1 Wilson & Carston 07 Advances

download 1 Wilson & Carston 07 Advances

of 42

Transcript of 1 Wilson & Carston 07 Advances

  • 8/10/2019 1 Wilson & Carston 07 Advances

    1/42

    A Unitary Approach to Lexical Pragmatics: Relevance, Inference and Ad Hoc

    Concepts*

    P!lished in "oel #rton$Ro!erts %ed&' ())&Pragmatics. Palgrave, London: (+)$(-

    Deirdre Wilson and Robyn Carston

    ABSTRACT:

    According to recent work in the new field of lexical pragmatics the meanings of words are fre!"ently

    pragmatically ad#"sted and fine$t"ned in context so that their contrib"tion to the proposition

    expressed is different from their lexically encoded sense% Well$known examples incl"de lexical

    narrowing &e%g% 'drink( "sed to mean A)C*+*),CDR,-./ approximation &or loosening/ &e%g% 'flat(

    "sed to mean R0)AT,10)23)AT/ and metaphorical extension &e%g% 'b"lldo4er( "sed to mean 3*RC035)

    60RS*-/% These three phenomena are often st"died in isolation from each other and gi7en !"ite

    distinct kinds of explanation% ,n this chapter we will propose a more "nified acco"nt% We will try to

    show that narrowing loosening and metaphorical extension are simply different o"tcomes of a single

    interpreti7e process which creates an ad hoc concept or occasion$specific sense based on interaction

    among encoded concepts context"al information and pragmatic expectations or principles% We will

    o"tline an inferential acco"nt of the lexical ad#"stment process "sing the framework of rele7ance

    theory and compare it with some alternati7e acco"nts%

    8This work is part of an A+RC$f"nded pro#ect 'A 5nified Theory of )exical 6ragmatics( &AR9;

  • 8/10/2019 1 Wilson & Carston 07 Advances

    2/42

    .& Introdction

    The relati7ely new field of lexical pragmatics explores the application of the semantics$

    pragmatics distinction at the le7el of indi7id"al words or phrases rather than whole sentences%

    The ad7antages of disting"ishing semantic and pragmatic aspects of word meaning ha7e long

    been recognised in pragmatically$oriented approaches to the philosophy of lang"age and

    were the starting point for @rice(s William James Lectures&@rice 9: ;$E9/% +owe7er

    the de7elopment of a separate field of lexical pragmatics was accelerated in the 9Fs by a

    series of p"blications by ling"ists psychologists and philosophers proposing more or less

    s"bstantial depart"res from @rice(s acco"nt%9

    The approaches disc"ssed in this chapter share the 7iew that lexical interpretation

    typically in7ol7es the constr"ction of an ad hoc conceptor occasion$specific sense based on

    interaction among encoded concepts context"al information and pragmatic expectations or

    principles% 5se of the term 'ad hoc concept( in this connection is often traced to the

    psychologist )awrence Barsalo" &9 9;/ whose work on categorisation showed that

    prototypical narrowing&i%e% the interpretation of a general term as picking o"t a s"bset of

    prototypical or stereotypical category members/ was m"ch more flexible and context$

    dependent than was standardly ass"med% ,n later work &e%g% by the psycholing"ist Sam

    @l"cksberg and his colleag"es and by pragmatists working within the rele7ance$theoretic

    framework/ it was s"ggested that the o"tcome of the ad hoc concept constr"ction process

    co"ld be either a narrowing or a broadeningof the ling"istically$specified meaning: that is

    the comm"nicated concept may be either more specific or more general than the encoded

    9 See for instance Bl"tner &9/ and )ascarides G Copestake &9/ in formal ling"istics Barsalo" &9

    9;/ and @l"cksberg =anfredi G =c@lone &9/ in psychology and Recanati &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    3/42

    concept% This opens "p the possibility of a "nified acco"nt on which lexical narrowing and

    broadening &or a combination of the two/ are the o"tcomes of a single interpreti7e process

    which fine$t"nes the interpretation of almost e7ery word% We will follow the standard practice

    of representing ad hoc concepts as starred concepts &e%g% +A6628 S+AR.8 BR0A.8/%

    The most radical 7ersions of this "nified approach arg"e not only that narrowing and

    broadening in7ol7e the same interpreti7e mechanisms and may combine in the interpretation

    of a single word b"t that there is a contin""m of cases of broadening ranging from strictly

    literal "se thro"gh approximation and other forms of loosening to 'fig"rati7e( cases s"ch as

    hyperbole and metaphor with no clear c"t$off points between them% S"ch f"lly "nified

    acco"nts re#ect the traditional distinction between literal and fig"rati7e meaning and claim

    that approximation hyperbole and metaphor are not distinct nat"ral kinds re!"iring different

    interpreti7e mechanisms b"t in7ol7e exactly the same interpreti7e processes as are "sed for

    ordinary literal "tterances% This is a s"bstantial depart"re from the standard @ricean

    acco"nt%E

    Whether or not they aim to pro7ide a "nified acco"nt of the f"ll range of cases most

    c"rrent approaches to lexical pragmatics also share the 7iew that narrowing andor

    broadening contrib"te to the tr"th$conditional content of "tterances &what is asserted or

    explicated/ as well as to what is implicated% That is the ad hoc concepts created by the

    pragmatic interpretation of indi7id"al words and phrases are seen as constit"ents of the

    proposition the speaker is taken to ha7e expressed rather than merely contrib"ting to

    implicat"res as in the standard @ricean acco"nt% 3ollowing Recanati &9;/ we will call

    s"ch approaches truth-conditional pragmatic approaches%;H

    E Among those who explicitly re#ect the traditional literal$fig"rati7e distinction are Atlas &EFF

  • 8/10/2019 1 Wilson & Carston 07 Advances

    4/42

    Altho"gh there is a growing consens"s that lexical interpretation in7ol7es the

    constr"ction of ad hoc concepts which contrib"te to the tr"th$conditional content of

    "tterances there is m"ch less agreement on the nat"re of the interpreti7e mechanisms

    in7ol7ed%

  • 8/10/2019 1 Wilson & Carston 07 Advances

    5/42

    The chapter is organised as follows% Section E introd"ces and ill"strates the basic data

    that a theory of lexical pragmatics sho"ld explain% Section ; arg"es that the data call for a

    radical 7ersion of the "nified tr"th$conditional pragmatic acco"nt and that ad hoc concept

    constr"ction is the "nitary process re!"ired% ,n that section we contrast f"lly inferential with

    partly inferential and p"rely associati7e acco"nts of the cogniti7e mechanisms in7ol7ed in

    lexical interpretation and in section H we propose an inferential acco"nt of lexical narrowing

    and broadening "sing the framework of rele7ance theory% Section < raises some f"rther iss"es

    and considers some possible ob#ections to a "nified inferential approach%

    (& /arieties of lexical ad0stment

    The aim of this section is to ill"strate the processes of lexical narrowing and lexical

    broadening "sing a 7ariety of examples which s"ggest that there is a contin""m of cases

    between literal "se approximation metaphor and hyperbole with no clear di7isions between

    them% We will try to show that narrowing and broadening are flexible highly context$

    dependent processes which cannot be ade!"ately handled in terms of code$like r"les and end

    by introd"cing a f"rther range of data that a "nified acco"nt of lexical pragmatics might help

    to explain%

    Lexical narrowingin7ol7es the "se of a word to con7ey a more specific sense than the

    encoded one with a more restricteddenotation &picking o"t a s"bset of the items that fall

    "nder the encoded concept/% -arrowing may take place to different degrees and in different

    directions% Some ill"strations are gi7en in &9/:

    9a% ,(m not drinking tonight.

  • 8/10/2019 1 Wilson & Carston 07 Advances

    6/42

    9b% B"ying a ho"se is easy if yo"(7e got money.

    9c% Ch"rchill was a man%

    ,n 7ario"s different circ"mstances the speaker of &9a/ might be "nderstood as con7eying that

    she will not drink any li!"id at all that she will not drink any alcohol &or any of a certain type

    of alcohol e%g% spirits/ or that she will not drink significant amo"nts of alcohol% 0ach

    s"ccessi7e interpretation is narrower than the pre7io"s one with a more restricted

    denotation%&9b/ s"ggests a pragmatic moti7e for narrowing% *n a literal interpretation the

    speaker wo"ld be "nderstood as making the blatantly false claim that b"ying a ho"se is easy

    for someone with any money at allI the effect of narrowing is to yield a more pla"sible

    informati7e or rele7ant interpretation on which the speaker is "nderstood as claiming that

    b"ying a ho"se is easy for someone with a s"itable amo"nt of money% &9c/ shows that

    narrowing may take place not only to different degrees b"t also in different directions: in

    different sit"ations of "tterance the speaker might be "nderstood as con7eying that Ch"rchill

    was a typical man or that Ch"rchill was an ideal man &where the notion of what constit"tes a

    typical man or an ideal man like the notion of what constit"tes a significant amo"nt of

    alcohol in &9a/ or an appropriate amo"nt of money in &9b/ is itself hea7ily context

    dependent/ &cf% Barsalo" 9 9;/% An ade!"ate pragmatic acco"nt of narrowing sho"ld

    shed some light on what triggers the narrowing process what direction it takes and when it

    stops%

    *ne way of bringing o"t the flexibility and context dependence of narrowing is to

    consider the 7ariety of interpretations that the same word wo"ld recei7e in different ling"istic

    Arg"ably the 7erb 'drink( has now ac!"ired an additional lexical sense as a res"lt of fre!"ent narrowing to the

    more specific sense 'drink alcohol(% +owe7er the f"rther narrowings to 'drink s"bstantial amo"nts ofcertain

    types of alcohol( do not seem to ha7e become lexicalised &and in any case the notion of what constit"tes a

    s"bstantial amo"nt or an appropriate type of alcohol is itself highly context$dependent/%

  • 8/10/2019 1 Wilson & Carston 07 Advances

    7/42

    contexts% Standard examples disc"ssed in the philosophical literat"re incl"de the 7erbs

    'open( 'c"t( and 'lea7e( as ill"strated in &E/:

    Ea% cutthe lawnsomeone(s haira cakeone(s fingera pack of cards J

    Eb% openc"rtainsone(s mo"tha booka bottlea roadthe mo"ntain J

    Ec% leavethe ho"sehomefood on a plateone(s spo"sea note J

    There is no standard or stereotypical method for cutting, opening or leaving to"t co"rt b"t

    there are standard methods for cutting hair, cutting a lawn, opening curtains, and so on each

    of which in7ol7es a narrowing of the more general concepts C5T *60-and )0A10&Searle

    9F/% A similar point is made in an experimental st"dy of ad#ecti7es by the psychologist

    @regory ="rphy &9/% Taking as an example the ad#ecti7e 'fresh( and "sing a 7ariety of

    experimental techni!"es he showed that it has inn"merable slightly different interpretations

    across contexts% *ne method was to ask participants to pro7ide antonymsfor its occ"rrence in

    different ad#ecti7e$no"n combinations% Some of the most fre!"ent responses are listed in &;/

    &="rphy 9: E;$;/:

    ;% fresh A-T*-2=S

    shirt dirty

    7egetables rotten

    fish fro4en

    sheets recently slept in

    water dirtysalt

    bread stale

    air poll"ted

  • 8/10/2019 1 Wilson & Carston 07 Advances

    8/42

    o"tlook tired

    assistant experienced

    idea old

    This clearly ill"strates the point that what is arg"ably a single lexical item encoding a general

    concept 3R0S+ gets specifiednarrowedfine$t"ned in slightly different ways in different

    ling"istic contexts and s"pports the more general claim that disco"rse context and pragmatic

    expectations strongly infl"ence the direction in which narrowing takes place%

    Lexical broadeningin7ol7es the "se of a word to con7ey a more general sense than

    the encoded one with a conse!"ent expansion of the ling"istically$specified denotation% As

    noted abo7e radical 7ersions of the "nified approach to lexical pragmatics s"ch as the one

    proposed in rele7ance theory treat approximation hyperbole and metaphor as s"b7arieties of

    broadening which differ mainly in the degree to which the ling"istically$specified denotation

    is expanded%Approximation is the case where a word with a relati7ely strict sense is

    marginally extended to incl"de a pen"mbra of items &what )asersohn &9/ calls a

    'pragmatic halo(/ that strictly speaking fall o"tside its ling"istically$specified denotation%

    Some ill"strations are gi7en in &H/:

    Ha% That bottle is empty.

    ,t is worth noting that the narrowings ind"ced by ling"istic context are not in7ariably accepted b"t may be

    o7erridden or pre$empted by salient feat"res of extraling"istic context% 3or example in a shop that sells ready$

    made lawn t"rf c"tting it will not in7ol7e the "s"al mowing b"t rather the action of slicing it into transportable

    strips &see Searle 9F/ or when we(re o"t on a picnic the bread may cease to be fresh not beca"se it is stale

    b"t beca"se it has fallen on the gro"nd and is co7ered with dirt% So as Bl"tner &9/ points o"t e7en in those

    cases where a partic"lar pragmatic narrowing is reg"larly deri7ed on the basis of a partic"lar ling"istic

    collocation the process is too flexible and context$dependent to be treated in code$like terms%

  • 8/10/2019 1 Wilson & Carston 07 Advances

    9/42

    Hb% This policy will bankrupt the farmers.

    Hc% The garden issouth-facing.

    ,n &Ha/ the word 'empty( which has a relati7ely strict sense might be intended and

    "nderstood as an approximation so that the speaker wo"ld be interpreted as claiming that the

    bottle in !"estion is 0=6T28: that is close eno"gh to being 0=6T2for the differences to be

    inconse!"ential for the p"rpose at hand &for instance collecting bottles for recycling/% ,n &Hb/

    the word 'bankr"pt( may be intended and "nderstood either literally &BA-.R56T/ or as an

    approximation &BA-.R56T8/ in which case the speaker wo"ld be interpreted as claiming that

    the policy will bring the farmers close eno"gh to bankr"ptcy for the differences to be

    inconse!"ential% Similarly in &Hc/ the term 'so"th$facing( may be "sed literally &to mean that

    the garden faces d"e so"th/ or as an approximation &S*5T+$3AC,-@8/ meaning that the

    garden faces in a generally so"therly direction%

    *n more radical 7ersions of the "nified approach hyperbole is seen as in7ol7ing a

    f"rther degree of broadening and hence a greater depart"re from the encoded meaning% 3or

    instance a parent might say &Ha/ hyperbolically while pointing to a three$!"arters$empty

    bottle intending to con7ey that a teenager has dr"nk too m"ch% Similarly an opposition

    member might "se &Hb/ to indicate hyperbolically that as a res"lt of the go7ernment(s policy

    the farmers will be s"bstantially poorer than might ha7e been expected or desiredI and a new

    ho"se owner might say &Hc/ hyperbolically of a ho"se described in the estate agent(s broch"re

    as facing east$so"th$east intending to implicate that she has made the right choice% The fact

    that in each of these examples &'empty( 'bankr"pt( 'so"th$facing(/ there seems to be a

    gradient or contin""m of cases between literal "se approximation and hyperbole makes it

    worth looking for a "nified acco"nt in which the same interpreti7e mechanisms apply

    thro"gho"t%

  • 8/10/2019 1 Wilson & Carston 07 Advances

    10/42

    Within the f"lly "nified acco"nt we are proposing metaphor is seen as a still more

    radical 7ariety of broadening than hyperbole in7ol7ing a greater depart"re from the encoded

    meaning% Consider &

  • 8/10/2019 1 Wilson & Carston 07 Advances

    11/42

  • 8/10/2019 1 Wilson & Carston 07 Advances

    12/42

    ,n &a/ '1ietnam(may be "nderstood as con7eying an ad hoc concept 1,0T-A=8 which

    represents the category of disastro"s military inter7entions% ,n &b/ 'Chomsky(might be

    "nderstood as con7eying an ad hoc concept C+*=S.28 which represents a broader category

    of forcef"l exponents of a partic"lar approach to ling"istics% ,n &c/ 3),C.$.-,10S8might

    represent the broader category of teenage weapons of choice and in &d/ M a typical piece of

    lifestyle writer(s disco"rse M 2*@A8might be seen as representing the category of fashionable

    pastimes for relie7ing stress% These cases of category extension are not analysable as

    approximations% The claim in &a/ is not that ,ra! is a borderline case close eno"gh to being

    1ietnam for it to be acceptable to call it '1ietnam( b"t merely that it belongs to a broader

    category of which 1ietnam is a salient memberI and so on for the other examples% What

    approximation and category extension ha7e in common is that they are both analysable as

    o"tcomes of a single pragmatic process of lexical ad#"stment which res"lts in an ad hoc

    category whose denotation is broader than that of the lexically encoded concept%

    -eologisms and word coinages pro7ide f"rther data for a theory of lexical pragmatics

    and shed some light on the nat"re of the mental mechanisms in7ol7ed% 0xperiments by Clark

    G Clark &9/ and Clark G @errig &9;/ show that newly$coined 7erbs deri7ed from

    no"ns as in &a/$&b/ and the recr"itment of proper names into compo"nd 7erbs or

    ad#ecti7es as in &c/$&d/ are no harder to "nderstand than reg"lar "ses:

    a% The boyporched the newspaper%

    b% She wristed the ball o7er the net%

    c% +e did a "apoleon for the camera%

    d% They ha7e a lifestyle which is 7ery #an $rancisco.

    3or disc"ssion of more complex cases s"ch as blends &e%g% 'swingle( 'fr"itopia( 'cattit"de(/ which re!"ire a

    greater amo"nt of processing effort see )ehrer &EFF;/%

    9E

  • 8/10/2019 1 Wilson & Carston 07 Advances

    13/42

    5nderstanding &a/ and &b/ depends on knowing the encoded meaning of the no"ns 'porch(

    and 'wrist( and ha7ing appropriate backgro"nd knowledge in the one case abo"t newspaper

    deli7eries in certain comm"nities and in the other abo"t the 7ario"s arm mo7ements of

    competent tennis players% *ther cases are m"ch more idiosyncratic and depend not so m"ch

    on general knowledge b"t on a specific contextI see for instance Clark(s &9;/ disc"ssion

    of '=ax tried to teapota policeman( where 'to teapot L( meant in the partic"lar scenario 'to

    r"b the back of the leg of L with a teapot(% This point is worth emphasising since it indicates

    that there is no principled limit on the possible interpretations of words in "se &i%e% gi7en that

    there are indefinitely many possible contexts there are indefinitely many possible

    ad#"stments of the encoded sense&s//% The interpretations of &c/ and &d/ again depend on

    ha7ing certain kinds of fairly general encyclopaedic information: abo"t -apoleon(s typical

    bodily stance in p"blic and abo"t the way in which people li7e or are rep"ted to li7e in San

    3rancisco &laidback leis"rely well$off/% The speed and apparent ease with which

    experimental participants "nderstand these neologisms s"ggests that lexical$pragmatic

    processes apply 'on$line( in a flexible creati7e and context dependent way%

    A f"rther range of examples that a theory of lexical pragmatics might help to explain

    are p"n$like cases in7ol7ing an element of e!"i7ocation or word play% Consider &/:

    % -ot all banks are ri7er banks%

    =ost 0nglish hearers int"iti7ely "nderstand this as both tr"e and informati7e: that is they

    int"iti7ely interpret 'bank( as picking o"t both the set of ri7er banks and the set of financial

    instit"tions% The !"estion is how the interpretation is best explained% *ne possibility is to go

    metaling"istic and treat 'bank( here as representing the set of things that are called 'banks(%

    An alternati7e possibility s"ggested by the ad hoc concept approach is to treat the

    9;

  • 8/10/2019 1 Wilson & Carston 07 Advances

    14/42

    interpretation as in7ol7ing the online constr"ction of an ad hoc concept BA-.8 whose

    denotation incl"des both the set of ri7er banks and the set of financial instit"tions% This

    approach might shed some light on the common "se of p"n$like comparisons s"ch as those in

    &9F/ in which ad#ecti7es s"ch as 'cold( and 'hard( ha7e to be sim"ltaneo"sly "nderstood in

    both physical and psychological senses:

    9Fa% +is mind was as coldas the ice forming on the windscreen%

    9Fb% +is eyes were as coldas polar ice%

    9Fc% +is 7oice was low and as coldas steel%

    9Fd% Kane is as hardas nails%

    9Fe% S"e is as toughas old leather%

    9Ff% Kimmy is assharpas a knife%

    As the range of cases s"r7eyed here indicate lexical ad#"stment may bea one$off

    process "sed once and then forgotten creating an ad hoc concept tied to a partic"lar context

    that may ne7er occ"r again &a 'nonce( sense as Clark &9;/ p"ts it/% +owe7er some of these

    pragmatically constr"cted senses may catch on in the comm"nicati7e interactions of a few

    people or a gro"p and so become reg"larly and fre!"ently "sed% ,n s"ch cases the pragmatic

    process of concept constr"ction becomes progressi7ely more ro"tinised and may "ltimately

    spread thro"gh a speech comm"nity and stabilise as an extra lexical sense% We wo"ld

    therefore expect a "nified acco"nt of lexical pragmatic processes to shed light on the nat"re

    of polysemy &the fact that many words ha7e a range of distinct tho"gh related senses/ and on

    3or a disc"ssion of the processes that acco"nt for the 'do"ble f"nction( &physical and psychological/ of these

    and a wide range of other ad#ecti7es see Wilson and Carston &EFF/ and for an interesting hypothesis abo"t the

    "nitary concept"al basis "nderlying the d"al "ses see Asch &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    15/42

    processes of lexical change more generally%9F,n fact it is often pointed o"t that pragmatic

    processes of broadening narrowing and metaphorical extension play a ma#or role in semantic

    change &)yons 9I Tra"gott and Dasher EFF9/% ,f o"r "nified acco"nt is correct the

    res"lting senses sho"ld all be seen as o"tcomes of the fre!"ent and widespread application to

    a partic"lar lexical item of a single pragmatic process of ad hoc concept constr"ction%

    +& Approaches to lexical ad0stment

    =any pragmatic or philosophical acco"nts seem to take for granted that narrowing

    approximation and metaphorical extension are distinct pragmatic processes which lack

    common descriptions or explanations and need not be st"died together% 3or instance

    narrowing is often analysed as a case of defa"lt inference to a stereotypical interpretation99

    approximation has been seen as linked to 7ariations in the standards of precision go7erning

    different types of disco"rse9Eand metaphor is still !"ite widely treated on @ricean lines as a

    blatant 7iolation of a maxim of tr"thf"lness with res"lting implicat"re%9;These acco"nts do

    not generalise: metaphors are not analysable as ro"gh approximations narrowings are not

    analysable as blatant 7iolations of a maxim of tr"thf"lness and so on% Separate analyses of

    9F *f co"rse a f"ll explanation of lexical semantic change wo"ld re!"ire in addition to the acco"nt of lexical

    pragmatic processes an acco"nt of how and why certain representations spread &catch on/ in a comm"nity or

    c"lt"re while others do not% 3or a nat"ralistic approach to the 'epidemiology( of representations &incl"ding

    semantic representations/ which takes acco"nt of a 7ariety of contrib"ting mental and en7ironmental

    phenomena see Sperber &9/%

    99 See e%g% )e7inson &EFFF/ Bl"tner &9 EFFH/% 3or disc"ssion see +orn &9H this 7ol"me/ )akoff &9/%

    9E See e%g% )ewis &9/ )asersohn &9/% 3or disc"ssion see @ross &EFF9/%

    9; See e%g% @rice &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    16/42

    approximations and fig"rati7e "tterances co"ld be #"stified by showing that there is a sharp

    bo"ndary between them b"t as examples &Ha/$&Hc/ and &a/$&c/ were designed to show in

    many cases no clear c"t$off point exists% We ha7e arg"ed in some detail elsewhere that there

    are also internal descripti7e and theoretical reasons for wanting to go beyond these existing

    acco"nts and de7elop a more "nified approach &see Carston 9 EFFEI Wilson G Sperber

    EFFEI Wilson EFF;/% We will not repeat those points here b"t ass"me that the concl"sion they

    point to is correct: that is that narrowing and broadening &incl"ding metaphorical cases/ are

    complementary processes one restricting and the other extending the category denoted by the

    ling"istically$encoded concept so that a "nitary acco"nt is well worth p"rs"ing%

    ,n the next section we will o"tline the rele7ance$theoretic 7iew that lexical

    comprehension in7ol7es a process of ad hoc concept constr"ction based on information

    readily accessible from the encyclopaedic entries of the encoded concepts and constrained by

    expectations of rele7ance%9H,n the rest of this section we want to establish two preliminary

    points: first that lexical pragmatic processes s"ch as narrowing and broadening contrib"te to

    tr"th$conditional content &what is asserted or explicated/ rather than merely affecting

    implicat"res as on standard @ricean acco"ntsI second that c"rrent approaches which agree

    on this point differ significantly on the nat"re of the cogniti7e mechanisms in7ol7ed and

    specifically on the extent to which they are properly inferential%

    3irst some arg"ments s"pporting the tr"th$conditional 7iew% 6erhaps the clearest

    e7idence comes from neologisms s"ch as those exemplified in &a/$&d/% 1erbs s"ch as

    9H The acco"nts of Barsalo" &9/ and @l"cksberg &EFF9 EFF;/ share the ass"mption that encyclopaedic

    information associated with a mentally$represented category or concept may be "sed to restrict or extend its

    denotation in an ad hoc occasion$specific way and they both mention the role of considerations of rele7ance in

    selecting an appropriate set of attrib"tes% +owe7er neither is aiming to de7elop a "nified pragmatic acco"nt:

    Barsalo" is mainly concerned with narrowing while @l"cksberg is mainly concerned with broadening &and

    specifically with metaphor and category extension/ and neither offers detailed s"ggestions abo"t what factors

    trigger lexical$pragmatic processes what direction they take and what makes them stop%

    9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    17/42

    'porch( 'wrist( and 'teapot( were coined for experimental p"rposes and ha7e no encoded

    meanings in 0nglish: th"s if the ad hoc concepts 6*RC+8 WR,ST8 and T0A6*T8 do not

    contrib"te to the proposition expressed or asserted by these "tterances there is no proposition

    expressed or asserted at all% This r"ns co"nter to the clear and widespread int"ition that these

    "tterances can be "sed to make assertions in the reg"lar way and wo"ld make it 7ery diffic"lt

    to pro7ide an ade!"ate acco"nt of how partic"lar implicat"res are warranted in these cases%

    =ore generally any analysis of how "tterances are "sed in comm"nication m"st take acco"nt

    of potential differences in the ac!"isition and organisation of lexical meanings among

    members of a speech comm"nity% A word that is familiar to the speaker of an "tterance and

    has a reg"lar encoded meaning in 0nglish may nonetheless be "nfamiliar to the hearer who

    will be forced to interpret it inferentially along similar lines to neologisms% Con7ersely a

    word that is "nfamiliar to the speaker and prod"ced as a neologism may be one the hearer has

    enco"ntered before and assigned a reg"lar encoded meaning% @i7en s"ch differences the idea

    that only encoded concepts can contrib"te to the tr"th$conditional content of "tterances

    appears rather arbitrary and "nworkable%

    There is also e7idence that it is the pragmatically ad#"sted meaning of a word or

    phrase rather than the ling"istically encoded meaning that falls within the scope of sentence

    operators s"ch as negation conditionals dis#"nction 'beca"se( imperati7e and interrogati7e

    moods etc% Consider the examples in &99/:

    99a% -o teenager is asaint%

    99b% ,f the bottle is empty lea7e it o"t for recycling%

    99c% Be an angeland pick "p the shopping for me on the way home%

    99d% 0ither yo" become a human beingor yo" lea7e the gro"p%

    9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    18/42

    ,n &99a/ if the encoded meaning of 'saint( is interpreted as falling within the scope of

    negation the speaker will be "nderstood as making the tri7ial claim that no teenager has been

    canonised% ,n fact the speaker of &99a/ wo"ld generally be "nderstood as making the more

    pla"sible informati7e or rele7ant claim that no teenager falls in the category of SA,-TS8 &that

    is ro"ghly people of o"tstanding 7irt"e/% B"t in that case it is the ad#"sted meaning rather

    than the encoded meaning that falls within the scope of negation and contrib"tes to the

    proposition expressed% Similar points apply to the other examples% ,n the case of &99d/ for

    instance gi7en that the addressee is a h"man being already it wo"ld be impossible for him to

    comply with the first dis#"nct literally "nderstood &and therefore pointless for the speaker to

    ask him to comply with it/% Clearly what the hearer is being enco"raged to do as an

    alternati7e to lea7ing the gro"p is to start beha7ing in a more reasonable or sensiti7e way

    &that is to become a +5=A-B0,-@8/% B"t in that case it is the ad#"sted meaning rather than

    the encoded meaning that falls within the scope of the dis#"nction operator and contrib"tes to

    the proposition expressed% &3or disc"ssion and ill"stration of these 'embedding tests( applied

    to cases of lexical ad#"stment see Recanati 9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    19/42

    does not in7ol7e ad hoc concept constr"ction b"t is a simple matter of disambig"ation

    &choosing which of two or more encoded senses sho"ld fig"re in the proposition expressed/%

    3or others who ha7e enco"ntered these metaphors less fre!"ently or not at all 'saint( and

    'angel( may ha7e only a single encoded sense &SA,-T A-@0)/ and the interpretation of &99a/

    and &99c/ wo"ld in7ol7e constr"cting &or re$constr"cting/ an appropriate ad hoc concept% ,n

    the first case where 'saint( and 'angel( are gen"inely polysemo"s and interpreted 7ia

    disambig"ation there is no do"bt that the encoded senses SA,-T8 and A-@0)8 wo"ld

    contrib"te to the tr"th$conditional content of &99a/ and &99c/ and hence to what is asserted or

    explicated% B"t it is impla"sible to s"ppose that &99a/ and &99c/ wo"ld be "nderstood as

    expressing or asserting entirely different propositions depending on whether the concepts

    SA,-T8 and A-@0)8 are reco7ered by disambig"ation or ad hoc concept constr"ction% *ne of

    the most important f"nctions of pragmatic inference is to compensate for grammatical and

    lexical differences among members of a speech comm"nity so that addressees with different

    encoded senses can end "p with the same interpretations albeit 7ia different ro"tes%9Th"s

    all the a7ailable e7idence points to the concl"sion that ad hoc concepts contrib"te to the tr"th$

    conditional content of "tterances rather than merely to implicat"res%

    As noted abo7e &section 9/ c"rrent tr"th$conditional acco"nts of lexical pragmatics

    differ significantly as to the cogniti7e mechanisms in7ol7ed in the lexical ad#"stment process%

    At one extreme are f"lly inferential acco"nts which treat "tterance interpretation in general

    and lexical ad#"stment in partic"lar as properly inferential processes taking a set of premises

    as inp"t and yielding as o"tp"t a set of concl"sions logically deri7able from or at least

    warranted by the premises% According to rele7ance theory for instance the interpretation of

    9 See R"bio 3ernande4 &EFF

  • 8/10/2019 1 Wilson & Carston 07 Advances

    20/42

    &

  • 8/10/2019 1 Wilson & Carston 07 Advances

    21/42

  • 8/10/2019 1 Wilson & Carston 07 Advances

    22/42

    Between these two extremes lie mixed associati7einferential approaches: for

    instance Recanati &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    23/42

    inferential rele7ance$theoretic acco"nt of the fast on$line pragmatic process of lexical

    ad#"stment% At the end of the chapter we ret"rn to the iss"e of inferential 7ers"s non$

    inferential processes and consider some possible ob#ections to a "nified inferential acco"nt%

    1& An inferential accont of lexical ad0stment

    An ade!"ate acco"nt of lexical narrowing and broadening m"st answer fo"r !"estions:

    I& What triggers the lexical ad#"stment process &why not simply accept the encoded

    sense/P

    II& What determines the direction that the ad#"stment process takesP

    III& +ow does the ad#"stment process work in detailP

    I/& What brings it to an endP

    Rele7ance theory treats lexical narrowing and broadening like "tterance interpretation in

    general as g"ided by expectations of rele7ance% Rele7ance is defined as a property of

    "tterances and other inp"ts to cogniti7e processes &e%g% external stim"li s"ch as sights and

    so"nds and internal representations s"ch as tho"ghts memories or concl"sions of

    inferences/% An inp"t is relevantto an indi7id"al when it connects with a7ailable context"al

    ass"mptions to yieldpositi7e cogniti7e effects &e%g% tr"e context"al implications warranted

    strengthenings or re7isions of existing ass"mptions/% 3or present p"rposes the most

    important type of cogniti7e effect is a contextual implication: an implication ded"cible from

    inp"t and context together b"t from neither inp"t nor context alone% *ther things being e!"al

    the greater the cogniti7e effects and the smaller the mental effort re!"ired to deri7e them &by

    E;

  • 8/10/2019 1 Wilson & Carston 07 Advances

    24/42

    representing the inp"t accessing a context and deri7ing any context"al implications/ the

    greater the rele7ance of the inp"t to the indi7id"al at that time%9

    Rele7ance theory makes two general claims abo"t the role of rele7ance in cognition

    and comm"nication% According to the Cogniti7e 6rinciple of Rele7ance h"man cognition

    tends to be geared to the maximisation of rele7ance% According to the Comm"nicati7e

    6rinciple of Rele7ance "tterances &and other acts of ostensi7e comm"nication/ are special

    among cogniti7e inp"ts in that they raise pres"mptions or expectations of rele7ance in their

    addressees% Typically an "tterance creates both a general pres"mption of optimal relevance

    &that the "tterance is at least rele7ant eno"gh to be worth the addressee(s processing effort

    and is moreo7er the most rele7ant one compatible with the speaker(s abilities and

    preferences/ and more occasion$specific expectations abo"t where the rele7ance of the

    "tterance will lie &what sort of context"al implications it will ha7e/% The central claim of the

    rele7ance$based acco"nt of pragmatic processing is that addressees take the fact that the

    speaker has "ttered a sentence with a certain ling"istic meaning as a cl"e to the speaker(s

    intentions and "se the following he"ristic to deri7e a warranted concl"sion abo"t the

    speaker(s meaning:

    Relevance$theoretic comprehension heristic

    &a/ 3ollow a path of least effort in constr"cting an interpretation of the "tterance &that is test

    interpreti7e hypotheses in order of their accessibility/I

    &b/ Stop when yo"r expectation of rele7ance is satisfied &or abandoned/%

    According to this he"ristic at each point in the on$line processing of an "tterance the

    addressee tentati7ely chooses the most accessible interpretation and reconsiders this choice

    9

    3or m"ch f"ller exposition of rele7ance theory and comparison with alternati7e approaches see Sperber G

    Wilson &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    25/42

    only if it seems "nlikely &on the basis of the a7ailable e7idence/ to lead to an o7erall

    interpretation that satisfies his expectation of rele7ance% The same proced"re applies to the

    f"ll range of pragmatic tasks: assigning referents to referential expressions disambig"ating

    ambig"o"s words or str"ct"res s"pplying context"al ass"mptions deri7ing implications etc%

    Th"s the fact that an interpretation is highly accessible gi7es it an initial degree of

    pla"sibility% A hearer "sing this he"ristic will stop at the first o7erall interpretation that

    satisfies his expectation of rele7ance: this is his best hypothesis abo"t the speaker(s meaning

    gi7en the e7idence a7ailable to him%

    This proced"re applies e!"ally to the ad#"stment of lexical meaning which &following

    the arg"ments of the last section/ we will treat as contrib"ting both to the proposition

    expressed by an "tterance and to its context"al implications or implicat"res &i%e% intended

    context"al implications/% Consider how the 7erb 'rest( might be "nderstood in the following

    exchange:

    9;% Bill: ,(m doing the 9Fkm circ"it r"n this afternoon% Wanna come with meP

    S"e: -o thanks ,(m restingtoday%

    The 7erb 'rest( has a rather general meaning which co7ers any degree of inacti7ity &physical

    or mental/ from sleeping to staying awake b"t not mo7ing m"ch to performing a range of not

    7ery stren"o"s tasks &with many more possibilities in between/% S"ppose now that S"e is

    !"ite an athletic person who exercises reg"larly: then her "se of 'rest( here is pla"sibly

    "nderstood as expressing the ad hoc concept R0ST8 which indicates a m"ch lower degree of

    physical acti7ity than she "ndertakes on her training days b"t is still !"ite compatible with

    her pottering abo"t the garden or walking to the shops% A hearer "sing the rele7ance$theoretic

    comprehension he"ristic wo"ld narrow the encoded concept R0ST#"st so far as is re!"ired to

    E

  • 8/10/2019 1 Wilson & Carston 07 Advances

    26/42

  • 8/10/2019 1 Wilson & Carston 07 Advances

    27/42

    A-@0) it is ded"cible that Sally is a S560R-AT5RA)B0,-@*3AC0RTA,-.,-D/% The decoded

    concept A-@0)also acti7ates a 7ariety of more or less strongly e7idenced encyclopaedic

    properties of different s"bsets of angels &good angels g"ardian angels a7enging angels dark

    angels fallen angels and so on/ enabling f"rther concl"sions to be drawn &e%g% the

    proposition that Sally is a &good/ angel if processed in a context containing the ass"mption

    that &good/ angels are exceptionally kind context"ally implies that Sally is exceptionally

    kind/% Some pla"sible encyclopaedic properties of &good/ angels are gi7en in &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    28/42

    SA))2,SWATC+35)A-DW,))+0)6W+0--00D0D

    SA))2W,)))**.A3T0RT+0C+,)DR0-,360T0RA-D=AR2@0T,))%

    *f co"rse Sally is not a s"pernat"ral being and therefore not an A-@0) so the context"al

    implications in &9/ are not yet properly warranted% +owe7er by narrowing the denotation of

    A-@0)to incl"de only good angels and broadening it to incl"de people who share with good

    angels some of the encyclopaedic properties in &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    29/42

    additional context"al ass"mptions/ and they in t"rn pro7ide the basis for a 'backward(

    inference to the ad#"sted propositions &Sallyxis an A-@0)8 A-@0)S8 are exceptionally kind

    &etc%// which #"stify their acceptance as part of an o7erall interpretation that satisfies the

    hearer(s expectations of rele7ance%EFBoth narrowing and broadening emerge as by$prod"cts

    of the search for rele7ance and the same encoded concept may be narrowed or broadened &or

    both/ to different degrees and in different ways across different occasions of "se%

    We s"ggested abo7e &section E/ that 'approximation( 'hyperbole( 'metaphor( are not

    distinct theoretical kinds re!"iring different interpreti7e mechanisms b"t merely occ"py

    different points on a contin""m of degrees of broadening% To ill"strate this point in more

    detail consider &9/:

    9% The water is boiling%

    This "tterance might be intended and "nderstood literally or as an approximation a

    hyperbole or a metaphor with no clear c"t$off points between these 7ario"s possibilities% *n

    the rele7ance$theoretic acco"nt o"tlined abo7e all these interpretations are arri7ed at in the

    same way by adding to the context encyclopaedic information made accessible by the

    encoded concept B*,),-@&and by other concepts acti7ated by the "tterance or the disco"rse/

    and deri7ing eno"gh implications to satisfy the hearer(s expectations of rele7ance% What

    makes the res"lting interpretation int"iti7ely 'literal( 'approximate( 'hyperbolical( or

    'metaphorical( is simply the partic"lar set of encyclopaedic ass"mptions act"ally deployed in

    making the "tterance rele7ant in the expected way%

    EF

    *n the role of 'backward( inference in the m"t"al ad#"stment process see Sperber and Wilson &9/ Wilson

    and Sperber &EFFE/ Carston &EFFE/ Wilson G Carston &EFF/%

    E

  • 8/10/2019 1 Wilson & Carston 07 Advances

    30/42

    )et(s s"ppose that the encyclopaedic properties sim"ltaneo"sly acti7ated by both

    'water( and 'boiling( &and therefore potentially highly accessible for the interpretation of &9//

    incl"de those in &9a/$&9d/:

    9% B*,),-@WAT0R: 2ncyclopaedic propertiesE9

    a% S00T+0SA-DB5BB)0S +,DD0-5-D0RC5RR0-TS 0=,TS1A6*5R etc%

    b% T**+*TT*WAS+*-0(S+A-DS,- T**+*TT*BAT+0,- etc%

    c% S5,TAB)03*R=A.,-@T0A DA-@0R*5ST*T*5C+ etc%

    d% SA30T*5S0,-ST0R,),S,-@,-STR5=0-TS etc%EE

    Then &9/ wo"ld be int"iti7ely 'metaphorical( if the implications that make the "tterance

    rele7ant in the expected way depend on &9a/ b"t not on &9b/$&9d/ &so that the speaker is

    not "nderstood as committed to the claim that the water is hot/I it wo"ld be int"iti7ely a

    'hyperbole( if these implications depend on &9b/ b"t not on &9c/$&9d/I it wo"ld be an

    'approximation( if these implications depend on &9c/ b"t not on &9d/ and it wo"ld be

    'literal( if the deployment of &9d/ is cr"cial to making the "tterance rele7ant in the expected

    E9 To sa7e space we present these simply as properties rather than as complete propositions% +owe7er since the

    f"nction of encyclopaedic information is to pro7ide premises for the deri7ation of context"al implications each

    property sho"ld be seen as a constit"ent of a complete proposition%

    EE +ere the 'etc%( is intended to co7er encyclopaedic properties of strictly B*,),-@water that do not hold for

    broader interpretationsI in &9c/ it co7ers encyclopaedic properties that hold both for strictly B*,),-@water and

    for water that is almost B*,),-@&i%e% B*,),-@8/ b"t not for water that is B*,),-@88 or B*,),-@888I and so on

    for &9b/ and &9a/% We are not claiming of co"rse that encyclopaedic information is neatly organised in this

    way: merely that the choice of a partic"lar set of ass"mptions in the co"rse of the m"t"al ad#"stment process

    will determine whether the "tterance is int"iti7ely 'literal( 'approximate( 'metaphorical( and so on%

    ;F

  • 8/10/2019 1 Wilson & Carston 07 Advances

    31/42

    way &so that the denotation of the concept expressed incl"des only items that are act"ally

    B*,),-@/% ,n each case the comprehension process works in the same way by selection of an

    appropriate set of encyclopaedic ass"mptions to act as premises for the deri7ation of the

    expected context"al implications% The appropriateness of different sets of encyclopaedic

    ass"mptions depends on the one hand on their degree of accessibility in the partic"lar

    disco"rse context and on the other on the potential context"al implications they yield% As

    always the hearer(s goal is to deri7e eno"gh implications at a low eno"gh cost to satisfy the

    partic"lar expectations of rele7ance raised by the "tterance in that disco"rse context%

    ,n this section we ha7e proposed a f"lly inferential acco"nt of lexical narrowing and

    broadening which answers the fo"r basic !"estions of lexical pragmatics as follows:

    I& -arrowing and broadening are triggered by the search for rele7ance%

    II& They follow a path of least effort in whate7er direction it leads%

    III& They come abo"t thro"gh mtal ad0stmentof explicat"res context"al ass"mptions

    and implications &or implicat"res/ so as to satisfy the expectations of rele7ance raised by

    the "tterance%

    I/& They stop when these expectations are satisfied%

    As s"ggested abo7e &section ;/ this acco"nt treats "tterance comprehension as do"bly

    inferential: it consists of an o7erall non$demonstrati7e inference process in which the

    ded"cti7e inference processes re!"ired to deri7e the context"al implications that satisfy the

    hearer(s expectations of rele7ance play an essential role% *7erall comprehension starts from

    the premise that the speaker has "ttered a sentence S &e%g% 'Sally is an angel(/ with a certain

    encoded meaning and arri7es at the warranted concl"sion that the speaker meant that 69J6n

    &e%g% that Sallyxis an A-@0)8 that she is good kind watchf"l and will help look after the

    ;9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    32/42

    children/% This o7erall non$demonstrati7e process is carried o"t by the rele7ance$theoretic

    comprehension he"ristic a domain$specific form of inference #"stified only for o7ert

    comm"nication &Sperber G Wilson EFFE/% ,n order to arri7e at a warranted concl"sion abo"t

    the speaker(s meaning the hearer m"st perform a range of ded"cti7e inferences &e%g% from

    A-@0)A-@0)8 to 10R2@**DA-D.,-D WATC+35) +0)635)&etc%// in order to deri7e

    eno"gh context"al implications to satisfy his expectations of rele7ance% According to

    rele7ance theory these two types of inference are intimately connected% ,t is only by deri7ing

    &ded"cti7e/ context"al implications as in &9/ abo7e that the hearer can #"stify a partic"lar

    hypothesis abo"t the speaker(s meaning as the best explanation of the fact that she "ttered

    sentence S%By contrast non$inferential acco"nts of lexical ad#"stment in7ol7e neither type of

    inference while mixed associati7einferential approaches s"ch as Recanati(s &9

  • 8/10/2019 1 Wilson & Carston 07 Advances

    33/42

    ,n this final section we will briefly consider some broader !"estions abo"t inferential

    acco"nts of lexical ad#"stment% While some &in o"r 7iew/ can be straightforwardly answered

    others raise gen"ine iss"es for f"t"re research%

    An ob#ection sometimes raised to inferential acco"nts of metaphor interpretation is

    that they cannot handle the so$called emergent property iss"e% To ill"strate consider the

    "tterance in &9/:

    9% That s"rgeon sho"ld be dismissed% +e is a butcher%

    The speaker of &9/ is pla"sibly "nderstood as con7eying that the s"rgeon in !"estion is

    extremely incompetent dangero"s not to be tr"sted with the li7es of patients and so on%

    These properties are not standardly associated with either S5R@0*-or B5TC+0Rin isolation

    b"t 'emerge( in the co"rse of the interpretation% 0mergent properties raise a problem for all

    acco"nts of metaphor interpretation% The challenge for inferential acco"nts is to show how

    they can be inferentially deri7ed &see Carston EFFE chapter

  • 8/10/2019 1 Wilson & Carston 07 Advances

    34/42

    ,n recent rele7ance theoretic acco"nts it has been arg"ed that emergent properties are

    analysable as gen"ine context"al implications which emerge in the co"rse of the m"t"al

    ad#"stment process based on context"al premises deri7ed from se7eral so"rces%EH,n &9/ for

    instance the speaker may be "nderstood as asserting that the s"rgeon in !"estion is a

    B5TC+0R8 where B5TC+0R8 is a reg"lar ad#"sted concept based on encyclopaedic

    information associated with the encoded concept B5TC+0R denoting people who share with

    b"tchers the encyclopaedic property of c"tting flesh in a certain way &"sing the same

    techni!"es as b"tchers with the same intentions concern for welfare degree of skill and so

    on/% 3rom the proposition that the s"rgeon is a B5TC+0R8 together with encyclopaedic

    information associated with the encoded concept S5R@0*- it follows straightforwardly that

    the s"rgeon in !"estion is grossly incompetent dangero"s not to be tr"sted with the li7es of

    patients and deser7es to be dismissed% Context"al implications of this type are highly

    accessible in the disco"rse context in &9/ and wo"ld help to satisfy the hearer(s expectations

    of rele7anceI they are therefore likely to be accepted as implicat"res by a hearer "sing the

    rele7ance$theoretic comprehension he"ristic% This acco"nt is gen"inely inferential and the

    'emergent properties( are straightforwardly analysable as both context"al implications and

    implicat"res%E