12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 1 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

Subject

Key-Topics

DOI:

29. Pragmatics and Cognitive Linguistics

GILLES FAUCONNIER

Theoretical Linguistics

»

Pragmatics

cognitive

10.1111/b.9780631225485.2005.00031.x

As we start a new century, rich and diverse evidence from the social and cognitive sciences continues to
provide an ever sharper conception of the way we humans think. It is now widely understood that brains and
cultures play a major role in constructing the world as we see it. They are not, and we are not, passive
interpreters of that world. Distinctive and powerful cognitive capacities such as analogy, framing, metaphor,
schematization, recursion, reference-point organization, and conceptual blending lie at the heart of
perception, conception, and action, and suffuse our lives with rich meaning - sometimes unbearably rich
meaning. For the most part, the operations and processes that correspond to those powerful capacities are
carried out massively below the level of our consciousness.

Language, which we

do experience consciously through sound, gesture, or writing, is a spectacular but far

from transparent window into this backstage cognition of our minds. Conversely, many of the seemingly
complicated, irregular, and “illogical” properties of language structures become intelligible when their
cognitive underpinnings are discovered.

In everyday life, when we use or hear words in a language we know, in a culture we know, meanings form
instantly. Consciously, we experience words, word combinations, and the automatic sensation of knowing
what they mean. Quite naturally, we feel that the meanings are in the words and in their combinations.

Twentieth-century attempts to study meaning most often reflect this intuition by assigning to every form its
“meaning” independently of context. But important work in pragmatics over several decades of the same
century has shown the assumption of invariant meaning to be simplistic. Context matters. A popular scheme
to emerge from work in pragmatics preserved the intuition that form had meaning

ex tempore, but added to

the arsenal of meaning theorists a very powerful pragmatic component that could use context and pragmatic
principles to derive from the pristine meaning of a form in isolation the messier meaning of that form when
mired in context.

In modern linguistic and philosophical research, the best-known efforts in this direction are the Gricean
scheme of maxims and implicatures and the theory of speech acts developed by Austin and Searle.
Interestingly, Grice's work was originally motivated by logical considerations - keeping the logic of semantic
meanings simple and standard by accounting pragmatically for apparent complexities. Linguists in the late
1960s, and through the 1970s and beyond, noted many features of grammar that were associated with
pragmatic properties, and that in turn led linguists to study pragmatics in its own right and for its own sake.
There was a convergence of issues of common interest to philosophers, linguists, and others who ponder
the mysteries of meaning. Pragmatics exploded into a rich, but not particularly well-defined, field of inquiry.

However, the very linguistic work that had contributed to an explosion of pragmatics was about to lead to its
reassessment. It became clear to an evergrowing number of researchers (e.g. Ducrot 1972, Fillmore 1982,
Jackendoff 1983, de Cornulier 1985, Fauconnier 1985, G. Lakoff 1987, Langacker 1987, Sweetser 1990) that
the line in the sand between semantics and pragmatics was not an easy one to draw. The “meanings”

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 2 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

the line in the sand between semantics and pragmatics was not an easy one to draw. The “meanings”
spontaneously associated by informants with isolated forms were not the pristine core meanings that we
yearned for; they were only defaults made extremely probable by culture, norms, and lack of imagination.

Just as a rigorous description of syntax in the 1950s, using the resources of computability theory, revealed
its immense complexity, a rigorous description of meaning using the resources of cognitive science reveals
far greater complexity than commonly assumed. Language is only the tip of a spectacular cognitive iceberg,
and when we engage in any language activity, we draw unconsciously on vast cognitive and cultural
resources, call up innumerable models and frames, set up multiple connections, coordinate large arrays of
information, and engage in creative mappings, transfers, and elaborations.

Thus we find that language does not “represent” meaning: language prompts for the construction of
meaning in particular contexts with particular cultural models and cognitive resources. To do so, it draws
heavily on “backstage cognition” that is not accessible to our consciousness. When scientific analysis brings
out backstage cognition explicitly, it highlights the contrast between the extreme brevity of the linguistic
form and the spectacular wealth of the corresponding meaning construction. Very sparse grammar guides us
along the same rich mental paths, by prompting us to perform complex cognitive operations.

Language forms do not “carry” information; they latch on to rich pre-existent networks in the subjects'
brains and trigger massive sequential and parallel activations. Those activated networks are themselves in
the appropriate state by virtue of general organization due to cognition and culture and local organization
due to physical and mental context. Crucially, we have no awareness of this amazing chain of cognitive
events that takes place as we talk and listen, except for the external manifestation of language (sounds,
words, sentences) and the internal manifestation of meaning, experienced consciously with lightning speed.
This is very similar to perception, which is also instantaneous and immediate, with no awareness of the
extraordinarily complex intervening neural events.

Within cognitive frameworks for studying meaning construction, many standard issues of pragmatics remain
as important as ever - we seek to account for scalar phenomena, speech acts and performatives,
presupposition, referential opacity, so-called figurative speech, metonymic pragmatic functions, and
implicature - but old problems are framed in novel ways.

There is a great wealth of recent work on these matters and there is not enough space in the present
handbook entry to even remotely do it justice. I will proceed by giving some illustrative examples and
pointing to larger bodies of work.

1 Language as a Prompt: Turner's

xyz Constructions

In

Reading Minds, Turner (1991) draws our attention to an extremely productive construction in English, x is

y of z, as in:

•

Vanity is the quicksand of reason.

•

Wit is the salt of conversation.

•

Money is the root of evil.

Turner notes insightfully that in the course of understanding such expressions, we build up a metaphorical
mapping of the form (

figure 29.1

).

So, for example, in

Vanity is the quicksand of reason, a plausible metaphorical source would have a traveler

(

w) who is sucked into quicksand (y), and that source would be mapped onto a target of human behavior

containing vanity (

x) and reason (z), in which vanity causes reason to falter and perhaps disappear. The

grammatical construction “NP be NP of NP” licenses this mapping scheme, but tells us nothing about the
domains themselves or the relevant inferences that will produce efficient understanding. In fact, the
linguistic form has nothing that actually corresponds to the crucial fourth element

w that must be mapped

onto z (in the quicksand example, this element could be the traveler who does not notice the quicksand and
falls into it). In

Money is the root of evil, the target domain is the complex social world of humans, money,

greed, good and evil. The source is the domain of plants.

y is “the root,” the missing w is the plant; the

counterpart of “the root” (

y) (in the source) is “money” (x) in the target domain. The counterpart of the plant

(

w) is evil (z). Some inferences transfer from the source to the target:

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 3 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

(

w) is evil (z). Some inferences transfer from the source to the target:

Figure 29.1Metaphorical mapping for

xyz constructions

root “causes” plant money causes evil

without a root there is no plant without money there is no evil

we see the plant, not the root we do not see the link between money and evil

There are two important points about the way in which the construction “NP be NP of NP” prompts for a
systematic

xyz(w) mapping scheme.

First, the grammar prompts us to look for two domains and a mapping between them, with a certain
counterpart structure (e.g.

root corresponds to money, and something (w) corresponds to evil). From this

very minimal information, we are supposed to retrieve the right domains and the efficient inferential
structure. Since the domains are not specified by language, their retrieval belongs to what is usually called
pragmatics. But no meaning, even partial or schematic, except for the

xyz mapping scheme itself, can

possibly be constructed without the domains needed for the mapping. It follows that there cannot be a
minimal, independent (“literal”) meaning with truth conditions, linked solely to the linguistic form, from
which the contextual meaning would be pragmatically derived. Rather, the form is a direct prompt to
perform the pragmatic operations needed to satisfy the mapping scheme licensed by the linguistic
construction. The truth conditions for the expression can only be computed after the connections and
domains have been selected.

Second, the mapping scheme is not specialized in any particular type of surface phenomenon. One might
think from the examples cited above that the

xyz construction has the function of setting up metaphors,

since indeed that is what it does in those examples. But in fact, the construction fits other kinds of
mappings. So, if “NP is the NP of NP” is the English sentence

Paul is the father of Sally, it will prompt for

exactly the same

xyz mapping scheme as in the other examples (see

figure 29.2

). This time, Paul and Sally

(

x and z) are mapped onto the kinship frame containing father, and the missing element w is child in the

kinship frame.

Choosing a mapping from a kinship frame to individuals in this case is the default, but there are many other
possibilities, all within the same mapping scheme, as in

Paul is the father of Sally for today (as when a

neighbor takes over some fatherly duties, e.g. taking Sally to school, reading her a story, cooking her lunch);
The Pope is the father of the church; Newton is the father of physics; Fear, father of cruelty, etc.
Constructing meaning in all these cases is a matter of finding domains and structure that best fit both the
context and the mapping scheme. It is what used to be called pragmatics, but there is no longer an
intermediate, core “semantic” content from which meaning in context is derived. Rather, the language form
is a prompt to go directly to the context, using cognitive resources like memory, schemas, and cultural
models (Fauconnier 2000). In fact, such examples highlight the often spurious nature of “core semantics”
and the sometimes dubious application of Gricean maxims: why go to the trouble of trying to process a
nonsensical “literal” meaning of

Money is the root of evil, when the grammatical construction invites you to

find and match separate domains directly, with

money and evil in one, and root in the other? Strikingly, the

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 4 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

find and match separate domains directly, with

money and evil in one, and root in the other? Strikingly, the

grammatical construction makes no distinction between the case where the required matching is felt to be
metaphorical (vanity and quicksand) and the case where it is not (father and Paul). The point is that the most
elegant and general account of meaning construction in cases like this does not require, and in fact does not
allow, an intermediate “literal” or “logical” or “core” meaning to intervene between the syntactic form and the
intended meaning. It is useful to appreciate this, because in much of standard pragmatics, pragmatic
principles apply to the semantic output of a grammar. The direct construction of meaning on the basis of
(often minimal) prompting by grammatical constructions is not just an exotic property of

xyz constructions;

it is a pervasive, and arguably constitutive, property of language and meaning.

Figure 29.2

xyz mapping scheme used for Paul is the father of Sally

To achieve the desired generalizations, it is useful to forget notions like “meaning of an expression,”
“semantic representation,” “truth function,” and the like, and to think instead of the meaning potential of a
language form. Meaning potential is the essentially unlimited number of ways in which an expression can
prompt dynamic cognitive processes, which include conceptual connections, mappings, blends, and
simulations. Such processes are inherently creative, and we recognize them as such when they are triggered
or produced by art and literature. In everyday life, the creativity is hidden by the largely unconscious and
extremely swift nature of the myriad cognitive operations that enter into the simplest of our meaning
constructions. It is also hidden by the necessary folk-theory of our everyday behavior, which is based quite
naturally on our conscious experience rather than on the less accessible components of our cognition. In
doing away with distinct semantic and pragmatic components, cognitive theories actually highlight the
centrality of pragmatic operations in meaning construction.

2 Opacity and Presuppositions

Opacity and presupposition projection are classical problems for the logic of language. A very general way of
looking at these problems within a cognitive approach to meaning construction is to study the connections
between the mental spaces that are set up in ongoing discourse.

Mental spaces are small conceptual packets constructed as we think and talk for purposes of local
understanding and action. They are partial assemblies containing elements and are structured by frames and
cognitive models. They are interconnected and can be modified as thought and discourse unfold.

Mental spaces proliferate in the unfolding of discourse, map onto each other in intricate ways, and provide
abstract mental structure for shifting anchoring, viewpoint, and focus, allowing us to direct our attention at
any time onto very partial and simple structures while maintaining an elaborate web of connections in
working memory and in long-term memory.

Consider the following simple example of mental space construction prompted by sentence (1) coming at a
certain point into a particular discourse in context (from Fauconnier 1998):

(1) Max thought the winner received $100.

The language form prompts for setting up two mental spaces. One is the Base space, B, the initial space with
partial structure corresponding to what has already been introduced at that point in the discourse or what

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 5 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

partial structure corresponding to what has already been introduced at that point in the discourse or what
may be introduced freely because it is pragmatically available in the situation. Another mental space, M,
subordinate to this one, will contain partial structure corresponding to “what Max thinks.” It is structured by
received $100. That form evokes a general frame of receiving of which we may know a great number of
more specific instances (receive money, a shock, a letter, guests, etc.). The expression

Max thought explicitly

sets up this second mental space.

Max and the winner are noun phrases and will provide access to elements

in the spaces. This happens as follows: the noun phrase is a name or description which either fits some
already established element in some space or introduces a new element in some space. That element may, in
turn, provide access to another element through a cognitive connection (a connector). In our example, the
name

Max accesses an element in the base space; the description the winner accesses a role in a frame of

winning appropriate for the given context (winning a particular race, lottery, game, etc.). Roles can have
values, and a role element can always access another element that is the value of that role. So we can say
The winner will get $10, without pointing to any particular individual; this is the role interpretation. Or we
can say

The winner is bald, where being bald is a property of the individual who happened to win, not a

condition for getting the prize; this is the value interpretation. The two mental spaces B and M are
connected, and there can be counterparts of elements of B in M, as diagrammed in

figure 29.3

. The ACCESS

PRINCIPLE defines a general procedure for accessing elements: If two elements

a and a′ are linked by a

connector

F (a' = F(a)), then element a′ can be identified by naming, describing, or pointing to its

counterpart

a.

Figure 29.3Mental space construction

What philosophers and linguists traditionally call the number of readings of a sentence like (1) is the number
of accessing possibilities compatible with the explicit instructions of the linguistic form. But that number
varies with the available types of connections between mental spaces and the available mental spaces in a
particular discourse and pragmatic environment. So the number of “logical” readings of a particular language
form is itself a pragmatic property, and constructing the appropriate reading in a particular context is itself a
pragmatic exercise: finding a particular access path through a mental space configuration.

In Fauconnier (1998), several accessing possibilities for example (1) are outlined that are all equally
sanctioned by the grammatical form. Typically, a particular context or discourse creates a strong bias for
following a particular accessing path, and other paths formally possible for the single sentence form are not
even considered consciously. These paths can remain available, however. Coulson (2001) has shown
experimentally some neural correlates of frame-shifting and reactivation of accessing paths in the case of
jokes. The punch line of a joke often forces a frame shift and a reconfiguration of the accessing path.

For a grammatical form like the

winner example above, the accessing possibilities in the simplest contexts

depend on whether the access principle is applied across spaces and from role to value. The grammatical

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 6 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

depend on whether the access principle is applied across spaces and from role to value. The grammatical
form does not force a particular access path.

In a pure role interpretation within space M with no application of the access principle, we have readings in
which Max thought there was a contest and thought it was a feature of that contest that $100 would be
awarded to whoever won. This accessing strategy is noncommittal as to whether the speaker also assumes
there was such a contest, whether an actual winner was ever selected, and whether Max thinks that a winner
was selected. The sentence under this strategy would be appropriate in mini-discourses like the following:

(2) The Boston Marathon will take place next week. Max thought the winner received $100, but it
turns out there won't be any prize money.
(3) My friends were under the impression that I was running a lottery in my garage. Max thought the
winner received $100. But they were all wrong -there was no lottery.

Suppose now that the access principle operates, but only within the subordinate space

M, linking the role

“winner” to a value of that role. As before, Max believes that there was a contest and, moreover, that
somebody won, and he has additional beliefs about the person who he assumes won. Although a likely
default is that the $100 was prize money, this interpretation is no longer obligatory. Max may believe that
something else happened, causing the person who had won to receive $100 independently.

In the two accessing possibilities just considered, a word (

winner) simply evokes a script within a single

space. but the access principle may also operate across spaces. The speaker may have a particular contest in
mind, for which there is a role “winner,” set up as an element

w in the base B. That role can have a value

(e.g. Harry) with a counterpart in mental space

M. The access principle allows the winner to access the role,

and then its value, and finally the counterpart of that value.

The interpretation, then, is that the speaker presents Harry as the winner, and says that Max thought Harry
received $100. This is compatible with Max knowing nothing about the contest, or believing that someone
other than Harry won and that Harry got the $100 for selling a used car or as a consolation prize in the
contest.

Other accessing paths are available for this simple sentence if counterpart roles are considered, or if the
extra space introduced by the past tense is taken into account, or if other spaces are accessible at that
point in the discourse to provide counterparts. Typically, an understander does not have to consider all
possibilities. The intended path will be favored by the space configuration in the discourse at the point at
which the statement is made: we might already have the role “winner” in space

B, or in space M but not in B;

we might already have different values in

M and B for the same role (w and its counterpart w'); and so on. In

other cases, of course, the understander will lack sufficient information and may have to revise a space
configuration or may simply misunderstand a speaker's intent.

Our example had a subordinate space corresponding to “belief.” There are many other kinds of mental
spaces, but they all share these complex accessing possibilities. For instance, time expressions are space-
builders and set up new spaces in discourse. Consider (as part of a larger discourse) the following sentence:
In 1968, the winner received $100. As before, we have a base B and a subordinate space M, corresponding
to 1968 and set up by the space-builder

in 1968. Also as before, the noun phrase the winner can access a

role in

M, or the counterpart role in M of a role in B, or the counterpart in M of a value in B, or a value of a

role in

M. Situations that fit these respective strategies might include:

• There was a certain type of game in 1968 (no longer played today) in which you got $100 for
winning.
• There is a certain sports competition, say the Boston Marathon, that exists today (role

w in space B)

and also existed in 1968 (counterpart role

w' in space M). In 1968 (as opposed to today), whoever

won got $100.
• The winner of the chess championship held today is Susan. Back in 1968, in unrelated
circumstances (e.g. selling her used car), Susan received $100.
• The winner back in 1968 of the contest we are talking about was Harry, and that year Harry received
$100 (perhaps for selling his used car).

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 7 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

The access paths available in this example involving time (1968) are exactly the same as the access paths in
the previous example involving belief. Even though time and belief are conceptually quite different, they give
rise, at the level of discourse management considered here, to the same mental space configurations. More
generally, we find that mental spaces are set up for a wide variety of conceptual domains that include time,
belief, wishes, plays, movies, pictures, possibility, necessity, hypotheticals and counterfactuals, locatives, and
reality. The connectors, the access principle, and the role/value distinctions work uniformly across this broad
range of cases.

The study of accessing strategies is developed in great detail in Fauconnier (1985, 1997), Fauconnier and
Sweetser (1996), and Fauconnier and Turner (2002). What we see is that phenomena traditionally viewed as
logical, such as opacity, attributivity, and presupposition projection, and others traditionally viewed as
literary, such as the

xyz phenomena discussed in section 1, are all part of the same uniform and powerful

prompting by language for connections across mental spaces.

Indeed, presupposition projection finds an elegant and general solution in these terms. The principle which
guides presupposition projection can be formulated as:

(4) Presupposition Float
A presupposed structure ? in mental space

M will propagate to the next higher space N, unless

structure already in

M or N is incompatible with ? or entails?.

Informally, a presupposition floats up until it meets itself or its opposite. To say that a structure
“propagates” from space

X to another space Y is to say that if it is satisfied for elements x, y,… in X, it is

satisfied for their counterparts

x', y',… in Y, via some connector C. As usual, we find structure mappings

involved in matching processes that transfer structure.

So, consider for example a familiar case like that presented in (5):

(2) Sue believes Luke has a child and that Luke's child will visit her.

In this example, grammatical clues prompt for the construction of three spaces: the base

B, a belief space M,

and a future space

W. The three spaces contain counterparts a, a′, a″ for Sue and b, b', b″ for Luke. The

presupposition in the future space W that

b″ has a child> meets itself in the form b has a child> in the

higher belief space, and therefore does not float up. Because the presupposition does not float up to

M, it

cannot, a fortiori, float up to the base: it has been halted at the

W level. This ensures that the piece of

discourse does not globally presuppose (or entail) that Luke has children.

Notice that although the presupposition does not float all the way up, it is not canceled, but rather remains
in force in the future space: a presupposition will float up into higher spaces until it is halted. It will then
remain in force for the mental spaces into which it has floated. In other words, inheritance is not an “all or
nothing” process. The general issue is: Which spaces inherit the presupposition? The vast literature on
presupposition projection typically focuses on asking if the “whole sentence” inherits the presupposition, i.e.
if the presupposition floats all the way up. To account for the full range of entailments, a more general
question has to be answered: How far up does the presupposition float? The answer to the general question,
given by the Presupposition Float Principle, subsumes the answer to the special question: Does it float to the
base?

3 Word and Sentence Meanings

Sections 1 and 2 illustrate the fundamental but counterintuitive idea that language forms do not in
themselves have core meanings that get modified, adjusted, or expanded through pragmatic elaboration.
Rather, language forms steer the construction of meaning along certain paths in particular contexts. Thus
there is no “autonomous” semantics separate from pragmatics, context, and cognition. The impossibility of
drawing this line is well documented: detailed and forceful arguments are given in Haiman (1980b), Moore
and Carling (1982), Sweetser (1999), and no doubt many others. Haiman finds that “the distinction between
dictionaries and encyclopedias is … fundamentally misconceived” (1980b: 331). Langacker (1987) is
unequivocal: “Certainly an autonomous semantics … can be formulated, but the account it offers of the
meanings of linguistic expressions is apt to be so restricted and impoverished relative to the full richness of

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 8 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

meanings of linguistic expressions is apt to be so restricted and impoverished relative to the full richness of
how we actually understand them that one can only question its utility and cognitive reality. Only limited
interest attaches to a linguistic semantics that avoids most of the relevant phenomena and leaves
recalcitrant data for an ill-defined ‘pragmatic’ component” (1987: 155). And he goes on to write: “The
assumption that language (and in particular semantics) constitutes an autonomous formal system is simply
gratuitous” (p. 156). He explains how entities have specifications that form a gradation in terms of their
centrality, and how the notion of centrality in turn correlates with the notions of conventional, generic,
intrinsic, and characteristic. Taking as an example the situation in which a wrestler triumphs over a tiger,
so that the cat is on the mat, Langacker writes: “it is not that the expression [

the cat is on the mat]

intrinsically holds or conveys the contextual meaning, but rather, that conventional units sanction this
meaning as falling within the open-ended class of conceptualizations they motivate through judgments of
full or partial schematicity. These conceptualizations may draw on any facet of a speaker's conceptual
universe.”

Coulson (2001), Sweetser (1999), and Fauconnier and Turner (1998, 2002) explore many aspects of the on-
line construction of meaning using conceptual blending theory. Conceptual blending is a very general
cognitive operation that partially matches two (or more) input mental spaces and selectively projects from
the matched spaces to create a blended mental space with emergent structure. This creates a conceptual
integration network of the form shown in

figure 29.4

. The generic space represents the structure shared by

the inputs. The square in the blended space stands here for the emergent structure which arises in the
blending.

So, for example, one way to understand the counterfactual in (6):

(6) In France, Watergate wouldn't have done Nixon any harm.

is to build a conceptual integration network that partially matches two input spaces with prominent aspects
of the American political system and the French political system, respectively, and develops an emergent
blended space framed by the French political and social system with Nixon as president and Watergate as a
French scandal. In the selective projection, Nixon loses his American characteristics (English language,
former vice-president, etc.). Watergate, similarly, loses most of its specific geographical and American
characteristics. It becomes a French scandal with some properties analogous and others disanalogous to
those of the American scandal. Emergent structure in the blended space includes the specification that Nixon
is unharmed and the inferences (and emotions) that this may lead to. For most understanders, it projects
back to the inputs, telling them something about the French mentality, or political system, and perhaps
inviting value judgments in one direction or the other.

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 9 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

Figure 29.4Diagram showing conceptual blending

But of course, this is only one of the very many ways to construct a counter-factual blend licensed by
example (6). Here are some other perfectly fine uses with completely different intended meanings,
corresponding to different choices of selective projection:

(7)

a. In France, Watergate wouldn't have done Nixon any harm because Nixon is loved by the

French.

b. In France, Watergate wouldn't have done Nixon any harm because French presidents are not

affected by US scandals.

c. In France, Watergate wouldn't have done Nixon any harm because spying on American political

parties is supported by public opinion.

d. In France, Watergate wouldn't have done Nixon any harm because he would never have been

elected president in the first place.

Here again, we see that there is no single special meaning of the expression that would serve as a core and
to which further pragmatic operations would apply to produce the other readings. Rather, the grammatical
construction is from the outset an invitation to set up a counterfactual conceptual integration network, but it
leaves highly underspecified much of the construction of that network - in particular, prominent aspects of
the input mental spaces and selective projection (Mandelblit and Fauconnier 2000). Building the right
network in context is clearly part of pragmatics; but there is no contentful meaning for the language form
independent of the pragmatic construction. Fauconnier and Turner (2002) show that conceptual blending is
highly constrained by numerous constitutive and governing principles. In addition to those cognitive
principles, appropriate constructions of blends for a given context will be guided by pragmatic principles of
the usual sort: goals and focus of the language activity, prior discourse, and general relevance.

Blending also plays a central role in constructing meanings for words in context, and once again it turns out
that contextual meanings are not derived from prototypical or basic meanings. Rather, it is the other way
around: what we take intuitively to be basic meanings are simply defaults for situations with minimum

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 10 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

context. These defaults are not a basis for constructing the more elaborate meanings, rather they are special
cases under special conditions (minimum, widely available context). They are psychologically real, but not
theoretically fundamental.

Consider now the morphosyntactically simple forms of language that consist of putting two words together:
e.g., noun-noun compounds like

boat house, adjective-noun combinations like angry man, and noun-

adjective combinations like

child-safe or sugar-free.

Complicated integration networks lie behind these kinds of words. Take for example

dolphin-safe, shark-

safe, and child-safe: in all of these cases, one input has the abstract frame of DANGER and the other input
has a more specific content, involving dolphins, sharks, or children. We build a specific counterfactual
scenario of HARM in which

dolphin, shark, or child is assigned to a role in the DANGER frame. In the

counterfactual scenario, a victim is harmed. The word

safe implies a disanalogy between the counterfactual

blend and the specific input.

Dolphin-safe, as it is currently used on cans of tuna, means that measures were taken to avoid harming
dolphins during the harvesting of the tuna.

Shark-safe, as applied to, say, swimming, means conditions

under which swimmers are not vulnerable to attack by sharks.

Child-safe as applied to rooms can mean

rooms that are free of typical dangers for children or rooms immune to children. In every case, the
understander must construct elaborate counterfactual scenarios from simple forms.

How the understander does this mental work may differ from case to case. In

dolphin-safe tuna, the role of

the dolphin as potential victim is taken to be useful. In

dolphin-safe diving, said of mine-seeking human

divers who are protected by dolphins who are not themselves at risk, the blend uses dolphins in the role of
agents of the safety. In

dolphin-safe diving, said of diving that imitates the way dolphins swim and is

therefore safe, the blend uses the manner of swimming associated with dolphins. If we assume that dolphins
eat goldfish, then

dolphin-safe goldfish casts dolphins in the role of predators. Genetic engineers who are

concerned not to produce anything resembling a dolphin might refer to a technique that is known never to
lead to a dolphin embryo as

dolphin-safe. The dolphin this time does not fill the role of victim, victimizer,

causal agent, or role model. And of course, a compositional theory of meaning immune to the dolphin
examples would be hailed as

dolphin-safe. There is a vast and insightful literature in linguistics dealing with

compounds and the obvious fact that their meaning is not derivable from their form (e.g. Downing 1977,
Ryder 1994). The deeper point, already made by Travis (1981), and Langacker (1987: 271–4) in his work on
active zones, is that whatever operations are needed to construct meanings in such cases are also needed
for the everyday cases like

black kettle, red pencil, and brown cow. Fauconnier and Turner (2002: chapter

17) analyze in some detail the role of conceptual blending in understanding such simple forms.

Examples like

dolphin-safe are useful because they highlight in a transparent and uncontroversial way the

nature of the blending process. Furthermore, they abound. Think of

cruelty-free on bottles of shampoo or

the variety of non-compositional integration running across

waterproof, tamper-proof, foolproof, and child-

proof or talent pool, gene pool, water pool, football pool, and betting pool.

Sweetser (1999) considers the case in which

likely candidate, in a political context, means not someone

likely to become a candidate or succeed as a candidate but, for instance, a candidate likely to grant an
interview. As she writes, “So long as we can think up a scenario relative to the candidate in question, and
evaluate that scenario for likelihood,

likely candidate can mean the candidate who figures in the scenario we

have labeled as likely.” On her analysis, conceiving of such a scenario and evaluating it consists of finding a
blend of the frame for likelihood, conceived of as probability of occurrence in a sequence, and the frame for
candidate. Like

safe above, likely prompts for a conceptual blend. Likely requires the recruitment of a

scenario of likelihood and another input space in which the noun prompts for a frame and an element filling
a role in that frame. In the blend, the element must play a role in the specific scenario of likelihood, but
there are many possibilities for the nature of this role. One general possibility is for the scenario evoked by
the noun to already be a scenario of likelihood, as in

likely candidate, meaning someone likely to become a

candidate or a candidate likely to be chosen.

Likely suspect is similar, meaning alternatively a person likely

to be deemed a suspect or a suspect likely to be guilty or a suspect likely to be convicted. However, the
scenario of likelihood in the blend need not be the one evoked by the noun. Sweetser's reading, “a candidate
likely to grant an interview,” does not use the likelihood of either becoming a candidate or being chosen. If
every year the governor pardons someone in the jailhouse, we might bet on who is the “likely suspect.”

Sweetser's examples make the point clearly that the scenario of likelihood need not be one evoked by the
noun itself, as we see, for example, in the case in which

possible textbook refers to a textbook that might

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 11 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

noun itself, as we see, for example, in the case in which

possible textbook refers to a textbook that might

possibly be chosen as the one to be used in a college course. Just as the different meanings of

safe may go

unnoticed, so the different meanings of

possible and likely may go unnoticed. But a possible textbook, in

the sense of one that may be adopted, is not the same as a

possible textbook in the sense of one that

might exist, or might be written, or a trade book that could double as a textbook. Humans have the crucial
ability to construct the elaborate integration networks appropriate for each of these cases. This is not a
special-purpose exotic capacity - it is the very foundation of meaning construction, and children master it
early in their development. A semantic analysis that merely said that the meaning of

possible ranges over

unspecified elements would miss the heart of the problem.

Coulson has shown that

fake is an adjective that carries an especially elaborate blending and mapping

scheme (Coulson and Fauconnier 1999, Coulson 2001): it calls for two input spaces with a counterfactual
connector, such that an element in one space is real, but in the other space its counterpart is not. For
example, the

fake of fake gun prompts a mapping between an actual scenario with an agent and an

instrument and a counterfactual scenario in which the counterpart of that instrument is a real gun and some
other participants, if there are any, react accordingly. In the blend, the reactions and beliefs of other
participants are projected from the counterfactual scenario, while the nature of the object and the beliefs of
the gunman are projected from the actual scenario.

In the case of

fake flowers meant to decorate a dining table, there is an input space in which the object is

flowers and a space in which it is not because it is, for example, silk or rubber. There is a counterfactual
connector between the flowers in one space and the silk or rubber object in the other space. This
counterfactual relation is compressed into a property of the object in the blend: it is now a fake flower. Note
that there need be no indication that any of the participants believes the object is a flower. It can be treated
as a flower, perhaps for aesthetic purposes, by projection from one space and as not a flower by projection
from the other space. One might enjoy it but not have to water it.

Fake money likewise requires two spaces:

one in which the object is money and the other in which it is just, for example, paper. There are many
possible projections to the blend. For example, the fake money might not be legal currency but still have
monetary value since it is issued by a gambling house for in-house use; it cannot be stolen since it can only
be redeemed by the in-house cashier. Alternatively, it could be used to fool someone by making them a
worthless payment. In that case, the space where it is money is also the belief space of the person fooled,
and the space where it is paper is the belief space of the crook. Whether the belief of the crook is shared by
the speaker of the utterance depends on overall aspects of the discourse mental space configuration. In (8) it
would be shared, but not in (9):

(8) The crook paid them with fake money.
(9) The crook thought he was paying them with fake money, but in fact it was the new European
currency.

4 Performatives

The general finding that much of language and grammar is devoted to prompting for mapping schemes
between mental spaces opens up an interesting way to rethink the key pragmatic questions raised by Austin
(1962).

Sweetser (2001) develops a unified and elegant approach to performativity. Her general definition of
performativity is that it involves a particular relation of fit between a mental space that is a representation
and the corresponding represented space. If the representation is taken as fitting the represented space,
then the relation between the spaces is depictive, or representational. It is the success or failure of depictive
fit (of representation to world) which is described as

true or false, accurate or inaccurate. If, on the other

hand, the represented space is taken as fitting (being causally influenced or changed by) the representation,
then the relation is performative. Sweetser's account maintains Searle's (1979, 1989) crucial distinctions
between using a representation descriptively and using it performatively, and between performing an action
by describing it and performing it by other means, but it generalizes the approach to non-linguistic
phenomena and reveals that performativity is a special case of compression of vital relations in integration
networks.

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 12 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

A non-linguistic example offered by Sweetser (2001) is that of a painting of a buffalo hunt on a cave wall,
discovered by archaeologists who know from associated artifacts that the members of the social group that
produced the paintings were also buffalo hunters. In the painting, buffalo fall prey to successful hunters: the
painting could be depictive (describing a successful past hunt) or performative (made with the intention of
bringing about success in a future hunt). (See Searle 1975b for a similar case, the “shopping-list parable.”)

In both cases, we have a conceptual integration over two inputs, the mental space of the image and the
mental space of the hunt (as it happened or as desired). In the depictive usage, the vital relation of
representation is compressed into identity in the blended space. In the performative usage, it is the vital
relation of cause-effect that is compressed into identity. These very compressions are widely attested
independently of speech act phenomena (Fauconnier and Turner 2002), and so Sweetser's account places
Austin's initial observations about performative vs. constative utterances in a significantly broader cognitive
and cultural framework.

In particular, as Sweetser (2001) and Sorensen (2000) observe, non-linguistic performative examples abound
in ritual and magic. An extreme example is voodoo death, in which manipulation of the representation
produces a desired effect in reality. The belief in the efficacy of the conceptual integration network is basic
for the community and for the victim, and so the effect of this deeply constitutive way of thinking is that the
social body and the physiological body conspire to make the death happen.

Sweetser (2001) also notes that names, like other representations, are susceptible to referential use for
either depictive or performative purposes. A name applied to an entity is depictive referentially if the entity is
present, but names are also used as invocations or evocations of the named entity. Names of gods, for
example, are powerful because they may invoke the presence of the deity. A second kind of performative
use of names is the one that establishes a naming convention by using a name or a nickname.

Sweetser generalizes her analysis to include common types of rituals, such as that of a newborn baby being
carried up the stairs of its parents' house as part of the celebration of its birth in a European ritual. The
ritual is meant, symbolically, to promote the child's chances of rising in life and is supported by a conceptual
integration network. One input in the network is the ordinary action of carrying a baby up the stairs. The
other input is the schematic space of life, already structured so that living a life is metaphorically moving
along a path, such that good fortune is up and misfortune is down. In a partial match between these inputs,
the path up the stairs corresponds to the course of life, the baby is the person who will live this life, the
manner of motion up the stairs corresponds to how the person goes through life, and so on. In the symbolic
ritual, the two inputs are blended so that the ascent of the stairs is the course of life: an easy ascent
signifies an easy rise in life for the person that the baby will become, and stumbling or falling might take on
extraordinary significance. Vital relations hold within the network: cause-effect (success in climbing the
stairs will (help to) cause success in life), part-whole (the ritual itself is a necessary part of the entire life),
time (the short time it takes to climb the stairs is mapped onto the much longer time it takes to live a life),
change (within the life input, the person will change considerably from birth till death), analogy (through
metaphor, “going up” the stairs is similar to “going up” in life). In the blended space, these vital relations are
compressed: cause-effect, part-whole, and change are all compressed into uniqueness - the cause (climbing
the stairs) is the effect (leading a good life), the part (being carried up the stairs) is the whole (the entire
life), and change becomes uniqueness. The baby doesn't change physically as it is carried up the stairs -
there is a unique, unchanging person. Time is scaled down from the duration of life to the time it takes to
climb the stairs. Such compressions are the rule in integration networks. The ritual of the newborn baby,
although non-linguistic, is both metaphorical and performative.

Much religious ritual, notes Sweetser (2001), is both metaphorical and performative in this sense. The
circular shape of a ring metaphorically represents the unending permanence of marriage, but its use in a
wedding ceremony is to bring that permanence into social being, not just to describe it. The white dress
worn by many Christian brides can metaphorically depict (truly or not) the bride's virgin status, but white
ritual garments can also be worn by penitents for purification, thus metaphorically purporting to

bring about

purity, not to describe it. Sorensen (2000), in a similar vein, offers a detailed analysis of Catholic
communion, which can be understood as describing spiritual union and also being intended to bring it
about.

5 Cognitive Approaches to Some Other Classical Pragmatic Issues

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 13 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531

Another topic that played a central role in the healthy development of modern pragmatics is the study of
pragmatic scales, stemming from the pioneering work of Horn (1972) and Ducrot (1972). Horn discovered
the exceptional importance of pragmatic scales for syntactic and semantic phenomena in language. Ducrot
developed an elegant theory of argumentative force to explain a wide range of phenomena, including the
meanings of notoriously difficult connectives such as

but, however, French puisque, parce que, and scalar

operators such as English

even, and French presque, à peine. Fauconnier (1978b and elsewhere) introduced

the powerful concept of implication reversal, proved a weak extension of De Morgan's laws for the general
case of implication reversal, and showed that the extended De Morgan's law was one of the keys to a unified
understanding of polarity phenomena of different stripes - grammatically entrenched, semantic-ally marked,
or pragmatically motivated. In his new book, Israel (to appear) explores the cognitive underpinnings of what
he calls the rhetoric of grammar, extends previous analyses of scale reversal in a number of insightful
directions, and achieves a unified cognitive semantic account of a broad spectrum of polarity phenomena.

Coulson (2001) is another cognitive scientist who takes pragmatic scales seriously and develops a realistic
account of the power and hidden complexity of everyday human reasoning. As she explains early in her
book, discourse participants, like paleontologists, “have the task of combining different sorts of information
to derive the overall meaning of the discourse event, to exploit their imaginative capacities and derive the
life of the organism from its grammatical bones” (2001: xii). Coulson emphasizes that, rather than compiling
meanings, people use linguistic information to help them assemble cognitive models of the discourse event.
Building pragmatic scales is part of the rhetorical strategies that participants deploy, often unconsciously, in
constructing and manipulating cognitive and cultural models relevant to particular discourse events. This is
not done in addition to the construction of some basic meaning, rather it is a necessary and central aspect
of the construction of any meaning. Coulson shows how scales can be linked, forced upon participants,
transformed into different scales, or rejected by switching frames and cultural models.

In the past ten years, cognitive linguistics has grown exponentially with findings and analyses that extend
far beyond the usual boundaries of linguistics. Applications to many fields of human inquiry include
musicology, art, literature, philosophy, political science, theology, and mathematics. To understand how
classical pragmatic issues are re-evaluated, readers will want to add to the illustrative examples in the
present entry the more general perspectives of seminal work on metaphor theory (Lakoff and Johnson 1999),
cognitive semantics (Talmy 2000), cognitive grammar (Langacker 1987, vol. 1), and conceptual integration
(Fauconnier and Turner 2002). The exemplary studies of anaphora by Van Hoek (1997), of signed languages
by Liddell (1998), and of gesture by McNeill (2000) are also of exceptional importance. Coulson (1994)
provides an excellent overview of the relevance of cognitive science to research in pragmatics. Mandelblit
and Zachar (1998) explore the notion of dynamic unit in the conceptual developments of recent cognitive
science.

The study of pragmatics in the second half of the twentieth century revealed many fascinating aspects of
human thought and action that we take for granted in everyday life. We are not inclined to pay much
conscious attention to pragmatics in everyday life, or even in the behavioral sciences, because pragmatics is
largely part of invisible backstage cognition, with exceptionally sparse formal or symbolic marking. And so it
belongs, like Erving Goffman's social framing (Goffman 1974), or Edwin Hutchins's distributed cognition
(Hutchins 1995), to the realm of what is obvious and pervasive and yet invisible to its expert users and
participants. A contribution of cognitive linguistics in recent years has been to show that this invisibility is a
general feature of meaning construction, linguistic or not, pragmatic or not.

Cite this article

FAUCONNIER, GILLES. "Pragmatics and Cognitive Linguistics."

The Handbook of Pragmatics. Horn, Laurence R. and

Gregory Ward (eds). Blackwell Publishing, 2005. Blackwell Reference Online. 28 December 2007
<http://www.blackwellreference.com/subscriber/tocnode?id=g9780631225485_chunk_g978063122548531>

Bibliographic Details

The Handbook of Pragmatics

Edited by: Laurence R. Horn And Gregory Ward
eISBN: 9780631225485
Print publication date: 2005

12/28/07 5:47 PM

29. Pragmatics and Cognitive Linguistics : The Handbook of Pragmatics : Blackwell Reference Online

Page 14 of 14

http://www.blackwellreference.com/subscriber/uid=532/tocnode?id=g9780631225485_chunk_g978063122548531