1
Towards a Critical Cognitive Linguistics?
Peter Stockwell: School of English Studies, Nottingham University, UK
peter.stockwell@nottingham.ac.uk
1. The Landscape of Modern Linguistics
At the turn of the millennium, the two most rapidly developing fields of modern linguistics are
Cognitive Linguistics (CL) and Critical Discourse Analysis (CDA). Both fields are relatively
new and innovative. Both claim to offer a radical new direction for the study of language and
communication, and both effect this by widening the traditional conception of language as an
area of study in itself. Both have arisen out of interdisciplinary studies with linguistics at the
core, and both are characterised by the missionary zeal of their proponents and their sense of
the inadequacies of other approaches to language. This paper asks whether the two disciplines
share any common ground, and whether any synthesis or at least accommodation is possible
between them. The reason for this inquiry is partly to use the comparative method to highlight
whether there are any proper or peculiar domains to which each discipline belongs, and partly
to use the occasion of the comparison to investigate some of the claims made within each
discipline.
CL is best exemplified in the work of Fauconnier (1994, 1997), Geeraerts (with Grondelaers
and Bakema 1994), Gibbs (1994), Johnson (1987), Lakoff (and Johnson 1980, 1999, Lakoff
1987), Langacker (1987, 1991), Sweetser (1990, Fauconnier and Sweetser 1996), and
Turner (1987, 1991, Lakoff and Turner 1989), among many others. Ungerer and Schmid
(1996) is a good introduction, and though in this paper I assume some knowledge of the
discipline, the key ideas are as follows.
CL represents an experientialist and thus anti-objectivist position, in describing the
relationship between the world on one hand and language and thought on the other. This has
far-reaching consequences for reference, anaphora, deixis, pragmatic force, categorisation,
lexicalisation and lexical semantics, many of which are in the process of being developed at the
moment.
The fundamental re-evaluation of CL involves a rejection of Cartesian dualism, reuniting mind
and body to see language and thought - and conceptualisation itself - as embodied. Embodied
experience finds expression functionally in metaphorical structures (idealised cognitive models,
or ICMs) which in turn are manifest in both conventional and novel metaphors and
expressions. Conventional communication involves shared (perhaps universal) ICMs and
image-schemas, through which we structure our understanding of the world and through
which we even structure new concepts. That is, knowledge of the world is constituted
through and by these conceptual metaphors, to the extent that even newly-encountered or
abstract concepts are isomorphically understood in terms of them.
2
The process of categorisation itself consists of basic-level schemas and is arranged in a way
that manifests prototype effects. Items within such radial categories can be judged as being
central or peripheral, and rated on the basis of ‘goodness-of-example’. CL has formulated
constraints on metaphorical mappings so that the framework matches intuitive senses of
linguistic usage. Among the many interdisciplinary applications of CL is the sub-branch of
cognitive poetics, which investigates the consequences for literary analysis of ideas from CL.
Critical Discourse Analysis (CDA) is exemplified in the work of Birch (1989), Caldas-
Coulthard and Coulthard (1996), Fairclough (1989, 1995a, 1995b), Fowler, Hodge, Kress
and Trew (1979, Fowler 1981, 1986, 1991, Kress and Hodge 1979, Hodge and Kress
1988) and Toolan (1996), among many others. The key points in CDA are as follows.
The approach mainly uses Hallidayan (1985) systemic-functional linguistics to examine the
rhetoric and ideology of institutions, such as the media, government, politicians, regulatory
bodies and popular influential texts from fictional romances to billboard advertising. CDA
developed from the linguistic criticism of the late 1970s and ‘80s, and has since broadened
into social semiotics and a variety of critical linguistic approaches to a whole range of
discourses. Fairclough’s (1995a) analysis is explicitly Marxian and emphasises the
responsibility of academic practice in unearthing the latent ideologies of controlling hegemonic
institutions. This is based on a tripartite analytical framework:
- spoken and written text analysis
- the analysis of the discourse practice of production and interpretation
- and a politically situated analysis of social practice (Fairclough 1995a: 133).
CDA is allied closely (especially in Toolan’s (1996) work) with integrationalism (after Harris
1981, 1987). This means that the dimensions of communicative experience - such as context,
power relations and background knowledge - are not sidelined as in traditional linguistic rule-
systems, but become part of a holistic integrated study.
The respective ancestries of CL and CDA seem to be significant. CL includes practitioners
who grew up as generativists studying transformational-generative grammar and language
universals in the 1970s. Though many of the current writers refute their earlier selves, the
search for universal or totalizing linguistic and conceptual structures is still a tendency in CL.
There is less oedipal angst in the development of CDA; its roots lie in left-wing politics and
systemic-functional linguistics and this is still largely the agenda for writers. It has always had a
concern to expose conservative or anti-democratic ideologies in persuasive, regulatory,
institutional, media and popular influential texts, though later CDA problematises the
truth/falsity issue of studying textual ‘distortions’ of a preferred reality, in response to criticism
often from within the movement (such as Pateman 1981, Richardson 1987).
3
In general, CL mainly has a continental European and US bias, while CDA is generally to be
found amongst academics working in Britain and Australia. In spite of these geographical and
historical differences, there are points of contact even at the theoretical level.
2. A Comparison of Theoretical Issues
Both CL and CDA are interested in suggesting deep structures that are made manifest in
linguistic expressions. CDA is focused on how individual utterances and sentences are
expressions of ideological discursive practices (such as analyses of women’s magazine articles,
tabloid newspaper reports, university regulations, and so on). CL is focused on how individual
utterances and sentences are expressions of conceptual metaphors (such as ‘he blew his top’
as an example of ANGER IS A CONTAINER OF HOT LIQUID, or ‘she rejected his
advances’ as LOVE IS WAR, and so on). Both traditions emphasise that linguistic
conventions are not just examples of social practice, but that linguistic usage is also
‘constitutive’ (Fairclough 1995a: 131) of social practice. CDA focuses on how hegemonic
institutions attempt to structure conventional thinking, and CL focuses on making explicit the
conceptual metaphors of everyday usage.
Although CDA claims to be interventionist (it wants to make explicit an awareness of control
in order to resist it critically) and CL aims to be descriptive (it wants to be simply a
methodological tool which can be used in a variety of ideological ways), there is no reason
why the linguistic procedures of CL cannot be used in the service of CDA. I will return to this
point later.
Both CL (centrally, Lakoff 1987) and CDA (latterly, Fairclough 1995a) are anti-objectivist in
their view of the conceptually constitutive power of language. Both place re-emphasis on
‘experientialism’. However, there are differences in definition and how thoroughly the
assimilation of the term is embraced. In CL, experientialism serves to situate conceptualisation
in the body (Lakoff and Johnson 1980, 1999, Lakoff 1987, Turner 1987, 1991), and it
emphasises the ‘embodiment’ of experience in idealised cognitive models (ICMs). Thus the
abstract concept TIME is figured as the human-scale and tangible SPACE, and emotions are
metaphorically directional in prepositions of being ‘up’ or ‘down,’ ‘high’ or ‘low’ in relation to
the conditions of being corporeally human. Where categories and concepts are shared in the
language system, the individual has learnt the convention experientially: though there is an
element of a social theory here, the focus is on the individual and their mental space being
imprinted with the culturally correct convention. The CDA understanding of experience is
more dynamic and interactive than this. Toolan (1996) (after pointing out the problems CL has
in dealing with creative and novel metaphors – see also Stockwell 1999 on this) argues that
the CL rejection of objectivism is not thorough enough:
... it is clear that his [Lakoff’s] rejection of abstract objectivism is in no way a
rejection of collective categorization itself but rather as [an?] emphasis on different
roots of categorization (experience, in the body) and a different kind of
4
categorization (prototypical etc. rather than absolute). As a shared mental
framework, categorization is crucial to the revision ... In fact, Lakoff presents
‘basicness in human categorization’ very much as if it is analogous, for things in
the world, to what Chomsky’s universal grammar is claimed to be for natural
language syntactic structures ...
Lakoff’s [1987] book reports revisions as to what counts as ‘membership of a
category’ (i.e., what the criteria are); it does not take the radical step of
confronting the possibility that ‘membershipping’ (categorization) is contingent,
varying from case to case according to criteria that may differ from case to case.
(Toolan 1996: 87-8)
There is a divergence in what ‘experientialism’ means and how thoroughly it can be assimilated
into an investigative methodology.
Nevertheless, Toolan’s book is primarily about integrationalism rather than CDA directly, and
in fact there are correspondences in the notions of prototypes and categorisations between CL
and CDA. An ICM is an experientially-accumulated knowledge structure that is always open
to new information, is connected with other domains of knowledge in a network, and is
omnipresent in cognitive activity (Ungerer and Schmid 1996: 48-9). The notion is a new
(Lakoff 1987) version of the notions of frames and schemas of knowledge developed in the
AI research of the 1970s. Fairclough links CDA to these ideas:
It may be useful to think of ideologies in terms of content-like entities which are
manifested in various formal features, and perhaps frame, schema, script and
related concepts are of value in this respect (Schank and Abelson (1977)).
(Fairclough 1995a: 75)
In his earlier work, Fairclough (1989) developed the notion of members’ resources (MR).
These are accumulated knowledge structures ‘which people have in their heads and draw
upon when they produce or interpret texts – including their knowledge of language,
representations of the natural and social worlds they inhabit, values, beliefs, assumptions, and
so on’ (Fairclough 1989: 24). Fairclough’s MRs are clearly experientialist ICMs. Earlier
(Fairclough 1989: 10) he refers to them as a set of ‘prototypes’, within a section entitled
‘Cognitive psychology and artificial intelligence’. He links these mental structures to the
cognitive dimension:
The MR which people draw upon to produce and interpret texts are cognitive in
the sense that they are in people’s heads, but they are social in the sense that they
have social origins.
(Fairclough 1989: 24)
It is the social aspect that is of primary interest for Fairclough, and he uses this concern to
argue against the emphasis within CL:
5
Not surprisingly, cognitive psychology and artificial intelligence have given little
attention to the social origins or significance of MR. I shall argue later that
attention to the processes of production and comprehension is essential to an
understanding of the interrelations of language, power and ideology, and that this
is so because MR are socially determined and ideologically shaped, though their
‘common sense’ and automatic character typically disguises that fact. Routine and
unselfconscious resort to MR in the ordinary business of discourse is, I shall
suggest, a powerful mechanism for sustaining the relations of power which
ultimately underlie them.
(Fairclough 1989: 11)
Fairclough’s first criticism here seems to have been met by CL in the 1990s. Furthermore, in
the same way as Fairclough (1989: 91) sees ‘common sense’ as the ‘naturalization’ of an
ideological set of assumptions so that they are not perceived as ideological any more, CL
regards cultural models as shared conceptual metaphors which operate as folk-theories and
structure our relationship with society (Holland and Quinn 1987). Though Fairclough argues
directly against early CL emphases here, it seems to me that the difference is based on a
different focus and is additional or complementary rather than being an antithesis. It seems that
Fairclough might prefer a framework which shifted the focus of idealised cognitive models to
being ideological cognitive models.
Fairclough points out that ‘frames, scripts, and schemata’ are all ‘a part of MR constituting
interpretative procedures ... and share the property of mental representations in general of
being ideologically variable’ (Fairclough 1989: 158). It is the awareness of ideology and the
status of linguistic analysis as scientific method or critical engagement that is at the heart of the
CL/CDA comparison. Though some work within the sub-branch of ‘cognitive poetics’
(Turner 1987, 1991, Lakoff and Turner 1989, D. Freeman 1993a, M.H. Freeman 1997) has
focused on the stylistic expression of linguistic metaphors, in general CL is concerned mainly
with the conceptual mappings which underlie metaphorical expressions. Fairclough recognises
the ubiquity of metaphor, and though he is interested in underlying ideological functions, his
systemic linguistic framework makes him sensitive to the ideological nuances of different
stylistic choices:
Metaphor is a means of representing one aspect of experience in terms of
another, and is by no means restricted to the sort of discourse it tends to be
stereotypically associated with – poetry and literary discourse. But any aspect of
experience can be represented in terms of any number of metaphors, and it is the
relationship between alternative metaphors that is of particular interest here, for
different metaphors have different ideological attachments.
(Fairclough 1989: 119)
This seems to me to be a difference in the current practices of the two disciplines, but it is not
one that is necessary to their distinctiveness as disciplines. As the cognitive poetics work
shows, CL can be used successfully to discuss stylistic variation in a way that would
correspond quite happily with CDA.
6
A more problematic divergence is in the question of what sort of disciplines CL and CDA
think they are, respectively. CL explicitly and unapologetically regards itself as a science with
the job of investigating a natural phenomenon (language) and producing the best possible
current account of the workings of that system in the mind. Freeman (1993b) has argued that
in itself CL is a method rather than a methodology, a tool with no inherent ideological
assumptions, and which could be used in the service of a whole range of other ideological
approaches. CL explains the detail of a range of interpretations, and excludes some readings
which are demonstrably without cognitive basis, but it cannot itself choose between different
interpretations of, for example, how a conceptual metaphor is applied. In relation to literary
interpretation, Freeman asserts:
Of course there is no one God’s-eye interpretation of a literary work, whether the
evidence for such a claim arises from cognitive metaphor or anything else. But
there is a range of plausible interpretations and a scale of valid ones.
(Freeman 1993a: 17)
Freeman’s challenge is to ask which part of CL is inherently ideological. Of course,
considered as a ‘pure’ theory, this cannot be answered, except in the very general sense that
there is an implicit ideological motivation in choosing the framework in the first place.
However, claiming that CL is exempt from ideological assumptions does not seem tenable to
me. There is a fundamental consensualism in the notion of cultural models, conventional
mapping of ICMs, shared prototypes, and so on, and the common method of discussion of
CL is to examine individual sentences asocially. Even if CL does not exclude a social
dimension, it tends to focus elsewhere and this is an applied ideological choice. It would be
true to say that it is the application that is ideological rather than the theoretical framework, but
since the only way of discussing CL is to apply it, this seems a bit mischievous.
As Gross (1997) points out, new disciplines always tend to overstate their radicalism,
innovativeness and novelty, and perhaps CL in this respect is a victim of its own hype.
Freeman (1993a) takes a more robust but charitable view:
I am often taxed ... with being ‘totalizing’ or ‘essentialist’. Guilty as charged. I
take these terms to mean ‘general, ignoring particulars that do not fit the theory’.
Noam Chomsky’s early work in linguistic theory is often held up to me as an
example of this ‘fault’ ... None of ... [the modern linguistic] developments would
have been possible, in my view, had not Chomsky been an unrepentant ‘totalizer’
from the start. Any theory of anything worth anything begins as totalizing,
essentialist, and universalist, and progressively qualifies its claims as research
proceeds.
(Freeman 1993: 18)
The process is certainly the institutional practice in sciences (where scientists compete for
funding by good self-marketing), but it could be argued that this ‘scientific’ method makes for
7
bad science. Certainly the history of science has involved the progression from one
explanatory framework to another, regarded as an improvement, but there is an inescapable
ideological background and motivation to science as to every other human activity.
This is what makes the following document from within CL all the more astonishing. This is the
(anonymous) referee’s comment on a proposed paper to be given by a colleague at a CL
conference:
Comments on:
‘Cognitive Linguistics and the Marxist approach to ideology’
This paper appears to have been submitted to the wrong conference. It is a
critique of Cognitive Linguistics from a Marxist perspective, which would be
entirely appropriate at a conference of Marxists giving Marxist critiques of things.
The paper ignores a fundamental difference between CL and Marxism: CL is a
scientific endeavor, part of cognitive science. It is not an apriori theory, as
Marxism is. At a conference on CL, an appropriate paper might be a cognitive
analysis of Marxist thought. Indeed, the abstract itself would make in interesting
subject for analysis.
One thing is clear from the abstract (actually, it's been clear for many years): The
consequences of empirical research on the mind in cognitive science in general
and CL in particular are inconsistent with Marxist ideology. That is not
particularly strange, since Lakoff and Johnson argue in their new book,
Philosophy in the Flesh, that most of Western philosophy is inconsistent with
results coming from cognitive science. Marx fits right in with Kant and Aristotle
and Descartes. Of course, from within Marxist ideology, CL would be an
example of ‘false consciousness’, as would anything disagreeing with Marxism,
whether it had scientific support or not.
I recommend rejecting this abstract. It should be given at conference of Marxists,
or perhaps paired on neutral turf with a paper analyzing the folk theories and
metaphors in the paper itself. [Thi
本文档为【towards a critical cognitive linguistics】,请使用软件OFFICE或WPS软件打开。作品中的文字与图均可以修改和编辑,
图片更改请在作品中右键图片并更换,文字修改请直接点击文字进行修改,也可以新增和删除文档中的内容。
该文档来自用户分享,如有侵权行为请发邮件ishare@vip.sina.com联系网站客服,我们会及时删除。
[版权声明] 本站所有资料为用户分享产生,若发现您的权利被侵害,请联系客服邮件isharekefu@iask.cn,我们尽快处理。
本作品所展示的图片、画像、字体、音乐的版权可能需版权方额外授权,请谨慎使用。
网站提供的党政主题相关内容(国旗、国徽、党徽..)目的在于配合国家政策宣传,仅限个人学习分享使用,禁止用于任何广告和商用目的。