Editor for this issue: Andrew Carnie <carnie@linguistlist.org>
What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Andrew Carnie at carnie@linguistlist.org
1. Francisco Yus, Turner: Semantics Pragmatics Interface
Date: Mon, 4 Oct 1999 09:35:16 +0200
From: Francisco Yus <francisco.yus@alc.es>
Subject: Turner: Semantics Pragmatics Interface
Ken Turner (ed.) The Semantics/Pragmatics Interface from
Different Points of View. (Current Research in the
Semantics/Pragmatics Interface, vol. 1). Oxford: Elsevier,
1999. x + 491 pages, 90.50 USD, 178 NLG.
Reviewed by Francisco Yus, University of Alicante, Spain.
"We human beings are odd compared with our nearest
animal relatives. Unlike them, we can say what we want,
when we want. All normal humans can produce and
understand any number of new words and sentences.
Humans use the multiple options of language often
without thinking. But blindly, they sometimes fall into
its traps. They are like spiders who exploit their
webs, but themselves get caught in the sticky strands"
(Aitchison, 1996).
This quote may well be applied to the study of language from
a semantic or pragmatic point of view. As speaking animals,
we could have just packed our messages in carefully word-
wrapped containers and send them out to be opened easily by
our addressees. As linguists, we could have built up only
idealized models of languages as transparent reflections of
people's thoughts. But, of course, that would have been too
easy. Instead, as human beings we have developed an ability
to speak figuratively, be ironic, understate, speak loosely,
create metaphors... and as linguists we have undertaken the
task of explaining how all these 'special effects' are
produced and understood, have taken the path of a context-
or use-bound pragmatic research and have finally got
trapped in its sticky strands, in "an area of linguistic
research with fuzzy boundaries - spreading like an
uncontrollable oil slick" (Parret, Sbis and Verschueren,
1981, quoted in Turner, this volume, Introduction, p. 14).
So many aspects of language production and comprehension
have had to be taken into consideration that eventually
pragmatics has ended up with no clear-cut research program
apart from a manifest interest in the study of context and
language in use.
However, at the same time, many researchers have
finally realized that no serious analysis of language can
be carried out without appealing to (at least some of) its
pragmatic aspects. This has put analysts of context-free
semantics in a difficult position, faced with an
increasingly ubiquitous body of pragmatic research. Exactly
how many (purely) semantic and (context-or-use-bound)
pragmatic aspects are there in ordinary language use? The
fifteen articles included in the volume under review (plus
the editor's introduction) aim at establishing where this
interface may be located. As the title of the book
indicates, the reader will find no 'philosopher's stone' to
satisfy everyone but, rather, "different points of view"
illustrating where in a hypothetical semantic/pragmatic
continuum this interface lies. It is precisely this
multiplicity of perspectives and instructive lack of
agreement that I find most interesting in the book.
Turner, in his introduction to the volume
("Introduction -from a certain point of view (seven inch
version)", 1-18) briefly discusses different proposals of
where (if ever) semantics and pragmatics meet, concluding
that the aim of the volume is "to take some steps to
reducing the heat of some of these discussions and to begin
to increase the light that might profitably be shed on some
of the problems of interdigitating content and context" (p.
14). In this book these steps are definitely taken, and
should lead the way to forthcoming volumes in the series.
Below, I will sketch the different views of the
semantics/pragmatics interface that are proposed in the
volume. In general, the reader will find not only general
accounts of theories dealing with this interface (as it
would be expected in a first volume in the series), but also
very specific research on sentential particles which shows
that a purely semantic analysis is limited or inadequate.
1. Semantics // Pragmatics = dynamic semantics // dynamic
pragmatics
N. Asher ("Discourse structure and the logic of
conversation", 19-48) reconsiders Gricean and Searlian
approaches to communication from a new dynamic point of view
relying on the claim "that discourse structure is an
essential component in discourse interpretation and results
from integrating pragmatic and semantic information
together" (p. 20). The outcome of his analysis is a
cognitive modeling in which discourse structure and
speakers' beliefs and goals interact in a more fine-grained
theory than Grice's (1975) Cooperative Principle and
Searle's account of illocutionary force attached to
propositional content.
However, his analysis also minimizes of the role of
hearers' recognition of speakers' intentions in
communication (e.g. pp. 25-26), currently essential in post-
Gricean approaches like 'relevance theory' (Sperber and
Wilson, 1986/95), which sounds a bit extreme to me. For
Asher, not recovering the speaker's intention in
communicating certain information "may not stop me from
understanding the story itself and assigning it a coherent
discourse structure and acquiring as part of the meaning of
the text the truth-conditional implications of that text
structure. In many cases, interpreters may not exactly be
sure of the speaker's intentions and beliefs, but they are
irrelevant to the content of what's said" (p. 26). In my
opinion, intentions may not be essential to determine what
sentences literally mean, but they certainly are essential
to recover what speakers are trying to communicate with
these sentences, which often differs drastically from the
literal truth-conditional content of the utterance.
2. Semantics // Pragmatics = polyfunctional semantics //
polyfunctional pragmatics
J. van der Auwera ("On the semantic and pragmatic
polyfunctionality of modal verbs", 49-64) addresses a very
specific linguistic item: modals. He sketches a general
typology of context- dependent meaning or
"polyfunctionality". There can be 'semantic
polyfunctionality', covering such cases as vagueness,
polysemy, homonymy and ambiguity, and 'pragmatic
polyfunctionality: "what we find when a word is meant in a
way that flouts the semantics. Either the intended meaning
is more general than the semantics, as when one uses
'grandfather' for any old man... or it is related
metaphorically, as when one uses 'mother' to characterize a
certain node in a generative grammar tree... or
metonymically, when one uses 'pamper' to denote a pamper-
wearing baby" (p. 53). He then goes on to analyze modals in
this dual- polyfunctionality view.
3. Semantics // Pragmatics = what is said // impliciture /
implicature
K. Bach ("The semantics-pragmatics distinction: what it is
and why it matters", 65-84) develops a concept somewhere
between 'what is said' (i.e., the purely linguistic content
of the utterance), which also includes some processes such
as reference assignment and disambiguation, and 'what is
implicated' (in the Gricean sense). The concept, called
'impliciture', integrates several pragmatic processes,
especially expansions of what is said (i.e., fleshings out
of the logical form of the utterance) in order to get a
'proposition expressed' intended by the speaker, as in (1a-
b) below (see Vicente, 1998 for a critical assessment):
(1) a. I haven't eaten [this morning]. b. She has
nothing [appropriate] to wear.
There is also another kind of pragmatic support -
completion- required in those cases in which the utterances
are semantically incomplete and not yet propositional, as in
(2a-b) below.
(2) a. The princess is late [for the party]. b. Tom
has finished [speaking].
Besides, Bach postulates two different degrees of
contextual support: one playing a limited role in
determining context (affecting such variables as the
determination of the speaker's identity or the spatial-
temporal location of interlocutors), and a broad notion of
context covering anything that the hearer can or has to
take into account in determining the speaker's
communicative intention (p. 72). Hence, Bach assumes a much
more important role of intentions in communication than
Asher does: "Pragmatic information concerns facts relevant
to making sense of a speaker's utterance of a sentence (or
other expression). The hearer thereby seeks to identify the
speaker's intention in making the utterance" (p. 74). This
leads to a clear-cut demarcation of semantics and
pragmatics, the former being associated with the linguistic
material uttered and the latter related to the (intentional)
'act' of uttering that material (p. 75).
4. Semantics // Pragmatics = logical form // explicature /
(higher-level) explicature / implicature
Recognizing intentions is also basic in the relevance
theoretic framework (Sperber and Wilson, 1986/95), the one
R. Carston sets out to explain in her article ("The
semantics/pragmatics distinction: A view from relevance
theory", 85-125). A basic assumption in this theory is that
language is decoded -logical forms- but it has to be
enriched pragmatically in order to reach a proposition
supposedly intended by the speaker (semantics/pragmatics is
decoding/inferring within relevance theory). Thus the
theory relies heavily on the importance of pragmatic import
in understanding, since "the decoded 'semantic'
representation is seldom, if ever, fully propositional; it
functions merely as a kind of template or assumption schema,
which necessarily requires pragmatic inference to develop
it into the proposition the speaker intended to express"
(p. 86).
The hearers' pragmatic import, supplied when enriching
linguistic content, varies according to a basic,
cognitively rooted search for relevance in the incoming
information, which amounts to finding the first
interpretation which offers the greatest interest -cognitive
effects- in exchange for the least mental effort.
Inferential mechanisms involved in the development of a
logical form include reference assignment, disambiguation,
enrichment, and loosening. The outcome is an explicature
which, when embedded in a higher-level (metarepresentative)
description (basically including the speaker's
attitude/belief in producing the utterance) results in a
higher-level explicature. The difference between
explicatures and implicatures lies in the amount of
inference and reliance on the logical form: "An explicature
is derived by inferentially developing the logical form of
the utterance. All other communicated assumptions are
implicatures; they are derived by inference alone, inference
in which the explicature is one of the premises" (p. 113).
5. Semantics // Pragmatics = what is said // proposition
expressed
B.S. Gillon ("English indefinite noun phrases and
plurality", 127-147) studies noun phrases such as those in
(3) and (4) below, in which some puzzling interpretations
result from singular/plural number in (b) examples when
embedded in (a) situations:
(3) a. [William, Dan and Reed pool their resources to buy a
single house].
b. The men bought a house.
(4) a. [William, Dan and Reed each buy their own house].
b. The men bought houses.
Surprisingly, (3b) can reflect the circumstance in (4a)
despite the singular number in 'a house'; and (4b) can also
reflect the circumstance in (4a) despite the fact that each
of them bought only one house and the sentence includes a
plural noun phrase 'houses'. These are the "two horns of
the dilemma" that Gillon sets out to explain.
His analysis is mainly centered upon isolated
sentences, and little pragmatic support (i.e., language-in-
use) is taken into consideration (apart from the fact that
certain arrangements of linguistic items -noun phrases-
tend to be understood in specific ways), or shown that
context plays any basic role in the determination of what
interpretation of singular/plural noun phrases is intended
-or eventually selected- in interaction. His analysis is,
rather, in a similar wavelength to pragmatic analyses of
scalar implicatures or referential/attributive
interpretations of definite descriptions.
6. Semantics // Pragmatics = what is said //
interpretation of situated and goal-oriented discourse
Y. Gu ("Towards a model of situated discourse analysis",
149-178) embarks on a very ambitious project (quite
different from Gillon's) of studying highly contextualized
and goal- oriented interactional discourse, and assuming a
wide-margin concept of 'situation'. Gu intends "to come to
terms with the actual use of language by actual people doing
actual things with language in actual social situations
[and] to advance... the view of the actual use of language
as goal-directed social process" (p. 150).
By actual use, Gu means all the spontaneous talk
produced by ALL the adult native users of a language for a
period of time and across all social situations
(comprehensive-all use) or at a particular time and social
situation (sample-all use); and some talk produced by TWO
OR MORE adult native users of a language for a period of
time and across all social situations (comprehensive-some
use), or at a particular time and social situation (sample-
some use). This starting-point terminology is then used in
a review of the bibliography on (the importance of)
language use (pp. 151-154), which is contrasted with one of
Gu's aim in the paper: To examine "the way actual users use
language to attain communicative and extra-communicative
goals in real-life social situations, with a full
recognition of actual users as discourse
developers/managers". A good number of pages are devoted to
this aim, which involves the inclusion of many contextual
parameters in the analysis of discourse.
7. Semantics // Pragmatics = semantics // semanticized
pragmatics
M. Hand's paper ("Semantics vs. pragmatics: ANY in game-
theoretical semantics", 179-198) focuses on a very specific
linguistic item: "any". He pictures straightforward cases
deserving purely semantic or pragmatic analysis and others
in which the semantics/pragmatics distinction is fuzzy, to
the extent that "[c]ompeting theoretical frameworks may
treat these borderline problems in different ways, so that
a linguistic problem that is treated by purely semantical
means in one theory is treated pragmatically by the other"
(p. 180).
His analysis fits "game-theoretical semantics" (GTS)
(initiated back in the early 70s) "with an eye toward
showing how the GTS apparatus can 'semanticize' a pragmatic
insight into the semantics of the notorious English
quantifier any" (ibid.). Basically, he proposes a
semantics-centered unitary account of the two traditional
uses of 'any' (universal and existential quantifier) that
shows why it manifests semantically in such dissimilar ways.
At first sight, this study seems to pay less attention
to contextualized pragmatic uses of language than to its
semantics. However, some aspects of GTS such as the
reference to the presence of two players who must make
selections when quantifiers are interpreted (even if these
players are not to be identified with actual flesh-and-blood
users) indicate some interest in how 'any' is understood
pragmatically.
8. Semantics // Pragmatics = intentions intruding into
semantic representation // context arranging
interpretations in scale of salience
K.M. Jaszczolt, one of the editors (together with K.
Turner) of the series "Current Research in the
Semantics/Pragmatics Interface" (CRISPI), to which this
first volume belongs, proposes an interesting hypothesis in
her paper ("Default semantics, pragmatics, and intentions",
199- 232) and applies it to definite descriptions. Apart
from three previous approaches to the interface (semantic
ambiguity, Grice's unitary semantics supplemented with
conversational implicatures, and semantic
underdeterminacy), Jaszczolt suggests a fourth view of "non-
ambiguous semantics achieved through the interaction of the
speaker's intention with the logical form of the
expression... There is one, discourse-level representation,
to which syntax, semantics and pragmatics contribute" (p.
201). Besides, the resulting interpretations of the
utterance, instead of being treated as equally possible
(context-free), are pragmatically ranked on a scale of
salience and predictability.
9. Semantics // Pragmatics = semantic model of entities //
contextual constraints
A. Kehler and G. Ward ("On the semantics and pragmatics of
'identifier so'", 233-256) focus on the establishment of
reference in discourse. Traditional studies have dealt with
reference to entities (a broad term used here to denote the
class of things that can be referred to linguistically)
through the use of lexical and pronominal noun phrases, and
much less often through the use of verb phrases.
In this paper, the authors present "an analysis of
identifier so based on the informational structure of the
discourse in which it is used... anaphoric expressions
containing so impose a set of constraints on the
information status of their referents that is not found for
any of the various types of NP anaphora in English" (p.
234).
Interestingly, the authors give context an important
role in the determination of reference, since during
comprehension hearers "build up a semantic model
representing the entities that have been introduced thus
far and the relationships that hold between them" (p. 235)
and in order to achieve this goal, they resort to various
contextual sources of entity-fixing including their
knowledge store, previous discursive strings, context of
utterance, immediate surroundings, etc., all of which may
vary the outcome of the referent-signaling procedures.
10. Semantics // Pragmatics = sentence meaning // scalar
implicatures
M. Krifka ("At least some determiners aren't determiners",
257-291) presents a formal- semantics analysis of
determiners that sheds light on classical problems such as
the scope of quantifiers (for example 'three' versus 'at
least three') and their scalar implicatures.
11. Semantics // Pragmatics = lexical meaning //
contextual meaning
S. Kubo ("On an illocutionary connective datte", 293-315)
studies this Japanese connective aiming at a unified view
of its semantic and pragmatic functions within speech act
theory. Previous research on the connective regards it as
being (a) an exponent of the speaker's intention to declare
his/her speech action in conversation; and (b) an
'explanation of reasons' at the propositional level
describing 'assertion' against a partner's intention. Kubo
develops this research in a proposal of datte as "compound
speech acts consisting of two consecutive speech acts,
namely the speech act with assertive illocutionary force of
'objection', and that with assertive illocutionary force of
'supplementary explanation' or 'justification'" (p. 294).
12. Semantics // Pragmatics = contrastive topic expressed
// pragmatic effects created by the contrast
C. Lee ("Contrastive topic: A locus of the interface
evidence from Korean and English", 317- 342) studies Korean
data for the determination of the semantic/pragmatic status
of Contrastive Topic. That variety occurs when the topic in
discourse is divided into parts, one of them contrasting
with the rest of the parts and the speaker has the
alternative contrast in mind. Typically, Contrastive Topic
-due to its cancellative function- gives rise to an
implicature concerning the alternative in contrast (often
accompanied by a marked accent). These effects, together
with the speaker's reliance on background and/or common
knowledge are typically analyzable within pragmatics.
13. Semantics // Pragmatics = semantics --><-- pragmatics
F. Nemo ("The pragmatics of signs, the semantics of
relevance, and the semantics/pragmatics interface", 343-
417) provides a lengthy and ambitious study of the
semantics/pragmatics interface. I find some of his claims a
bit extreme, but he is to be congratulated for a daring
attempt to shed light on this interface.
The paper (strikingly) starts with its final
conclusion, namely, that "even if semantics and pragmatics
are two disciplines with two clearly distinct objects and
separate goals... what they actually find - and not what
they search - is often very similar" (p. 345). This is why
in the above heading semantics and pragmatics converge (--
><--): "the main interface between the two disciplines
could be their own results" (ibid.).
Nemo initially distinguishes two types of semantics
(S) and pragmatics (P): S1 (what is directly attached to
signs), S2 (what is truth-conditional), P1 (a 'cognitive
approach' centered on communication, implicitness and
inference), and P2 (a 'biotic approach' with emphasis on
action, conversation, interaction). These four ends cover
many areas of linguistic research (as illustrated on p.
349). After that, he goes on to show how similar semantics
and pragmatics are despite the traditional debate on their
different scope of analysis and sometimes he arrives at
surprising conclusions. In general, a certain bias towards
pragmatics is felt, as shown by these sample quotes:
(a) Semantics and pragmatics share a not directly
accessible object and explore more or less the same fields
and data (p. 352); (b) "it is not only the origin of the
semantic meaning which is pragmatic, but the semantic
meaning itself" (p. 355); (c) "[pragmatics] has gradually
become [semantics] car handyman mechanic, with so many
reparations to be paid for as to pretend becoming the owner
of the car" (p. 356); (d) "human languages just cannot point
to a people-free world... not only because there are
relations between signs and their users, but mainly because
there are direct relations between the latter and the
objects" (p. 357); (e) "semantic meaning must explain use
in general, and not primarily (apparently) conventional
uses" (p. 364); (f) "the semantic content of a word is not a
description of what it conventionally refers to, but a
comparison of different things or states to isolate the
specificities of something" (p. 376); (g) "a word can be
used to refer to anything, as soon as it locally has a
sufficient discriminative capacity... its semantic content
will depend on the discriminative scene" (p. 377). These
are but a few in-between claims in the article.
The general conclusion of this long study is that
semantic content has a marked pragmatic motivation, and
that pragmatics is closely dependent on semantic
representations, which means that rather than looking for
an interface between semantics and pragmatics we should
rather accept (a) that semantics cannot find purely semantic
answers to purely semantic questions, and (b) that
pragmatics cannot find purely pragmatic answers to purely
pragmatic questions: "whatever affects the saying process
actually leave (sic) traces in semantic in semantic
content, which allows semantic contents to produce pragmatic
effects" (p. 414).
14. Semantics // Pragmatics = invariant meaning //
indexical meaning / meaning as interpretational construct
J. Peregrin ("The pragmatization of semantics", 419-442)
criticizes the three-fold (and clear- cut) Carnapian
distinction between syntax (expressions-expressions),
semantics (expressions- meaning) and pragmatics
(expressions-users). Several challenges to this division
include the 'internal challenge' ("development of
linguistics and logic which extends Carnapian semantics far
beyond its original boundaries to swallow up much of what
originally counted to pragmatics", p. 420) and 'the
external challenge' ("questioning the whole model, a
development... which casts doubt on the entire Carnapian way
of viewing language", ibid.).
Peregrin shows how semantics in the 50s could not
cling to a logic-based null-context view of content and at
the same time analyze natural language phenomena such as
indexicals, which the author calls "context consumers".
Even traditional notions such as 'subject' and 'predicate'
demand a dynamic approach (p. 422). This evidence led
semanticists to see meanings of natural language sentences
as 'context-change potentials'. Indeed, "the very working
of language is essentially oversimplified if meanings are
explicated in a way which does not account for how
utterances interact with each other via contexts" (p. 424).
Later, this evidence leads to Peregrin's view of meaning as
an 'interpretational construct': "Assigning meaning is
specifying a role, or possible roles, within a co-operative
enterprise; it is to state how an expression could be
useful for the purposes for which we use language" (p. 432).
However, this emphasis in language use sounds like a
rest-in-peace approach to semantics. Later in the article,
Peregrin restates the role of semantics by proposing a
concept of 'invariance': "Meaning of an expression is,
roughly speaking, that which is invariant among various
cases of employment of this expression... Thus when I say 'I
am hungry' and you say 'I am hungry too', the fact that the
first I refers to me, whereas the second to you, is a
pragmatic matter. What is invariant is that it always refers
to the speaker" (p. 436-37).
15. Semantics // Pragmatics = literal meaning //
contextualized implicit information compatible with
interlocutors' beliefs
A. Ramsay ("Does it make any sense? Updating = consistency
checking", 443-478) aims at a computational explanation of
how users of language extract meaning from what they say.
The emphasis is on significance as 'entailed' by utterances
in the contexts in which they are uttered. This provides a
picture in which "some of the apparent content of an
utterance will emerge from interactions between its literal
meaning, the information embodied in general knowledge...
and information available in the current discourse
situation" (p. 444).
The author's picture sounds very pragmatic, but it is
also very content-centered, stressing the importance of
information 'already' contained in a proposition in order to
build up further information, that is, "to obtain new
propositions from old ones" (p. 445). In general, the paper
combines an interest in the semantic representation of what
speakers say and also pragmatic factors (e.g.
interlocutors' common ground - despite the problematic
status of this notion nowadays -) influencing
interpretation. This is reflected in the three-fold task of
a hearer H when the speaker S utters U: (a) Construct a
representation of what U says about S's view of each
participant's beliefs and their common ground; (b) check
that this representation is consistent with H's own view of
each participant's beliefs and common ground in order to
construct a model; (c) take this model to be H's view of the
new common ground.
Conclusion
This volume is, no doubt, highly recommendable reading for
anybody interested in the roles that semantics and
pragmatics play in the production and comprehension of
language. The book avoids reductionist views of semantics
and pragmatics and acknowledges a more adequate role of
pragmatics beyond the cliche "wastebasket" of semantics. It
offers fifteen interesting views of where a hypothetical
interface between them might possibly occur, and no claim is
made (by the editor) about the superiority of any of these
views upon the rest. The reader is thus invited to take a
personal position in this on-going debate. If, after reading
the book, some readers think that the semantics/pragmatics
interface should be located elsewhere, they are also invited
to take part in this discussion and contribute to this
series (by entering the official page at
http://www.elsevier.nl/locate/series/crispi) which has just
- and fortunately - been initiated.
References
Aitchison, Jean (1996). "Word traps and how to avoid them".
The Independent, 6 March 1996.
Grice, Paul (1975). "Logic and conversation", in P. Cole
and J. Morgan (eds.), Syntax and Semantics, Vol. 3. London:
Academic Press, 41-58.
Sperber, Dan and Deirdre Wilson (198695). Relevance Theory.
Communication and Cognition. Oxford: Blackwell.
Vicente, Bego¤a (1998). "Against blurring the
explicit/implicit distinction". Revista Alicantina de
Estudios Ingleses 11: 241-258.
Francisco Yus teaches linguistics at the Department of
English Studies of the University of Alicante, Spain. He
has specialized in the application of pragmatics to media
discourses, but his latest research has had to do with the
application of relevance theory to the analysis of
misunderstandings and irony in conversation.
Dr. Francisco Yus
University of Alicante
Department of English Studies
http://www.ua.es/dfing
Tel: +34 9653400 Extension: 3027 (university)
e-mail: francisco.yus@ua.es (university)
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue