Structure Metaphor. This discovery has attached
even greater importance to metaphor in human cognition.
The fact that a nonmetaphorical definition of love such as in (9) is inadequate
for human conceptualization of love and 也.at metaphorical projection of
partial s位ucture 企om a source domain is essential is also evidenced by Wierzbicka's
(1986: 300-30的 discussion ofsuch prepositional phrases as ‘in love,'
‘沁 pain，' and ‘ìn despair.' Wierzbicka contested Lakoff and Johnson's (1980)
çlaim that nonphysical is typically conceptualized, metaphorically, in terms of
the physical, as demonstrated in (10) (企om Lakoffand Johnson 1980: 59):
(10) a. H缸ry is in the kitchen.
b. H缸ry is in the Elks.
c. H缸ry is in love.
According to Lakoff and Johnson (1980: 59-60), there is a difference among
these three sentences with respect to conceptual structuring. In (10a) the concept
IN "emerges directly 企om spatial experience in a cle寻rly delineated fash-
36 THE CONTEMPORARY THEORY OF METAPHOR
ion"; it is not a metaphorical conc叩t. But (10b, c) are instances of conc叩ts
that are metaphorica11y structured (i.e. A SOC1AL GROUP 1S CONTAINER and A
EMOTIONAL STATE 1S A CONTAINER). ''The word ‘in' and the concept IN 町e 也e
same in a11 three examples; we do not have three different CORC叩ts of IN or
three homophonous words ‘in\"
Wierzbicka, however, disagreed that the word ‘in' is used inthe 也ree sentences
in the same sense; particl,llarly, she disagreed that the expression ‘in
love' is metaphorical. She argued, "Plain common sense indicates that expressions
.such as in love. in pain or in despair don't refer to place. They refer to
certain psychological states" (p. 302). But，仕lere exists ample evidence that abstract
states are conceptualized in terms of bounded locations in 叩ace， as in
the Event Structure Metaphor (e.g. Lakoff 1990, 1993a, 1993b, 1994, Taub
1996). Now let us look at Wierzbicka's own interpretation of the emotional
弛， as opposed to the locational ‘in' (p. 305):
1 would speculate that something Iike the following (subconscious) teasoning
may operate here. First，也e form ‘ in Nemotion' is more marked, more unusual than
the ordinary Adj.motion or Verb.刷on'τbis in itself may suggest a marked form of
the emotion. Second, the idea ofplace ('in') evokes the image of something surrounding
the person 企.om all sides. Normally, an emotion (designated by an adjective
or a verb) is viewed as something that takes place within a person. If the
Iinguistic form seems to suggest a reversal ofthe image (a person surrounded by
something instead of something ‘ enclosed' within a person) the impression can
easily arise that the feeling is intense，由at it is stronger and more ov，町whelming
than would normally be expected, and also that the normal relationship between a
person and his emotions is reversed, so 也at the emotions, instead ofbeing subject
in some measure to the p町80n's control,assume power over him. Finally,the image
of a person being ‘in' something (rather than out) mayiilvite the idea ofrestrictions
on this person's movements (and, possibly, on his field ofvision): when
we are outside we can go whenever we like; but when we are inside a house, our
仕eedom ofmovements is circumscribed by the four walls.
As can be seen, the passage is packed with spatial metaphors. And 仕lat is typical
metaphorical conceptualization ofemotions. What Wierzbicka was doing is
mapping the structure of space onto the structure ofemotions, and that is metaphorical
mapping in the sense of Lakoff and Johnson (1980). This me即horical
mapping，部 illustrated by Wierzbicka's interpretation in the above p部sage，
AREVIEW 37
is cognitively constitutive of the understanding of emotions in a partial way.
Wierzbicka denies the metaphorical nature of the phr田e ‘in love' because she
maintains a different definition ofmetaphor from Lakoffand Johnson's. Lakoff
and Johnson believe 由at metaphor is primarily conceptual in nature whereas '
linguistic expressions are byproducts of the conceptual level. on the other
h四d， Wierzbicka treats metaphor as "a linguistic device ... which by definition
can't convey meaning in a fully explicit manner" (p. 294). To her, metaphor
only provides ways of"talking about" things. It is therefore prim缸ily linguistic
rather than conceptual in nature.
In their review ofLakoff and Turner (1989), Jackendoff and Aaron (1991)
insisted that the term metaphor is used too broadly in the book. To them, what
is called conventional metaphors should not be taken as metaphorical. Rather,
世也le'町:y believed t伽ha'刽t
p协hoωrs i旭sw附o毗巾1 pres阳町v悦in鸣g" (p. 326). Thus, 也ey drew a distinction between
"LT-metaphors" (i.e. what Lakoff and Turn衍 count as metaphors) and "1metaphors"
(i.e. "the more standardly acknowledged as metaphorical" based on
the criterion of "literal incongruity") (p. 326)，缸'guing that only I-metaphors
are rea1 metaphors. For instance, DEATH IS DEPARTIJRE is an LT-metaphor, but
its status 幽 an I-nietaphor is relativized across cultur四. In many cultures
where death is viewed litera11y as the soul (or person) passing on to its next
existence, DEATH IS DEPARTIJRE is not an I-metaphor, but a literal belief.
Jackendoff and Aaron also argued that some LT-metaphors should not
count as metaphors in the first place. They believed 也at Lakoff and Turner's
conc叩tual metaphors such as STATES ARE LοCATIONS, PURPOSES ARE DES.
TINATIONS, TIME MOVES are not metaphor亩， but instances of "Thematic Relations
Hypothesis" (see a1so Jackendoff 1983: Ch 10),which claims that
the conc叩饥.Ull structures expressed by naturallanguage are organized in tenns of
a set of abstract parameters that are most cIearly revealed in language about
space, but that apply to many other semantic fields as well. It is not 也at space is
taken as a METAPHOR that supplements or enriches the conc叩tualization of these
fields; rather, this common organization is the ONLY way we have of conceptualizing
them. In oth町 words，位lis basic skeletal organization of conceptual structure
... receives many parallel realizations, among them the conceptualization of
space. (p. 328)
According to Jackendoff and Aaron, the statements such as STATES ARE
LOCATIONS, PURPOSES ARE DESTINAτ10NS， TIME MOVES (i.e. LT-metaphors) are
38 THE CONTEMPORARY THEORY OF METAPHOR
not based on "mapping of incongruous domains" but on "thematic parallelism"
(pp. 329-330), so they are not I-metaphors. In short, they could not agree with
the "excessively broad notion of metaphor" of Lakoff and Tumer (1989), who,
"having drained ÍÌom the term ‘metaphor' much of its traditional content,"
"have created a theoretical construct so broad and unstructured that the term
‘ metaphor' may no longer be appropriate" (p. 331).
Regarding the equation DEATH IS DEPARTURE, Jackendoff and Aaron were
certainly right when they pointed out that there exists some cross-cultural relativity
as to whether it is a metaphor. In cultures where DEATH IS DEPARTURE is
taken as literal belief, namely the be!ief that spatial movement is involved in
death, this equation certainly is not a metaphor. The key to the definition of
metaphor in the contempor缸y theory is the notion of ‘semantic autonomy'
(Lakoff and Tumer 1989). If a concept is understood in its own terms and
hence semantically autonomous, it is not metaphorical. If, on the other hand, a
concept is understood in terms of another concept or concepts, it is not semantically
autonomous and therefore is a metaphor. Jackendoff and Aaron of
course would not accept this definition for being "excessively broad." They
preferred, instead, "the standard sense" ofmetaphor which, as can be seen ÍÌom
the above quotation, cannot be "the ONLY way we have of conceptualizing"
things, but in one way or another "supplements or enriches 出e conceptualization."
This is also apparent in Jackendoffs (1983: 209) response to the view
that the thωry ofthematic relations reveals widespread systems ofmetaphor in
our language and thought: "But 1think this debases both the theory of thematic
relations and the conc叩t ofmetaphor, for, un!ike metaphor, thematic relations
are not used for 缸tistic or picturesque effect." However, as Ortony (1975) and
Fainsilber and Ortony (1987) have shown, metaphors serve at least 伽ree communicative
functions, of which one is ‘inexpressibility' (namely, metaphors
may allow one to express 也at which would be difficult or impossible to express
if one were res位icted to literal uses of language), the other two being
‘compactness' and 'vivi也less.' Jackendoff and Aaron's view 也at metaphors
merely provide altemate ways of talking about something nicely is similar to
that ofWierzbicka's. Holding this view they would believe that many conventional
metaphors are no
AREVIEW 3~
one of them. On this view, time, for instance, is not metaphorically conceptu.
alized in terms ofspace, but simply holds a thematic p缸allel with space.
Gibbs (1994a: 167-169) contains a critique ofthe Thematic Relations Hy.
pothesis, which is based on Lakoffs unpublished reply to Jackendoff and
Aaron (1991). It is argued that the so-called thematic parallels actually do n01
exist between the spatial and tempNal domains and that the seeming parallel.
isms are indeed consequence of metaphorical mapping from the former to the
latter domain. Specifically, in the spatial domain an observer has an inheren1
企ont and back. For instance, when 1stand with my back against a wall, looking
at the mountain in 企ont of me, 1 can say something to the effect that ‘ The
mountain is ahead of me and the wall is behind me,' regardless of which culture
1 am 企om or what language 1 speak. And that is reality. ln the temporal
domain, 1 can also say that ‘The fuωre is ahead of me and the past is behind
me,' and what 1 say indeed ref1ects the way 1 think of the future and the past.
But in reality, the future does not really exist in 企ont of me, nor does the past
behind me. What 1 say in this case is only the consequence of a metaphorical
way of thinking or conceptualization that maps the spatial domain onto the
temporal domain. Fur位lermore， the way of conceptualizing the future and the
past is relative across cultures (s饵， e.g., Alverson 1994). Hence, a person from
a different culture standing next to me side by side may say that ‘The future is
behind me and the past is ahead of me,' even though we both are facing the
same direction. Moreover, there are different ways of conceptualizing time
even within a single culture or language. In English, for example,‘Christmas is
coming up on us' and ‘We're coming up on Christmas' ref1ect two special
cases of how time is metaphorically conceptualized. Times are understood as
moving entities in the first case and as fixed locations in the second. These two
special cases 缸e inconsistent to each other, and cannot be accounted for by the
notion of thematic p缸allels since parallels do not really exist between the spatial
and temporal domains. Chapter 4 below will lend further support to this
argument. Here and now, 1 want to further emphasize that the notion of ‘ thematic
parallels or parallelisms' cannot solve the problem of asymmetry and
directionality observed between conceptual domains. That is, metaphorical
mapping is usuall
40 THE CONTEMPORARY THEORY OF METAPHOR
and abstract domains on the other.ls
Mac Corm挝、 (1985) criticism of Lakoff and Johnson (1980) also
stemmed from a different view of metaphor. He believed that conventional
metaphors (or ‘ dead metaphors' in his terms) should be regarded as literal
rather than metaphorical. This view is apparent in his distinction between literal
and metaphorical language:
Literallanguage seeks to use estab!ished categories (including those derived 企om
dead metaphors) to describe the natural world in common terrns that can be universally
comprehended. MetaphoricaJ language seeks to create new suggestive
ways of perceiving and understanding the world and involves a conceptual process
different from that ofliteral description. (p. 78)
He argued that most of the examples of metaphor presented by Lakoff and
Johnson are ‘ dead metaphors' which have already faded into literallanguage. If
they are also taken as metaphorical, it amounts to saying that all language is
metaphorical, whereas he believed that a literal-metaphorical distinction is essential
in explaining how new meanings ar巳 derived via metaphor with literal
meanings as given and how language changes in such a process.
Regarding this criticism Indurkhya (1992: 296) pointed out that the main
objective ofLakoff and Johnson's study is to show that metaphor is a powerful
tool in shaping the cognitive world that we experience. For this objective, conventional
metaphors "are quite important, since they bring evidence that even
what we take to be the conventional and ordinary description ()f the world is
actually brought about by a metaphor." But "Mac Corm饵， s criticism of Lakoff
and Johnson utterly fails to appreciate this important point." Indurkhya's critique
is very much to the point. While MacCorm饵 's view that only novel
metaphors are real metaphors served his purpose to show how novel tnetaphors
derive their meanings, he failed to realize that there exist at least four different
senses of ‘ literal' (see Lakoff 1986b for a critique).
In addition to the above criticism, Mac Cormac also opposed Lakoff and
Johnson's account of how abstract concepts are structured metaphorically in
terms of spatial concepts which are directly emergent from our bodily experience.
He argued that even the delineation of the spatial is cultural, emerging
"in linguistic forms that are already cu1turally mediated and transmitted" (p.
67). For illustration, he cited Lakoff and Johnson's (1980: 161) own e宜ample
of 丘ont-back orientation relative cross胸cu1turally:
A REVIEW 41
Given a medium-sized rock in our visual field and a ball between us and the rock,
say a foot from 扰， we would perceive the ball as being in front ofthe rock. The
Hausas make a different prQjection than we do and would understand the ball as
being in back ofthe rock.切1US， a front-back orientation is not an inherent property
of objects like rocks but rather an orientation that we project onto them, and
the way we do this varies from culture to culture.
Mac Corrnac (1 985: 68) then asked: "If some spatial concepts vary from culture
to culture, how can we have any certainty that spatial concepts emerge directly?
It seems more like a mediated, indirect emergence, which Lakoff and
Johnson would call metaphorical" (see also Indurkhya 1992 for a similar criti-
cism).
The answer to Mac Corrnac's question is that, when the experientialist
view of cognition claims that spatial conc叩ts emerge directly, it means that
these concepts are derived direct1y 企om the physical wor\d, as opposed to abstract
concepts which are often mediated by metaphor. It is important to note
that cognitive linguistics n巳ver c1aims that spatial concepts are universally interpreted
and that some of them are indeed interpreted differently across cul-.
tures. Just as shown by the above example 企om Lakoff and Johnson (1 980), a
rock that lacks an inherent 企ont and back is assigned one by culture-specific
interpretation. Researches within the cognitive paradigm have shown that "although
the physical configuration and neurophysiological apparatus of human
beings give us all a common starting point for the way we experience the
wor\d, our perccptions of it are differentiated by individual cultures" (AlIan
1995: 13). While the spatial domain is surely the source domain for metaphors
mapping into more abstract domains, it is also the domain currently under investigation
fot possible cross-linguistic and cross-cultural universals and differences
(see, e.g., Allan Ì995, Ameka 1995, Brown and Levinson 1993, Heine
1995, Levinson 1991, 1992a, 1992b, Pederson 1995, Regier 1995, Sinha et a1.
1994, Sinha and Thorseng 1995, Svorou 19Ç,4, Wilkins and HillI995).
Allan's 1995 study is a case in point. It is shown in this study that the word
卫班， in English is defined as "that part of a body opposite the interactiveside"
(p. 11). It is originally defined "on an anthropomorphic model of the
prototypical human being in upright stance con企onting the wor\d by looking
forward and walking forward," and is metaphorically "extended to the corresponding
proper parts of vertebrate and invertebrate animals and to inanimate
objects such as houses, cupboards, and computers" (p. 11). With cupboards or
computers, the 仕ont is always the door or screen side, which is the interactive
42 THE CONTEMPORARY THEORY OF METAPHOR
side, while the back is the opposite side. With houses, the back is usually opposite
the side with the main entrance. But in some cultures, it can be the house
roof, conceptualized on a "zoomorphic" model in which the back is "roughly
horizontal and facing skyward like an animal's back" (p. 19). Relevant to Lakoff
and Johnson's example of the rock is the following conceptualization: "If
a static concrete inanimate object is assigned no intrinsic front, the part or region
of the object facing the human viewer is contingently named 的efront.
and the part or region on the opposite side or end of the contingent front is
named the back" (p. 22). This is because "The characteristics of the canonical
encounter between humans are transferred to the encounter between a human
being and a nonhuman object, with the result that the viewer faces the 企ont of
the object, and it con企onts himlher"ω. 22). This is the case in most languages,
but op.e altemative, selected in Hausa (Chadi), Kiswahili (Bantu), and Maasai
(Eastem Nilotic), is to imagine the object 吁acing the same direction as (i.e.
aligned with) the human viewer," n缸nely， "the human viewer faces the backs
of things" (p. 22). "It is a matter of cultural convention whether the viewer is
facing the 仕ont ofthe 0均 ect or its back" (p. 22). Allan's conc1usion is that "the
uses and meanings of English back are motivated by our cognitive modelling
of the world and that they evidence a powerful anthropocentric image of ‘the
body in the mind' of humankind" (p. 11). A similar study of spatial terms ‘m
仕ont of and ‘behind' is in Kalisz (1990), which also shows "an experiential
grounding ofterms expressing space orientation" (p. 167).
Finally, 1 consider the criticism raised by Noami Quinn, a cognitive anthropologist,
in her ‘The cultural basis of metaphor' (1991). She criticized Lakoff
and Johnson for the tendency "to neglect altogether the organizing role of
culture in human thought, or to grant culture, at best, a ~f;':sidual or epiphenomenal
place in their accounts" (p. 57). She was oppo时 to their assi民ing a
constitutive role to metaphor in human understanding. Instead, she argued that
metaphors are used only to fit preexisting cultural models, which are "presupposed,
taken-for-granted models ofthe world that are widely shared (although
not necessarily to the exclusion of other, altemative models) by the members
ofa society and that play an enormous role in their understanding ofthat world
and th
AREVIEW 43
tural model of American marriage (see also Quinn 1987), she insisted that it is
the cultural model that commands the selection ofmetaphors:
particular metaphors are se!ected by speakers, and are favored by these speakers,
just because they provide satisfying mappings onto alrcady cxisting cultural understandings-that
is, because elements and relations between elements in the
source domain make a good match with elements and relations among them in
the cultural model. (p. 65)
Thus, "conventional metaphors" have become "conventional" only because
"they are satisfying instantiations of a ‘conventional' or culturalJy shared
model, capturing Q1ultiple elements of that model" (p. 79). ln conclusion,
Quinn claimed that "metaphor plays 1 comp缸atively minor role in constituting
our understanding of our world, and that a relatively major role in constituting
this understanding is played by cultural models ofthat world" (p. 91).
So, while Quinn's conclusion did not completely deny the constitutive role
of metaphors in human understanding, which is at least in part in accordance
with cognitive semanticis饨， claim that metaphors pa叫y constitute our understanding
of the wor1d, she laid special emphasis on the major role of cultural
models in constituting this understanding, claiming that cultural models actually
constrain the selection of metaphors. She was certainly right to some ex-
tent.
The contemporary theory claims that metaphors are rooted in our bodily
experience. Here,‘.bodily experience' should be interpreted in a broad sense,
referring to our bodily function and interaction with the outside world, and our
knowledge so derived. However, the bodily experience can only tell what are
possible metaphors. Whether these potential metaphors are actually selected in
a given culture is largely dependent upon the cu1tural models shared by individuals
living in this culture. The use of sports metaphors, which are typically
rooted in our bodily activities, should serve as an illustration. In American
English, sports have always been a favorite source domain for metaphors in
everyday and political discourse (see, e.g., Hardaway 1976, Howe 1988, Ching
1993, Shore 1996).16 The reason is apparent, as Hardaway (1976: 78) saw it:
飞
Nobody would argue the place of sports in American life; they are big business.
And they are big business because they fit philosophically with the widely accepted
American dream of open competition in a free market economy. Americans
beJieve in competition, foster it, and encourage it. They live by its rules. No
44 THE CONTEMPORARY THEORY OF METAPHOR
wonder the language of athletic competition has found its way as metaphor into
every aspect ofAmerican life.
In Chinese, sports metaph'Or百 have enj'Oyed increasing p'Opularity in the past 15
years 'Or so when ref'Orm has remarkably raised the nati'On's level 'Of c'Ompetiti'On
in b'Oth ec'On'Omy and sp'Orts. That is, the increasing p'Opularity 'Of sp'Orts in
an improving ec'On'Omic envir'Onment has br'Ought ab'Out increasing p'Opularity
'Of sp'Orts metaph'Ors. H'Owever, there exist s'Ome variati'Ons between the American
and Chinese cultures as t'O which athletic events are more likely t'O be
s'Ource d'Omains f'Or sp'Orts metaph'Ors. Acc'Ording t'O H'Owe (1988), f'Or instance,
the m'Ost p'Opular sp'Orts metaph'Ors in American p'Olitical disc'Ourse are 企om
American f'O'Otball, baseball, and b'Oxing. On the 'Other hand, my 'Own 'Observati'On
in Chinese has singled 'Out v'Olleyball, s'Occer, and pingp'Ong 臼 c'Omm'On
s'Ource d'Omains f'Or sp'Orts metaph'Ors in Chinese p'Olitical and everyday disc'Ourse.
It is hard t'O imagine, f'Or instance, that American football metaph'Ors
sh'Ould enj'Oy p'Opularity in Chinese while m'Ost Chinese pe'Ople do n'Ot kn'Ow
whata ‘ t'Ouchd'Own' is.
While 1 am in c'Omplete ~gr~~!llent with Quinn's c'Onclusi'On that cultural
m'Odels play a maj'Or r'Ole in c'Onstit1,lting 'Our understanding 'Of the w'Orld and
c'Onstrain the selecti'On 'Ofmetaph'Ors,lI w'Ould like t'O raise a questi'On that challenges
her betief that metaph'Ors 'Only play a min'Or r'Ole in c'Onstituting our understanding
'Of the w'Orld and are 'Ordinarily selected t'O fit 'Or satisty the preexisting
cultural m'Odels. My questi'On is: C'Ould the cultural model, or culturally
shared understanding, itself be metaph'Orical 'Or 企ee 'Of metaphor? When it is
said that "An underlying metaph'Or f'Or life in the United States is LIFE IS
PLAYING A GAME" (Ching 1993: 43), is this metaph'Or the culturally shared understanding
in the American culture that c'Ontr'Ols the GAME and PLAY metaph'Ors
which are pervasive in American English? Is 也is metaph'Or entrenched in
the middle 'Of the cultural m'Odel sh缸ed by American people in their understanding
'Of American life? A p'Ositive answer seems t'O make m'Ore sense. If, in
sh'Ort, a cultural m'Odel 'Or culturally shared understanding c'Ould be metaph'Orical
in nature, then the r'Ole played by metaph'Or in human understanding w'Ould
consequently be maj'Or as well. Chapter 3 'Of this b'O'Ok will provide further supp'Ort
t'O this claim. Readers are also referred to Gibbs (1994a) and Kövecses
(1995c, 1997) f'Or critiques 'OfQuinn's view.17
Finally, 1 w'Ould like t'O digress and discuss a little 'On the retati'Onsbip between
anthr'Op'Ol'Ogy a
AREVIEW 45
two disciplines have different focuses of study and their approaches,may proceed
in opposite directions. For linguists, "cultural models promise the key to
linguistic usage," whereas for anthropologists, "linguistic usage provides the
best available data for reconstruction of cultural models" (p. 24). In the past, as
Keesing (1992: 593), an anthropologist, has pointed out, "Linguistics and anthropology
have had a curious, <!i.~lectical relationship, sometimes coming together
in productive conjunctures, sometimes drifting apart." He suggested that
"after a long period of estrangement, anthropology and linguistics are overdue
for a reconciliation which could be productive for both disciplines" (p. 593).
The period of reconciliation and dialogue has come after the emergence of
cognitive anthropology and cognitive linguistics which are now sharing an area
of common ground and working "toward each other's positions 企om opposite
directions" (MacLaury 1995: x). Yet, the gap that divides them still exists.
Palmer (1 996), another anthropologist, has recently pointed out that cognitive
linguistic studies often seem to lack an essential cultural dimension, whereas
few of cognitive anthropological studies have incorporated recent developments
in cognitive linguistics.18
Both cognitive anthropologists and cognitive linguists are trying to account
for human understanding and human cognition. Assumably, however, there are
at least two kinds of human understanding. One is culturally shared understanding-ì)äsed
on cultural models of the 飞~õfld received across individual
minds within a particular culture. The other is universally shared understanding
based on common human experience of the world. But these two kinds of understanding
are merely "two sides of the sarne coin, each side illuminating the
other" (Keesing 1992: 601). If cognitive anthropologists and cognitive linguists
have their emphasis on one side over the other, their contributions to
their common endeavor to account for human cognition are complementary,
rather than con企ontational， to each other. It is simply what MacLaury (1995:
x) calls "a natural division of labor" within the sarne task. For instance, it
seems reasonable to argue for a distinction between the ‘enculturated' meaning
and the ‘embodied' meaning based on the distinction between culture and biology.
Nevertheless, the distinction between cultur
If meaning is structured and guided by the mapping of the body in the brain, then
it is not possible to separate human culture from human bodies. Culture is pattems
of activity in brains; brains are structured in accord with their bodies; therefore
cuHure, which is activity in brains, is structured in accord with the bodies in
46 THE CONTEMPORARY THEORY OF METAPHOR
which it resides. Conversely, brains are in various ways developed under cultural
experience, such as experience of language. A certain amount of our ac阳al neurobiology
is inseparable 企om culture.
That is, culture and biology .are mutua11y dependent and coexistent. For this
reason, a complete study of human meaning must inc1ude both ‘四culturated'
and ‘embodied' meaning, so as to reveal the whole picture ofhuman cognition
in terms of how it is relative across different cultures and universal among a11
human beings. Cognitive linguistics and cognitive anthropology should join
their hands closely in their search into human cognition. The.coalition between
these and other cousin disciplines is crucial for the success of the current
"Cognitive Revo1ution" (Bruner 1996, Shore 1996).'9
2.9. Questions faced by the contemporary theory
In his review of Metaphors We Live By (Lakoff & Johnson 1980), Lawler
(1983: 205) had this to say about the significance ofthe book to the field of
metaphor research:
From the viewpoint ofa metaphor researcher,this book is ciearly a milestone,but
it does not point in any particular direction for further research一-rather， it points
in many directions. One obvious suggestion is that other languages than English
should be treated along the lines which Lakoffand Johnson lay down.
Keesing (1985: 201) also argued that Lakoff and Johnson and many others'
discovery in English of the systematic paradigmatic nature of conventiona1
metaphors and their experiential bases ca11ed for more systematic exploration
"to m旦p the metaphoric schemata of nçm-Westem peoples." He stressed that
"the mapping and analysis of metaphoric schemata in non-Westem languages
must be given a high research priority" (p. 214). More recent1y, Mühlhäusler
(1995) expressed the urgent need to intensively study ‘metaphors others live
by.' He believed that intensive study of non-Westem metaphorica1 systems
would even help solve problems such as social, technological, environmental,
and philosophical, in Westem cultures by "generating a1temative ways of
looking at things" (p. 282).20 In spite of a11 these ca11s, intensive studies of
metaphor systems in non-Westem languages, and even in Westem languages
other than English, have not been carried out satisfactorily, to say the least.
AREVIEW 47
As 1 see it, the contemporary theory of metaphor stil1 faces two major
questions that require cross-linguistic and cross-cultural research. The first is :
whether abstract human reasoning is at least partial1y a metaphorical version of
imagistic reasoning. Lakoff (1990: 39) sees 也is as "a major question for future
research in cognitive linguistics." There exists some evidence suggesting that
such abstract concepts as time, states, changes, causes, purposes, quantity
scales, categories, as well as emotions, are characterized metaphorically. But to
what extent? The existing evidence is mainly 丘om English. But is it true in
other languages as well? ,\J1d to what extent is it true in other languages?
The second question is that of universality versus relativity. According to
Lakoff(1993a: 245), "Metaphorical mappings vary in universality; some seem
to be universal, others are wid巳spread， and some seem to be culture specific."
But it is still unknown as to what and how conceptual metaphors are universal,
widespread, or culture-specific.
Since, as argued, human understanding, meaning, and reasoning are
grounded in our embodied experience, and since basic bodily experience
should be common among al1 human beings, it can be hypothesized that there
exist cognitive universals, as well as linguistic universals. On the other hand,
since bodily experience always interacts with specific physical, social, and
cultural environmeIits, it is also expected that there should be cognitive variations
across cultures and languages. However, to what extent and in what manner
cognitive universals and variations exist across cultures and languages is
still a largely unexplored area.
Regarding the lack of cross-linguistic and cross-cultural studies, Johnson
(1992: 354)21 has made the following incisive statement:
Given the nature of our bodies and brains, and given the kinds of physical and
cultural interactions we engage in because of the kinds of interests and purposes
we have, there may well be universal image schemas, metaphorical concepts, or
cognitive structures. Whether there are such universals is an empirical issue. The
cross-cultural studies 出at could identify such empirical universals have simply
not been carried out extensively enough at the present time. So, we cannot make
any strong assertion along these lìnes. Neither, however, can we deny their existence.
We wi11 only know the answer when we do the necessary cross-cultural re-
search.
Whether certain conceptual metaphors, image schemas, or cognitive prototypes
are universal, as 1believe, awaits further study.
48 THE CONTEMPORARY THEORY OF METAPHOR
Johnson's statement summarizes the situation of cognitive semantics: cognitive
semantics has reached a point where it has to be supported by cross-cultural
research.
My studies presented in the next three chapters a忧empt to make a contribution
to this end.