Figurative Language in a Modern Theory of Meaning Construction:
A Lexical Concepts and Cognitive Models Approach
Vyvyan Evans and Jörg Zinken*
University of Brighton and University of Portsmouth
[Draft September 2006. To appear in Art, body and embodiment, edited by C. Makris and R. Chrisley.
To be published by Cambridge Scholars Press.]
1. Introduction
The purpose of this volume is to explore the relationship between art, body and
embodiment. One of the central propositions associated with the interdisciplinary
approach to language and mind known as ‘cognitive linguistics’ is that the nature of
human embodiment affects both what we can experience and how we experience it.
This view, known as the thesis of embodied cognition, holds that the range of
concepts that populate the human conceptual system bears the imprint of human
embodiment (see Evans 2004; Evans and Green 2006; Johnson 1987; Lakoff 1987;
Lakoff and Johnson 1999). As forms of expression, such as art, make use of the
human symbolic ability, which relies upon the conceptual system, it follows that
artistic expression must reflect the nature of human embodiment, and thus bodily
experience. One way in which art manifests itself is through language, particularly
literary language. Here we will be concerned with one form that literary language
takes, namely figurative language.
In this paper we are not primarily concerned with studying figurative language
qua art. Nor do we directly focus on the bodily basis of figurative language. Rather,
*
Contact details for Evans: School of Languages, University of Brighton, Falmer
Brighton, BN1 9PH, UK. E-mail: [email protected] . Contact details for Zinken: Department of
1
our concern is to tackle the nature of two particularly well-studied tropes: metaphor
and metonymy, and offer a fresh perspective on the nature of these literary forms. We
do so in the light of recent advances in cognitive linguistics. Our rationale in adopting
this somewhat narrower focus is as follows. We take the view that it is only by first
providing an account of figurative language which accords with what we now know
about the nature of language as a usage-based system, and moreover, what we also
know about the nature of the conceptual system and resources upon which language
and meaning-construction processes draw, that we can be in a position to study the
relationship between the bodily basis of figurative language and its role in literary
languague.
Accordingly, our aim in the present paper is to formulate a new model of
figurative meaning construction, which builds on recent advances in cognitive
semantics on lexical representation.1 This account also builds on and develops Croft'
s
(1993) domain-highlighting account. The new model can explain similarities and
differences between literal and figurative language as well as between metaphor and
metonymy. It also allows us to address the difference between literal and figurative
language use as a matter of degree, rather than as a principled distinction in meaning
construction processes. We outline a theory of lexical representation and apply it to
typical examples discussed in the literature on metaphor and metonymy. Alternative
current approaches to figurative language which we comapre and contrast our
approach with include Relevance Theory (Sperber and Wilson 1995) and Conceptual
Metaphor Theory (Lakoff and Johnson 1980, 1999) as well as recent proposals in the
Philosophy of Language literature. While all of these approaches make important
Psychology, University of Portsmouth, King Henry Building, King Henry I Street, Portsmouth, PO1
2DY, UK. E-mail: [email protected]
1
Cognitive semantics is a sub-branch of cognitive linguistics. For a comprehensive overview of both
cognitive semantics and cognitive linguistics see Evans and Green (2006).
2
contributions to our understanding of figurative language, none provide, we argue, a
cognitively realistic account of figurative meaning construction, in the light of recent
developments in terms of our knowledge of lexical representation and meaningconstruction. We attempt to make a step towards such an account by modelling: the
relation between semantic representation and conceptual representation; the difference
in meaning construction involving metaphorically used forms on the one hand, and
conventionally used forms, including conventionalised '
metaphors'
, on the other hand;
and the processes involved in meaning construction in general, and figurative
meaning construction in particular.
2. Metaphor and Metonymy
Our aim in this paper is to develop an account of the range of linguistic phenomena
often described as constituting ‘metaphor’ and ‘metonymy’. These are exemplified
by expressions of the following kind:
Metaphor
(1)
a. Achilles is a lion
b. Time whizzed by
Metonymy
(2)
a. France beat New Zealand in the Rugby world cup
b. The ham sandwich asked for the bill
3
In modern linguistics, metaphor is often understood as involving the interpretation (or
conceptualisation) of one entity in terms of something else, as in Achilles in terms of
a lion, or time in terms of an object in motion. Metonymy on the other hand is often
taken to relate to a referent other than the one literally designated. For instance, in
(2a), France refers to the national Rugby team of France, a group of 15 players, rather
than the nation-state. Similarly, ham sandwich refers to the customer who ordered the
ham sandwich rather than to the sandwich.
The received view in both philosophy of language and (formal) semantics has
been to assume that figurative language of this kind is deviant. After all, the
examples in (1) and (2) don’t literally mean what they say. It is often assumed that
literal language is the norm and figurative language requires a ‘special’ kind of
additional processing.2
In contrast, recent work in cognitive linguistics, in particular, as developed by
Lakoff and Johnson (e.g., 1980, 1999), argues that figurative construals such as
metaphor are in fact fundamental to conceptual organisation, and that figurative
language reveals systematicities in the way conceptual knowledge is structured. For
instance, Lakoff and Johnson argue that in an example such as (1b) time is understood
in terms of motion precisely because the ‘domain’ of time is systematically structured
in terms of motion, as evidenced by a welter of examples:
(3)
a. Christmas is approaching
b. The time for a decision has come
c. Time flies when you’re having fun
d. My favourite part of the piece is coming up
2
For a review of this position and difficulties with it see Gibbs (1994).
4
However, Conceptual Metaphor Theory as developed by Lakoff and Johnson
with their various collaborators (e.g., Lakoff and Johnson 1980, 1999; Lakoff 1987,
1993; Lakoff and Turner 1989), while extremely influential, suffers from a number of
problems. One of these is that while Conceptual Metaphor Theory studies and draws
conclusions based on linguistic phenomena, it is not a theory about linguistic
organisation, nor is it a theory which is primarily concerned with the nature of
linguistic meaning construction. Rather, it is concerned with using language as a
methodological tool for inferring the existence of language-independent knowledge
structures, ‘conceptual metaphors’. Yet, while these are inferred on the basis of
linguistic expressions, they are claimed to license the very linguistic patterns which
provide the evidence for the conceptual metaphors in the first place, a circular form of
reasoning. One of our main objectives in the present work is to develop a cognitive
linguistic account of the functions that the semantic structure associated with wordforms has in figurative meaning construction. We will argue that metaphor and
metonymy can be more parsimoniously accounted for on the basis of an approach to
linguistic semantics (a theory of lexical representation and meaning construction)
which accounts for both literal and figurative language with a single set of
mechanisms. Accordingly, we argue that positing an additional level of stable subsymbolic knowledge structures is not warranted.
The second problem is that while Conceptual Metaphor Theory represented
one of the earliest perspectives in cognitive linguistics, it is an approach which has
been superseded by more recent developments. While this is not itself problematic,
for some researchers, particularly in cognitive linguistics and other cognitive sciences
directly influenced by this movement, the existence of conceptual metaphors appears
5
sometimes to be taken as being axiomatic. As Conceptual Metaphor Theory was
developed prior to a number of crucial advances in our understanding of the nature of
semantic representation, it is premature to conclude that anything like conceptual
metaphors may actually approach a plausible model of conceptual organisation. This
is even more so given that Conceptual Metaphor Theory remains programmatic.
One of the recent developments in our understanding of semantic
representation is the work on encyclopaedic representation (e.g., Langacker 1987; see
Evans and Green 2006 for a review). Croft (1993) has begun to show how an
encyclopaedic approach to linguistic meaning provides a way of understanding the
distinction between metaphor and metonymy, and, moreover, how figurative language
is continuous with literal language.
Our aim in this paper is to continue this line of research, developing it on the
basis of current research on the nature of lexical representation, on the way lexical
structures interact with encyclopaedic knowledge, and on sentential meaning
construction. From the present perspective, the model proposed by Lakoff and
Johnson in 1980, and with relatively minor revisions in 1999, oversimplifies, and to
some extent misrepresents the nature of the conceptual processes involved in
meaning-construction, and, in particular, the nature of meaning-construction as
evident in figurative language use.
In order to be able to present our account of metaphor and metonymy we must
first develop and present an account of lexical representation and meaningconstruction. This is the task of the next section, section 3. We argue that linguistic
units such as words and grammatical constructions are associated with semantic units
which we refer to as lexical concepts. These lexical concepts provide access to
larger-scale conceptual knowledge structures termed cognitive models. Meaning is a
6
function of combining lexical concepts in a number of ways, which give rise to a
situated, i.e. utterance-specific, interpretation. Thus, we argue that meaningconstruction constitutes a situated means of expressing a local communicative
intention. From this perspective, meaning is a process, and moreover, one which
results from the way in which conceptual knowledge structures are accessed. Thus,
the distinction between literal versus figurative language relates not to distinct kinds
of knowledge, or even language use. Rather, figurative language provides different
sorts of access to the cognitive models which lexical concepts are relativised with
respect to. We develop our notion of figurative language and the differences between
the two types of figurative language we consider in section 4 of the paper. In section
5 we compare and contrast our account with Croft’s domain highlighting model,
Conceptual Metaphor Theory, and the Relevance Theory account of metaphor.
Conclusions are presented in section 6.
3. Lexical Concepts, Cognitive Models and Meaning-construction
In this section we present a brief overview of the Theory of Lexical Concepts and
Cognitive Models, or LCCM Theory for short. LCCM Theory underpins our account
of figurative language. Our presentation of LCCM Theory is based on Evans (2006,
To appear).
The starting points for LCCM Theory are the observation that word
‘meanings’ are protean (flexible, indeterminate and context-dependent), and the
position that linguistic meaning is a situated phenomenon, a function of language use,
rather than an inalienable property of language. These starting points give rise to a
number of desiderata for a theory of lexical representation and a theory of concept
integration, which together should contribute to a descriptively adequate and
7
psychologically realistic account of meaning-construction. Firstly, lexical entries
need to be relatively detailed and flexible. Secondly, our theory of lexical
representation should interface with conceptual knowledge. That is, we require a
theory that takes an ‘encyclopaedic’ perspective on linguistic meaning. Thirdly, we
also require an account of how lexical representations combine in order to provide
situated meanings.
Finally, as the meanings associated with words are a function of specific
utterances, and thus a consequence of discrete usage-events, our theories of lexical
representation and concept integration must be usage-based. In view of these
desiderata, LCCM Theory has two components: i) an account of lexical
representation, involving two theoretical constructs: lexical concepts and cognitive
models, which are developed in more detail below, and ii) mechanisms which
facilitate the composition of lexical concepts (see Evans To appear for further details).
The understanding of composition to be developed is quite distinct from the Fregean
notion of compositionality.
3.1. Lexical Representation
In this section we sketch the view of lexical representation we will be assuming in the
rest of the paper.
MEANING AND USE
Language use is integral to our knowledge of language, our ‘language system’ or
‘mental grammar’. The organisation of our language system is intimately related to,
and derives directly from, how language is used (Croft 2000; Langacker 2000;
8
Tomasello 2003). Through processes of abstraction and schematisation (Langacker
2000), based on pattern-recognition skills and abilities in understanding other minds
(Tomasello 1999, 2003), language users derive linguistic units. These subsist in the
mental grammars of adult language users as relatively well-entrenched mental
routines consisting of conventional pairings of form and semantic representation.
However, the range of linguistic units available to the language user
massively underdetermines the range of situations, events, states, relationships and
other interpersonal functions that the language user may potentially seek to use
language to express and fulfil. One reason for this is that language users live in a
socio-physical ‘matrix’ that is continually shifting and evolving. No two situations,
feelings or relationships, at any given point in time, are exactly alike. We are
continually using language to express unique meanings, about unique states of affairs
and relationships, in unique ways. While language has a range of ‘ready made’
schemas, or linguistic units which can be combined to express a certain range of
scenarios we may wish to refer to and describe, these necessarily underdetermine the
mutability of human experience. As Langacker puts it, “linguistic convention cannot
provide a fixed, unitary expression for every conceivable situation that a speaker
might wish to describe” (1987: 278). Accordingly, the linguistic units employed by
language users can only ever partially sanction (in Langacker’s e.g., 2000 terms) the
situated way in which they are used. Language use involves solving a co-ordination
problem, in which language users must employ non-conventional co-ordination
strategies and devices (Croft 2000). Language users employ the conventional
repertoire of linguistic units, including patterns of assembling linguistic units (such as
9
word order conventions, which are themselves linguistic units), in non-conventional
ways.3
On this view, meaning, which is associated with the utterance (or usageevent), is a consequence of combining the conventional schemas or linguistic units in
novels ways in order to solve the particular co-ordination problem at hand, thereby
facilitating communication. The reason, then, for the protean nature of the
‘meanings’, or semantic representations, associated with words is that linguistic units
are only ever realised as part of linguistic utterances. But in being so realised, they
have necessarily undergone context-induced ‘shifts’ in their semantic value. As with
the distinction between allophones and phonemes in phonological theory, linguistic
representations, by analogy akin to phonemes, are never actually perceived, but are
inferred based on how ‘sense-shifted’ words appear to behave in situated usageevents, as judged over many instances of use. In this, then, the job of the lexical
semanticist is to employ meaning in usage-data, by analogy akin to allophones, in
order to infer the existence of the underlying lexical concepts which partially sanction
the semantic contributions which surface.
The essential distinction between lexical representation and meaning is that
while meaning is a property of the utterance, lexical representations are the mental
abstractions which we infer must be stored as part of the language user’s knowledge
of language, in order to produce the range of novel uses associated with situated
instances of a particular word (or construction). The meaning associated with an
utterance we refer to as a conception. Thus, conceptions are a function of language
use. Lexical representations, or rather more technically, lexical concepts, represent
3
As Croft (2000) notes, this is precisely why language change is possible, and proceeds as rapidly (in
relative terms) as it does.
10
the semantic pole of linguistic units, and are the mentally-instantiated abstractions
which language users derive from conceptions and the specific semantic contribution
perceived to be associated with particular forms.
LEXICAL CONCEPTS
Having examined the distinction between conceptions (meanings) and lexical
concepts (lexical representations) we now examine the notion of lexical concepts in
more detail.
A lexical concept is a conventional semantic unit which is form-specific.
Lexical concepts constitute the semantic pole in any given linguistic unit, such as a
word.4 By ‘linguistic unit’ we have in mind any conventionalised symbolic formmeaning pairing, in the sense of, for instance, Langacker’s symbolic assemblies
(1987), or Goldberg’s notion of construction (Goldberg 1995; 2003; see Evans and
Green 2006 for a review).
Semantic representations are distinct from conceptual representations. They
constitute the conventional form that conceptual structure takes for being encoded in
language. While semantic structure is a subset of conceptual structure, there are many
more concepts than we have conventionalised linguistic resources for expressing. A
crucial function of meaning-construction, which arises by virtue of language use, is to
prompt for novel conceptions, as we will see in detail in the following section.
Given the notion of a ‘linguistic unit’ that we are assuming, it follows that
lexical concepts are conventionally associated with a wide range of forms. The range
of forms with which lexical concepts are conventionally associated include overt
forms: those with resolved phonetic form, such as cat, and implicit forms: those with
4
Cf. the related notion of lemma (Levelt 1989).
11
unresolved phonetic form, such as the ditransitive construction (SUBJECT VERB OBJ1
OBJ2),
e.g., John baked Mary a cake; John gave Mary the cake; John refused Mary
the cake, etc. (see Goldberg 1995). Moreover, overt forms that have distinct lexical
concepts conventionally associated with them include bound morphemes, ‘simplex’
words, ‘complex’ or polymorphemic words, and idiomatic expressions and phrases.
In addition to grammatical constructions, implicit forms include grammatical
‘relations’ such as subject and object, and lexical classes such as noun and verb.
A single form can be conventionally associated with a potentially large
number of distinct lexical concepts which are related to degrees as attested by the
phenomenon of polysemy.5 Forms are not concept-specific. A consequence of this is
that the lexical concepts which share the same form can be modelled in terms of a
semantic network. As any given lexical concept potentially provides access to other
lexical representations it is associated with, we refer to the lexical concepts B, C, D...
in the same semantic network as lexical concept A as the semantic network profile of
that particular lexical concept.
A lexical concept has a unique lexical profile, its unique ‘biometric’ identifier.
A lexical profile is an extension of criteria presented in Evans (2004a), and akin to the
notion of an ID tag (Atkins 1987) and behavioural profile (Gries 2005). While a
lexical concept associated with a particular form can be glossed by a particular
semantic value—we will use the mnemonic of small capitals inside square brackets—
determining whether a particular usage of a form relates to one lexical concept rather
than another is a matter of examining the selectional tendencies (i.e., semantic,
collocational and formal patternings) associated with a given usage. While any given
usage of a lexical concept will have its own unique selectional specifications, general
5
See Evans (2005) and Tyler and Evans (2001, 2003) for detailed discussion of polysemy.
12
patterns (‘tendencies’) can be established, and form part of the conventional
knowledge associated with a particular lexical concept.6 General selectional patterns
in terms of semantic, collocational and grammatical tendencies are what we refer to as
a lexical profile. Some examples of lexical concepts related by polysemy are given
below (for detailed discussion of time, see Evans 2005; for over, see Tyler and Evans
2003; for fly, see Evans and Green 2006):
time (noun)
(4)
a. The relationship lasted a long time
[DURATION]
b. The time for a decision has come
[MOMENT]
c. Her time [=death] has come
[EVENT]
d. British Summer Time begins today
[MEASUREMENT-SYSTEM]
e. The sales figures improved for the third time in the quarter
[INSTANCE/OCCURRENCE]
over (preposition)
(5)
6
a. The picture is over the sofa
[ABOVE]
b. The picture is over the hole
[COVERING]
c. The ball is over the wall
[ON-THE-OTHER-SIDE-OF]
d. The government handed over power
[TRANSFER]
e. She has a strange power over me
[CONTROL]
f. The relationship evolved over the years
[TEMPORAL]
Identifying such selectional tendencies is ultimately an empirical question. Important techniques in
13
fly (verb)
(6)
a. The plane/bird is flying (in the sky)
[SELF-PROPELLED AERODYNAMIC
MOTION]
b. The pilot is flying the plane (in the sky) [OPERATION BY AGENT OF ENTITY
CAPABLE OF AERODYNAMIC MOTION]
c. The child is flying the kite (in the breeze) [CONTROL OF LIGHTWEIGHT
ENTITY BY AGENT]
d. The flag is flying (in the breeze)
[SUSPENSION
OF LIGHTWEIGHT OBJECT]
To get a sense of how the lexical profile can be used to identify distinct lexical
concepts which underlie specific usage-events, we present the lexical profiles for a
number of distinct lexical concepts encoded by the polysemous form time (based on
Evans 2004a, b, 2005):
(7)
Time flies when you’re having fun
(8)
The time for a decision is getting closer
(9)
The old man’s time [= death] is fast approaching
(10)
Time flows on (forever)
These instances of the lexical form time all appear in the ‘subject’ phrase. Moreover,
the verb phrase which complements the subject phrase relates to a motion event.
Thus, motion is being ascribed to the entities time contributes in prompting for. Yet,
the semantic contribution associated with time appears to be distinct in each example.
this regard have been developed recently in corpus linguistics. See for instance Gries (2006) and
14
In (7), the semantic contribution associated with time appears to relate to an
assessment of temporal magnitude. Thus, we might gloss the semantic value
associated with this instance of time as [DURATION]. In (8), the semantic contribution
of time might be glossed as [MOMENT]. This follows as the conception associated
with the utterance as a whole relates to a specific temporal moment when a particular
decision is to be taken. Thus, the contribution of time in this example appears not to
relate to a durational elapse, but rather a discrete instant. In (9), the semantic
contribution associated with time appears to relate to an event, which extra-linguistic
context informs us is death. Thus, this instance might be glossed by [EVENT].
Finally, in (10), the semantic contribution associated with time appears to relate to an
unending temporal elapse. In Evans (2004a) this was described as the ‘matrix’ sense
associated with time, in which we understand time to be ‘the’ event in which all other
events occur. Thus, the gloss we might apply to describe this instance of time is
[MATRIX].
Based on the quite distinct semantic contributions associated with this range of
usages of time, Evans (2004a) argued that there is a range of distinct lexical concepts
associated with time identified by the glosses introduced above. Moreover, each of
these distinct usages has a distinct lexical profile associated with it which supports
this perspective. Two sorts of information form a lexical concept’s lexical profile.
The first relates to grammatical selectional tendencies.
The second concerns
semantic selectional tendencies.
We begin by examining the grammatical tendencies associated with each use
of time. To do this, let’s consider the kind of noun phrase in which each semanticallydistinct use appears. We first note that the examples in (7) and (10) appear, on the
Stefanowitsch and Gries (2006).
15
face of it, to be similar. Neither is pre-modified by a determiner. However, further
examples reveal that the [DURATION] lexical concept of time as in (7) can be
determined by the definite article when the assessment of temporal magnitude is
specific rather than generic, while the [MATRIX] lexical concept cannot be:
(11)
During the dinner date, the time seemed to fly
[DURATION]
(12)
*The time flows on (forever)
[MATRIX]
This patterning appears consistent with the semantics of these uses. While the
[MATRIX] lexical concept already relates to a unique referent, the event which
subsumes all others, and thus further specification with the definite article would be
superfluous, in assessing temporal magnitude both specific and more generic readings
are available, encoded by determiner patterns exhibited in (11) and (7) respectively.
Thus, we can say that while both the [DURATION] and [MATRIX] lexical concepts
appear to pattern formally like mass nouns (see Evans 2004a for evidence that they
fail to allow determination by the indefinite article), the [DURATION] lexical concept,
but not the [MATRIX] lexical concept, allows determination by the definite article.
The examples in (8) and (9) also exhibit unique patterns in terms of
grammatical collocational tendencies, both from each other and from the examples in
(7) and (10). The [MOMENT] lexical concept appears to pattern straightforwardly as a
count noun, allowing determination by the definite article, as in (8), or by the
indefinite article, as in (13) below:
(13)
A time will come when we’ll be forced to make a decision
[MOMENT]
16
The [EVENT] lexical concept in (9) appears to require a pre-modifying genitive noun
phrase followed by the enclitic possessive ‘-s’, or else an attributive pronoun, serving
a similar function:
(14)
His time [=death] is fast approaching.
Thus, in subject position, these uses of time all appear to have quite distinct formal
selectional tendencies.
Let’s now turn to the semantic selectional tendencies of these uses of the
lexical concepts associated with time. The point here is that the nature of the motion
event encoded by the verb phrase is distinct for each of the semantic uses. Moreover,
the choice of motion event type is compatible with the semantic value of the lexical
concepts which underlie each use (see e.g., Evans 2004a, for detailed discussion). For
instance, the [DURATION] lexical concept, and this particular variant—which in
previous work was referred to as [TEMPORAL COMPRESSION] as this instance relates to
an assessment of temporal magnitude which proceeds more ‘quickly’ than usual—is
complemented by verb phrases which encode motion events which are rapid in nature,
as evidenced by the example in (7).7 In contrast, the [MOMENT] lexical concept
appears to allow a wider range of motion events, including imperceptible motion as in
(15), rapid motion, as in (16), and terminal motion, as in (17):
(15)
The time for a decision has gone/vanished/disappeared
(16)
The time for decision is racing towards us/fast approaching
(17)
The time for a decision is approaching/getting closer/has arrived
17
The [EVENT] lexical concept appears to restrict the range of motion events which can
collocate with it to terminal motion events, i.e., motion events which terminate ‘at’
the experiential locus, typically a human experiencer. Finally, the [MATRIX] lexical
concept appears to require motion events which are non-terminal in nature. That is, it
requires motion events which are on-going, a paradigm example being ‘flow’. Thus,
each of the examples represent specific instances of distinct lexical concepts which
exhibit distinct semantic and formal selectional tendencies: distinct lexical profiles.
An important property associated with lexical concepts relates to their
semantic value. While the lexical concepts mentioned thus far have been provided
with semantic glosses, given in small capitals inside square brackets, these are simply
shorthand labels for the complex conceptual (or “encyclopaedic”) knowledge
structures that lexical concepts provide access to. 8 We refer to the conceptual
knowledge which lexical concepts provide access to as cognitive models.
COGNITIVE MODELS
This section is concerned with introducing and describing the construct of the
cognitive model. Our claim is that cognitive models, related to the notion of frame
(Barsalou 1999), semantic frame (Fillmore e.g., 1982; 1985; Fillmore and Atkins
1992) and base (Langacker 1987), but distinct from all three, are necessary for
understanding the way lexical concepts contribute to meaning-construction. The main
claim is that lexical concepts provide sites of access to, and thus access routes
7
The TEMPORAL COMPRESSION variant of DURATION associated with time can also be structured in
terms of motion events which relate to a lack of perceptual awareness, such as the following: Where
has the time gone? The time seemed to have vanished, etc.
8
Langacker (1987) refers to points of access.
18
through, cognitive models, are relativised with respect to them, and in part constitute
cognitive models.
The case for thinking that lexical concepts presuppose, and are thus relativised
to, background knowledge structures has been made by a number of scholars.
Perhaps the most compelling arguments are associated with the work of Langacker
(e.g. 1987) and his theory of domains, and with Fillmore’s (e.g., 1982, 1985) theory
of Frame Semantics. Fillmore’s point is that any given lexical concept is relativised
with respect to and thus can’t be understood without the other lexical concepts which
collectively comprise the knowledge structure, or semantic frame, that it forms part
of. In related fashion, Langacker argues that part of the meaning of any lexical item is
a function of the knowledge structure or base that is presupposed by it. For instance,
the lexical concept [HUMAN MAJOR ARM JOINT] associated with the form elbow cannot
be properly understood without knowledge of the arm which is necessary for
understanding the nature and function of the elbow.
The reason for preferring the term ‘cognitive model’ over the related notions
of base or semantic frame is that a cognitive model, in our sense, is a coherent, in
large part non-linguistic, knowledge structure, similar to what Langacker seems to
have in mind, and in some statements to what Fillmore has in mind, particularly when
he seems to tend towards the view of a semantic frame as incorporating an
(experiential) scene. That is, it is a richly specified conceptual entity, akin to what
Barsalou (1999) refers to by his use of the term ‘frame’. A cognitive model represents
an interface between richly-specified conceptual knowledge and nodes of access at
particular points in the cognitive model provided by specific lexical concepts. Thus,
lexical concepts provide conventional perspectives or construals with respect to the
cognitive model, in part, constituting it. Yet, a cognitive model is richer than the sum
19
of the lexical concepts which provide access sites to it. This follows as while lexical
concepts are conceptual units specialised for symbolic representation in language,
cognitive models are not. Rather, they are multi-modal conceptual entities, which can
be used as a basis for what Barsalou and others (e.g. Prinz 2002, and Zwaan 2004)
refer to as simulations.
Cognitive models relate to coherent bodies of knowledge of any kind. For
instance, they include knowledge relating to specific entities, such as the complex
knowledge associated with a specific entity such as a ‘car’, or a more specific entity
such as ‘my car’, including information such as whether it needs filling up and when I
last cleaned its interior. Cognitive models can relate to ‘procedural’ bodies of
knowledge, such as ‘cultural scripts’ which form templates for how to interact in
restaurants in order to be seated and secure a meal (cf. Sharifian’s (2003) notion of
cultural conceptualisations which are culturally ‘distributed’; Musolff’s (2004) notion
of a metaphor scenario). Cognitive models also include bodies of knowledge relating
to more abstract entities such as containment, love and physics. They operate at
varying levels of detail, and while stable, are dynamic being in a perpetual state of
modification and renewal by virtue of on-going experience, mediated both by
linguistic and non-linguistic interaction with others and one’s environment.
Lexical concepts may be relativised with respect to more than one cognitive
model. As cognitive models are necessarily connected to and inherit structure from
many others, a lexical concept can potentially be connected to a vast network of
conceptual knowledge. Those cognitive models which are directly accessed by a
lexical concept are referred to as primary cognitive models. The set of such primary
cognitive models for a given lexical concepts is termed the primary cognitive model
profile. For instance, the lexical concept [BOOK] is relativised to at least two primary
20
cognitive models. The first is a cognitive model relating to the physical structure
associated with books, termed BOOK. That is, part of our knowledge associated with
books relates to our knowledge that they have a particular organisation and physical
structure, such as a binding, a hard-back or paper cover, pages with text on, etc. It is
also related to another cognitive model which relates to the activity associated with
books, termed READING. This constitutes an understanding of how we interact with
books and the sorts of skills and activities involved in this interaction. Note that
cognitive models are glossed by mnemonics without square brackets, in contrast to
lexical concepts. Thus, the lexical concept [BOOK] provides an access site in two
primary cognitive models: BOOK and READING. It is for this reason that the form
book, which encodes this lexical concept, can be modified providing distinct
interpretations:
(18)
a. a heavy book
b. a long book
These interpretations are due to different access routes afforded by the access sites in
two cognitive models provided by this lexical concept. Modification of book by the
lexical concept encoded by long provides access to the READING cognitive model, and
our knowledge that some books require extended periods of time to be read.
Modification by heavy relates to the physical organisation of books, and our
knowledge that some such entities consist of more pages and thus are heavier.
3.2. Lexical concept integration
21
The discussion of lexical concepts and how they relate to cognitive models in the
preceding section now allows us to examine how meaning-construction occurs. That
is, we are now in a position to see how lexical representations contribute to the
formation of a conception.
This process of meaning-construction we refer to as lexical concept
integration. We discuss two sub-processes of this, termed selection and fusion.
SELECTION
Selection is the process in which linguistic or extra-linguistic context selects for a
particular lexical concept. Selecting the ‘correct’ lexical concept is required by
fusion, the operation in which lexical concepts are integrated and the resulting
integrations are interpreted.
An example of extra-linguistic context giving rise to selection is the following:
(19)
That recent hike is killing me
The form hike has at least two lexical concepts associated with it. One relates to a
long, typically cross-country, walk, while another relates to an increase in financial
charges or payments of some kind. In the context of a recent cross-country walk, the
example in (19) might relate to sore body parts. In the context of, for instance, a
recent central bank base-rate increase, the speaker might be referring to the difficulty
of keeping up with mortgage repayments. Thus, the extra-linguistic context provides
a means of selecting the most appropriate lexical concept.
Now consider the examples below, which illustrate the role of linguistic
context in selection:
22
(20)
The judge asked the defendant to approach the bar
(21)
The customer ordered her beer at the bar
The form bar has a number of distinct lexical concepts associated with it, including
the ‘bar of a court’ at which the judge sits, and a ‘bar in a public house’ at which
alcohol is purchased and served. The appropriate lexical concept is selected in these
examples due to the linguistic context.
Despite being able to differentially identify the separate contributions of extralinguistic and linguistic context with respect to lexical concept selection, the typical
arrangement appears to involve a process we refer to as co-selection, involving both
linguistic and extra-linguistic context. To illustrate, consider the following utterance
adapted from a recent newspaper headline:
(22)
France shot down the EU constitution
One of the points made by Croft (1993) in discussing a similar example was that
words often appear to take on ‘meaning’ only when it is clear what the ‘meaning of
the whole’, what we refer to as the conception, relates to. That is, it is only by
knowing what the entire utterance relates to that the ‘parts’ can be interpreted. For
instance, France might relate to the geographical landmass, the political ‘nation-state’
entity, the government, the head of state, the people, a national sports team, a
delegation from France, or something else. Similarly, shot down can be interpreted
conventionally as involving the use of a weapon, or figuratively as ‘forcefully
rejecting’ with civil means. Similarly, EU constitution could relate to the membership
23
of the EU, the health of the EU, or the new treaty and charter of rights and social
provision recently presented to European Union member states for ratification.
Co-selection relies upon selecting the most mutually appropriate readings
associated with each of these expressions. Selecting the most appropriate lexical
concept associated with a given form is a mutually-involving ‘symbiotic’ process.
Nevertheless, how is this co-selection process guided? We suggest that it
proceeds by virtue of a discourse “anchor”, which we refer to as the utterance topic.
This is related to what Fauconnier (1997) refers to as a ‘base space’. The topic might
be inferred based on the preceding discourse, the extra-linguistic context, or indeed,
the utterance context itself. However, the topic constitutes a ‘mini-theory’ concerning
the general nature of the conception, and is informed by the hearer’s assumptions
regarding the speaker’s communicative intention. Thus, co-selection crucially relies
on knowing the topic, which guides the co-selectional process. This is what Croft
(1993) refers to when he talks of the whole leading to our understanding of the parts.
Unless we understand that the topic of the conception illustrated by (22) relates to
European Union politics, we will be unable to make the most appropriate lexical
concept co-selections, and thus be unable to build the conception that the newspaper
headline writer has in mind.
FUSION
Fusion concerns the process in which selected lexical concepts are composed such
that they give rise to a particular conception. Fusion involves two component
processes: integration and interpretation. We address each in turn.
Integration is the process in which selected lexical concepts are incorporated
into larger structures. One way in which this process occurs is elaboration
24
(Langacker, 1987). For instance, the semantic representation associated with a verb
such as kick encodes schematic roles for ‘kicker’ and ‘kickee’. These roles can be
integrated with lexical concepts encoded by other lexical forms, e.g., He kicked me.
Of course, integration can become more complex when it involves lexical
concepts which have more than two elaboration sites. A case in point is the so-called
ditransitive or double-object construction. Goldberg (1995) shows that this
grammatical form has, in present terms, a conventional lexical concept associated
with it which can be glossed as [X CAUSES Y TO RECEIVE Z]. The letters x, y and z
correspond to distinct elaboration sites:
(23)
She kicked him the ball
There is more to integration than elaboration. Composite lexical structures
can themselves be conjoined with other composite lexical structures. This process we
refer to as constituency reanalysis. Essentially, this is the process in which composite
lexical structures once established are treated as unitary entities or constituents at the
next level of processing. For instance, the lexical concept encoded by and is
specialised for integrating composite lexical structures. Thus, integration involves
elaboration (building of composite constituents), and constituency reanalysis
(building utterances from smaller ‘meaning’ constituents). However, and crucially,
what licenses these processes, providing coherence to the integration, is compatibility
of the lexical concepts involved, rather than, for instance, semantically ‘blind’
syntactic processes, as in many formal approaches (see Evans and Green 2006, for a
review).
25
Interpretation is a process that proceeds in conjunction with integration.
While integration serves to conjoin lexical concepts, interpretation serves to highlight
part of the (primary) cognitive model profile that each lexical concept provides access
to, in a way that is consistent with the other lexical concepts of the composite lexical
conceptual structure. In other words, it is not enough for meaning-construction to
select an appropriate lexical concept, and integrate lexical concepts into a composite
lexical-conceptual structure. In addition, the selected lexical concept must then be
interpreted within the composite lexical-conceptual structure, the new linguistic
context, in which it occurs. This process of interpretation, then, provides the crucial
break between lexical representation and meaning-construction. It is as a
consequence of interpretation that a conception arises.
Lexical concepts provide different ‘access routes’ through the cognitive model
profile. This is facilitated as lexical concepts serve as access sites to typically more
than one primary cognitive model, and are thus associated with a primary cognitive
model profile, and indeed, secondary cognitive models as we shall see in the next
section when we turn to figurative language.
To illustrate access we reconsider the examples in (18), reproduced below:
(18)
a. a heavy book
b. a long book
The distinction between the examples (18a) and (18b) relates to access in different
primary cognitive models. That is, the lexical concept [BOOK] encoded by the form
book in these examples is interpreted slightly differently by virtue of the contextdependent lexical concepts which serve to differentially modify the form book.
26
Hence, a ‘heavy book’ requires accessing conceptual knowledge relating to the
structure and physical make-up of books, while a ‘long book’ requires accessing
knowledge relating to the activity of reading, and knowledge concerning the length of
time required for reading books of different kinds. The distinction between the
cognitive models involved and the attributes they subsume are captured in figure 1.
TOME
TEXT
BOOK
DURATION
READING
[BOOK]
Level of attributes
Level of cognitive models
Level of lexical concepts
Figure 1: The relationship between lexical concepts, cognitive models and attributes
In figure 1, lines between boxes denote conceptual connections. The lexical
concept [BOOK] can access either of its distinct primary cognitive models: BOOK and
READING,
as evidenced by the examples (18a) and (18b), or a particular attribute of a
single cognitive model can be differentially perspectivised.
4. Figurative language
In the previous section we saw that meaning-construction is a process which relies on
lexical representation, modelled here in terms of lexical concepts and cognitive
27
models. We also presented an account of semantic composition which involves the
constituent processes of selection and fusion. Selection involves the contextdependent process of identifying an appropriate lexical concept: the lexical
representation conventionally associated with a particular form. However, meaningconstruction is a function of situated language use, and thus results not from the
concatenation of lexical concepts but rather from their fusion, which is to say
integration and interpretation.
Integration is a process in which lexical concepts are ‘embedded’ in nested
fashion in other lexical concepts which, thereby, affects the semantic contribution of
the lexical concepts involved. Interpretation relates to the access route that the lexical
concept provides in the cognitive model profile associated with the lexical concept.
In this section we are concerned with presenting an account of figurative
language. We will focus on two types, metaphor and metonymy. What is common to
both, we shall argue, is that they constitute instances of meaning-construction
employing basically the same principles as those outlined for literal language. The
distinction between literal and figurative language, as a principled one, dissolves.
Still, the intuition that there is a distinction is correct. We argue that the intuition
arises because the kind of access route in ‘figurative’ language use differs from
meaning-construction relating to language use that is commonly identified as ‘literal’.
Both metonymy and metaphor involve a form of interpretation which is distinct from
that involved in literal language use, as we explicate below. The access routes evident
in both metaphor and metonymy involve access to secondary cognitive models, which
is to say cognitive models not directly accessed by the lexical concept. Such
cognitive models are accessed via the primary cognitive models which are more
closely associated with the lexical concepts in question.
28
However, there is an important distinction between the sorts of situated
meanings identified as relating to metaphor and metonymy. The distinction relates to
the other aspects of meaning-construction identified earlier. We propose that
conceptions often identified as involving metaphor differ from metonymy in terms of
integration. As metaphor involves saying something about something else, and thus
involves attribution or predication, attribution or predication is a function of the
linguistic units in which the interpretation process occurs. That is, the linguistic unit
is associated with a lexical concept which itself provides instructions as to how the
constituent lexical concepts which ‘fill’ the construction should be integrated. This
process of integration thus guides the interpretation of the lexical concepts involved.
Metonymy, on the other hand, while also involving interpretation, is guided primarily
by co-selection. That is, metonymy involves an interpretation of a lexical concept
which is in keeping with the other lexical concepts selected for in the utterance. Thus,
metonymy involves an ‘identity’ or ‘referential’ function. Interpretation comes
directly from the access route provided by the lexical concept itself. Thus, the
distinction between metaphor and metonymy can be viewed as one relating to access
to secondary cognitive models mediated by integration (metaphor), or secondary
model access primarily mediated by co-selection (metonymy).
4.1. Metonymy
We begin by examining the sort of interpretation resulting in the formation of a
conception commonly referred to as metonymy. It is important to emphasise that we
use the term metonymy only in the sense that it is a convenient and short-hand label
for a form of interpretation (i.e., interpretation mediated by an access route to a
29
secondary cognitive model profile), resulting in situated meaning-construction. To
illustrate, we will consider examples relating to the lexical concept [FRANCE]:
(25)
France is a region of contrasting landscapes.
(26)
France is one of the EU’s pivotal states
(27)
France beat the New Zealand Rugby team in the World Cup competition
(28)
France rejected the EU constitution
In order to understand metonymy as an access route phenomenon which is no
different in principle from the routes of access provided by the lexical concept for
[BOOK] discussed above, we must first get some sense of the complexity of conceptual
knowledge we have of France. Consider figure 2.
CONSTITUTIONAL
SYSTEM
NATIONAL
SPORTS
GEOGRAPHICAL
LANDMASS
ELECTORATE
POLITICAL
SYSTEM
NATION
STATE
HEAD OF
STATE
CUISINE
HOLIDAY
DESTINATION
[FRANCE]
Figure 2: Partial cognitive model profile for [FRANCE]
30
Figure 2 represents a partial cognitive model profile associated with the lexical
concept [FRANCE]. Each of the individual cognitive models involves a hugely
complex body of interrelated knowledge. In figure 2, which is meant to be suggestive
only, we list three cognitive models which make up a very partial primary cognitive
model profile. We suggest that [FRANCE] directly provides access to a cognitive
model relating to France as a geographical region, including knowledge concerning its
terrain and geo-physical attributes, a cognitive model relating to France as a political
entity, and France as a holiday destination. Each of these cognitive models, in turn,
provides access to further cognitive models. For instance, we diagram a number of
cognitive models which are accessed from the NATIONSTATE cognitive model. Each
of these provides access to further cognitive models. We’ve given three which relate
to the POLITICAL SYSTEM model. Only the first three cognitive models:
GEOGRAPHICAL LANDMASS, NATION STATE
and HOLIDAY DESTINATION constitute the
(partial) primary cognitive model profile. The other models comprise the secondary
cognitive model profile, as these are not directly accessed by the lexical concept
[FRANCE]. They are accessed indirectly, via the primary cognitive models. In other
words, the primary cognitive models are presupposed by the lexical concepts, while
the secondary cognitive models provide knowledge which is presupposed by that
contained in the primary cognitive models.
Now let’s return to the examples in (25) to (28). What we see is that the
examples divide between those which provide access to primary cognitive models,
(25) and (26) and those which provide access to secondary cognitive models, (27) and
(28). For instance, in (25), the lexical concept [FRANCE] provides access in the
GEOGRAPHICAL LANDMASS
NATION STATE
cognitive model, while in (26) access is provided in the
cognitive model. In (27) and (28), usages which are typically judged
31
as being more metonymic in nature, access is provided in the secondary cognitive
model profile. In (27) for instance, access is provided in the NATIONAL SPORTS
cognitive model. Crucially, this attribute is accessed by the same lexical concept that
is employed in the examples in (25) and (26). Similarly, the example in (28) provides
access in the ELECTORATE cognitive model, which is accessed via the NATION STATE
primary cognitive model.
Thus far this discussion has illustrated that metonymy involves access via a
lexical concept to the cognitive model profile in the same way as literal language
usage. The difference between metonymy and literal language use, however, is that
the access route involves secondary cognitive model access, in the case of metonymy,
whereas access involves the primary cognitive model profile in the case of literal
language.
Access, of course, results from the interpretation of the lexical concept in
question, here [FRANCE]. But what drives the process of interpretation in each of
these examples is the process of selection. That is, it is only by virtue of selecting
appropriate lexical concepts, that the lexical concept [FRANCE] receives the utterance
specific interpretation it does, and thus the route through the cognitive model profiles
that it provides access to. The ‘identity’ associated with each usage of [FRANCE] is in
part determined by the other lexical concepts selected, which comprise the utterance
context in which the [FRANCE] lexical concept is embedded. The importance of
selection for lexical concept interpretation distinguishes metonymy from metaphor.
Metonymy involves an ‘identity’ or ‘referential’ function, in the sense that the
‘identity’ of the lexical concept is subject to interpretation. But interpretation is
guided by the surrounding lexical concepts. For instance, we know that [FRANCE] in
(28) provides access in the ELECTORATE cognitive model precisely because the
32
linguistic and extra-linguistic context guides selection of the other lexical concepts in
the utterance. These in turn co-determine the access route that [FRANCE] in this
utterance facilitates.
4.2. Metaphor
Metaphor has often, particularly in the philosophy of language and literary traditions,
been associated with utterances of the following kind, reproduced from above:
(1) a. Achilles is a lion
What is strikingly figurative about examples such as this is that Achilles is not
normally taken as being a member of the class of lion. However, the nominal
predicate construction typically does have the function of class-inclusion associated
with it:
(29) Sally is a pianist
This sentence is intransitive. That is, it features the copula ‘be’, which takes not a
direct object, but rather an intransitive nominal known as a (nominal) predicate. The
function of be in this construction is to signal a stative relation (Langacker 1991):
Sally is a member of the class of pianists.
The same cannot hold for the Achilles example as Achilles was a man, and
moreover a famed Greek warrior. Rather, we understand this utterance to mean that
Achilles is courageous and valiant, much as our folk model for lions holds them to be
courageous and valiant.
33
The received view of metaphor then considers examples such as the one in
(1a) to be quite distinct from the example in (29). 9 The present approach to meaningconstruction allows us to see the similarities and differences between metaphor and
the ‘literal’ nominal predicative type examples such as (29). An important point of
similarity relates to the process of fusion crucial for meaning-construction. As
presented earlier, this involves two constituent processes: integration and
interpretation. Just as we saw with metonymy, metaphor diverges from ‘literal’
language use in terms of the sorts of access routes it provides, which is to say in terms
of interpretation.
In an utterance such as ‘Achilles is a lion’, the lexical concepts selected
include the nominal predicate construction itself. That is, this construction which has
the form [NomSubj IS NomPred] is conventionally associated with the lexical
concept which we might label as: [X IS A MEMBER OF THE CLASS OF Y]. In the example
in (29), integration of the nominal (or subject) and (nominal) predicate lexical
concepts results in interpretation of the individual lexical concepts in the primary
cognitive model profile of each. That is, the forms ‘Sally’ and ‘pianist’ provide
access in their primary cognitive models. ‘Sally’ provides access to a number of
primary cognitive models, including HUMAN, while ‘pianist’ provides access to
similar cognitive models. Thus, the integration of Sally ‘as a member of the class of’
pianist provides interpretations of the lexical concepts ‘Sally’ and ‘pianist’ which are
compatible with the integration provided. That is, given encyclopaedic knowledge, it
9
The nominal predicate construction has often been assumed to be symptomatic of metaphor,
particularly in rhetoric and philosophy of language. Indeed, and perhaps strangely, while Conceptual
Metaphor Theory largely excludes examples which exhibit this construction from its discussion of
metaphor (such examples have been argued to relate to similarity or perceptual resemblance and thus
are not central to the concerns of Conceptual Metaphor Theory), it uses this form as a basis for the A IS
B formula it employs as a shorthand for identifying putative conceptual metaphors.
34
is semantically acceptable to state that Sally is a pianist because Sally is a human
female and human females can be pianists.
In the ‘Achilles’ example however, the process of integration, which also
involves the nominal predicate construction, gives rise to potential interpretations of
Achilles and lion which clash in their primary cognitive model profiles. This is where
metaphor differs from literal class-inclusion statements. A partial primary cognitive
model profile for Achilles is provided below:
GREECE
TROY
WARRIOR
[ACHILLES]
Figure 3: Partial primary cognitive model profile for [ACHILLES]
For instance, the primary cognitive model profile for [ACHILLES] includes at
least cognitive models relating to GREECE, in particular the Greece of the historical
period in which Achilles lived, as it is presented in Homer’s Iliad perhaps, or as
portrayed in a Hollywood blockbuster. Another salient primary cognitive model
relates to TROY and the decade long siege in which Achilles and his Greek comrades
eventually triumphed. It also includes a large body of knowledge relating to Achilles
as the supreme WARRIOR. Each of these cognitive models subsumes a large body of
knowledge, including attributes of many kinds. For instance, the cognitive model for
WARRIOR
may include attributes relating to personal facts such as knowledge
35
concerning Achilles sole weakness, his heel, the only place in his body where he
could be fatally wounded.
In addition to these primary cognitive models, each is connected to a large
network of other secondary cognitive models. For instance, the cognitive model for
GREECE
may be linked to cognitive models for the dominant Greek city states of
Achilles’ time. Similarly, TROY may be linked to a cognitive model associated with a
particular sort of political system, a kingdom, including the key protagonists who
incurred the wrath of the Greeks. Crucially for the present analysis, the WARRIOR
cognitive model is related to further cognitive models including KILLER, and attributes
such as courage and valour. After all, we have detailed knowledge that warriors
engage in killing, the range of ways in which killing occurs, that this engenders risk to
the warrior, and thus that warriors might be courageous.
Now let’s consider the lexical concept [LION]. Like [ACHILLES], this is
associated with both primary and secondary cognitive model profiles. For instance,
primary cognitive models might relate to ANATOMY, PHYSIOLOGY, HABITAT, DIET,
PATTERNS OF BEHAVIOUR,
COURAGE
etc. A salient secondary model includes the folk model of
associated with lions. Given the apparent fearlessness they manifest in
seeking out prey often many times bigger, a cognitive model of COURAGE is
associated with them, that is evident in somewhat cliched aphorisms such as: As brave
as a lion. This cognitive model is secondary in the sense that it is associated with and
accessed via the patterns of behaviour lions manifest. Bravery is a human attribute
ascribed to lions by virtue of behaviours such as their willingness to tackle prey often
many times their size.
As with metonymy discussed above, what makes metaphor a form of
figurative language is that it involves interpretation by virtue of providing access to
36
secondary cognitive models. That is, [LION] provides access to the secondary
cognitive model relating to COURAGE, for instance. However, the distinction between
metaphor and metonymy is that in the ‘Achilles’ example, the function of the
interpretation is not to provide an appropriate ‘identity’ for [LION]. Rather, the
interpretation is driven by the requirement, imposed by the larger linguistic unit, the
nominal predicate construction, that [LION] be interpreted in such a way as to say
something ‘about’ [ACHILLES], the nominal element. The interpretation of [LION] as
providing access to a secondary cognitive model is motivated by the process of
integration. The language-user is guided by virtue of integration to understand
[ACHLLIES] as, in some way, a kind of lion. As Achilles cannot be a lion in the sense
of the knowledge available in the primary cognitive model profile, i.e., a four-legged
animal with a mane and tail, access is provided to the secondary cognitive model
profile. The informational characterisation here provides relational information
concerning ferocity and courage, which are compatible with some of the relational
knowledge that [ACHILLES] provides access to. Thus, the distinction between
metaphor and metonymy is that while both rely on an interpretative process in which
access is provided to the secondary cognitive model profile, interpretation is primarily
guided by integration in metaphor. Metonymy, on the other hand, also involves
access to the secondary cognitive model profile, but interpretation is primarily guided
in this case by lexical concept selection.
A partial cognitive model (primary and secondary) profile for [LION] is given
below:
37
COURAGE
HABITAT
DIET/EATING
HABITS
PATTERNS OF
BEHAVIOUR
SOCIAL
GROUPING
[LION]
Figure 4: Partial cognitive model profile for [LION]
An important point that falls out from this account is that ‘Achilles is a lion’
means something quite different from the ‘literal’ paraphrase ‘Achilles is
courageous’. This follows as lexical concepts provide routes of access. In other
words, [LION] provides an access route which gives rise to an informational
characterisation which is distinct from that provided by any other lexical concept.
This is a function of lexical concepts providing unique access sites to conceptual
knowledge. Thus, in providing access to the cognitive model of COURAGE, the notion
of a lion and its various qualities are also accessed, although not highlighted.
Having examined an instance of metaphor which relates to interpretation of
nominal forms, let’s now examine how this applies to lexical concepts associated with
other forms. To do this we consider the verb phrase whiz by as in:
(30) The evening whizzed by
cf. The plane whizzed by
38
While an entity such as a plane can literally ‘whiz by’, which is to pass by emitting an
audible sound, it is less clear that a temporally protracted entity such as ‘the evening’
can. On some accounts then, this example is metaphoric, as we are understanding
‘evening’ as an entity which can undergo physical motion, and thus can emit an
audible sound as it passes.
The conception which results from this utterance relates to an assessment of
temporal magnitude, which is to say duration, associated with the temporal period
designated by ‘the evening’. In particular, the assessment of temporal magnitude is
understood as being ‘shorter’, or proceeding ‘more quickly’ than usual, a
psychologically real phenomenon known as temporal compression (see Evans 2004a).
How does this conception arise? As in the nominal predicate construction, this
construction is intransitive. That is, some condition is predicated of the subject.
Thus, integration stipulates that the lexical concept which is associated with the verb
applies to the subject, the evening. As an interpretation of the evening is incompatible
with an interpretation of [WHIZ] in its primary cognitive model profile, interpretation
results by affording access in the secondary cognitive model profile. That is, while
the primary cognitive model profile of [WHIZ] relates to entities capable of rapid
motion, and the release of sound, one secondary cognitive model might relate to
perceptual accessibility. A consequence of rapid motion is that the entity undergoing
locomotion is subject to relatively impoverished perceptual access. Another is that
the duration of perceptual access, from the perspective of a localised vantage point is
reduced. In other words, encyclopaedic knowledge associated with rapid motion
provides access to further knowledge structures. These can be accessed via the
primary cognitive model profile, and integrated with the primary cognitive model
profile for [EVENING] in a way which provides a coherent conception. Thus, to say
39
that the evning ‘whizzed by’ is not to imply that the evening is capable of undergoing
rapid locomotion. Rather, it is to highlight the fact that one consequence of temporal
compression is that perceptual accessibility is reduced. The evening ‘feels’ as if it
lasted for a reduced period, and thus fewer details were experienced.
4.3. What is not figurative language (in the strong sense)
Having considered how LCCM Theory accounts for conceptions associated with
utterances interpreted as ‘metaphor’ and ‘metonymy’, we now consider examples
which have been assumed, by some scholars, to be figurative. Such examples
include:
(31)
This happened a long time ago.
(32)
Anchorage is a cold city. (Leezenberg 2001)
We argue that examples such as these are not figurative in the sense explicated so far.
That is, they do not afford access to secondary cognitive model profiles. Let us
consider the example in (31). Our lexical concept based account, which focuses on
the conceptual structures conventionally prompted for by linguistic units, posits that
long, like other linguistic forms, has several coded meanings associated with it. In
addition to the [EXTENDED IN HORIZONTAL SPACE] lexical concept, it also has a
conventional [EXTENDED IN TIME] lexical concept associated with it.
Evidence for this comes from examples such as the following:
(33)
a. A long kiss
b. A long book
40
‘Long’ in ‘long kiss’ relates to extended duration, not to physical length. Similarly in
(33b), we are not, or at least not typically, dealing with an oversize book, but rather
extended reading time. Understanding the form long as relating to [EXTENDED
DURATION]
relates to the process of selection. We select the [EXTENDED DURATION]
lexical concept in conjunction with ‘book’ as providing the most coherent
conception.10 Of course, we are helped by the frequency with which these two forms
collocate and are associated with this very conception. Collocations of this kind
which provide a pre-assembled conception we refer to as concept collocations. In the
same way, long time represents a concept collocation. Originally it would have been
motivated by lexical concept selection.
In view of the LCCM Theory account, concept collocations such as ‘long
time’ are not appropriately thought of as involving ‘metaphor’. In so far as, for
instance, Conceptual Metaphor Theory takes such examples to be metaphoric, the
analysis is invoking interference from the spatial lexical concept associated with long
in order to motivate a metaphor account.
From our perspective, the interesting question in such cases is how an
[EXTENDED DURATION] lexical concept became conventionally associated with the
form long. Recent work on semantic change pioneered by Traugott (e.g., Traugott
and Dasher 2002) has argued that situated implicatures (or invited inferences) can
become ‘detached’ from their contexts of use and reanalysed as being distinct
meaning-components associated with a given form. Intuitions, by some scholars, that
these examples are figurative are based on interference of contextually irrelevant
lexical concepts, in the case of our example the lexical concept [EXTENDED IN
41
HORIZONTAL SPACE].
We assume that such interference can occur when the temporal
restrictions on language processing are relaxed, as is the case in the theoretical
practice (Bourdieu, 1977) of scientists. In actual conversation, we would argue, talk
of a ‘long time’ is hardly ever felt to be figurative.11
The [EXTENDED DURATION] lexical concept associated with long might be
historically derived from contexts of communication in which reference to length can
be understood as reference to duration without harming expression of the
communicative intention, as in communication about ‘long journeys’. Through
repeated use of this form in such bridging contexts (Evans and Wilkins 2000), it is
plausible that long developed an [EXTENDED DURATION] lexical concept by virtue of
decontextualisation (Langacker 1987).
This example and the account we suggest are reminiscent of recent
philosophical approaches to the context-dependence of metaphor suggested by
Leezenberg (2001) and Stern (2000). Leezenberg discusses the example in (32),
reproduced here:
(32)
Anchorage is a cold city
Leezenberg argues that, when cold is intended in its '
emotional'meaning, such an
understanding does not have to be derived via the '
temperature'meaning of cold, but
can be accessed directly. It is the thematic context that allows the listener to directly
access the intended meaning. We agree with this approach for the example in (32) and
10
See Evans (To appear) for a discussion of the notion of ‘semanticality’. This relates to the issue of
coherence of conceptions.
11
Interference can be generated, however, on the basis of post hoc analysis, as is done, for instance, by
Conceptual Metaphor Theorists. This is possible by virtue of ‘semantic network profiles’ (see Evans
2005). That is, polysemy relations which mediate distinct lexical concepts and a single form, facilitate
42
other cases of ‘conventional metaphor’ discussed in the literature. In our terms, such
direct access is possible because the form cold has several lexical concepts
conventionally associated with it. Linguistic and – in this example crucially – nonlinguistic context guides the selection of the relevant lexical concept, and the process
of fusion can proceed as a straightforward class-inclusion. Leezenberg regards this
process of meaning construction as lying at the heart of metaphor understanding. We
would prefer not to regard such cases as metaphorical, on the grounds that the
relevant understanding in question is afforded by convention. However, this might be
a terminological question.
5. Comparison With Other Approaches
In this section we briefly compare and contrast the LCCM Theory approach to
figurative language with Croft’s domain-highlighting approach, Lakoff and Johnson’s
Conceptual Metaphor Theory, and Relevance Theory.
5.1. Croft’s domain-highlighting approach
Croft (1993) sought to integrate findings on metaphor and metonymy from conceptual
metaphor theory (discussed below) with the new work on domain theory developed
by Langacker (1987). One major innovation of Croft’s paper was to treat lexical
‘ambiguities’ such as those illustrated above relating to ‘book’ and metonymy as
being commensurable. That is, by applying an encyclopaedic account of semantic
organisation to ‘literal’ language use, and certain figurative phenomena, Croft was
able to provide a unified account. Croft argues that metaphor and metonymy result
analyses which (mistakenly) assume that a ‘literal’ sense-unit is motivating a metaphoric’
43
from, and differ from literal language in terms of, domain-highlighting. While
metaphor involves the ‘highlighting’ of an attribute across domain matrices,
metonymy involves a shift in reference within a single domain matrix.
Another major innovation was his position, in the light of a more sophisticated
approach to encyclopaedic knowledge than that available to (and developed by)
Lakoff and Johnson, that metonymy, for instance, involves a shift in reference across
domains. In this respect, metonymy shares certain affinities with an encyclopaedic
approach to metaphor, as we emphasise in the model we have presented in this paper.
One way in which the LCCM Theory approach is distinct from Croft’s is that
we distinguish between primary and secondary cognitive models which lexical
concepts provide access to. Thus, attribute focus within a single cognitive model
versus access within the primary cognitive model profile gives rise to different
conceptions from those apparent in metonymy. Another difference is that we treat
metaphor as involving essentially the same sort of access phenomena as metonymy,
which is to say, access to secondary cognitive profiles. For us, the distinction
between metaphor and metonymy relates not to inter- versus intra-domain matrix
highlighting, but to the important distinction between the processes we have described
as integration versus selection, and how they impact on the resulting process of
interpretation. That is, interpretation of a ‘metonymic’ lexical concept is influenced
by selection of the other lexical concepts in the linguistic utterance. In metaphor,
interpretation is influenced by integration. The consequence of this is that metonymy
typically involves reference while metaphor involves attribution or predication.
Croft’s (1993) work provided the beginnings of a fully-fledged encyclopaedic account
of figurative language, which we have attempted to outline here.
interpretation.
44
5.2. Conceptual Metaphor Theory
We briefly compare our approach with Conceptual Metaphor Theory (CMT) as
developed in Lakoff and Johnson (1980) and in subsequent works. The appeal and
success of this theory lies in its simplicity. CMT provides a two-domain model in
which structure from a more concrete domain of experience, the source, is mapped
onto a more abstract domain, the target. However, this model is not descriptively
adequate.
The basic suggestion of CMT is that there are stable long-term knowledge
structures, ‘conceptual metaphors’, which inhere in memory. These knowledge
structures both organise our conceptual systems and license symbol use in language.
They are thought of as sub-symbolic, in that they develop in the mind independently
of the particular symbol systems (language, gesture, pictures) used in the respective
community. We argue, on the contrary, that figurative language use requires symbols,
i.e., lexical concepts, with their various characteristics as discussed earlier.
For CMT, figurative meaning-construction involves mapping the forms
encountered in an utterance onto the relevant sentence type (Stern 2000), i.e. a
conceptual metaphor. For us, figurative language can best be thought of in terms of a
situated creative act, providing a novel conception. Whereas there is a sharp
distinction between literal and figurative language and thought in CMT, we maintain
that lower and higher degrees of figurativity are a symptom of general meaningconstruction processes: selection and fusion. CMT takes a narrow view of meaning,
ignoring the dynamic aspects of meaning-construction which are crucial for
understanding the way conceptual resources are marshalled by language users.
45
A further difficulty we see with CMT is that it is, in essence, not a usagebased account of language. This follows as in large measure it is not really an
account of language at all. Rather, it is an attempt to understand conceptual
organisation from the perspective of a narrow view of the nature of embodied
cognition. As our approach sees figurative language as symptomatic of more general
meaning-construction processes which are grounded in situated language use,
communication and context, it follows that our view of figurative language treats nonliteralness as an outcome not of ‘special’ dedicated knowledge structures, i.e., crossdomain mappings, but rather the same sorts of knowledge structures which are, in
principle, accessible to literal uses of language.
There are a number of recent perspectives which corroborate the LCCM
Theory approach to figurative language developed here. The first relates to the
linguistic evidence. In a detailed study of conceptualisations of time, Evans (2004a)
showed that the CMT account of time is descriptively inadequate. That is, the
linguistic evidence employed by CMT in order to support its claims in fact misses
many patterns and generalisations which are more appropriately stated at a symbolic
rather than a sub-symbolic level. In other words, the level of the lexical concept rather
than the conceptual metaphor provides a more descriptively adequate account of the
linguistic data.
A second perspective relates to recent work on the nature of figurative
language from the perspective of public discourse. For instance, work by Zinken et
al. (To appear), which introduces and develops the construct of ‘discourse metaphor’,
shows that successful figurative language, which gets taken up and replicated in a
discourse community, uses rather specific vehicle concepts. Corpus evidence suggests
that these are not symptoms of more general mappings. Successful metaphors are
46
closely linked to the limited and specific communicative goals of the proponents of a
given discourse. In this, metaphor is wholly dependent on the symbolically-mediated
resources of language, both in terms of its development, use and propagation.
5.3. Relevance Theory
The view of figurative language we have suggested here is to some extent consistent
with the approach in Relevance Theory. We briefly discuss the similarities and
differences between the two approaches.
Relevance Theory suggests that understanding figurative and literal language
relies, in a fundamental way, on the same processes of meaning construction. This
seems consistent with our proposal that figurativity is a matter of degree.
Furthermore, the two proposals are consistent in that they suggest that semantic
representations always underspecify the intended meaning of an utterance, and
therefore need to be enriched by encyclopaedic knowledge in the process of meaning
construction. The difference between the two approaches lies in the reasons for
assuming the necessity of such encyclopaedic enrichment. While our argument rests
on linguistic evidence relating to the semantic '
flexibility'of linguistic forms, the
Relevance Theory argument ultimately rests on an attempt to maintain a truth-based
semantics. In such an approach, semantic representations must be defined in such a
way that they can be assigned a truth value in a '
language of thought'
. Encyclopaedic
enrichment, on this account, becomes necessary because everyday communication
usually does not adhere to the requirements of formal logic. For example, a concept
such as SQUARE must be geometrically defined in order to be useful as an argument in
a logic-based language of thought. However, when a speaker says that "our garden is
square", it is understood that the garden most likely is not exactly square. According
47
to Relevance Theory, an encyclopaedically enriched concept *SQUARE is derived in
the process of establishing the most relevant meaning intended by the speaker. This is
referred to as '
loose use'of concepts. Metaphors such as "my son is a baby" – in a case
where the son in question is in fact a grown man – are treated as just another kind of
loose use. A contextually relevant concept *BABY is derived, which contains only
those properties that might literally be attributed to grown men: '
requires constant
attention'
,'
behaves immaturely'etc. (Carston, 2002).
This truth-based approach might be appropriate for the questions Relevance
Theory addresses, but it would not necessarily be appropriate for a cognitively
realistic account of meaning construction, which is what cognitive linguistic
approaches attempt to provide. While propositional logic can be a useful tool in some
contexts of semantics, it cannot define the explanandum of a cognitive science
(Gigerenzer 1996). If cognitive semantics were to explain why people'
s use of
language does not adhere to the rules of formal logic, it would be elevating formal
logic to the status of a model of how the human mind actually works.
A further question in which the Relevance Theory approach to figurative
language and the one we have suggested here partially overlap is the role of literal
meaning in '
arriving at'a figurative understanding. Both approaches suggest that the
literal meaning of, for example, '
pigsty'in the metaphor in (34) plays a role in
figurative meaning construction.
(34)
"This room is a pigsty" (Sperber & Wilson, 1995)
It is for this reason that figurative expressions have a richer communicative effect
than literal paraphrases, such as "this room is untidy" (Sperber & Wilson, 1995).
48
In Relevance Theory, the properties associated with the loosely used vehicle concept
get sorted into two disjoint sets: the set of irrelevant properties and the set of relevant
properties (Carston, 2002). The relevant properties of PIGSTY would include
'
untidiness'
,'
smelliness'and other properties that can literally be asserted of a room
inhabited by a human being. If we understand the Relevance Theory account correctly
it is only this set of properties that is used in the actual process of compositional
meaning construction, what we have referred to as fusion. If this were the case,
though, we wonder where the '
fun'of using a metaphor rather than a literal statement
would come from. On this account, figurative language understanding, in the case of
the nominal predicate construction, would ultimately be a more complicated literal
class inclusion.12 On the account we have proposed, the nominal predicate
construction itself is one of the semantic representations that contribute to the '
added
value'of a metaphor in the process of fusion. Instead of the metaphors of '
broadening'
and '
loose'use, we have suggested the alternative metaphor of '
access routes'for
talking about the relations between more literal and more figurative meanings
associated with a form. What this metaphor underlines is that the literal meaning of,
for example, pigsty, stays in peripheral consciousness (Chafe, 1994) throughout the
process of meaning construction.
A final issue on which Relevance Theory and LCCM Theory seem to be close,
but actually currently diverge, is the understanding of what an encyclopaedic
approach to meaning construction means. Carston (2002) discusses cases such as the
one in (35):
(35)
12
Robert is a bulldozer
It seems to us that this also applies to Glucksberg'
s (2001) model of metaphor comprehension.
49
Cases like this, not to mention more poetic metaphors that Carston also discusses, are
problematic for an account that restricts encyclopaedic knowledge to property lists
'
written'in an a-modal language of thought. The problem is that what we most easily
understand to be asserted of Robert in (35) – that he acts out his intentions without
much regard for others'preferences – is not actually a property of bulldozers. How
this works is not well understood, and we cannot claim that we have an answer at the
moment. However, there seems to be a growing consensus that in understanding more
creative metaphors, modal, and possibly multi-modal, simulations are involved
(Barsalou, 1999; Carston, 2002; Indurkhya, 1999; see also Coulson, 2000).13 The
mental image of Robert as a bulldozer might evoke proprioceptive '
mimetic schemas'
(Zlatev 2005) relating, for example, to forces and the negative feelings associated
with being unable to resist a physical force. In any case, this seems to be a contingent
difference between the Relevance Theory account and the one that we propose, and
further discussions should benefit advances in both Relevance Theory and cognitive
semantics.
6. Conclusion
Two main metaphors currently exist for understanding metaphorical language. The
first is the idea of a ‘horizontal’ mapping or a projection across domains (e.g.,
Gentner, 2001; Lakoff & Johnson, 1980). This comes close to the original idea of
'
carrying over'a name, expressed in the term '
metaphor'
. The second metaphor is the
idea of a ‘vertical’ class-inclusion (e.g., Carston, 2002; Glucksberg, 2001). The
vertical metaphor might go a bit too far in likening metaphorical utterances of a
50
particular kind to literal class-inclusion statements. The horizontal metaphor leads to a
sharp distinction between metaphorical and literal meaning construction.14 We have
tried in this paper to start to develop a new metaphor for figurative language
understanding, involving terms such as '
access route'and '
fusion'
. We hope that our
explication of these terms has provided initial steps towards a model that can integrate
different forms of figurative understanding – metaphor and metonymy – with more
'
literal'processes of meaning construction, without loosing the ability to make
distinctions between them.
Our account of meaning construction attempts to model how figurative
cognition is constrained by the conventional representations of the particular semiotic
system that the cognition is carried out for: In the case of figurative language, it is
constrained by the conventional semantic representations associated with lexical
concepts. This seems to us to be an important innovation in the context of previous
cognitive semantic approaches to figurative language and thought.
Our proposal for a new theory of figurative language has – unavoidably –
raised as many questions as it has provided answers. It is not clear at present what
exactly makes a cognitive model '
primary'
, i.e. directly accessable, as opposed to
'
secondary'
. The nature of the interaction between semantic representations and
conceptual knowledge in other modalities is not well understood. Also, to name one
final example, the integration of lexical concepts associated with syntactic
constructions with lexical concepts associated with words needs to be clarified. We
hope that such issues, which are crucial for a cognitively realistic account of
figurative meaning construction, can be addressed in discussions with related
13
It has been proposed that a hybrid model of metaphor understanding, marrying Relevance Theory to
Conceptual Metaphor Theory might provide a solution for such cases. However, this brings with it the
problems associated with Conceptual Metaphor Theory discussed above.
14
It might, however, be more appropriate as a model for analogy rather than metaphor.
51
approaches in cognitive pragmatics, the philosophy of language, and
psycholinguistics, as well as specific theories which attempt to address how figurative
language ‘prompts’, and access phenomena interface with ‘backstage’ conceptual
processes as developed in the work of Fauconnier and Turner (2002).
References
Atkins, Beryl T. (1987). ‘Semantic ID Tags: Corpus Evidence for Dictionary
Senses’. Proceedings of the Third Annual Conference of the UW Centre for
the New Oxford English Dictionary, 17-36.
Barsalou, Lawrence. (1999). ‘Perceptual symbol systems’. Behavioral and Brain
Sciences, 22, 577-609.
Bourdieu, Pierre (1977). Outline of a Theory of Practice. Cambridge: Cambridge
University Press.
Carston, Robyn. (2002). ‘Metaphor, ad hoc concepts and word meaning - more
questions than answers’. UCL Working Papers in Linguistics 14. Accessible
at: http://www.phon.ucl.ac.uk/home/robyn/pdf/uclwpl142002.pdf
Chafe, Wallace. (1994). Discourse, Consciousness and Time: The Flow and
Displacement of Conscious Experience in Speaking and Writing. Chicago:
Chicago University Press.
Coulson, Seana. (2000). Semantic Leaps: Frame-shifting and Conceptual Blending
in Meaning-construction. Cambridge: CUP.
Croft, William. (1993) ‘The role of domains in the interpretation of metaphors and
metonymies’. Cognitive Linguistics 4: 335-70.
Croft, William. (2000). Explaining Language Change: An Evolutionary Approach.
London: Longman.
52
Evans, Nicholas & David Wilkins. (2000). ‘In the mind'
s ear: the semantic extensions
of perception verbs in Australian languages’. Language 76.3: 546-592.
Evans, Vyvyan. (2004a). The Structure of Time: Language, Meaning and Temporal
Cognition. Amsterdam: John Benjamins.
Evans, Vyvyan. (2004b). ‘How We Conceptualise Time’. Essays in Arts and
Sciences, 33, 2: 13-44.
Evans, Vyvyan. (2005). ‘The Meaning of Time: Polysemy, the Lexicon and
Conceptual Structure’. Journal of Linguistics, 41.1, 33-75.
Evans, Vyvyan (2006). ‘Towards a Cognitive Compositional Semantics: An
Overview of LCCM Theory’. In H. Kardela, A. Glaz and U. Magnusson (eds.).
New Insights in Semantics and Lexicography. Lublin: Marie Curie-Sklodowska
University Press, Poland
Evans, Vyvyan. (To appear). ‘Lexical Concepts, Cognitive Models and MeaningConstruction’. 18/1, Cognitive Linguistics.
Evans, Vyvyan and Melanie Green. (2006). Cognitive Linguistics: An Introduction.
Edinburgh: Edinburgh University Press.
Fauconner, Gilles and Mark Turner. (2002) The Way We Think: Conceptual
Blending and the Mind’s Hidden Complexities. New York: Basic Books.
Fillmore, Charles. (1982). ‘Frame semantics’. The Linguistic Society of Korea (ed.),
Linguistics in the Morning Calm, pp. 111-137.
Fillmore, Charles. (1985). ‘Frames and the semantics of understanding’. Quaderni
di Semantica, 6, 222-254.
Gentner, Dedre, Bowdle, B. F., Wolff, P. & Boronat, C. (2001). ‘Metaphor is like
analogy’. In D. Gentner, K. J. Holyoak & B. N. Kokinov (Eds.) The
53
Analogical Mind: Perspectives from Cognitive Science (pp. 199-253).
Cambridge: MIT Press.
Gibbs, Raymond W. (1994). The Poetics of Mind. Cambridge: CUP.
Gigerenzer, Gerd (1996). ‘From tools to theories: Discovery in cognitive psychology’.
In C. F. Graumann & H. J. Gergen (Eds.) Historical Dimensions of
Psychological Discourse (pp. 36-59). Cambridge: CUP.
Glucksberg, Sam (2001) Understanding Figurative Language: From Metaphors to
Idioms. Oxford: Oxford University Press.
Goldberg, Adele. (1995). Constructions: A Construction Grammar Approach to
Argument Structure. Chicago: Chicago University Press.
Goldberg, Adele. (2003). ‘Constructions: A new theoretical approach to language.’
Trends in Cognitive Science.
Gries, Stefan Th. (2006). ‘Corpus-based Methods and Cognitive Semantics: The
Many meanings of to run’. In: Gries, S. and A. Stefanowitsch (eds.). Corpora in
Cognitive Linguistics: The Syntax-Lexis Interface. Berlin: Mouton de Gruyter.
Indurkhya, Bipin (1999). ‘Creativity of metaphor in perceptual symbol systems’.
Behavioral and Brain Sciences 22 (4), 621-627.
Johnson, Mark. (1987). The Body in the Mind. Chicago: University of Chicago
Press.
Lakoff, George. (1987). Women, Fire and Dangerous Things: What Categories
Reveal About the Mind. Chicago: University of Chicago Press.
Lakoff, George. (1993). ‘The contemporary theory of metaphor’. In A. Ortony
(ed.)., Metaphor and Thought (second edition). Cambridge: Cambridge
University Press.
54
Lakoff, George and Mark Johnson (1980). Metaphors We Live By. Chicago: Chicago
University Press.
Lakoff, George and Mark Johnson (1999). Philosophy in the Flesh: The Embodied
Mind and its Challenge for Western Thought. New York: Basic Books.
Lakoff, George and Mark Turner (1989). More Than Cool Reason: A Field Guide to
Poetic Metaphor. Chicago: Chaicgo University Press.
Langacker, Ronald (1987). Foundations of Cognitive Grammar, Volume I. Stanford:
Stanford University Press.
Langacker, Ronald (1991). Foundations of Cognitive Grammar, Volume II.
Stanford: Stanford University Press.
Langacker, Ronald. (2000). ‘A dynamic usage-based model’. In M. Barlow and S.
Kemmer (eds.), Usage-based Models of Language, 1-64. Stanford, CA.: CSLI
Publications.
Leezenberg, Michiel. (2001). Contexts of Metaphor. Oxford: Elsevier.
Musolff, Andreas. (2004). Metaphor and Political Discourse : Analogical Reasoning
in Debates about Europe. London: Palgrave Macmillan.
Prinz, Jesse J. (2002). Furnishing the Mind: Concepts and their Perceptual Basis.
Cambridge: MIT Press.
Sharifian, Farzad. 2003. ‘On Cultural Conceptualisations’. Journal of Cognition and
Culture, 3, 3, 187-207.
Sperber, Dan and Deirdre Wilson. (1995). Relevance: Communication and Cognition,
second edition. Oxford: Blackwell.
Stefanowitsch, Anatol & Stefan Th. Gries. (eds.) (2006). Corpora in Cognitive
Linguistics. Berlin: Mouton de Gruyter.
Stern, Josef. (2000). Metaphor in Context. Cambridge, MA: MIT Press.
55
Tomasello, Michael. (1999). The Cultural Origins of Human Cognition. Harvard:
Harvard University Press.
Tomasello, Michael. (2003). Constructing a Language: A Usage-based Theory of
Language Acquisition. Harvard: Harvard University Press.
Traugott, Elizabeth Closs and Richard Dasher (2002). Regularity in Semantic Change.
Cambridge: Cambridge University Press.
Tyler, Andrea and Vyvyan Evans. (2001). ‘Reconsidering prepositional polysemy
networks: The case of over’. Language, 77, 4, 724-65.
Tyler, Andrea and Vyvyan Evans. (2003). The Semantics of English Prepositions:
Spatial Scenes, Embodied Meaning and Cognition. Cambridge: CUP.
Zinken, Jörg, Iina Hellsten and Brigitte Nerlich (To appear). ‘Discourse metaphors’,
in J. Zlatev, T. Ziemke and R. Frank (eds), Body, Language and Mind:
Sociocultural Situatedness. Berlin: Mouton de Gruyter.
Zlatev, Jordan. (2005).
Zwaan, Rolf. (2004). ‘The Immersed Experiencer: Toward an Embodied Theory of
Language Comprehension. In: B.H. Ross (Ed.), The Psychology of Learning
and Motivation, Vol. 44, 35-62. New York: Academic Press.
56