Cognitive Linguistics Internal Dynamics and Interdisciplinary Interaction Cognitive Linguistics

Чтобы посмотреть этот PDF файл с форматированием и разметкой, скачайте его и откройте на своем компьютере.
Wiggs, Cheri 232, 246
Wilson, Deirdre 10, 262, 263, 273,
274, 275, 280, 353, 354, 380,
124, 125, 127, 131, 132, 133,
Len, David de 211, 219
Leontev, Aleksej A. 209, 211, 219
100, 167, 187, 221, 353, 354,
356, 357, 380, 381, 384
Hume, David 179, 180, 181, 186
175, 176, 177, 178, 179, 181,
157, 354, 355, 381, 384
Jensen de Lpez, Kristine 167, 188,
56, 57, 59, 61, 63, 68, 167, 185,
Aliaga, Francisco 202, 216
Textual corpora 7, 163, 183
embodied action in 225, 227,
prototypical 10, 314, 315, 356,
289, 291, 293, 297299,
Deontic modality 73, 83, 85, 86
Discourse analysis (see Critical
283, 287, 290, 296, 302, 310
Discourse (see Elementary Dis-
253, 269, 314, 373, 374
253, 256258, 269
perception and language 8, 234,
339, 350
Autopoiesis and Cognition: The Realization of the Living
Foot and Mouth Disease Doesnt Affect Humans:
the human health and social consequences of the 2001 UK epi-
Selling Science. How the Press Covers Science and Technology
Agriculture and Human Values
1999The influence of
popular cultural imagery on public attitudes to-
wards cloning.
Sociological Research Online
2000Clones and c
rops: The use of stock characters and word play in
two debates about bioengineering.
. London: Thames and Hudson.
13.Without Roslyn Franks prodding this section would never have been written.
There is obviously much more to say here, but this will have to wait for another
14.On blending and its dynamic features, see e.g. Fauconnier and Turner (2002).
15.A rather arbitrary dichotomy that has dogged western thinking for a long time
back to nature, we need science to restore a balance that has been destroyed
gradually since humans began to cultivate the land, but a science that dares
to confront big business, where dreams of conquering, subjugating, con-
Carson took the opportunity to remind the world of the wider implications
of her work: We still talk in terms of conquest. We still havent become
mature enough to think of ourselves as only a tiny part of a vast and in-
credible universe. Mans attitude toward nature is today critically important
simply because we have now acquired a fateful power to alter and destroy
nature. But man is part of nature and his war against nature is inevitably a
Since 1962 this power has increased manifold, especially through de-
The inter-textual potential of
silent spring
emerges from the way
the blend resonates with other scientific and fictional narratives,
which filled the literary and cultural space around it throughout the
The co-textual potential of
silent spring
derives from the way this
The story of Silent Spring not only helped to start the environmental move-
ment, it also coincided with the intensification of agriculture. We may have
come full circle in that environmental concerns will reshape agriculture
and in so doing, reshape the countryside.
But promoting the value and po-
It should be stressed however that while a simple contrast of powerful
images makes a good story, most conservationists would acknowledge that
the reality of habitat restoration and maintenance is much more complex
disease under control and thus to control Nature and the ecology of com-
Nerlich, Clarke, and Dingwall 2000), which all tie in, in one way or an-
As hinted at in the epigraph used at the beginning of this chapter, the start
of 2001 was a turbulent time in the UK: This year, like the enactment of
some apocalyptic, millennial fantasy, we have already had storms, foods
and blizzards. Agriculture is still linked to BSE, e-coli, salmonella, bovine
tuberculosis and swine fever. Now theres a visitation from a virus, reap-
pearing from a painful, long-ago memory, and burning through the ecology
of commerce like wildfire. This is how Paul Evans, the
s coun-
tryside diarist, described the situation on March 7, 2001 in his article enti-
tled The silent spring. The last major outbreak of FMD in Britain had
occurred in 1967. At that time over 250,000 million animals were killed,
mainly in one part of the country. This time about 10 million, mostly unin-
fected, animals, were killed all over the UK. It should be stressed that FMD
the countryside, as one writer, Graham Harvey, has put it in a book pub-
There was once a town in the heart of America where all life seemed
to live
in harmony with its surroundings
. The town lay in the midst of a checker-
board of prosperous farms, with fields of grain and hillsides of orchards
Along the roads, laurel, viburnum and alder, great ferns and wildflowers
delighted the travelers eye through much of the year. Even in winter the
roadsides were places of beauty, where
came to
feed on the
countryside was, in fact, famous for the abundance and
intensification is about making as great a proportion of primary production
as possible available for human consumption. To the extent that this is
atomic bomb
nuclear winter
Gulf war
cold war
risks to animals, humans, food, environment, ecology
The period under study in this chapter covers the years that followed the
cloning debate in 1997 and the GM food debate in 1998 and goes up to the
outbreak of FMD in the UK in 2001 and its consequences, with the shadow
For this pilot study I used the available online material from four British
industrial farming
'the silence of the lambs'
'terminator seeds'
'seeds of hope'
(pastoral idyll)
animal noises
physical sense but also a collective consciousness capable of reason, com-
This was the beginning of a new literary tradition of apocalyptic narra-
tives and of the new genre of the ecocatastrophe, inspired both by the threat
At the same time Paul Ehrlich published another book that played with the
image of the bomb:
(Ehrlich 1968). Many of the topics
tackled by Ehrlich overlapped with Carsons interests, especially the effects
that humans have on nature, such as deforestation, overfishing, chemicals in
the atmosphere, the toxification of the environment and the human body (what
Nicola Baird called a toxic time bomb in an article for
Silent Spring
, 25/09/02), and, of course, the exponential growth of the
Cognitive operations and pragmatic implication. In
Constructions: A Construction Grammar Approach to Argument
13.Events are conceptualized here as idealized cognitive models (ICMs) that con-
14.Yorick Wilks (p.c.) has argued that the sense of (17) is General Motors was forced to
stop production. If this were correct, the proposition General Motors stopped pro-
a free TV!
Whats that
Whats that
Whats that
Shes got
Bills a
.The taxonomic structure of the
(41)In his earlier incarnation he
to veil his power (as Gandalf
) and could appear as a commanding figure of great strength of
to be coreferential with the
presidents speech
. However, there is clearly a
The president
was brief�

the president
the presidents speech
: conceptually backgrounded
conceptually prominent
The reading of (31) would thus be that the
given by the
anaphor to be used is therefore
The president

speak briefly

: conceptually backgrounded
conceptually prominent
himself did not drop the bombs on Hanoi, but he was ultimately responsi-
ble for this military action. In other words, the referent designated by the
source meaning is the
of the action. However, it is not the
U.S. Air Force Pilots
: conceptually backgrounded
: conceptually prominent
the sa&#x-6.2;&#xt-7.;he1;.7 ;牐&#xsa15;&#x.7x0;x
the sax
the saxophone player
: conceptually backgrounded
conceptually prominent
.Conceptuallyprominent targetmeaning
North Koreas willingness to publicly
flout its commitmen�ts
ngness to
N.K. publicly
flout its commitments

signifier-signified relation
metonymic relation
non-activated metonymic links
: conceptual prominence
For example, the
tied to the
An alternative approach to sentence (15)

more in line with construc-

would assume that the imperative con-
cal meaning might also nibble at constructional meaning and change it
However, there are also naturally sounding utterances like (13) that contain
(13)Be wealthy in ten months.
One should normally not expect to find a stative predicate like
be wealthy
in an imperative construction. Nevertheless, sentence (13) receives an ac-
ence point in Ronald Langackers terminology. The linguistic context of
of implicature. For example, in their introductory textbook to
whole chapter (ch. 4) to the significance of pragmatic inferencing in the
emergence of grammatical meanings from lexical meanings, do not make a
of being both abstract and specific enough to guide actual reasoning to the
inson 2000: 6
7, 28). Levinson proposes that, from the speakers per-
spective, the solution to the bottleneck problem is to encode only the strict
(but sufficient) minimum of information and leave the recovery of the full
richness of intended meaning (including default meaning) to inferencing
To quote Levinson (2000: 29): [] inference is
cheap, articulation expensive, and thus the design requirements are for a
In modern pragmatic theories inferential meanings are accounted for by
1996Mental sp
aces and the grammar of conditonal constructions. In
, Gilles Fauconnier and Eve
2002bOn the ubiquity and mu
2. Instructions for choosing the right general route.
That is,
is counted as one compound word,
) is counted as one word (thus inflectional morphemes are not counted as
is counted as two words, despite the contraction, since they
have independent citation forms (
) and they do not constitute an estab-
16.To save space, as a rule only overall statistics for the five case studies are of-
has in its frame a role structurally equivalent to the role of its counterpart in the
source (e.g. in the
The brief conversation on which this study has been done was slightly adapted
from Radden (2000: 94
The text is a well-known joke, attributed to W.C. Fields and borrowed from At-
The text chosen for this study is the
first sentence
of the first paragraph of an
authentic narrative-descriptive text (comprising nine paragraphs, 660 words) ran-
A similar example occurs in Case Study D (noun phrase
tence 2 in the text; see appendix), where
10. The inference (afforded by epistemic conditional) Under
normal conditions, the reader that has driven west on that section
links in the chain linking this non-prototypical oblique form of she to
the prototypical form. Two likely links are these two weak forms before
vowels: /h
The prototypical referential meaning of the
the Continental Divide
(i.e. a phrasal name), namely the Rocky Moun-
tains, a sense registered both by the
and by the
A long dialogue, again chosen at random (consisting of the final para-
graph of the opening stage directions and the first scene of Act I of
The texts are presented in the final Appendix. They should be checked
constantly when reading the analyses of parts of them which are presented
trix clauses may be more dependent than complements. In
Representation: Linguistic and Psycholinguistic Aspects
Speaking: From Intention to Articulation
roducing spoken language: a blueprint of the speaker. In
Neurocognition of Language
2001The analysis of disc
The Handbook of Discourse
2002Speaking in
Mappings in Thought and Language
1996Practices in the construction of t
Constituency and the grammar of turn increments. In
guage of Turn and Sequence,
Cecilia Ford, Barbara Fox, and
Sandra Thompson (eds.), 14
38. Oxford: Oxford University
1996Interactional units in c
onversation: syntactic, intonational and
pragmatic resources. In
Emanuel A. Schegloff, and Sandra Thompson (eds.), 138
Intentions in the Experience of Meaning
1995Sentence construction within interacti
Aspects of Oral
Media Texts: Authors
Basic discourse units are Basic Discourse Acts. They are multidimen-
sional wholes which consist of subacts at the four dimensions of conceptu-
alization, formulation, materialization, and communication. This configu-
writing also involves acting upon an addressee, albeit at some remove. This
duction intended as (multidimensional) instructions to update a (multidi-
mensional) Current Discourse Space, as we have seen above; the corre-
sponding acts of reception or comprehension provide evidence that they are
generally taken as such by language users. That is why I propose to see
In conversation, Basic Discourse Acts may turn out to be the crucial
salient value at each of the dimensions. It is the typical illustration of lan-
guage use, consisting of one proposition, one clause, one intonation or
Typical basic discourse units may be contrasted with less typical basic
discourse units by either adding or removing salient values (propositions,
intonation or punctuation units, and clauses, illocutions) from their struc-
ture. I have shown that it is possible to have more than one salient value
within a basic discourse unit for each of the four dimensions, except for the
communicative one. Similarly, I have shown how it is possible to have a
lack of a salient value at each of the four dimensions, with the possible
Pending research on the latter issue, deciding about the status of a dis-
(2) What is the weight of the values of the other dimensions in making
If a unit has salient values at all three other dimensions, it is probably
easier to come up with an illocutionary act for that unit than if it has non-
salient values for all three other dimensions. In this light it will be interest-
ing to study the nature of the basic discourse units identified in RST and
One of the assumptions that plays an important role in making these de-
The mere presence of salient categories for each of the four dimensions,
then, does not necessarily point to basic discourse unit status. When salient
categories are used independently, they may provide such indication. But
when they are used dependently of more encompassing units at their own
dimension, they cannot be taken as signals for basic discourse units. In-
The second variation on the theme of non-basic discourse units takes us
away from the salient categories highlighted so far, and turns to the other
possible values for each of the dimensions. If salient categories may be
used as
parts of
basic discourse units, as we have just seen, this is even
more typically the role for non-salient categories. Thus, non-basic dis-
course units for the conceptual dimension are all possible components of
propositions, such as concepts designating relations, entities, and attributes.
And non-basic discourse units for the vocalization dimension are all possi-
ble components of intonation units, such as pausal segments and so on.
Non-basic discourse units of the lexico-grammatical dimension are con-
stituents of clauses, which, apart from dependent clauses, are phrases. And
finally, non-basic discourse units at the level of communication are all
componential acts of illocutions, such as many of the secondary and com-
Non-basic discourse units are parts of basic discourse units. They are
typically realized by non-salient categories for each of the four dimensions.
However, they may also be realized by salient categories, provided that
these are used in a dependent fashion. And finally, in some cases, even
independently used salient categories may still function as parts of basic
discourse units, as is the case, for instance, for such substantial intonation
units as On the desk and Conceivably in (1) and (4a). These observa-
tions merely point out the need for further modeling and explication in
in which an entity can be conceptual unit for discourse building. Therefore
A basic discourse act that does not contain a clause may be exemplified
by phrases of affirmation or denial, as in the two No turns in our conver-
nds I think he
Took him for a ride on thatn Bill said that he was at lea
st goin
hty miles an hou
r, which is supported by an analysis of
gaze. For that reason, Gary adds an Extension in order to offer another
However, by the same argument, it does not only illustrate a discourse
unit with two illocutions, but also with two propositions, two clauses, and
probably two intonation units. In other words, the first turn may also be
analyzed as consisting of two typical basic discourse units which combine
to produce one higher-level discourse unit, the latter functioning as another
whole in the sequential action structure. In that case, the first turn of (5)
does not illustrate a less typical basic discourse unit with two illocutions,
but two typical basic discourse units in one higher-order discourse unit, the
If discourse units are only measured with reference to the dimension of
turn-taking, which is not quite equivalent with the communicative dimen-
sion as I conceive of it, then the first turn of (7), almost by definition, is
one unit, because it is one turn. However, since turns can consist of more
than one act, it is still possible to see two basic discourse acts in the one
sequential discourse unit called turn. Here is another illustration of the
fication of the boundaries of basic discourse units. But we shall come back
A second variation on the theme of less typical basic discourse units in-
have to be taken into account to arrive at an analysis of basic discourse
units as opposed to units at the various distinct levels of discourse organi-
The last variation on the theme of less typical basic discourse units con-
It is true that these parts are separate units at one dimension of dis-
course, in this case the material dimension. This may even mean that they
have a role of their own to play in updating the current discourse space.
This is the basic point that Langacker is making in his paper. From that
However, the fact that they are independent intonation units does not
automatically turn them into separate basic discourse units as intended
here. For that, some similar degree of independence at one or more of the
other dimensions is required. It is the central question implicated by this
another discourse-analytical approach which invokes different combina-
Both approaches provide indications that analysts may have oriented
likely that these categories play a major role in the structure of these units.
If a proposition, an intonation unit, a clause, and an illocution are all used
to achieve closure at their respective dimensions at the same time, this is a
powerful signaling device to the addressee that a discourse act has been
How such typical basic discourse units are distributed across various
classes of discourse and how they are used for which purposes in compari-
son with less typical basic discourse units is an unanswered question. Ex-
amples of typical basic discourse acts from the article by Langacker (2001)
(3)On the desk, he noticed an important-looking document.
would require more than one proposition to account for the structure of the
clause. Clauses would then not be equivalent to propositions. An approach
stantive intonation unit, as opposed to regulatory intonation units. For for-
mulation, the typical basic discourse unit has the value of the clause, as
opposed to other groupings of phrases, or the phrase by itself. And com-
munication is the dimension of basic discourse units that Langacker does
not discuss as a separate dimension. However, he has used the relation
dimension manifested by
. These references and others that
will follow below may serve to indicate that my proposal is not totally
At the same time, though, these references also suggest that there are
too many issues in the modeling of basic discourse units for extensive
analysis and discussion at this moment of the deeper reasons of these dis-
quotation from pages 144
145 above, he writes that the focus of attention
is what an expression profiles (by means of its lexico-grammatical ele-
ments), and he adds that this focus is included in the frame of attention
(which by implication is a wider conceptual structure). This seems per-
fectly fine to a discourse psychologist, but it also suggests that we are
To illustrate what may be at stake, consider the following line from a
From my point of view, Langacker assigns too restricted a function to
discourse units, as merely serving to update the conceptualization of what
is being talked about. In discourse-psychological terms, Langacker con-
centrates on the role of discourse units for updating situation models (or
event models, see Van Dijk 1999: 125), at the expense of their simultane-
ous function for updating context models (Van Dijk 1999; Van Dijk and
Kintsch 1983). This is where my communicative dimension and its typical
To rephrase my view in the terms Langacker uses in the above quota-
tion, in successful communication, people do not only manage to coordi-
nate their action and focus attention on the same conceived entity, but they
also manage to coordinate their interaction and focus attention on what the
other language user wants them to do in conceptualizing that entity. For
instance, when Langacker (2001: 174
177) analyzes an extremely simple
short narrative text, he treats the consecutive clauses of the text as instruc-
tions to build a sequentially organized conceptual structure. However, he
does not include within his representation the communicative aspect that
these discourse units are meant to be accepted as so many truthful asser-
tions, with implicit temporal and causal interrelations (e.g. Mann and
Thompson 1988, 1992). Moreover, these assertions are intended to contrib-
ute to an overall coherent message with a divertive or informative or per-
suasive function that is part of a specific genre of discourse, such as a con-
versation (e.g. Steen 1999, 2003). Even though Langackers analysis is
compatible with this type of discourse analysis, it leaves out an important
dimension of the role of discourse units in the psychology of communica-
tion, namely their communicative role, which in typical cases is manifested
That Langackers approach is compatible with discourse analysis as I
see it is suggested by one example in a section on augmented linguistic
units. When dealing with the conventional use of declaratives as assertive
speech acts, Langacker (2001: 166) writes: It is only when a finite clause
is embedded in a larger interactive frame

involving the speakers intent to
portray the proposition as true and offer it for the hearers assessment

producing it constitutes an assertion. The point of the present approach is
that language use never takes place outside such a larger interactive frame,
and that its role should be systematically accounted for in the description of
discourse units. This should not only take place in those cases where pat-
terns of pragmatic inference have become entrenched to produce aug-
mented linguistic units, for instance to the effect that declaratives can con-
ventionally be taken as assertive acts. It should happen for all usage events.
nation units, clauses, and relations
. At the same time, in emphasizing
these are associations, they also suggest that other values and con-
There are two refinements that I wish to add to Langackers comments.
First, intonation units require an equivalent category for written language
use, for which I will use the term punctuation units (Hannay and Kroon
is a semantic term in Cognitive Linguistics
which, when used in discourse-analytical contexts, may be ambiguous be-
which is Langackers (2001) paper on discourse in Cognitive Grammar.
But as far as I can see, none of the other proposals that I know of is entirely
identical with what I am proposing here. When it is relevant and possible, I
If we take the notion of a typical basic discourse act as valid, turns 1 and 2
are typical, while turns 4 and 5 as well as 8 and 10 are less typical. The
units in the other turns have more subtle properties that may make them
more or less typical, such as the embedded clauses and propositions in turn
6. Moreover, turn 7 may be considered as a case where we may have to do
with two instead of one basic discourse acts. It will be a major concern of
this chapter to discuss some of the ways in which basic discourse acts may
This chapter will propose, develop, and discuss a simple idea, that there is such
consists of an illocutionary act, a proposition, a clause, and an intonation unit.
In writing, the intonation unit has to be replaced by an equivalent, the punctua-
tion unit. Basic discourse acts may be thought of as utterances in the full be-
havioral sense of the term, that is, as verbal acts requiring production and com-
prehension in speech or writing. However, since
is too closely
associated either with pragmatics as opposed to discourse analysis, or, within
discourse analysis, with conversation analysis as opposed to text analysis, it is
Basic discourse acts may be typical and exhibit the four properties of
illocution, proposition, clause, and intonation/punctuation unit, or they may
be less typical. Consider the following short dialogue, taken from a novel
by Julian Barnes and slightly edited for our purposes. A father and his
(1)1F: Was that it?
10D: No.
A discourse-oriented Cognitive Linguistics
toru, Daniela Tuchel, and Michaela Praisler (eds.). Bucarest: Di-
2002Patterns of conceptual interacti
9 (2):124162.
Mental Spaces: Aspects of Meaning Construction in Natural
Cambridge/NewYork: Cambridge University Press.
1996Blending as a central
Giving an adequate explanation of how such mechanisms work endows
conceptual projection theory with a large degree of parsimony. Note that
emergent structure theorists need to postulate the existence of a dynamic
blend which levels out inconsistencies simply because they believe that the
blend only incorporates structure which has been correlated in a careful
way. In our own hypothesis, Turner and Fauconniers blend is replaced by
a projection space, which is different from other mental spaces, like input
spaces, in that it does not supply information but receives and combines it.
It resembles the other spaces, however, in that it is not a dynamic construct.
What is dynamic in conceptual interaction is the different cognitive mecha-
hands at a taxi as a signal
to make it stop, then got
into it and hired a ride to
the airport. Speaker B got
to the airport
sion involves much more than simply seeing a certain angry person in
terms of a container with burning contents. For example, smoke serves as a
figurative indicator of internal combustion because of extreme heat. Real
combustion generates consumption of energy and materials. In a similar
eventually deprived of energy and vitality. That is why we can say that a
person is being consumed with anger. This is just one out of several po-
tential meaning implications of
You could see the smoke coming out of his
The person has lost control of his anger (i.e. the figurative fire has
The person is potentially harmful for other people (in the same way
These implications are independent of the context but need to be com-
public confession

public confession
of misdeed
marital infidelity
and its
in politics

s public
confession of
infidelity to avert
the negative
of his misdeed
the politician

s public
confession of his marital
infidelity is an exaggerated
demonstration of sorrow,
schema: journeys can have good and bad moments
The principle of interaction we have just presented has an important role to
play in the creation of combined spaces. A clear case is our discussion of the
You couldseethe smoke coming out of his ears
(figure 3 above),
(the burning substance which gives off smoke and heat) becomes
part of input
(the container) and not the other way around. Another clear case

phoric element in the protasis and the apodosis does not describe some-
Note also that both impossible conditionals and what we call pure
counterfactuals convey some sort of speakers emotional reaction. But this
reaction derives from different sources in each case. In an impossible con-
where Clinton merely seems to be surviving the scandals. In the blend,
instead, the scandal-iceberg is the greatest conceivable threat and the
Clinton-Titanic survives even this kind of threat. This structure is, accord-
ing to Turner and Fauconnier, constructed in the blend and projected back
ances (see Lyons 1977: 795). Counterfactuals take the form of unreal con-
ditional statements (e.g.
If I had tried harder, she would still be alive
weigh a few pounds). In the mapping, we see the physical and psycholo-
cal effects of the weight of the suitcase on the protagonist (i.e. the person
who has to carry its actual weight) in terms of the effects that we believe
Scalar concepts are not only amenable to mitigation but also to the con-
verse operation, which may be called
Sperber and Wilson (1995) have already identified this cognitive operation
under the label of enrichment in the context of Relevance Theory. For
An enriched representation contains the same information and more than
the initial representation. Rcanati (1989) uses the label strengthening to
Counterfactual sentences have traditionally been considered a subcase of
, i.e. expressions that commit the speaker
to the falsity of the proposition or propositions expressed by one or more of
its constituent clauses.
are the other subcase of contrafactive utter-
highlighting is an economical operation for the speaker: it is the ad-
combined input hypothesis
preserves some of the most relevant
Projection spaces (i.e. the counterpart of Turner and Fauconniers
blends) do not contain structure inconsistent with the structure
projected from the input spaces. The apparent existence of emer-
gent structure is explained in terms of the activation of multiple
conceptual spaces that interact according to a range of possible op-
Projection spaces are not dynamic in the sense that Turner and
Fauconnier postulate for blends, i.e. by being able to create their
own emergent structure independent of the structure provided by
the input spaces. Instead, we consider projection spaces to be the
Cognitive operations like integration, correlation, and contrast play
a prominent role in regulating the outcome of the projection proc-
ess. Other operations are also considered (see section 4 below):
input x]
input y]
input z]
input x]
INPUTx (container)
substance burns
smoke is produced
and released
heat is given off
interior a burning substance
which releases smoke
through an opening in the
container walls and which
gives off heat
in his interior,
perspires profusely,
and his skin is hot
(as evident from its
probably reached its limit
before the person loses
You couldseethe smoke coming out of his ears
The problem with this analysis is that it makes the strong claim that in the
blend there is actual integration of conceptual elements from the source and
Northen Light
journey of the
Great America
of the Northen Light
and the Great America
imaginary combined
gent structure inconsistent with that of the input spaces. The four-space
Generic Space
By way of illustration of how this model works, consider the following
example taken from Fauconnier and Turner (2001). In it, a clipper,
, which currently sails from San Francisco to Boston, is involved
in an imaginary race against the
, which did the same jour-
ney in 1953. In order to understand this situation, we need to combine the
following mental spaces: one input space for the passage of the Northern
Light in 1953; another for the passage of the present run by the Great
America; a generic space, which extracts structure common to the two in-
put spaces (i.e. a ship makes a journey of a certain duration from a source
to a destination); and the blended space into which the Northern Light and
the Great America are projected as taking part in a race. The blended space
has emergent structure that does not exist in any of the input spaces, where
action patterns are only one kind of constraint on conceptual projection.
They allow us to know about interaction possibilities but they reveal noth-
ing about the kinds of cognitive operation that underlie the projection and
integration of conceptual structure from different mental spaces. So we
shall also address the issue of how such operations govern the flow of in-
and projection spaces
M. Sandra Peña
The question of conceptual interaction has been a relatively important area
of interest in Cognitive Linguistics, especially in work carried out by Mark
notion of ‘blending’ or ‘conceptual integration’ (see Fauconnier and Turner
1996, 1998, 200
; Turner and Fauconnier 1995, 2000). Blending theory has
had an enormous impact on the cognitive linguistic community. It is gener-
ally seen as complementary to the standard cognitive model theory pro-
pounded by Lakoff (1987) and his associates (see Lakoff and Johnson
Bodily expression of emotion.
European Journal of Social Psy-
you squeeze a tomato? The role of motor representations in
semantic sensibility judgments.
Journal of Memory and Lan
f-recognition: Body and action
. Trends in Cognitive Sciences
2001Predicting the effects of action: Interaction of perception and ac-
1985The motor the
ory of speech perception revised.
Finally, the work described in this chapter provides another source of
evidence in support of the cognitive commitment within Cognitive Lin-
guistics, in which cognitive linguists seek explanations of linguistic struc-
ture and behavior in line with contemporary empirical findings on human
cognition from cognitive science. Not only does much of the work on em-
Linguistics, and its emphasis on embodiment. But Cognitive Linguistics,
through its demonstrations of the embodied grounding for many aspects of
language, has created a intellectual climate for cognitive scientists, like
als actions, including those associated with overt communication. One
ries above. Immediately after hearing the story, the participants were blind-
folded, and asked to walk out to the yellow ball while they were thinking
about the story they had just heard. At that point, students began to walk
out to where they thought the yellow ball was and then stopped when they
thought they were right at the ball. Once they stopped, an experimenter
nearby asked each participant to rate on a 7-point scale how they felt at the
moment. After this, the blindfold was removed, and the experiment was
over. The experimenters then measured how close the student actually was
to the yellow ball, and how far away from a straight line each participant
Did students walk differently having heard the smooth journey as op-
posed to the interrupted journey story? In fact they did. According to the
the relationship progressed further than the one in the interrupted journey.
Indeed, in the walking study, blindfolded students who heard the smooth
real walking. It appears, then, that thinking about the two stories differen-
tially affected peoples imagined walking, and it did this for both real and
This line of research is still in its infancy. But the findings observed in
these experiments strongly suggest that image-schematic reasoning in nar-
rative comprehension involves the construction of embodied simulations.
These simulations are embodied because of the functioning of as-if body
loops that are part of peoples immediate understanding of other individu-
not by merely activating a RELATIONSHIPS ARE JOURNEYS concep-
should verify that this phrase is meaningful faster than when they first per-
movements associated with these phrases should enhance the simulations
mental images for idiomatic and proverbial phrases (e.g., spill the beans
stood simply in terms of the physical actions they depict without any men-
tal mapping of knowledge from a different domain of experience. The
and phrases mean what they do, and in how people use and immediately
understand of language. An important theme of this work is its suggestion
1978). Diary studies show that when people imagine
specific emotions, these often correspond to very specific changes in body
movements and posture (W
tt 1998). One way to characterize the felt
dimension of emotional experience is in terms of affective space, or the
space we move through as we experience different emotions (Cataldi
1996). Consciousness is also clearly associated with basic sensory and
sensorimotor imagery in working memory, consisting of sensory and motor
associations, distributed across the cortex, in combination with ongoing
somatosensory input. Even higher-order consciousness is closely tied to
This overview touches on only a few selected topics in cognitive sci-
ence, but suggests that there is a clear trend in which various aspects of
human perception, cognition, and language are tightly linked to embodied
2000) specifies how language referring to some situation becomes mean-
ingfully embodied. First, words and phrases are indexed or mapped onto
objects or perceptual, analogical symbols (Barsalou 1999). Second, affor-
dances are derived from the objects or perceptual symbols, not the words
themselves, which are then meshed under the guidance of intrinsic con-
Research consistent with this perspective demonstrates, for example,
that literal action in one direction will interfere with the comprehension of
a sentence implying action in an incompatible direction (Glenberg and
Kaschak 2002), and that making appropriate hand shapes (pinching ones
matic perceptual representation becomes established in memory, it can
Perhaps the most interesting aspect of perceptual symbol theory is the
idea that conceptual processing involves sensorimotor simulations. Under
this view, concepts are not understood, and stored, as abstract, disembodied
symbols, because crucial elements of relevant perceptual and sensorimotor
information are used in conceptual processing. Much evidence is consistent
with this claim, such as cognitive neuroscience work that concepts are
grounded in sensory-motor regions of the brain (Stamenov and Gallese
2002). For example, functional imagery studies demonstrate that process-
watching video clips of themselves throwing the dart than when watching a
clip of someone else tossing the dart (Knobli
h and Flach 2001). People
ration of the visual world, for example, is directed by anticipatory schemes
There is various empirical research that supports this idea of a tight link
tions of the cognitive unconscious. All three are present, and explanations
at all three levels are necessary for an adequate account of the human mind.
physiology or kinesiology (i.e., the body as object), but demands recogni-
tion of how people dynamically move in the physical and cultural world
(i.e., the body experienced from a first-person, phenomenological perspec-
Cognitive linguists have already made an important contribution to un-
derstanding the embodied mind through their recognition of least three
levels of embodiment: the neural, the cognitive unconscious, and the phe-
Neural embodiment concerns the structures that characterize concepts
and cognitive operations at the neurophysiological level. Our concepts and
perceive depends upon what they are able to do, and what they do,
in time,
what they perceive. Perception, cognition, and action are
Cognitive Linguistics is a special discipline within the cognitive sciences
because it explicitly seeks explanations of linguistic structure and behavior
not as if these were distinct from cognition, but as if they arise from, and
continue to be a part of, human cognition and experience. I claimed back in
Much of the work in cognitive linguistics is unique because it attempts to
A mental-process-oriented Cognitive
2000aPrimate c
2000bFirst steps toward a usage-based the
ory of language acquisition.
2003What makes human c
ognition unique? From individual to shared
to collective intentionality.
18 (2): 121
2003Animal behavi
our: how self-organization evolves.
1996Perspective and the representation of sp
eech and thought in nar-
rative discourse. In In
Spaces, Worlds, and Grammars
[1976]ro Villa; Rev. de Luis Cebrin Tornos). Madrid: Akal.
Problemas Actuales de la Traduccin
. Granada: Granada Lin-
1998The notion of Dynamic Un
it: Conceptual developments in cogni-
ace accessibility and mood in Spanish. In
Spaces, Worlds, and
2002The Faculty of La
nguage: What is it, who has it, and how did it
2002An ecological a
pproach to embodiment and cognition.
2002Nurturing a view of human nature (A review of S. Pinker:
Iacoboni, Marco, Roger P. Woods, Marcel Brass, Harold Bekkering, John C.
Cortical mechanisms of human imitation.
286 (5449):
2003Adversity causatives in Port
1997Why statistical universals are better than absolute universals.
Raisons Pratiques. Sur la Thorie de laction
. Paris: ditions du
1996Mental sp
aces, constructional meaning, and pragmatic ambiguity.
Spaces, Worlds, and Grammars
, Gilles Fauconnier and Eve
1998Espacios mentales y ac
titudes epistmicas: la alternancia indica-
tivo/subjuntivo en espaol. In
Estudios de Lingstica Cognitiva
Vol. 2, Jos Luis Cifuentes (ed.), 441
El enigma de la esfinge. Las causas, el curso y el propsito de la
La Especie Elegida. La Larga Marcha de la Evolucin Humana
Usage based models of language
. Stanford, CA.: CSLI Publica-
2002The c
orpse of a wearisome debate. (A review of Pinker 2002).
1997aA failed c
ognitive analysis of the Spanish subjunctive. Ms., Uni-
1997bA partial synergetic model of deagentivisation.
Journal of Quan-
1998El subjuntivo espaol y los espacios mentales. In In
Estudios de
. Vol. 2, Jos Luis Cifuentes (ed.), 451
466.Alicante: Servicio de Publicaciones de la Universidad de Ali-
cognition are just two faces of the same thing (a familiar paraphrase:
This chapter has been mainly devoted to posing a problem and exposing a
danger for the Cognitive Linguistics enterprise. We have to be fully con-
scious of the need to (1) study linguistic variation, while avoiding exces-
4.3.What is to be compared in cognitive linguistic typology: the
Our proposal also shows what the comparanda have to be: not simple lin-
guistic forms or pairings of form and meaning, but the whole system of
cially Len 2002; also Alterman and Garland 2000; Clark 1999; Hirose
2002; Ratner 2000; Semin and Smith 2002). But what is more, the (indi-
vidual) cognition of all the individuals participating in similar collective
activities will develop in similar directions, in dependence of the activities.
We can say that in relation to a particular activity, all the participating indi-
viduals will collaborate in such a way that one can speak of distributed or
collective cognition. Mandelblit and Zachar (1998: 254) say on distributed
Cognitive activity may involve processes internal to the single individual,
impossible to dissociate from interaction, understood as social activity.
Michael Tomasellos recent work (1999, 2000a, 2000b; Tomasello and
Following the lead of Vygotsky , Bruner , Cole , and other cultural
psychologists, my view is that what makes human cognition unique, more
That is, all of the many artifacts that enable and empower human cognition
() are the joint product of many people working over many years, com-
In fact, Tomasellos view of imitation, attention to other peoples ac-
tions and development of a theory of mind as the central element in the
acquisition of language by children, also as opposed to the shortcomings of
those same social activities in apes, witnesses the extraordinary importance
of social, i.e. collective activity, for the development of individual cogni-
tion, both in the ontogeny and the philogeny. In a similar vein, palaeon-
tologists point to the richness of the social interaction of Homo Sapiens as
opposed to Neanderthals as the main reason for the prevalence of the for-
mer (Arsuaga 1999, 2001; Arsuaga and Martnez 1998), instead of the
after all, impossible to demonstrate pre-eminence of any a priori cognitive
In Cognitive Linguistics, mainly but not only when dealing with the
multiplicity of human languages and their nearly limitless variability, the
situatedness of cognition has to be integrated in our research. According to
Behavior can only be understood in the context of complex real-world
situations. An important focus of research should therefore be the relation-
convenience of taking usage as a central element of language, including the
Now, it is this usage-based grammar that has to be the focus of typological
research, instead of the traditional emphasis on form or the mere pairing of
form and meaning in absence of any context or conditions of use and usage.
a universal, global perspective on human emotions is misconceived: yes,
we usually hit each other in certain preferred ways, so to speak, and this is
Another example: Cristbal (in press) elicited Basque speaking chil-
drens version of the story used by Slobin and colleagues in their study on
thinking for speaking. Translation was not used in this case, however, but
free verbal construction on the basis of a series of drawings.
She found
that whereas most children began their telling using stative constructions
(as There is a boy on a tree and the like), in a few cases the children be-
gan with an active construction: a child has climbed a tree. The key for
this apparent anomaly was that the child was then trying to tell a tale, in-
stead of just verbalizing the drawings. Symptomatically the typical story
... one day..., which did not appear in the
more canonical, stative narratives. In this case, the selection of the stative
or the dynamic expression is a matter of text type, i.e. of the socially sanc-
tioned verbal activity to be carried out. A context-free elicitation of data,
i.e. one in which no attention is paid to the possible use that linguistic ex-
Moreover, frequently the translator is seen as free from errors and in-
consistencies (mere wishful thinking!). Of course, this does not apply to
anyone who has been confronted with the elicitation of data in the context
Both expressions are quite different in usage: in (17), the event is seen as a
whole and Spanish speakers do not focus on the manner of the action; this
is the most usual case. In (18), however, the speaker wants to emphasise
the manner of the verb; one can make up a context, for instance in a narra-
tive, where (18) could be the adequate, even the most convenient or even
Spanish could be equally classifiable as verb-framing or as satellite-
framing, usage being the decisive factor for the selection of one or the
other. Of course, we know that this distinction is not an absolute one, but
normally a matter of degree, and in these terms we could just say that
Spanish usually occupies a position towards the satellite-framing end in a
continuum, while under certain conditions being able to move towards the
closest to the original. This will be taken as the linguistic form of expres-
sion of transitive constructions in that language, at a pair with its English
counterpart, and consequences will then be drawn. Quite frequently, how-
ever, the elicited construction can just be the literal rendering of an English
another language, taken as an unproblematic, unlimited, general form of
Now, what kind of universal principles could we posit on the basis of
the Portuguese inflected infinitive? Should we begin discussing it in the
terms of its cognitive representation which necessarily has to be universal,
So English is a typologically rare language, but Portuguese, and cer-
tainly all languages, also features some quite infrequent constructions.
But... which language (type) is the most frequent one? Or, which indi-
vidual construction reflects basic cognitive processes in relation to lan-
guage? The tendency to take English as a sufficient basis for the positing of
universal features of human language and cognition has been recently criti-
cally analysed by Golumbia (2004) in connection with nonconfigurational-
ity as defined and studied by generative grammarians. From his paper one
ways in which a subject can be affected by an action carried out by others
I we take English as the tertium comparationis, we may be comparing lan-
guages on the standard of some linguistic rarity. Maybe this claim needs
Certain English grammatical constructions are extremely infrequent
(3) This pen has been written with (the subject is an instrument).
(4)This table has been slept on (the subject is a locative).
A similar construction does occur in other languages, for instance in the
Philippines. There is a significant difference, however, because in these
languages as opposed to English, the verb bears an explicit mark of its
having an instrument as its subject or, more appropriately, as the sentence
topic (see Siewierska 1991: 4.3.1); as is well known, the grammatical
Also the use of dummy auxiliaries is quite infrequent, as is the obligato-
riness of an expressed syntactic subject; or the resultative constructions
which, while having their counterparts in other Germanic languages, are far
(6) She laughed herself sick.
The same happens in other areas of language, where English has apparently
This construction could seem rather straightforward; its structural analysis
is fairly simple and apparently unproblematic. We could be tempted to
While much work on emotions has assumed that (scientific?) English pro-
has no internal variation in itself. However, it also was a bad choice because it
belongs to a particular type of language, it is linguistically non-neutral and the
reality of other languages was frequently distorted by our trying to understand
exotic phenomena through the Latin lenses. To this, another danger was
added: the Latin language was taken to reflect logical thinking, in such a way
that whatever did not fit the Latin structures, was also incorrect from the point
of view of thinking itself. That is, universal thinking we would now prefer
To avoid this, a number of more or less formalized comparison stan-
man language, i.e., to go from languages to language, from idioma to len-
All this is well known, as things have been basically like this for a long
time. But more recently the need has also arisen for the identification, de-
scription and explanation of the cognitive bases of language and linguistic
as the primary object of research. Maybe a consequence of the pervasive
folk theory of essences (Lakoff and Johnson 1999), or of the traditional
tendency to see what is most abstract, less in direct touch with reality, as
Social cognition: variation, language, and culture in
This chapter proposes to take linguistic variation and linguistic typology as
central elements for the theory of Cognitive Linguistics. This implies the
need to avoid taking one single language, English, as the point of reference
and the fundament for the identification and definition of general, univer-
sal, linguistic and cognitive principles and processes. The need for a socio-
cognitive view of language is emphasized, implying a view of cognition
not limited to the individual and including its role for human activity,
Constructing a Language. A Usage-based Theory of Language
inpressUsage-based linguistics. A technical state of the art. In
1954Is a structural dialectol
Usage-based Models of Language
Barlow and Suzanne Kemmer (eds.), vii
xxviii. Stanford: CSLI
2000De distributie van niet-ana
forisch er buiten de eerste zinsplaats.
Sociolectische, functionele en psycholinguistische aspecten van
ers status als presentatief signaal. Ph. D. diss., University of
2002Regressing on
. Statistical analysis of texts and language varia-
6th International Conference on the Statistical Analysis
of Textual Data
. Cambridge: Cambridge University Press.
2001The acquis
ition of finite complement clauses in English: A cor-
1988The conceptualisation of vertical sp
ace in English: the case of
Topics in Cognitive Linguistics
, Brygida Rudzka-Ostyn
social norms, but if the very nature of these norms as a coordinative struc-
ture reveals itself in their uneven, clustered distribution over numerous
Second, if Cognitive Linguistics embraces a social conception of lan-
guage (as it is compelled to do if it is to avoid the pitfalls of epistemologi-
that the logic of the argument is of the greatest common factor rather than
a lowest common multiple kind: the relevant clear cases are those that are
shared by all speakers of the language, precisely because they are supposed
to master
language. Following this logic, however, the common vo-
cabulary might be very small, and a number of words that at least some of
us would intuitively expect to be part of the language (
), might fall outside of it. Conversely, if we
try to salvage the vocabulary, we seem to sacrifice the size of the linguistic
community, in the sense that we might be left with just a few speakers with
The only way out of the dilemma (forfeiting the speakers or forfeiting
the words, downsizing the linguistic community or downsizing the vo-
cabulary) would be to accept a diasystemic view in which, as mentioned
earlier, the linguistic community is defined not in terms of a single reper-
toire of linguistic means of expression shared by all members of the com-
munity, but rather as a conglomerate of overlapping repertoires. But de-
scribing such a situation would
imply adopting an empirical,
such. With regard to the global norm itself, the deviant expression is en-
tirely legitimate (assuming of course that it is successful). Innovation in
language deviates from the existing conventions, but the deviation may be
While this type of underspecification of norms may indeed be the back-
ground of unclarities, there are two additional points to recognize. First, the
This however raises the question only marginally dealt with by It-
konen what the status of borderline cases in the normative system could
be. Why do borderline cases occur? I will argue that there are two sugges-
tions in Itkonens book of ways to explain the existence of unclear cases,
but that a closer scrutiny of each of them leads to the conclusion that an
4.2.1. Borderline cases (as reflected in the intuitive uncertainty of the lan-
of the language, i.e. even when he would make all the right, correct, nor-
mative decisions in actual contexts of use, he need not be able to recall or
reconstruct those contexts in an vitro situation. And when it is difficult to
come up with a sanctioning context, speakers may hesitate to pass an un-
ambiguous grammaticality judgment. In this sort of situation, the uncer-
tainty would not be a primary datum, but would itself result, ironically
grammaticality judgments (differ-
But such pre-experimental notions could just as well be termed pre-
guistics (see Labov 1972). From the perspective of Cognitive Linguistics,
however, which is only gradually turning towards variational linguistics
and a social conception of language, it may be useful to defend the socio-
social norms, not to the world of psychological states and hence intuition
In actual practice, however, the intuitive approach advocated by It-
konen does not seem to differ all too much from the introspective ap-
proach of generativism: the basic data of linguistics consists of grammati-
cal sentences, that is, sentences judged grammatical on the basis of ones
knowledge of the linguistic conventions. Itkonens rejection of
the contrary. According to Itkonen, linguistics as the study of language as
a social phenomenon is based on intuition rather than observation. So
Counterarguments are not difficult to find, however. While the overall
structure of the argument is correct, its practical application relies on a
questionable assumption, viz. that linguistic systems are easy to delineate.
Where exactly does one linguistic system end and where does another be-
Given this classification of types of meaning, the general argument will
be clear. First, lectal variation lies at the basis of a specific, non-
expressive or affective) involves that part of the meaning of an item that
communicates the speakers evaluation of, or his attitude towards the refer-
ent of the expression. According to the second criterion, an expression of
pain such as
, or an expression of disgust such as
(which both
have no identifiable denotational meaning) also illustrate the emotive type
Existing attempts at classification generally suffer from unclarity about the
principles lying at the basis of the classification. Furthermore, the existing
classifications often differ considerably, terminologically as well as with
The denotational meaning of a word is also referred to as its denotative,
referential, descriptive, cognitive or logical meaning. The denotational
meaning of a word does not coincide with the words reference (the extra-
sary for any Cognitive Linguistic attempt to analyse the usage data even
The work by Grondelaers mentioned at the end of the previous para-
graph is a case in point. Referring to Ariels
Accessibility Theory,
he analyses the intriguing Dutch particle
as an inaccessibility marker: it
signals that the clause-initial adjunct contributes less to accessibility than
continued in works such as Geeraerts, Grondelaers, and Speelman (1999),
2.2. Language variation has been studied in Cognitive Linguistics (and
adjacent approaches) primarily from three points of view: from a dia-
chronic perspective, including grammaticalization research (see e.g. Bybee
2001; Geeraerts 1997; Heine 1991; Hopper and Traugot
1993), from a
comparative and anthropological point of view (see e.g. Kvecses 2000;
Levinson 2003; Palmer 1996; Pederson 1998), and from a developmental
point of view (see e.g. Diesel and Tomasello 2001; Tomasello 2003). Lan-
guage-internal variation and sociolinguistic diversity has been much less
studied, but still, we may note a number of developments within Cognitive
Linguistics that are likely to contribute to an increased interest in sociolin-
First, there is the interest in cultural models and the way in which they
October 2004 all of these testify to the growing importance that the Cog-
With the increasing success of Cognitive Linguistics, the research frame-
work is developing in different directions. Some of the current develop-
ment of Cognitive Linguistics. The argument consists of two parts. On the
one hand, starting from the increasing tendency towards the use of empiri-
Current developments in Cognitive Linguistics include a growing interest
in empirical models of linguistic analysis, and a heightened awareness of
A usage-based Cognitive Linguistics
1977Reference-related and role-related
properties of subjects. In
Vol. 8:
Grammatical Relations
1995The symbolic alternative. In
New Trends in Semantics and Lexi-
, Henryk Kardela and Gunnar Persson (eds.), 89
(Ume Studies in the Humanities 127). Ume: Swedish Science
search 14).Berlin/New York: Mouton de Gruyter.
Assessing the cognitive linguistic enterprise. In
Constructions: A Construction Grammar Approach to Argument
ouns and verbs in Hocank (Winnebago).
International Journal
Patterns in the Mind: Language and Human Nature
oun and verb in Nootkan. In
, Barbara S. Efrat (ed.), 83
2000The conceptual basis of grammar: a c
ognitive approach to Japa-
nese clausal structure. Ph. D. diss., University of California at
4.These can be seen as two sides of the same coin. Conceptualizing a relationship
are independent factors that can be ascertained and even measured objec-
tively (e.g. by experimentation and corpus studies). Given real-world con-
conventionalization of a verbs occurrence in the caused-motion construc-
tion is equivalent to the verb having a caused-motion sense. This follows as
abnormal and is possibly quite prevalent for verbs to be used in construc-
tions that do not precisely match their already established meanings. I do
though have trouble accepting the notion that most verb uses are compara-
ble to that of
in (12). I have to regard the caused-motion use of
as lying toward one extreme of a continuous spectrum in terms of
how familiar it is for the verb to occur in the construction and how well the
verbs meaning fits the constructional meaning. The caused-motion use of
lies toward the opposite extreme of the spectrum. In contrast to (12), a
He sent a package to his cousin
represents a usage of
that is thoroughly entrenched and fully conventional. There is no reason
whatever to suspect its occurrence in this construction is in any way re-
markable, unanticipated, or inconsistent on the basis of its meaning. Use as
a caused-motion predicate is utterly normal in the case of
lies somewhere in the middle of the spectrum. Al-
though it may not be a prime example of a caused-motion predicate, and its
use in this construction might not be the first that comes to mind, expres-
sions analogous to (13c) are natural and familiar. I am certain I have en-
employed as a caused-motion expression on many occa-
Suppose it is granted that
are indeed conventionally es-

the shortest possible grammar may not be psychologically realis-
tic. We cannot just assume that language processing and linguistic repre-
sentations are maximally efficient and non-redundant. Moreover, positing
particular verb senses is not in principle the circular affair that Goldberg
makes it out to be. There is also a fundamental conceptual issue that needs
to be addressed: what precisely does it mean to say that a verb has a par-
ticular sense? This notion is hardly self-explanatory, but obviously crucial
Succinctly stated, Goldbergs strategy envisages a minimalist lexical
semantics and favors treating every verb use as being analogous to the
, as in (12). Insofar as possible, the semantic
nuances associated with occurrence in a particular construction (e.g. choice
of landmark, or the extent of what is profiled within the conceptual base)
are analyzed as being inherited from the construction and thus excluded
from the single, minimal meaning ascribed to the verb itself. Now actually,
I have no idea just how far insofar as possible might be. Although (in
fairness) the phrase is mine, I am left with no clear notion of the range of
cases for which Goldberg is willing to accept construction-appropriate verb
meanings, or on what basis she would draw the line. It does however ap-
pear that she resists positing such meanings in certain cases where the
verbs occurrence in the construction is conventionally well established.
These include the occurrence of
in the constructions represented in
Despite my admiration and affection for the new editor of
(with upper-case
), I believe this attitude to be problem-
is used in this construction, it is categorized by the sche-
matic caused-motion verb. Since the two conflict in the nature of their pro-
, for example, occurs in the ditran-
sitive construction resides in the constructional subschema shown in the
middle in Figure 16. This constructional variant

stantiating structures. Because the same structure is often categorized in
The constructional schemas on the left are reasonably posited as en-
trenched units abstracted from countless usage events. They are of course
category, nor is such a category universally definable in terms of distribu-
As I recall from the early days of transformational grammar, the failure
of distributional classes to coincide with basic categories came as a surprise
(given the presumption of fully general and productive rules) and engen-
idiosyncratic vis--vis other units. In this case, the most parsimonious de-
scription which accurately describes speakers linguistic knowledge is one
that explicitly includes these units despite their predictability. Ignoring the
possibility that certain familiar structures might coalesce as established
linguistic units despite conforming to general patterns constitutes what I
One motivation for the requirement, I suspect, is that non-predictability
gives the analyst a way of demonstrating that a construction exists and has
to be described as such. If a structure has properties not predictable from
other, independently established constructions, it must be a construction in
its own right. That is certainly valid. One must not however confuse the
to being a construction inside the grammar? Why should this structure
originally have no linguistic standing despite being used repeatedly, only to
is fully analyzable and does instantiate the schemas in question. Moreover,
it lacks any evident idiosyncrasies that would preclude its being con-
structed anew in accordance with the schemas. It is therefore not a con-
struction as defined in Construction Grammar. It is however a conventional
linguistic unit as defined in Cognitive Grammar. Being a fixed expression,
At one level, this point of difference is just a matter of how the terms
amples include
Our speaker needs no introduction
, as well as
that expression describing the official economic policy of the Bush admini-
With respect to compositionality, the orthodox view is that lexical items
have to be listed because they exhibit some measure of semantic non-
predictability, whereas novel expressions productively derived by syntactic
rules are fully compositional. It is true that lexical items have to be listed
that follows from the fact that they are fixed expressions, hence psycho-
logically entrenched as conventional units. I would however deny that lexi-
cal status correlates with semantic compositionality in this manner. On the
one hand, many fixed expressions come as close to full compositionality as
hot coffee
). On the other

actually irregularity

in defining lexicon
goes back to Bloomfield (1933: 274): The lexicon is really an appendix of
the grammar, a list of basic irregularities. This notion was carried over
into generative grammar, for positing this appendix provided a conven-
ient place to sequester anything that did not conform to the overall vision
In my view, the only sensible way to define lexicon

a way that is both
useful and approximates what is traditionally understood by the term

is to
not in any way diminish the interest and importance of typological investi-
{A/The} woman will take the rice out of {a/the} sack for {a/the}
The element selected as trigger can instantiate any semantic role, such as
agent, theme, location, or beneficiary, respectively exemplified in (9). The
choice is marked in two ways. First, the nominal so identified takes
the verb is inflected to indicate the triggers semantic role. Here I use the
labels AF, TF, LF, and BF for the inflection serving to focus the agent,
So what is this element called topic or trigger, which cannot be
identified as either a discourse topic or a subject in the classic, SAE-based
sense? In my view, it virtually begs to be called the trajector, i.e. primary
focal participant. It seems like the perfect example of a spotlight of focal
prominence that can be directed at different elements within a scene to
render them salient. This focusing, moreover, obtains at the clause level
and is thus distinct from a discourse topic, as well as being independent of
any particular semantic role. I conclude, then, that Tagalog does have sub-
jects, in accordance with the schematic conceptual characterization pro-
posed in Cognitive Grammar. Schachter, of course, is perfectly correct that
I am therefore suggesting that SAE-type subjects represent just one
manifestation of a more general linguistic phenomenon which has a unified
conceptual characterization at the schematic level. The unity becomes evi-
dent once we look beyond the associated, language-specific properties
(particular grammatical behaviors, correlation with discourse topic or se-
If valid, apprehending this schematic commonality enriches our under-
standing of structural variation rather than replacing it. Obviously it does
sensitive analysis of numerous languages typologically very different from
English and Standard Average European (SAE), arriving at optimal and
well justified descriptions from the Cognitive Grammar perspective. While
there has been some excellent work along these lines (e.g. Cook 1988;
Kumashiro 2000), it is far too limited for any conclusions to be drawn
about universality. I am not a typologist, and my own knowledge of cross-
linguistic variation is also quite limited, so at this point I can only specu-
late. For whatever it is worth, I would speculate that some basis can ulti-
mately be found in every language for ascribing trajector status (primary
focal prominence) to certain nominal elements, although in some languages
it plays very little role in grammar or has a very different role than in SAE.
I suspect that landmark status (secondary focal prominence) will turn out
not to be a motivated descriptive construct for all languages. Whatever
Bear in mind that my expectations are based on a particular view of
what constitutes the essence of subjecthood, even in English and SAE
schematic conceptual characterization in terms of focal prominence. Usual
statements to the effect that there are no grounds for positing subjects in a
certain language are based instead on semantic properties that are merely
prototypical (agent, discourse topic), and more fundamentally on a cluster
of characteristic grammatical behaviors (e.g. controller for verb agreement
and various kinds of anaphora). The situation looks entirely different if
If it is granted that impersonal
definitions of subject and object, applying to all instances, not just the pro-
totype. It stands to reason that primary focal prominence would tend to be
conferred on entities that are salient in other ways, hence the correlation
with such notions as agent and discourse topic. These, however, pertain to
Unlike agent and patient, definitions based on focal prominence are not
limited to canonical transitive clauses like (7a). They are equally applicable
to the subject and object of imperfective clauses like (7b) and the subjects
of special clause types motivated by various functional considerations,
same relationship:
Once more, then, Cognitive Grammar makes a radical claim not em-
braced by either Radical Construction Grammar or the less radical Con-
struction Grammar. But does this claim have any chance of being valid? To
assess this, we must first be clear about its nature. The terms subject and
object are much less clear-cut than noun and verb in regard to their cross-
linguistic application and even their applicability. I am not claiming that
my definitions are appropriate for everything which might have been re-
ferred to by these labels in any language, nor are they limited to things so
labeled. For obvious personal reasons, my point of departure is English, as
well as other languages where it is generally agreed that subjects and ob-
jects need to be posited, and for similar reasons. I am not however saying
that this classic conception of subject and object is universal, or that the
standard way of defining them is adequate. Instead, I argue for schematic
conceptual characterizations based on focal prominence, which are quite
different from traditional characterizations based primarily on syntactic
behavior. This grammatical behavior I take as being symptomatic of the
conceptual import of subject and object, not definitional. It is on the basis
object that I look for their manifestation in other languages and frame the
trajector of a profiled relationship; and an object is a nominal expression
specifying the landmark of a profiled relationship. More precisely, the
thing profiled by the nominal expression corresponds to the relational tra-
jector or landmark, as shown in Figure 7. The notions trajector and land-
mark are themselves characterized in terms of focal prominence, as the
achieved when the profile is shifted from the constitutive relationships to
In sum, I suggest that the conceptual characterizations proposed for
noun and verb ought to be taken seriously, for several reasons. First, they
offer explicit and intuitively appealing conceptual characterizations of no-
tions that have previously resisted semantic description. Second, they ap-
peal to only obvious and independently established cognitive abilities.
Third, they are descriptively successful, providing the basis for revealing
analyses of related phenomena. Minimally, I hope to have shown that the
impossibility of schematic conceptual characterizations has not in fact been
established. Cognitive linguists in particular ought to be quite resistant to
any claim that such fundamental and universal categories as noun and verb
do not have conceptual definitions, and should do their best to come up
with plausible candidates. If my own candidates are not accepted, they at
least suggest the requisite level of abstractness and the kinds of mental
Having resolved the issue of nouns and verbs, I turn now to subject and
object. My treatment will have to be more limited, even though matters are
more complex. The same two basic questions can be posed: Are these
grammatical notions susceptible to schematic semantic characterization?
Considering first the semantic question, the Cognitive Grammar posi-
tion is simply stated: a subject is a nominal expression which specifies the
mary scanning, for instance, when we watch the flight of a ball and then
We can now define a process as a relationship that is manifested
through time and scanned sequentially. A verb profiles a process. This
schematic definition makes no reference to any particular cognitive domain
(other than time) or any specific conceptual content. It is thus abstract
enough to apply to all members of the category. Moreover, it relies on only
clearly evident cognitive abilities: the capacity to apprehend a relationship,
At least as it applies to English, I have shown the descriptive elegance
of this characterization in various published but widely ignored analyses
(e.g. Langacker 1982, 1987a, 1991). Here I will merely observe that it af-
fords a simple, natural, and explicit way to describe the semantic contrast
I can offer only a few comments on how the proffered characterization
applies to specific nouns. It can first be noted that large numbers of nouns
clearly do designate groups of entities. A few examples are given in (4).
anew with an open mind, with full appreciation of the richness of our
mental capacities. My own proposals for schematic definitions (Langacker
1987a, 1987b) have to be accepted in the proper spirit. First, I would not
claim to have actually proved that these particular proposals are correct
(though I personally believe they are valid at least as first approximations).
They should rather be thought of as plausible candidates which minimally
and afford some idea of what they might look like. Second, the proposals
are psychologically plausible in the sense that they rely on mental phenom-
ena that are either well known or readily demonstrated. These phenomena
are real, and there is no reason not to expect them to have a basic role in
linguistic meaning

that we have the mental capacity to reify events and construe them as
abstract things, or (perhaps equivalently) that we are capable of ontological
[Langackers] conceptual analysis of parts of speech ... is broadly compati-
ble with the universal-typological theory ... [T]here is no inherent conflict
bold to indicate the relative degree of autonomy thereby achieved. Recur-
rence in multiple configurations reinforces its status as an independently
Moreover, I will argue that occurrence in some array of structural
experience of apprehending the content as an instance of the category de-
fined by the profiling. Indeed, the factoring in Figure 13 directly reflects
the notion that the category schema is immanent in its instantiations. The
categorizations shown in Figure 6 could thus be redrawn in the manner of
Figure 13, with the class schema depicted as one of the factors in the ex-
Lexical Stem
Lexical Stem
Lexical Stem
This kind of development is bound to affect substantial numbers of fre-
quently occurring lexical stems. Even if a language follows the basic strat-
egy of leaving lexical items unspecified for profiling and hence for gram-
matical category, many lexical items will come to be conventionally
employed in structural frames which impose on them categorization as a
would not have many hundreds, and probably thousands, of conventional-
We can then pose the following question: does such a language have lexical
nouns and verbs? But that only raises another, more fundamental question:
what precisely does it mean to say that a language has lexical nouns and verbs?
From my standpoint, this implies the existence of established lexical units
which incorporate the requisite profiling (of a thing or a process, respectively).
It follows, then, that the language does have lexical nouns and verbs. Along-
side the uncategorized lexical stem we started with in Figure 9, the language
has an augmented lexical stem, a conventional unit consisting of the basic stem
In fact, the same lexical stem, through use in different constructions,
might give rise to augmented stems representing both the noun and verb
categories, as shown in Figure 12(a)-(b). In this case, the basic (uncatego-
rized) lexical stem can be regarded as a schema abstracting away from their
difference in profiling, as in diagram (c). We might posit an analogous
schema for an English lexeme such as
noun and verb uses. The schematic sense of
would neutralize the
noun/verb distinction, invoking their shared conceptual content without
T as S
the equivalent diagram (b), which shows the result of categorization. It
represents more directly the effect of apprehending the conceptual content
of the lexical stem in the manner specified by the constructional schema in
I have been assuming the situation where the uncategorized lexical stem
occurs in this construction for the very first time, or at least occurs infre-
quently enough that the configuration in Figure 9 can be regarded as novel.
Clearly, though, if it is useful to employ the stem in this structural frame on
one occasion it might prove useful to do so on many other occasions as well.
Suppose this does happen with some frequency. Structural assemblies that
occur with any frequency are likely to coalesce as familiar, prepackaged
units. They become psychologically entrenched for individual speakers, and
conventional within a speech community if this happens for enough indi-
viduals. In this fashion, the entire assembly in Figure 9

trenched, conventional element of the language). This is shown (using the
relationship is enclosed in a box with rounded corners, on the assumption
that this particular stem has not previously been employed in this manner
(i.e. the categorization is novel, not itself an established conventional unit).
In similar fashion, the entire assembly (where the stem is so categorized in
Lexical Stem
by constructional schema
are assuming the case of a stem which lacks inherent categorization. This
translates into the absence of inherent profiling. Its semantic pole consists
of a conceptualization which may have any degree of internal complexity;
it suffices for our purposes to show it as including a thing which partici-
pates in some relationship, giving it the potential to be construed as either a
noun or a verb. Here it is construed as a noun by virtue of being employed
the language in question. So such languages are not counterexamples to the
claim that we need to talk about nouns and verbs in a grammatical descrip-
tion and in the characterization of particular complex expressions. Instead,
what the argument bears on is the more specific claim that nouns and verbs
are universal lexical categories. The argument only makes sense, I believe,
3. Basicgrammaticalconstructs
The full reduction of grammar to symbolic assemblies depends crucially on
the semantic definability of certain fundamental and universal grammatical
constructs, notably noun, verb, subject, and object. In contrast to both Con-
struction Grammar and Radical Construction Grammar, Cognitive Gram-
mar claims that these notions are indeed susceptible to semantic characteri-

not just at the prototype level (which is widely accepted), but at
the schema level, covering all instances. At the prototype level, each can be
its profile is inherited at the composite structure level. The other three
component symbolic structures are all noun phrases, serving to specify
central participants of the verbal process. In this construction they are dis-
tinguished by temporal order. Preceding the verb is the subject nominal,
whose profile corresponds to the verbs trajector. Directly following the
verb is the nominal whose profile corresponds to the landmark (the recipi-
ent), which I would identify as grammatical object. The third nominal,
Semantic Pole
Phonological Pole
The only substantive elements posited are semantic structures and
phonological structures, which are linked to form symbolic structures.
Symbolic structures are in turn connected by correspondences and catego-
rizing relationships to form assemblies of symbolic structures. There is no
separate level of grammatical form, nor are there any irreducible, specifi-
The subject and object relations are also definable in terms of symbolic
We can now examine a representative constructional schema. Figure 8
represents the English ditransitive construction in approximately the same
degree of specificity as the schemas in Figure 2. For sake of simplicity, I
have not indicated constituency, nor have I shown the composite structure
at any level of organization. I have also omitted the time arrow from the
representation of the verb. The schematic verb is one that profiles a canoni-
cal act of transfer (e.g.
). The trajector exerts some kind of force
grounded, i.e. a grammaticized specification is made of its epistemic status
Symbolic assemblies are connected in what Goldberg calls an inheri-
symbolic relationship. For diagrammatic ease, I have not shown the com-
posite structures separately, as integrated structures (they are given below
in Figure 6). Semantically,
are integrated by virtue of a corre-
Semantic Pole
Phonological Pole
Shown in Figure 5 are schemas for a few grammatical categories (to be
discussed more fully below). A noun profiles a thing (abstractly defined),
represented as a circle. Observe that the noun schema is bipolar

an expression, with both a form and meaning, hence a symbolic structure.
particular phonological content, so the schemas phonological pole is
maximally schematic [...]
A verb profiles a process, i.e. a relationship
scanned sequentially in its evolution through time (t). I have shown the
profiled relationship as an arrow with ellipses (for schematicity); a circle
indicates its primary participant, and a box represents whatever other enti-
ties may be involved. The bar along the time arrow indicates sequential
scanning. Certain higher-level categories can also be characterized seman-
tically. In particular, a noun phrase (or nominal) profiles a thing which is
the extent of their conventionalization in a speech community. These last
Some lexical items are represented in Figure 4.
are mor-
phemes. At the semantic pole I use crude diagrams as abbreviations for
what, in a serious description, would have to be highly elaborate, multifac-
these grammatical constructs tend to correlate with certain meanings or
functions, there is no definite claim that such constructs are fully definable
conceptually, nor even any inclination to consider this a possibility worth
exploring. To the extent that these constructs are not reduced to anything
more fundamental, they represent a vestige in these frameworks of strong
Why is this reduction

if achievable

Grammar and Radical Construction grammar, the form part of a form-
meaning pairing does include grammatical form. Thus Goldberg (1995: 51)
Grammatical Form
Symbolic Structure
is Newmeyer 1983). Conversely, functionalists would be more effective in
arguing against strong autonomy if they distinguished it more clearly from
the weak version and were more explicit about what a viable alternative
Weak autonomy is generally accepted in Functional and Cognitive Lin-
guistics. While strong autonomy is usually rejected, there is certainly no
unanimity concerning an alternative, or even the need to explicitly formu-
late one. Cognitive Grammar provides what I call a symbolic alternative to
strong autonomy (Langacker 1995). Some of its central ideas are as fol-
lows. Lexicon, morphology, and syntax form a continuum, divided only
more, Kay, and OConnor 1988; Michaelis and Lambrecht 1996), but
emerges by default as a comprehensive
and readily accessible statement. Any criticism I might offer of the works
by Croft and Goldberg should not obscure my great respect for these schol-
I will not compare and contrast these three kinds of construction gram-
less so
Ronald W. Langacker
A comparison of three formulations of “Construction Grammar”
(2001), Goldberg (1995), and myself (Langacker 1987a, 1990, 1991,
broadest issue concerns the putative autonomy of grammar. Central here is
the status of certain basic constructs for grammatical description, such as
noun, verb, subject, and object. These pose two fundamental questions.
First, are they universal? And second, are they semantically definable?
2002Grounding and the system of epistemic expressions in Dutch. In
Over de (beperkte) combineerbaarheid van deontische, epistemische
1985Frames and the semantics of understanding.
Quaderni di Semantica
1998Emergent grammar. In
The New Psychology of Language,
Consciousness and the Computational Mind
. Cambridge, MA: MIT-
1996Intra-speaker relativity. In
ing to, or throwing new perspectives on what CL contributes to this subject
matter. At the same time, the discussion has brought up another issue that
threatens to divide CL and FL, viz. the matter of the construction concept
versus process concept of a grammar. In this regards, I have argued that
In sum, one can only hope that the future will see an increasing ten-
dency for functional and cognitive linguists to communicate across their
borders, and to share and integrate their complementary perspectives, such
that we will ultimately arrive at a coordinated and coherent joint attempt to
tackle the complexities of language as a cognitive system for communica-
tion. The cognitive-functional approach which was implicitly present in the
discussions throughout this chapter, at least, is an attempt to contribute to
1.This research has been sponsored by the Fund for Scientific Research
Flanders (VNC-project G.0470.03) and by the Research Council of the Uni-
versity of Antwerp (Geconcerteerde Onderzoeksactie 2003/4). The author is
a member of the Center for Grammar, Cognition and Typology at the Univer-
2.CL is functionally oriented, too, of course (see below), but, for the sake of the
argument, in this chapter I will use the label FL to refer to functionalist ap-
3.In a way, nearly the entire European branch of CL is an illustration of this
category: most European cognitive linguists have their roots in FL, and have
as a reaction against the generative tradition, i.e. to a large extent it arose out
of the ashes of the generative semantics movement, which in turn was a re-
action to the orthodox transformational generative syntax tradition. FL how-
ever, essentially continues an old (mainly European) tradition which existed
American Bloomfieldian counterpart, predominantly functionalist in orien-
5.Obviously, due to these different accents the two strands also show, at least in
part, a tendency to zoom in on different phenomena for investigation, and to
to selecting and if necessary composing a constructional pattern which he
can utter as an adequate communicative act. And it is hard to see what else
this could involve than in general terms to start from a conceptual
meaning to be expressed and while taking into consideration the contextual
factors select the adequate lexical and grammatical structures (in language
production), and vice versa (in language understanding). In other words, it
seems quite unavoidable to assume that in language processing conceptual
meaning and linguistic form are applicable at different moments in time,
strual of the representational meaning in context). (The discussion in sec-
tion 3 obviously boils down to the claim that functionalist models should
even go further than they do at present in these regards.) For these reasons,
needed, and this is certainly another area in which functionalists must ac-
symbolic units, that is, pairings of form patterns (of different complexity,
up to the level of the full clause) and their meanings. This contrast was
actually already forecast in Langackers (1987) rejection of a process con-
kind in (2). But the way to deal with this is to refer to the procedural links
an essentially one-dimensional articulatory pattern. To put it squarely, in
linguistic structure layering is everywhere and nowhere, and as such it
would seem an excellent example of an emergent domain of grammar,
more or less in Hoppers (1998) sense. But, of course, precisely that fact
If the tense marker does syntactically affect an epistemic expression, how-
ever, different things can happen. In the case of a predicative adjective, the
past tense unavoidably turns it descriptive, i.e. it situates the epistemic
evaluation somewhere in the past cf. (17). The tense form thus remains a
(17)It seemed likely that the butler had killed her.
With an epistemic mental state predicate the situation is different, however.
b.I used to think that ...
c.I presently tend to think that ...
d.Ich glaubte, der Butler htte den Mord begangen.
In Dutch (18a), the past tense can do two things. Either it makes the predi-
cate descriptive, triggering the meaning in (18b). I.e., like in (17) it situates
the epistemic evaluation in the past. Alternatively, the past tense can
weaken the epistemic evaluation, as compared to the present tense form,
i.e. the meaning in (18c). In this case, the conceptual status of the epistemic
form its performativity remains unaltered, but the tense form no longer
expresses time, but contributes to the epistemic meaning. However, this
latter reading is less obvious for the English variant of (18a) (cf. the trans-
lation). And it is simply impossible in the German counterpart in (18d),
In the modals, however, English even strongly prefers the weakening
form, or they can be expressed separately, by means of a negative (or, more
rarely, a positive) marker for the polar and an epistemic marker for the
b.I think the butler did not kill her.
But the adverbial and grammatical forms in them do not allow an inte-
grated expression, since there are no negative modal adverbs or epistemic
auxiliaries. There is, for example, no such thing as
. So only
(13)a.The butler probably did not kill her.
b.The butler may well not have killed her.
Some Dutch and German modals are moreover more restrictive regarding
their combinability even with a separate negative marker than the English
modals, or than the adverbs in these languages, as shown in (14).
, do allow an epistemic reading when com-
normally do not: they then only allow a dynamic situ-
ational impossibility reading (cf. Nuyts 2005) cf. (14b). But there is an
exception to the latter, viz. when they are used in a complementing pattern,
and with the negation in the subordinate clause cf. (14c). (The reasons for
case: their combinability requires recourse to concepts and notions per-
taining to the status of the qualificational hierarchy as a whole which can
it in no way affects the speakers commitment to it. In other words: while
evidentiality, epistemic modality and deontic modality, each in its own
way, concern the questionability of the state of affairs, this is not true any-
more for time marking and other qualificational categories lower in the
It is not surprising, then, that precisely these dimensions, but not the
are strongly subject to culturally defined conventionalized generalization,
among others in terms of the clock and calendars. Nevertheless, there is
a fair amount of flexibility in how a speaking subject can handle the situa-
If we jump to epistemic modality, then, making an assessment in these
terms does not involve any perception of the state of affairs anymore. For,
obviously, if there is direct perception, there is no need for an epistemic
judgment. An epistemic evaluation is purely a matter of relating and com-
paring other bits and pieces of information about the world including
is not linguistic, but conceptual in nature, we possibly canfind an answer,
First note that, although the hierarchy is developed on the basis of lin-
The combination in (7a), featuring the auxiliary-like inferential predicate
and the epistemic adjective
the pattern also occurs with other epistemic and also with deontic adjec-
tives), occurs frequently in the Dutch data and also feels very normal, not
only in Dutch but also in English. Neither of the expressions in this combi-
nation can be called descriptive. Even in minimal variants of this pattern,
however, such as those in (7b) with an alternative auxiliary-like evidential
predicate, and in (7c), with an equivalent evidential adverb replacing the
auxiliary-like predicate, thesituation changes: the Dutch variants at least
do not occur in the data at all, and to the extent that one can imagine a
context in which these alternatives are usable at all, the evidential predicate
is special, then, in that a speaker using it is, overall, making an epistemic
evaluation of the state of affairs, but not an evidential one. It looks like the
evidential predicate is not evidential in (7a), but actually serves another
function, viz. as a subjectivizer of the epistemic expression. This may
sound ad hoc but it is not if viewed in the context of the functional profile
of the epistemic expressions (cf. Nuyts 2001). In terms of the functional
dimensions at work in the epistemic paradigm (see section 3.3), the adjec-
tival pattern is specialized in focalized expressions of the epistemic qualifi-
cation. But in its default pattern of the type
it is probable that
it also carries
intersubjectivity, due to its syntax. Hence the pattern in (7a) offers a sub-
jective alternative which can nevertheless still be focalized. In further sup-
port of this assumption, note that this subjective variant readily takes the
to me
, which in fact also appears in 75% of the
Dutch corpus cases. The default intersubjective pattern however hardly
allows this subjectifier, and there are in fact no instances of it in the Dutch
One final observation, beyond the corpus data. If one tries to construe
additional combinations of the qualifications at stake, it turns out that many
of them simply sound awkward though clearly not for any syntactic rea-
In those combinations that do occur, 2 types of cases turn up. In the
most common type, only one of the expressions is used performatively,
while the other one is descriptive. (6) offers some examples which mirror
cases in the Dutch data. For instance, in (6a), the epistemic form
performative, it does involve speaker commitment. But the deontic form
is descriptive: the speaker is not obliging the hearer to pay a fine,
he is rather speculating about an obligation which someone else may im-
b.The three of them
e.Apparently they maynot do that anymore.[evidential + deontic]In the other type of case in the data, which only occurs with a few specific
lemmas, but then frequently with those, one of the two expressions does
not have its normal deontic, epistemic or evidential meaning, but a derived
meaning or use, it serving as a modifier of the other expression, or as a
discourse marker, or as a speech act modifier (e.g., a politeness marker).
involves long term non-linguistic representations of world knowledge (see
Nuyts 2002 for discussion). Unfortunately, I cannot explore these matters
These representations do capture the intuition that these predicates all
invoke the entire commercial exchange, i.e. both the transfer of goods and
the countertransfer of money, with the participants in alternative roles, even
if only part of this cluster is actually made explicit in the expression. The
cluster is represented differently in the CR of each predicate, however,
depending on what is realized syntactically. For example, in
the trans-
fer of goods is the syntactic core, so its conceptual structure codes the
transfer as the main conceptual clause, and the countertransfer of money
is embedded in a modifying conceptual clause marked by a subordinating
function called exchange. But in
the countertransfer is syntactically
central, so in its conceptual structure the two events switch positions. So,
although Jackendoffs representations clearly come close to revealing what
is conceptually shared by the alternative expressions, his conceptual se-
mantics still does not offer a level to directly represent what they share, as

i TO



i TO







transaction, and C is (usually) money). As Fillmore (1977: 105) points out,
all these alternatives essentially go back to one and the same scene (infor-
mally) rendered in (4), and they differ in terms of how they perspectivize
Other examples of semantic paradigms pertaining to the object world
These different factors are then also responsible for many of the gram-
matical properties of these alternative expression forms see Nuyts (2001)
for an analysis. It is an educated guess that the situation is comparable for
the alternative expressions of the qualifications lower in the system, al-
though the functional factors at stake will certainly differ in certain ways,
How do these observations offer support for the assumption that the
qualificational categories are not (only) linguistic, but conceptual? Here is
the (very simple) argument. In terms of a process perspective (specifically,
a production perspective), since the alternatives in a semantic paradigm all
involve the same core qualificational notion, they must unavoidably all
have their cognitive origins in one and the same semantic core level at
which the qualification is conceived by the speaker. Claiming otherwise
would mean giving up meaning as a coherent cognitive category. When
expressing a qualification, then, a speaker works his way from this level to
a linguistic form via procedures sensitive to the relevant functional factors.
Now, the major expression types in (3) are lexically basic, hence basic in
terms of linguistic representation and processing. Hence they cannot be
productively derived from each other in the course of language production.
For if one were to take one of them as the core form, one would have to
invoke radical lexical exchanges and syntactic transformations to arrive at
one of the alternative forms. No one in present day linguistics and language
:the degree of moral acceptability or desirability of a state
A more formalized version of such a system has been introduced in Van
Valins (1993) Role and Reference Grammar and Diks (1997) Functional
c.Deontic modality and time:
d.Time and quantificational aspect:
e.Quantificational and qualificational aspect:
question if one does not explicitly take recourse to much more abstract
notions. Often functionalists do notice this problem, of course, but then,
rather than looking in the right place for a solution, viz. in general princi-
ples of human conceptual knowledge and processing, they tend to stuff all
kinds of concepts and notions into the types of linguistic structures they do
accept in their models which clearly do not belong there. On the other
hand, if functionalists would take the cognitive import of their analyses
serious, they would hold the key to some important new insights into the
nature of human conceptualization, thereby adding to what is done in this
regards in CL. This would bring us a bit closer to more intensive interac-
mantic issue with a strongly cognitive touch of how humans under-
stand or conceptualize the world i.e., how they make and represent
meaning and how the linguistic surface relates to this i.e., how it ren-
ders conceptual categories. In FL, however, a few important exceptions
(e.g., Wierzbicka 1992) aside, this specific focus is largely absent. Func-
tionalists predominantly concentrate on the more pragmatic issue of how
language is structured in view of how it externally functions in communi-
cation: what types of structures exist in languages, what is universal and
variable in them, and how does all of this relate to the dynamics of com-
munication, including the speakers intentions, the hearers knowledge, the
Probably the result is a position which is not considered functional linguis-
tic by hardcore functional linguists, but not cognitive linguistic by hardcore
in recent years the cognitive awareness among them is rapidly increasing.
And there are of course also quite a few functionalists who have explicitly
adopted a cognitive perspective since a very long time, including, very
have to do with the dominant research foci of the two strands, as coded in
The hallmark of CL is, of course, its concern with the se-
among the most complex and maybe even the most strained ones.
In many
A Clearing in the Forest: Law, Life, and Mind
. Chicago: Univer-
1976The sec
ond time around: cognitive and social strategies in lan-
1998The middle construction in English: a c
ognitive linguistic analy-
Linda Waugh (eds.), 205219. Amsterdam/Philadelphia: John
2000First steps toward a usage-based the
ory of language acquisition.
The New Psychology of Language: Cognitive and Functional
Approaches to Language Structure
2003Collostructions: investigating the interaction of w
ords and con-
International Journal of Corpus Linguistics
8 (2):
Applied Cognitive Linguistics II: Language Pedagogy
tive Linguistics Research 19.2). Berlin/New York: Mouton de
onymic sense shift: its origins in hearers abductive construal
of usage in context. In
Cognitive Approaches to Lexical Seman-
, Hubert Cuyckens, Ren Dirven, and John Taylor (eds.),
2000How met
The semantics of an impeachment. In
Language and Ideology.
Volume II: Descriptive Cognitive Approaches
, Ren Dirven,
Roslyn Frank, and Cornelia Ilie (eds.), 77106. Amster
Women, Fire, and Dangerous Things: What Categories Reveal
1993A contem
1999Grammatical constructions and li
nguistic generalizations: the
2000Intended t
1995The effects of meta
phorical and literal comprehension processes
on lexical decision latency of sentence components.
Journal of
1971The meta
1975Logic and c
Syntax and Semantcs. Volume 3:
Speech Acts
1995The c
ognitive psychological reality of image schemas and their
On Our Mind: Salience, Context and Figurative Language
1998Understanding meta
Current Directions in Psychological
guageand Ideology. Volume II: Descriptive Cognitive Ap
Ren Dirven, Roslyn Frank, and Cornelia Ilie (eds.),
Ren Dirven
1990Dating a fi
gure of thought.
7.Pustejovski (1995) offers an attractive attempt to reduce the so-called polysemy
of words and opts for a monosemist view in his qualia theory. While this may
work well for artifacts such as
, it cannot solve the polysemy
8.This possibility was pointed out to me by Beate Hampe (Jena).
work in typology, historical semantics, grammaticalization studies (Heine,
Traugott), cross-cultural semantics (Wierzbicka, Goddard), and now even
cognitive therapy (van Hoek). It remains to be seen how robustly these
sycholinguistics, e.g. by Glucksberg, Brown, and McGlone
other researchers working in the field of Cognitive Linguistics, neither
from those outside it. In comparison with this relative stability, it is striking
Ren Dirven
even if they only manage to realize one or two segments of them. (3)
Whereas according to Langackers view, adult speakers can either apply a
rule schema or reproduce a single instantiation of it, e.g. a plural noun,
for two English-speaking children. The reasons for the selection of
these particular prepositions were their frequency, monosyllabic shape,
Ren Dirven
throughout sensorimotor activity and from our perceptual understanding of
actions and events in the world. Thus image-schemas and their transforma-
tions provide part of the foundation for thought, reasoning, and imagina-
tion. Since there had been little opportunity for research in the relatively
short time since the rise of CL, Gibbs and Colston (1995) looked for em-
pirical evidence from earlier work in psycholinguistics, cognitive psychol-
ogy, and developmental psychology that is consistent with the idea that
image-schemas and their transformations play important roles in human
cognition. This experimental research was not originally conducted and has
therefore not generally been considered in terms of the cognitive linguistic
ideas on image-schemas. However, according to the authors, a large body
ceremonies involving the bear ancestor and other guardian animals were or
are carried out on a regular basis. As the use of the past tense in various
sentences above suggests, these Basque CCMs are under threat of erosion.
Frank (2003) studies shifting identities in Basque cultural models of self,
whereby the influence of the general European model on a specific aspect
of the Basque model is investigated. The socio-cultural identity, especially
that of the younger Basque generation, appears to be increasingly affected
False statements such as those of Clinton fall at the most marginal end of
the least culpable genre of this continuum of offences (Morgan 200
provided with answers to essential questions such as Where is this speaker
from? and What is this speaker like? This basic, cognitive need for so-
cial categorization thus in turn effects a need for dialectal variants, which
As Lakoffs answer confirms, he identifies a cognitive sociolinguistics
with ideology research. Moreover this ideology research tends to be associ-
ated with negative connotations, which is not astonishing given that ideol-
ogy is often closely intertwined with power relations. According to
Grdenfors (1998) the study of power relations is an intrinsic part of Cog-
nitive Linguistics. This trend is also dominant in most of the cognitive ap-
proaches to ideology in the twin volumes
Language and Ideology
giving expressions. The present investigation picks up where others have
left off and may offer insights for a renewed lexis-oriented sociolinguistics
ociolexicology, and equally, as pointed out before, for a
more corpus-based orientation of Cognitive Linguistics. More specifically,
Ren Dirven
(b) and (c) lies in the degree of subjectivity. With
thats why
there is a
Ren Dirven
theory can be applied complementarily to explore how mind styles are lin-
Also non-literary discourse study has received vital impulses from CL,
namely from a group of European scholars who concentrate on systematic
Cognitive discourse study established itself as a full-grown subdiscipline
within Cognitive Linguistics by the publication of the collective volume by
van Hoek, Kibrik, and Noordman (1999). A fast growing number of schol-
ars, e.g. Giora (2003), are currently doing research in this area from per-
ceptual viewpoints. Others start from coherence relations and the role of
consciousness in text representation. These are the central themes of the
research group organized around Noordman with Maes, Sanders, Schil-
peroord, Spooren, and several others. We select two types of research from
Noordman and de Blijzer (2000) examine the differences in reading
time and hence the difficulty of understanding for the different types of
peper (2002), Shen (1995, 1997, 2002), Steen (2002), Stockwell (2000
other cognitive processes such as analogical mappings, discourse manage-
activation principles, space tracking by viewpoint and focus shift, match-
ing, and structure projection. Given their rapid occurrence, these processes
are not consciously noticeable or linguistically encoded. Rather, language
provides underspecified contextual clues that prompt cognitive configura-
is part of the base space and the verb
is a space-
builder opening a new space (space 1) of an imagined world in which the
I was Marilyn Monroe
) is no longer identical with the first
) in the base space; rather it is part of a new knowledge frame in
which Marilyn Monroe is not kissing herself, but the speaker, i.e. the
the base space. Fauconnier (1997)
may indicate a possible
Affordances are nothing more than possibilities for action and use offered
by the local environment to a particular type of embodied agent, equipped
with specific bodily features. In this way perception is always contextual-
ized and constructed : the world is essentially perceived by some given or-
ganism endowed with its own intentions in some given context, and is seen
as affording opportunities for goal directed actions. Perception is therefore
always connected to action, and both perception and action are always con-
acts. In their view, indirect speech acts of the types that Searle (1979) calls
directive, commissive and expressive such as promises, offers, requests,
their environments. Instead, Zlatev proposes a situated embodiment and
presents an approach which he calls situated embodied semantics in
which meaning emerges from a pairing of linguistic expressions with
situations. He uses connectionist modeling to test the feasibility of the ap-
proach and for gaining insights into such issues as learning categories
without necessary and sufficient conditions for membership, the context
dependence of meaning and the ability to utter and comprehend novel ex-
Steels, Ikegami, Doering, and Emmeche are now preparing a volume
adopts a phenomenological approach as its philosophical basis (Lakoff and
Johnson 1980: 181, 1999). Geeraerts (1985: 355) characterizes this phe-
nomenological approach as follows: All individuals have an intentional
relationship to the world and their access to the world or their conscious-
ness is realized by their bodily experiences of that world. Everyday dis-
course, as well as highly abstract discourse is largely bodily- and spatially-
Ren Dirven
actions, purposes, means, difficulties. All of these are conceptualized in
be in doubt
as participating parts of the situation rather than either as representations or
sons life.
But things are far more complicated in the case of words de-
noting natural entities such as
. In its prototypical use,
refers to
Ren Dirven
ing booksellers exchange:
Where shall we put the new travel book?
Well, the corner shop window sells very well.
Obviously, we can observe
prototypicality effects in this construction too, demonstrating that we wit-
Having been in close direct contact with Chomsky and the generative para-
sophical basis of the traditional Aristotelian conception of categorization
underlying most human sciences, including linguistics
(Lakoff 1972)
known as phenomenology and based on human
and perception in the conception of the world, had won part
tive Grammar considers constructions to be reducible to symbolic rela-
tionships, Construction Grammar assumes that grammatical classes and
other constructs are still thought of as a separate level of organization.
This separate level of linguistic organization has become the focus of re-
search by Lakoff (1977, 1987: 467, 538), Goldberg (1992a, 1992b,1995,
perspective involves the choice of a direction of the mental scanning as in
specified (Talmy 1988a: 165). The lexicon contains content words and
reflects the tens of thousands of individual phenomena as single, concep-
tual categories, whereas the grammar develops more abstract, schematic
categories. Thus the schematic meaning of the plural morpheme, that is, a
meaning applying to all possible contexts, is the notion of multiplexity.
This is found not only with count nouns (
), but also with abstract
nouns (
fears, misgivings
), uncountable nouns (
ashes, waters
nouns (
a speaker makes use of principles similar to those of gestalt perception.
Wanting to describe a conceived situation, (s)he makes choices as to the
scope of elements from a situation and to the perspective adopted on the
situation. The speaker assembles the things and relations selected into
higher relationships and ultimately into sentences and texts. Such compos-
ite structures may become conventionalized as grammatical gestalts, also
known as patterns or constructions, as also pointed out by Lakoff (1977) in
his paper Linguistic Gestalts. Such constructions are systematically in-
vestigated by Goldberg (199
) and several others in
mar. Here it is assumed that constructions such as the caused motion con-
He blew the paper off the table
) cannot be seen as resulting from
compositional assembly, but constitute a separate level of organization and
exist as gestalt-like patterns or established configurations which have
meaning relations independent of the lexical components. As Michaelis
(2003) puts it, there are thus three types of meaning: word meaning, (com-
positional) sentence meaning, and constructional meaning, which is the
Ren Dirven
Major strands in Cognitive Linguistics
In contrast to most previous linguistic paradigms, which saw meaning ei-
ther as less relevant or else as an autonomous linguistic module, CL ap-
proaches language as an integrated part of human cognition which operates
in interaction with and on the basis of the same principles as other cogni-
tive faculties. CL is therefore defined as a linguistic theory which analyzes
language in its relation to other cognitive domains and faculties such as
bodily and mental experiences, image-schemas, perception, attention,
viewing frames, categorization, abstract thought, emotion, rea
namic evolution and its co-evolution with neighboring disciplines. This is
ing, but the whole system of form-meaning conditions of use. This pro-
posal is in full consonance with state-of-the art knowledge about coordi-
In the section on psycholinguistics and cognitive processing, Gibbs ar-
gues in favor of the embodiment of cognition, and consequently of mean-
ing, since language is regarded in CL as an essential part of cognition. The
chapter emphasizes the importance of whole-body action in the genesis and
development of perception, cognition, and language use, and suggests that
human thought and language, most generally, must be studied and under-
is seen in terms of simultaneous constraint satisfaction, among other char-
acteristics. Goldbergs Construction Grammar, Crofts Radical Construc-
tion Grammar, and Langackers Cognitive Grammar are constructional in
this sense. The discussion is then focused on three crucial issues: the ques-
tion of the putative autonomy of syntax, of which there is a strong and a
weak version (the latter usually subscribed to in cognitive and functional
the complexity of the comparison is partly due to the fact that both orienta-
As is evident from this brief overview, all sections cover two general
topics with wide-ranging implications which are crucial to future develop-
ments of research in CL and in linguistics in general: (i) the relationship
The Cognitive Linguistics (CL) agenda has always had a clear interdisci-
plinary concern. However, until very recently cognitive linguists have
mostly addressed interdisciplinary issues in terms of the connections be-
tween CL and other branches of cognitive science, especially artificial in-
telligence and the brain sciences, as evidenced by recent work in Embodied
Construction Grammar (Bergen and Chang 2002; Chang, Narayanan and
Articulating the dynamic nature of conceptual and grammatical structure
leads us inexorably to the dynamics of discourse and social interaction.
Table ofcontents
Basic Discourse Acts: towards a psychological theory of
discourse segmentation
Gerard Steen
The multilevel operation of metonymy in grammar and
with particular attention to metonymic chains
The role of conceptual metonymy in meaning construction
Tracking the fate of the metaphor
in British
environmental discourse
Brigitte Nerlich
Subject index
Table of contents
Introduction: as strong as its foundations, as wide as its scope
nitive an
Functional Linguistics
Construction Grammars: cognitive, radical, and less so
Ronald W. Langacker
Lectal variation and empirical data in Cognitive Linguistics
Dirk Geeraerts
Social cognition: variation, language, and culture in a
cognitive linguistic typology
Embodied action in thought and language
Raymond W. Gibbs Jr.
nitive o
Francisco J. Ruiz de Mendoza and M. Sandra Peña
To Víctor for his patience, cheerful encouragement,and love

Приложенные файлы

  • pdf 14772776
    Размер файла: 2 MB Загрузок: 0

Добавить комментарий