Experimental Pragmatics

Experimental Pragmatics

Ira Noveck

Laboratoire sur le Langage, le Cerveau et la Cognition (L2C2) &

Centre de Recherche Français à Jérusalem (CRFJ),

Centre National de Recherche Scientifique (CNRS)

&

Nicola Spotorno

Frontotemporal Degeneration Center,

University of Pennsylvania

Introduction

Imagine sitting at a bar where someone woefully says, “Some jobs are jails.”

This can be readily understood to implicate that some jobs are confining (or

dispiriting, demoralizing etc.) and, if one wants to be even more precise, it could be

taken to mean some but not all jobs are confining and so on. Both of these

propositions go beyond what was literally said and -‐-‐ remarkably -‐-‐ arriving at

these interpretations appears relatively routine. However, as is the case for many

everyday experiences, it is a challenge to understand how it works. This explains,

at least partly why there is a field of study, linguistic-‐pragmatics, devoted to

investigating how such intended interpretations come about. Experimental

Pragmatics can be viewed as a subdiscipline devoted to testing and advancing

pragmatic accounts through psychological experiments.

As has been described elsewhere (Sperber & Noveck, 2004; Noveck &

Sperber, 2007), the approaches that inspire or are tested by Experimental

Pragmatics accept as foundational two ideas that were defended by the

philosopher Paul Grice. One is that interlocutors share the goal of having the

speaker’s meaning—as opposed to the sentence meaning—recognised by the

hearer and that they use utterances to that end. The second foundational idea is

that, in inferring the speaker’s meaning on the basis of the decoding of her

utterance on the one hand and of contextual information on the other, the hearer is

guided by the expectation that the utterance should meet some standards. The

standards Grice envisaged were based on the idea that a conversation is a

cooperative activity and that interlocutors are guided by expectations of

informativeness, truthfulness, relevance, and so on. Over the course of the chapter,

it will become clear how these Gricean ideas and distinctions infuse this literature.

One generally accepted distinction is between what is said in a sentence and

what the speaker means by saying it. The gulf between the two can be made more

explicit by distinguishing between linguistic decoding and pragmatic inferencing.

As Robyn Carston (2002) put it in describing the underdeterminacy thesis there is a

great distance between the linguistic semantics of the utterance and the speaker’s

meaning:

What is meant by this (the underdeterminacy thesis) is that

the linguistic semantics of the utterance, that is, the meaning encoded

in the linguistic expressions used, the relatively stable meanings in a

linguistic system, meanings which are widely shared across a

community of users of the system, underdetermines the proposition

expressed (what is said). The hearer has to undertake processes of

pragmatic inference in order to work out not only what the hearer is

implicating but also what proposition she is directly expressing.

As will become clear, this way of describing the semantic-‐pragmatic divide is

central to all of the topics in Experimental Pragmatics.

In the first part of this chapter, we will summarize research on scalar

inferences, which concerns the way relatively weak expressions such as some

(which is compatible with all but can be interpreted as some but not all) are

interpreted. This will be divided into two parts, corresponding to two phases in

scalar inference research. The first showed that participants’ responses can be

used to experimentally distinguish between semantic and pragmatic readings of

weak scalar utterances (such as the one in 1a below). The second, more current,

line of research is devoted to determining how pragmatic inferences are folded

into speaker meaning in real time. We will then turn to referring expressions (e.g.

Point to the big star), which shares many, though not all, of the concerns of the

scalar literature. Finally, the last two sections address figurative language as it

summarizes work on metaphor and irony.

As will become clear, each of these areas is in a different stage of

development and has a unique history. Scalar inferences – which is the drosophila

of experimental pragmatic research – is arguably the most developed area in terms

of integrating current theoretical concerns with experimentation. Other topics,

such as reference and metaphor, have been the focus of research in the

psycholinguistic literature for some time, but their link to linguistic-‐pragmatic

theory-‐making has been more tenuous and partly because experimental work on

these topics predated the Gricean turn (see Noveck and Reboul 2008). Irony

represents yet another unique case because from early on innovative pragmatic

theoretical discussions gave rise to original experimental work about a role for

attitude-‐ascription, but many of the experiments that followed gave way to

research that focused on issues concerning literal meaning. This has resulted in a

disconnect between the (early) theoretical discussions on irony and the extant

experimental psychological literature. Before ending, we will complete our

panorama with brief discussions of other related areas.

Scalars

2.1 Background

Imagine you organized a night out at a restaurant for yourself and friends

when the waiter asks whether your guests enjoyed the meal. Now, imagine that

you respond by saying (1a):

(1a) Some of the guests enjoyed the meal.

While the semantic meaning of (1a) is compatible with everyone having enjoyed

the meal, your choice of words readily conveys information of the sort found in

(1b):

(1b) Not all of the guests enjoyed the meal.

It is this sort of inference, which goes beyond what was explicitly said, that has

been the focus of much work in experimental pragmatics.

As one can see, there is a distinction to be made between the linguistically

encoded meaning and the speaker’s meaning. That is, one can determine a sentence

meaning from an utterance by decoding that sentence, i.e. discovering the semantic

properties that the grammar pairs to its acoustic form. Retrieving the speaker’s

meaning involves attributing to the speaker an intention. More technically, the

speaker’s intention is to produce a cognitive effect in her audience and to cause

that audience to recognize that very intention. Although linguistic communication

is partly code-‐based, it cannot be reduced to a mere encoding-‐decoding process. It

involves the attribution of mental states to the speaker.

According to Grice’s original account (1975), a listener decodes the literal

meaning of an utterance (such as 1a) as a step in a process that ultimately arrives

at the speaker’s intended meaning. Given that interlocutors are guided by

expectations of informativeness, truthfulness, relevance, and so on, a violation of

an expectation would compel a listener to look for alternative interpretations that

lead to the speaker’s intended meaning. In the example above, such an expectation

is expressed through the Maxim of Quantity by which speakers should: i) make

their contribution as informative as is required, and ii) not make their contribution

more informative than is required. When the literal meaning of an utterance does

not satisfy these expectations (e.g., when an utterance is either overinformative or

underinformative), hearers are entitled to assume that the speaker intended

something more by way of an implicature. Thus, when a speaker says (1a), it could

appear underinformative (since a speaker could have said All of the guests enjoyed

the meal). This sort of explanation remains dominant in one form or another today

(e;g., see Geurts, 2010).

The earliest studies to explore the above distinction with these sorts of

materials were developmental (meaning, children figured prominently in the

study). Noveck (2001) showed that 7-‐to-‐10-‐year-‐olds do not make scalar

inferences in reasoning tasks associated with some and might to the same degree

as adults. In one of Noveck’s (2001) experiments, most children accepted

sentences such as “Some giraffes have long necks” while adults tended to reject

them as false (on the grounds that all giraffes have long necks). This effect has

proven to be robust (with a wide variety materials across dozens of languages)

through multiple studies that specifically aimed to investigate it (e.g., Pouscolous et

al., 2007; Guasti et al., 2005). The paper also offered a theoretical framework for

appreciating classic developmental studies that (often inadvertently) investigated

equally underinformative cases, including three where or was used to describe

scenarios where and would have been more appropriate (Paris, 1973, Sternberg,

1984; Braine and Rumain, 1981). For example, Paris (1973) showed a picture of a

boy with a bicycle and monkey in a tree; 8-‐year-‐olds were more likely than older

children and adults to accept as true statements such as “There is a boy with a

bicycle or a monkey in a tree.” While many researchers’ intuitions considered

exclusive or more ordinary for cases like this, the developmental data showed that

younger children were more willing to accept logical interpretations of or than

adults. Collectively, these developmental studies represented the opening of the

field of experimental pragmatics by seeking to, not only uncover novel robust

phenomena but to, account for lingering (and what had been mysterious) data in

the literature.

More intense discussion arises once theorists describe how this process is

carried out. Aside from approximations based on the outlines of Grice’s theory,

there are several post-‐Gricean explanations for such conversational phenomena.

Levinson (2000), for one, argues that there are inferential heuristics (or default

inferences) that make utterances informative (e.g. the Q-‐heuristic What isn’t said,

isn’t is applied to the utterance), transforming the linguistically encoded meaning

in (1a) to (1b)). That is, inferences are automatically generated and can be

cancelled if there are contextual reasons to do so. This has the advantage of making

the inference a relatively light one-‐step process, which needs to access neither

contextual premises nor the full Gricean rationale for their derivation.

Another prominent account comes from Relevance Theory (RT), which

describes how intended meanings are the result of the interplay between cognitive

effects and effort in a particularized way each time (Sperber and Wilson

1986,1995). According to Sperber and Wilson’s (1995) principle of Relevance,

determining the range of meaning of the quantifier some arises as part of

understanding the speaker’s intended meaning and that meaning is always

inferred (even when it consists in a literal interpretation of the linguistic

expressions used). The inferences involved, however, differ in the time and effort

they require. Both the sentence meaning and the context contribute to making

some interpretations more easily derived than others. If only sentence meaning

were involved, one prediction would be that the smaller the distance between it

and the speaker’s meaning, the lesser would be the time and effort required to

infer the speaker’s meaning (Noveck and Sperber, 2007).

Another middle-‐way strategy for uncovering scalar inference-‐making has

been to break it down into identifiable stages (Geurts, 2010; Breheny et al., 2013;

Barner et al., 2011). This has led to proposing substeps that include a) producing a

scale, or set, of alternative terms or propositions that contain a stronger

alternative, and b) assuming that the speaker is well informed about access to the

stronger alternative, so that; c) the listener can determine that the speaker is not

using the stronger alternative because she knows it is not true or because she lacks

evidence for it. These are important assumptions to keep in mind as we turn to the

processing question later.

This summary does not exhaust the range of recent discussions in the

“scalar” literature. For example, Katsos and Bishop’s (2008) pragmatic tolerance

hypothesis proposes that differences between children and adults with respect to

pragmatic abilities are due to children’s being more pragmatically tolerant than

adults (see also, Davies & Katsos, 2010). Davies and Katsos write that “[children]

do not consider pragmatic violations to be grave enough to warrant a rejection of

or objection to the offending utterance” (i.e. using some to describe a scenario that

would have been more appropriately described with all).

It is also important to point out that there is a class of accounts that

attribute narrowed readings to grammar (Chierchia 2001; see also, Chierchia, Fox

& Spector, to appear), i.e. it is assumed that scalar inference does not draw on

speakers’ intentions but on compositional semantics. According to this approach,

(1a) is understood as similar to its linguistically encoded reading when in a

downward entailing context (e.g. in question forms, under the scope of negation

etc.) and as (1b) otherwise. This is a largely non-‐Gricean way of viewing scalar

inferences and is the source of much discussion mostly in the semantics literature.

However, it has also led to cross-‐border discussions that concern the way Gricean

concepts are introduced into sentence processing (see Geurts, 2010).

2.2 How have experimentalists approached the processing question?

In a way analogous to the developmental data, evidence from sentence

processing has largely shown how extra processing is associated with the

narrowing of the more refined interpretations. While linguistically encoded

readings are often sufficient for making on-‐line interpretations, more refined

readings are associated with extra costs, even if they may be relatively light (e.g.,

see Bott, Bailey, and Grodner, 2012; Bott and Noveck, 2004; Breheny, Katsos, and

Williams 2007; De Neys and Schaeken 2007). For example, in a self-‐paced reading

task, examples of which are illustrated in (3) and (4), where the slashes indicate

where participants manually advance the text through a keyboard, Breheny et al.

showed that when a disjunction (class notes or summary) arises in an Upper-‐bound

context (as in 3), it takes longer to read because the context encourages a

narrowed reading that excludes the more informative conjunctive reading (class

notes or summary but not both), than in a Lower-‐bound context (in 4), where the

conjunctive reading remains compatible with the disjunctive phrase :

(3) Upper-bound context

John was taking a university course / and working at the same time. /

For the exams / he had to study / from short and comprehensive

sources./ Depending on the course, / he decided to read / the class

notes or the summary.

(4) Lower-bound context

John heard that / the textbook for Geophysics / was very advanced. /

Nobody understood it properly./ He heard that / if he wanted to pass

the course / he should read / the class notes or the summary.

That the disjunctive phrase takes significantly longer to process in (3) than in (4) is

consistent with results showing that narrowed readings are more effortful to

process than those that do not call for this sort of pragmatic enrichment.

It is not just contextual factors related to the sentence that prompt

narrowed readings. Cognitive factors -‐-‐ such as extra time or fewer distractions -‐-‐

affect narrowing as do individual differences. Bott and Noveck (2004), using

underinformative out-‐of-‐blue statements such as Some cows are mammals (to be

evaluated as true or false), showed how false responses (which are revealing of

narrowed readings of the sort not all cows are mammals) increased as allowed

response times did. More specifically, false responses were significantly more

likely when participants had up to three seconds to respond as opposed to roughly

one second to respond. Other more psychological features can also affect

narrowings. When participants are distracted by demanding secondary tasks they

are less likely to produce narrowings of statements with existential quantifiers (De

Neys & Schaeken, 2007). As far as individual differences are concerned, Noveck &

Posada (2003) showed that one can isolate groups of individuals with respect to

scalar inferences, with some seeming to prefer semantic readings and others

pragmatic readings.

Individual differences can be seen in on-‐line studies as well. Nieuwland et

al. (2010) showed that the moment at which underinformativeness becomes

apparent is generally not noticeable to most participants on line through EEG.1

Consider the word “lungs” in sentence (5), which is taken from Experiment 2 of

their study and is arguably underinformative at the point at which it is expressed

(since all people have lungs). Taking the participants as a group, lungs prompts an

EEG profile that is indistinguishable from the word “pets” in (6) when it is

1 Electroencephalography (EEG) refers to the study of electrical activity on the scalp. Evoked Response Potentials (ERP) studies investigate characteristic patterns in these electrical activities. As far as language processing is concerned, words (in the context of a sentence) provide characteristic components that are described by their negative- or positive-going voltage fluctuations and the time at which they occur in terms of milliseconds. Thus, felicitous sentences typically have components that can be described as P300’s or N400’s etc. on certain parts of the scalp. By introducing anomalies that are compared to control sentences, one can characterize these components and make claims about types of experimental sentences. For example, Kutas and Hillyard (1980) discovered how semantic anomalies give rise to a negative-going component that peaks about 400 msec (and most obviously on the central parietal part of the scalp) after the appearance of an inappropriate word, such as socks in John buttered his bread with socks. Since then, the N400 has been known as a marker of semantic processing.

expressed, even though the latter does not have the potential to appear

underinformative:

(5) Some people have lungs that require good care.

(6) Some people have pets that require good care.

However, another Experiment –– with the same paradigm –– shows that EEG

profiles appear more reactive when (a) participants are broken down into two

groups based on their “Communication Score” in Baron-‐Cohen’s Autism Quotient

(Baron-‐Cohen et al. 2001), where a higher score indicates that an individual is

higher on the Autism Spectrum, plus when; (b) a comma is inserted prompting a

sort of “clausal wrap-‐up.” For example, (5) and (6) were presented as (5’) and (6’):

(5’) Some people have lungs, which require good care.

(6’) Some people have pets, which require good care.

With these sentences, those participants who are considered more “pragmatically

skilled” (with low AQ scores) had EEG profiles with an N400 of greater magnitude

to the underinformative item in a sentence such as (5’) when compared to (6’); this

indicates that some sort of expectation was violated on line among this half of

participants. Those who scored higher on the scale had N400’s that were less

remarkable. The upshot here is that it can be demonstrated that a subgroup of

participants who appear to be more pragmatically skilled go further in narrowing

the meaning of a sentence on-‐line. The rest do not.

The above summary shows that the narrowing of (an utterance containing)

a weak scalar term is not a common occurrence nor automatic. Children are

generally less likely to carry them out than adults, adults do not necessarily carry

them out routinely and context is critical to making them occur. This provides the

grounds for discharging Levinson’s account that argues that the enriched reading

(the narrowing) occurs by default and that it can be potentially cancelled later.

This is an important development. It is hard to argue – based on a wide range of

data – that scalar enrichments arise by default. However, other accounts remain

and, in the meantime, new accounts have emerged. This leads to the literature’s

second more contemporary phase.

2.3 The on-‐line processing of scalars: How do pragmatic readings emerge?

Huang and Snedeker (2009) explicitly put forward the idea that semantic

readings are generally first (i.e., a definitional reading of “some” needs to be

processed and before pragmatic enrichments can take place). In order to test this

claim, they asked participants to point to a single character, among 4, based on

objects associated with it and the way a quantity was described under ambiguous

circumstances. For example, participants heard “Point to the girl that has some of

the socks.” Given the circumstances, this sentence can be definitively

disambiguated at its end because there were two candidates for “girl” -‐-‐ one with

(a subset of the) SOCKs in the trial (with the remaining socks being held by a boy)

and another with all the featured SOCCer balls. If the pragmatic interpretation

were to be accessed on line, then one would expect eyes to alight on the girl with

socks (because a fast narrowed reading of some ought to point to the girl with the

subset of items). Instead, Huang and Snedeker found that it took a remarkably long

time (600-‐800 msec) for participants to look at the girl who was associated with

the partitive some (the girl with socks in this example) when compared to control

items, e.g. a case where there was only one girl in the trial that was associated with

a set of objects (i.e. the other girl had nothing). Huang and Snedeker’s assumed

that the slowdown was due to the All case (the girl with all the SOCCer balls),

which arguably remains an attractive alternative for “the girl with some of the…”

until the participants got to the end of the sentence. The argument in support of

this approach is that one needs to process the word before further work can be

carried out, even if there are cascading bottom-‐up and top-‐down processes to

contend with.

There are some concerns about this influential paradigm because the study

used (a) quantities that were in the subitizing range and; (b) actual numbers (2, 3)

among its control items. Unlike in previous tasks, then, the potential acceptability

of some to represent all is further weighted down by the possible articulation of “2”

or “3” (see Degen & Tanenhaus, in prep). Although these additions can be viewed

as a kind of confound, it is also reveals an important fact: that the presence of

another attractive option – once it is made part of the paradigm’s common ground

-‐-‐ slows down the process of considering some as an adequate descriptor for a case

demonstrating all. This implies that slowdowns are due to having other attractive

options and not specifically due to a competition between some and its scalemate

all (which many assume follows from the production of this scale and the rejection

of the stronger sounding member). That numerals in clearly demarcated contexts

can appear as or more attractive as the quantifier all lends doubt to the claim that

scales are involved at all. The appeal of alternatives need not come from stronger

quantifiers, but from better descriptors (once they have raised their heads in the

context). It is arguably the case that a paradigm encourages participants to

anticipate certain alternative adjectives and there are some that are better fits than

others.

While aiming to show that one can get participants to interpret some

partitively without tempting participants to look at potential, alternative cases that

cover some (e.g. cases representing all), Grodner et al. (2010) used Huang and

Snedeker’s eye-‐tracking paradigm while presenting trials in which participants

saw, e.g., three girls each representing a particular case: One girl figure would have

no items associated with her, a second one would have all four mentioned items

(say, four soccer balls) and a third would have two of four socks associated with

her (there were also three boys, one of which would be associated with two of the

other socks). So, it should be noted that there are basically three choices – All

(expressed as “Alla”), Some (“Summa”), or None (“Nunna”) – each time; this makes

the paradigm unique because it concretizes a lower bounded case – as in Point to

the girl who has none of (pronounced nunna) the socks and makes the labeling of

Summa (as well as Alla and Nunna) quite straightforward. Another feature of the

paradigm is that its participants have several seconds to see and hear about the

distribution of items. Overall, Grodner et al. show that one can construct cases in

which partitive cases can be understood quickly under conditions when a) there

are three options of which the two others (All and None) are unambiguously

labeled and routinely employed; b) there are no other attractive options such as

numbers; c) when participants have a relatively long time to anticipate which of 10

possible statements could be made (there are 18 total possible combinations and 8

are removed from consideration in the critical trial by virtue of the trial’s

parameters), and d) when the meaning of Some is used uniquely (i.e., entrained) as

a partitive throughout in a paradigm with a fixed number of possible utterances.

2.4 At what point do mental states figure into scalar inferences?

Given that an enriched reading of a scalar expression ultimately means

having access to the speaker’s intention, another way to put it is that this means a

listener has access to the speaker’s epistemic state. Breheny and colleagues (2013)

have focused on the time at which a listener considers a speaker’s mental states as

a scalar term is voiced through an eye-‐tracking paradigm. For example, they had

participants listen to confederates describe a simple filmed scene in the following

way after witnessing a man (from the shoulders down) pour an entire jug of water

containing oranges into a bowl and part of another jug containing limes into

another bowl:

(7) The man has poured some of the water with limes into the bowl on tray A and

all of the water with oranges into the bowl on tray B.

The question is how early in the sentence do participants focus on the intended

jug? More specifically, does the word “some” in the above context provide enough

of a clue for the listener to fixate on the water with limes? In order to answer this

question, participants were provided baseline conditions too. For example, they

would see a man pour part of a jug of water containing limes and another part of

another jug containing oranges before receiving the following:

(8) The man has poured some of the water with limes into the bowl on tray A and

some of the water with oranges into the bowl on tray B.

In this case, (8), the word “some” ought not help identify the intended jug of water

until the word “limes” is stated and the data confirm that. Breheny et al. report that

participants were able to anticipate their looks to the correct target in the Some-

early condition (in 7) as soon as they heard the quantifier ‘some’ and long before

the disambiguating expression (“limes”/ “oranges”) was encountered. Their results

“suggest that there is no difference in the time course of access to scalar

implicatures relative to aspects of ‘what is said.’” This is an important development

which shows that, in situations that allow some amount of anticipation, pragmatic

inferencing occurs in parallel with linguistic decoding. This would imply that

epistemic steps can occur on-‐line, or at least earlier than has commonly been

claimed.

Referring

Referring is a second major area that fits under the umbrella of

Experimental Pragmatics, which is basically concerned with two kinds of

inferential processes. One arises when a speaker is referring specifically to one

object among many and adds an adjective to set it apart from the others (e.g.

imagine calling an object the big star, which allows one to eliminate from

consideration other small stars). Another line of research concerns the extent to

which a listener indelibly links a speaker to a (newly) labeled object. We will look

at each of these in turn.

Work in both of these domains was inspired by Herb Clark who developed

notions of common ground, which refers to shared knowledge between two people

that is constantly being updated and revised. To make this concrete, if I use the

name “Bart” in a conversation with a friend, I mean the Bart that is the most

prominent for us, in the particular context in which I am speaking. This sort of

shared knowledge is crucial for all exchanges and has been investigated for several

decades now. Referring is an ideal case for discovering what is common to two

interlocutors.

3.1 Referring and adjectival modification

Just like a scalar expression, a referential statement could be

underinformative. For example, in an investigation that predated Gricean

investigations, Ironsmith & Whitehurst (1978) showed that 5-‐ to 8-‐year-‐old

children rarely ask for clarification when an utterance instructs them to select one

referent that could in fact refer to two (the Experimenter would say “I am thinking

of the person with antennae” when there were two representations of people with

antennae). Meanwhile, 9-‐ to 12-‐year-‐olds ask for clarification (“which one?”)

reliably more often in such situations (see also Ackerman, 1981; Ackerman,

Szymanski & Silver, 1990; for a review see Flavell et al, 1981). In more modern

studies, Davies and Katsos (2010) had 5-‐year-‐old children evaluate under-‐

informative referential expressions. When a cartoon character said Pass me the

star when there were two (making the statement underinformative), children –

unlike the adult controls -‐-‐ considered the instruction “natural” significantly more

often. All of this is comparable to the findings on scalars where children are more

likely than adults to find a minimally satisfying interpretation to an

underinformative statement.

However, there are two differences between scalar cases and these sorts of

experimental phenomena. One is that scenarios in referring tasks are routinely

used to test over-‐informativity as well, while scalars cannot be used for that (at

least not obviously). For example, imagine a paradigm that has a single star that is

called the big star; the adjective here is essentially gratuitous for identifying the

intended object. The upshot is that this literature is more concerned about the

level of informativeness and felicity and less so about the separate contribution

made by pragmatics with respect to the utterance’s linguistically encoded

meaning. The other difference is that referring in these tasks does not necessarily

(at least not intuitively) engender propositions about “what could have been said”

and “scales” the way scalar phenomena do because interpretations of reference are

ad-‐hoc and more contextually dependent than scalar cases (which appear to have

generalized rules attached to them).

Despite these impressions, it should be pointed out that one can distinguish

between the linguistically encoded meaning of statement such as Point to the big

star, which would be about a singular star that is considered large (for whatever

reason), and a pragmatic reading, where the interpretation of the speaker’s

meaning coincidentally eliminates other stars from consideration. While the

modifying adjective perhaps allows for a gain in information in the contrastive

case on-‐line (where there are two like objects), it is nevertheless optional since the

listener can have a sufficiently meaningful interpretation of the speaker’s sentence,

and point to one of the stars, without further pragmatic enrichment. However,

tasks in the developmental literature have not fully explored contrastive inference

within a framework that makes this distinction (though see Kronmuller et al., in

press).

3.2 Referring and authorship

How integral is the speaker to reference attribution? That is, when someone

refers to an object with an original label, how personal is that label? Is the

listener’s (acceptance of the speaker’s) assigned name for that object linked with a

specific speaker?

Before answering these questions, it is important to provide some

background by way of terminology. We begin with lexical entrainment which refers

to the way interlocutors attribute names to objects and maintain them for the

length of a conversation (Brennan & Clark, 1996). When a speaker refers to a shoe

as a loafer (to distinguish it, say, from a sneaker) the hearer is likely to maintain

that description for the remainder of an exchange, even if later the more general

term shoe would do. This indicates that a conceptual pact between interlocutors

takes precedence over being just informative enough on each occasion. This leads

to the last remaining coinage – preemption – which refers to the way in which an

accepted label for an item blocks listeners from accepting new names for it

(Kronmüller & Barr, 2007).

Keysar and colleagues (2000) argued that a listener does not immediately

take into consideration the speaker’s point of view. This egocentrism claim is based

on experimental work that presents several objects distributed in the slots of a 4X4

grid. Whereas addressees can see all the grid’s objects, speakers can see only 12 of

these (the remaining four are blocked by screens). Keysar et al.’s eye-‐tracking data

indicate that -‐-‐ for situations where a speaker’s intended reference could only be

an object that is commonly viewed -‐-‐ listeners cannot prevent themselves from

fixating to some extent on a non-‐intended item (one that matches the speaker’s

description even though it is out of the speaker’s view) before fixating on the

intended referent.

Metzing and Brennan (2003) argue that the sharing of intentions between

interlocutors is so central to the referential process that the speaker herself is part

of the preemption process. So, pragmatic expectations are linked to specific

partners in a conversation and thus the partner’s epistemic states (E. V. Clark,

1990; Metzing & Brennan, 2003; Brown-‐Schmidt, 2009; Brennan & Hanna, 2009).

Evidence cited in favor of this approach includes the prediction that a listener

should not experience preemption when another speaker provides a newly coined

name for a previously-‐named object. In testing this claim, Metzing & Brennan,

(2003) employed an eye-‐tracking paradigm in which a confederate (a “Director”)

would provide a participant (a “Matcher”) with instructions to move small objects

to new locations within a vertical 5-‐by-‐5 grid of entirely open cubbyholes. By

measuring the duration of time between the onset of a referring expression and

first-‐looks to the target (as well as touches to the target), they could determine

how fast a listener makes referential commitments when the same (“original”)

Director comes up with a new name for a previously-‐named object as opposed to

when a new Director does so. They reported a slowdown only when the same

Director came up with a new name and took this as support for their hypothesis.

The debate does not end there. In support of a perspective-‐free account

(meaning that the listener incorporates the named object without necessarily

linking the speaker to the label), Kronmuller and Barr (2007) recorded eye-‐

movements linked to referent assignment as the process unfolds in a situation

similar to Metzing and Brennan’s (2003). They reported that a tendency to avoid

mapping a new expression onto an already-‐named referent emerged soon after a

new referring expression is heard, independently of whether the speaker uttering

the expression was new to the situation or not. In other words, when a new name

for an old object was offered, there was an identifiable recovery process regardless

of who the speaker was (the same speaker or a different one). Furthermore,

Shintel and Keysar (2007) report that listeners took more time to map an old

referent into a new expression, independently of whether the conversational

precedent for that object was established by the speaker to the listener or by

another speaker to the listener. Future work will no doubt clarify whether

speakers are integral to recognizing labels.

Metaphor

The study of metaphor comprehension has a long history in the

psycholinguistics literature with most of it revolving around its temporal

signature. On the one hand, there has long been the assumption that metaphor

processing is exceptional which implies that it should be linked with extra

processing and slowdowns. Formally, many cite Gricean models as a theoretical

basis for this approach since he argued that a sentence is recognized as non-‐literal

once that it violates the maxim of quality (“Do not say what you believe to be false”)

and engenders a process needing resolution. Thus, in order to understand that a

speaker did not literally mean “my son is a pig,” one must search out an alternative

while maintaining Grice’s Cooperative Principle, which assumes that interlocutors

are indeed cooperative (Gerrig and Healy, 1983). On the other hand, several

researchers have shown that there is nothing exceptional about metaphor

processing and that reaction times can be shown to be comparable to literal

controls when given the right context (see Gibbs, 2002 for a review). This has been

the basis of much back and forth in the psycholinguistic literature. Here, we take a

look at three theories that have dominated discussion on this matter in the

experimental literature.

4.1 The Standard Pragmatic Model

The Standard Pragmatic Model (SPM) evolved from the work of Grice

(1975) and Searle (1979) and is one of the earliest and most influential approaches

to the comprehension of figurative language. At its simplest, the so-‐called SPM is a

three-‐step process that involves 1) the computation of the semantic/literal

meaning; 2) the recognition of a violation of a maxim; and 3) the computation of an

implicature. One can see how the architecture of Grice’s seminal work can be

directly translated into a psychological explanation for understanding many non-‐

literal uses of language, including metaphor.

It has been difficult to establish that these three steps actually occur (let

alone in such an order) and, more fatally, such a three-‐step process seems too long

and slow for explaining the rapid on-‐line pragmatic processing of an utterance,

which has led to conflicting results. On the one hand, studies like Janus & Bever’s

(1985) are compatible with the Standard Pragmatic Model as are ERP studies that

reveal greater processing cost for a metaphor over a literal utterance (e.g., Pynte et

al., 1996; Lai et al, 2009). On the other hand, more recent accounts (e.g., see a

summary of the Direct Access view below) tend to argue against it. These latter

data have made it easy for critics to rail against the SPM and, in so doing, the entire

Gricean approach.

It should be pointed out, however, that Grice never intended his model to be

used as a model of actual language processing. The way Gricean theory is

transformed into “the SPM” is emblematic of a common pitfall in the Cognitive

Sciences (see Marr, 1982); one can advance theoretically at three different levels –

often referred to as the computational, algorithmic and implementational levels of

analysis – and that progress can be made by keeping the three levels separate and

complementary.2

4.2 The Direct Access View

More recently, Gibbs (1994, 2002) has offered an approach to figurative

language comprehension that directly contrasts with the SPM. Gibbs’s approach,

the Direct Access View, suggests similar processing mechanisms for both figurative

and literal language. According to this view, the comprehension of figurative

language does not involve special cognitive processes (Gibbs, 1994; Gibbs & Moise,

1997). This assumption is based on the notion that comprehending literal as well

as non-‐literal meanings of a sentence largely depends on pragmatic knowledge and

listeners’ figurative modes of thought (Gibbs, 1994, 2002). Furthermore, Gibbs

suggests that literal and non-‐literal meanings are not distinct from each other since

they are both determined by contextual information. By using pragmatic

knowledge together with contextual information, the analysis of some aspects of

word meaning should be sufficient for understanding intended figurative

meanings (Gibbs, 1999, 2002). Evidence in favor of the Direct Access View also

comes from some classic comprehension studies showing that latencies for

comprehending literal and figurative readings of similar target-‐sentences can be

rendered comparable as a function of context (e.g., Ortony, 1978; Gildea and

Glucksberg, 1982; Blasko and Connine, 1993).

2 The computational level makes explicit the input and output of the process as well as the

constraints that would allow a specified problem to be solved. The algorithmic level describes how to get from input to output, and specifically determines which representations have to be used and which processes have to be employed in order to build and manipulate the representations. The implementational level provides a description of the physical system that should realize the process at, say, the neuronal level. Grice’s theory was designed at the computational level; the SPM was invented to practically mimic it at the algorithmic level (see Noveck & Spotorno, in press). However, this amount of resemblance, as Marr argued, is neither necessary nor recommended. One can do theoretical work at the computational level without recourse to the algorithmic or implementational levels and similarly for any level with respect to the other two.

For example, Inhoff et al (1984), have shown that metaphorical expressions are

understood as quickly as literal ones when they are preceded by an extensive

context that practically primes the metaphorical interpretation (see the target

sentence in 9.c in the context of 9.a versus 9.b).

(9.a) Severe criticisms of Europe's oldest dictator came from within Spain and

without. For 35 years Franco was barraged with these constant criticisms.

To the end, the struggle continued between the ruthless dictator and his

critics.

(9.b) Iceland's coastal region was windswept. The angry arctic winds often

buffeted the fishing villages located on the numerous bays and inlets. The

darkening skies and the rising winds announced the onset of another

storm

(9.c) Target: The waves beat relentlessly against the rugged coastline.

On the contrary, when just a short prior context is provided (essentially, the first

sentences of 9.a and 9.b), the comprehension of a metaphoric reading requires

more effort than that of a literal sentence. This sort of finding is a central to the

metaphor-‐processing literature and calls for explanation. We will return to this

shortly as we summarize the remaining approaches.

4.3 The Graded Salience Hypothesis

The Standard Pragmatic Model and the Direct Access view can be

considered two extremes of a spectrum that other proposals can then fill. One of

the most influential accounts that fills the gap is the Graded Salience Hypothesis

(Giora, 1997). According to this proposal, the initial processing of lexical

information is an encapsulated and graded process in which salient meanings of

words or expressions are retrieved from the mental lexicon (Giora, 2003). During

initial processing, contextual information is processed in parallel but neither

interacts with lexical processes nor inhibits salient meanings when contextually

incompatible (e.g., Giora, 2002; Peleg, Giora, & Fein, 2001). Salience is a function of

properties such as familiarity, prototypicality and frequency; in order to be salient,

the meaning of a word has to be encoded into the mental lexicon. In case words or

expressions have multiple meanings varying in salience, Giora (2003) suggests that

this process is graded: more salient meanings are accessed earlier than less salient

meanings. Thus, most salient meanings are always accessed initially irrespective of

their literality or contextual support. This implies that the processing of figurative

sentences only diverges from that of literal sentences during later phases of

processing if accessed salient meanings cannot be integrated with contextual

information. In that case the salient meanings have to give way to less salient but

contextually appropriate meanings. As opposed to the Direct Access View,

contextual information is proposed to have a very limited impact because they

cannot restrict initial access of salient meanings that might be contextually

incompatible. Evidence for the Graded Salience Hypothesis comes from behavioral

studies that investigate the comprehension of metaphors. For example, Giora &

Fein (1999) have shown that familiar metaphors can be processed as easily as

literal remarks, while unconventional metaphors seem to require more effort.

Further support comes from a study by Lai and colleagues (2009) who showed

that conventional and unconventional metaphors differentially modulate the ERP

component known as the P600.

4.4 Metaphor as viewed by Relevance Theory

As we highlighted, metaphor represents an interesting case for

Experimental Pragmatics (see also Bambini & Resta, 2012) because utterances

need not be literally true in order for a listener to draw implications effectively.

For example, consider a scenario in which a swimming instructor says to a 5-‐year-‐

old, you are a tadpole. The utterance is a) not literally true while effectively

conveying information from teacher to student and it; b) goes further than its

literal equivalent (which one can imagine if one is talking directly to a tadpole, for

some reason); at the very least, the instructor's expression is arguably endearing

when speaking to a young child whereas the literal equivalent is not. This sort of

analysis has been fruitful in describing at least three kinds of experimental results

with respect to metaphor.

First, Noveck, Bianco, & Castry (2001; also see Almor et al., 2007) showed,

through reading times, that children are less likely than adults to fully appreciate

metaphoric references and that adult reading times of metaphoric sentences,

which take longer to read than literal controls, come with a deeper processing (as

determined by responses to comprehension questions). That is, children -‐-‐ having

fewer cognitive resources available than adults -‐-‐ are weaker than their older

cohorts at accessing metaphoric meanings while adults draw out more cognitive

effects from metaphoric references than they do from literal ones.

Second, the enhancement of relevant properties of the metaphor vehicle,

and suppression of those that are irrelevant for the figurative interpretation, is a

necessary process in metaphor comprehension. Consider a cross-‐modal lexical

priming study by Rubio-‐Fernandez (2007), which presented vignettes aurally and

concluded with a metaphor. At the end of a vignette, a probe word appeared on a

screen immediately or either 400 or 1000 msec later. These probes required

participants to determine whether they were spelled correctly (with the idea being

that a relatively quick lexical decision –– a “yes, it is spelled correctly” –– reveals

that the metaphor activated the probe word at the time of decision). For example,

participants would read (10):

(10) Nobody wanted to run against John at school. John was a cheetah.

These sentences were followed with what was referred to as a superordinate

probe such as cat or an unrelated probe, e.g. plant (a superordinate from another

trial). The decision latencies for those who got the probe at 400 msec showed a

preference for processing the superordinate probe over the unrelated one (644 vs.

698 msec, respectively), which disappeared when measured against those who got

the probe at 1000 msec (791 vs. 799 msec, respectively). One can see that the

categorical information no longer had much impact on the listener’s processing a

full second after finishing the sentence. In contrast, when the probes presented

were akin to fast (which is the relevant feature of being a cheetah in the example)

and it was compared to an unrelated one, e.g., tall (a non-‐relevant feature pulled

from a metaphor in another vignette), the decision latencies for those who got the

probe at 400 msec showed a slight preference for the former over the latter (658

vs. 710 msec, respectively) and it remained disparate at 1000 msec (598 vs. 651

msec, respectively). Both of these remained faster than all the so-‐called

superordinate cases. As Rubio Fernández wrote, her results support the notion

that “metaphor interpretation involves enhancing relevant properties of the

metaphor vehicle while suppressing irrelevant ones.”

These results are in line with the Relevance idea of loose talk. According to

the Relevance-‐theoretic account the process of adjustment of the meaning encoded

in a linguistic constituents requires the interaction among the lexical encoded

concepts, the concepts that are encoded in other constituents of the utterance and

contextual information. The outcome of this process is an ad hoc concept which has

been derived by inferential processes under the specific circumstances of an

utterance and which is valid only under those circumstances. In this way, for

example, the word “raw” in the utterance: “This steak is raw” communicates the

concept RAW* (the symbol * indexes the ad hoc concept) that is slightly different

from the lexical concept RAW. With the deflationary account of metaphor (to

indicate that metaphor is not special), Relevance theorists have proposed that the

same process of adjustment allows one to derive the metaphorical interpretation

in an expression such as “John is a cheetah” (Sperber and Wilson, 2008; Wilson and

Carston, 2007). In the end, the comprehension process of a metaphor should come

along the same path of all other loose uses of language, such as approximations

(e.g., “The water is boiling” when the water very hot) and hyperboles (e.g., “The

water is boiling” when the water is just warm).

The third sort of finding that shows how an apt metaphor is dependent on

context comes from more classic studies mentioned earlier, showing how the

amount of background information available for the addressee affects the

processing of a metaphorical expression (e.g. see, Gildea and Glucksberg, 1982;

Gibbs and Gerrig, 1989; Pynte et al. 1996; Inhoff et al, 1984). Findings such as

these suggest that one ought to place metaphor along a continuum of cognitive

effort. At one end, there are simple literal expressions that demand little effort

since they represent no unnecessary, extra load when processing sentences. At the

other end, one can imagine anomalous sentences, apparently too effortful to be

resolved because neither the concepts engendered nor the context are helpful in

processing the sentence. Metaphor floats between these two extremes according to

the conceptual load and the contextual support.

While Relevance Theory describes these data well, a lot of work remains.

For one thing, the complex notion of context has to be further disentangled.

Another is that the link between conceptual load and context needs to be explored

further because these two main variables have to interact in order for a listener to

grasp the intended meaning of a metaphoric utterance.

Irony

From the perspective of Paul Grice, irony, like metaphor, is a departure from

a norm of literal truthfulness used to convey a related figurative meaning or

implicature. Irony is said to be understood because (as with metaphor) the speaker

is blatantly violating the first maxim of Quality, which then triggers implicatures

that can be calculated from the literal meaning of the sentence. Although initially

influential, the Gricean account is now increasingly in doubt due mostly to its

emphasis on the presence of maxim-‐violations that are not always apparent. For

example, consider the utterance: “I love sunny days!” when it is said during a

downpour. The utterance is ironic and is probably true as well, but there are no

identifiable maxim violations.

5.1 Echoic-‐mention theory

The first challenge to Grice’s approach came from the Echoic mention

theory (Sperber & Wilson, 1981; Wilson, 2009; Jorgensen, Miller, & Sperber, 1984).

According to this account, the speaker of irony is not expressing her own thoughts,

but echoing a thought that can be attributed to some real or prototypical speaker

while expressing a dissociative (mocking, skeptical or contemptuous) attitude to

that thought. When our fictive interlocutor says “I love sunny days!” on a cold and

rainy day, what makes the utterance ironic is that the speaker is harking back to a

possible and an appropriate remark for desired weather. As Wilson (2009, page

197) writes:

… the point of irony is not to commit the speaker to the truth

of the proposition expressed but, on the contrary, to express a certain

type of derisory or dissociative attitude to a thought with a similar

content that she attributes to some source other than herself at the

current time. In other words, the speaker of irony is not expressing

her own thoughts, but echoing a thought she attributes to someone

else, and expressing her mocking, skeptical or contemptuous attitude

to that thought.

Through echoing the speaker makes her attitude apparent.

One of the earliest experimental studies on irony investigated Echoic

Mention Theory. Jorgensen, Miller & Sperber (1984) investigated how an explicit

antecedent remark can impact the later comprehension of an ironic remark. For

example, imagine two opera singers who are anticipating their evening’s

performance when one says: “Tonight will be great!” which is uttered before the

evening takes a negative turn. A later ironic remark of the sort “That was a superb

performance” is harking back to the speaker’s earlier expectations. Jorgenson et al.

(1984) predicted that participants would find the same ironic target sentence

harder to understand if the context does not include such an explicit antecedent

reference to the ironic speaker’s expectations. The results of their questionnaire

confirmed these predictions and provided the literature with the first set of data in

support of the Echoic Theory. Later, Francesca Happé, in her seminal paper on

pragmatic abilities in Autism Spectrum Disorders (1993), revealed that only

autistic people with intact high level mindreading abilities (namely the second-‐

order Theory of Mind, which is the ability to interpret someone’s thought about

someone else’s thought) are able to correctly interpret ironies. Happé’s study

provided strong evidence in favor of Sperber and Wilson’s analysis of linguistic

communication, which includes the Echoic account of irony under the umbrella of

Relevance theory (e.g., Sperber & Wilson, 1986; Wilson & Sperber, 2012b).

5.2 Pretense Theory

Clark and Gerrig (1984) proposed an alternative view which they called a

“Pretense Theory of Irony.” The main idea behind the Pretense account is that the

speaker of an ironical utterance is not herself performing a speech act (e.g. making

an assertion or asking a question) but pretending to perform one, in order to

convey a mocking, skeptical or contemptuous attitude. This pretense mood would

make an utterance ironic. With "I love sunny days," the speaker is pretending to be

an unseeing person, perhaps a weather forecaster, exclaiming to an unknowing

audience how much she likes beautiful weather. She intends the addressee to see

through the pretense—in such rain she obviously could not be making the

exclamation on her own behalf—and to see that she is thereby ridiculing the sort

of person who would make such an exclamation (e.g., the weather forecaster). The

addressee can take "delight" in "the secret intimacy" shared with the speaker in

recognizing that ignorance (Clark & Gerrig, 1984, 122).

While the Echoic-‐mention and Pretense accounts each give primacy to the

accessing of a speaker’s attitude, they differ with respect to the object of the ironic

remark. For Sperber & Wilson, it is some previously stated remark or shared

cultural norm and for Clark & Gerrig, it is practically a parody of the speaker who

would have made such a remark. It should also be noted that while Grice focused

on the blatant violation of truth in irony, he also took attitude into consideration

(Grice, 1989, p. 53) for he recognized that a “hostile or derogatory judgment or a

feeling such as indignation or contempt” had a role to play in understanding it;

nevertheless, as Deirdre Wilson pointed out to us, Grice never “officially”

introduced attitude ascription into his theory). Point is that all three of these

seminal accounts agree that the communication of the speaker’s attitude is crucial

to the comprehension of irony.

5.3 Irony as viewed by the psycholinguistic literature

Despite these early views on irony, which clearly took into account concerns

about Theory of Mind, a different debate arose in the psycholinguistic literature on

the immediacy with which ironic interpretations are made when compared to

literal readings. This is because Gibbs, in his processing paper (Gibbs, 1986),

argued strongly against Grice’s “Standard Pragmatic Model” by pointing out that

irony (and especially sarcasm) gives rise to the intended reading without requiring

a contrast between the literal reading and a given situation that blatantly violates a

maxim. Central to Gibbs’s account, known as the Direct Access View, is the

assumption that a figurative interpretation is constructed “directly” by the early

integration of lexical and contextual information. Much recent work presents

evidence as support for the Graded Salience Hypothesis (Filik & Moxey, 2010;

Giora, 1997; Giora & Fein, 1999; Giora, Fein, Kaufman, Eisenberg, & Erez, 2009)

and sets itself up as being in opposition to the Direct Access view also in the case of

Irony processing.

Remarkably, the debate about the comprehension process of irony mimics

the debate about metaphor processing; however, the similarities between the two

pragmatic phenomena are largely superficial. It is true that both of them are

instances of figurative language and both irony and metaphor are usually viewed

as literally false utterances, but mechanisms behind the two phenomena are rather

different. In the section on metaphor we described a process of enrichment and

adjustment of the concepts conveyed by critical words such as “cheetah” in “John is

a cheetah” as crucial to metaphor comprehension. In contrast, the hallmark of

irony comprehension is the interpretation of the speaker’s attitude instead of the

enrichment of the literal meaning of the utterance. This is quite different.

Let us return to our two opera singers who recognize that they gave an

awful performance and that after the show one singer tells the other:

(11) “Tonight we gave a superb performance.”

In order to reach the ironic interpretation of (11) one has to go beyond the

linguistic code and grasp the speaker’s dissociative attitude (see e.g., Wilson and

Sperber, 2012; Clark and Gerrig, 1984), but this process does not pass through the

adjustment of the concepts that are conveyed by the words “superb performance”.

The lexical meaning of those words is good enough to trigger the ironic

interpretation. The more recent debates are about the surface features of the

words used in irony and their effect on the speed of processing. This is partly due

to the effort of transferring the Direct Access versus Graded Salience debate to

irony. The net result is that the processing literature hardly considers a role for

attitude ascription in irony.

5.4 Reconciling theory and experimentation

In order to right this perceived imbalance, we -‐-‐ the authors of this chapter -‐

-‐ have embarked on a series of investigations that aim to show the extent to which

irony depends on Theory of Mind (ToM). One central study (Spotorno et al., 2012)

uses fMRI because the cerebral areas associated with ToM have been extensively

investigated in the cognitive neuroscience literature. In that study, we compared

participants’ comprehension of 18 target sentences (such as the one in 11) as

contexts make them either ironic or literal (to render (11) literal all one has to do

is say that they hit their notes with success). That work demonstrated that the ToM

network becomes more active while a participant is understanding a verbal irony

when compared to activity of those same areas while reading literal equivalents.

Moreover, they demonstrate -‐-‐ through Psychophysiological Interactions (PPI)

analyses -‐-‐ that ToM activity is directly linked with language comprehension

processes. The paradigm, its predictions, and the reported results contrasted

dramatically with those from seven prior fMRI studies. More importantly, it re-‐

establishes the central role played by ToM in irony research (also see Spotorno et

al., in press).

Summary

We have covered four different areas of Experimental Pragmatic research –

scalar inference, referring, metaphor and irony. While each may have a different

historical development, each investigates the gap between the linguistically

encoded meaning of an utterance, on the one hand, and its intended meaning and

its effects, on the other. Besides guiding research, this distinction highlights

different aspects of pragmatic processing. Scalar inference involves enrichments of

a potentially underinformative lexical item. Questions about referring are aimed at

determining the range of implications about what objects are excluded and who

made the reference. Metaphor is concerned about the way a literally false sentence

can become meaningful by making salient aspects of a word’s meaning and irony is

concerned about understanding the speaker’s attitude.

Despite our best efforts to be inclusive, we had space to present only four

areas that are arguably among the most well developed. Of course, this does not

cover everything. To name a few areas that we did not discuss but could have,

consider how prosody can influence speaker meaning (Chevallier et al., 2010) as

well as topics such as metonymy (Schumacher, in press) and indirect requests

(Van der Henst & Sperber, 2004; Basnakova et al., in press). Experimental

Pragmatics can also be the forum for understanding the way exceptional

populations, such as those on the autism spectrum, deal with pragmatically

interesting cases (Noveck et al., 2007; Pijnaker et al., 2009; Chevallier et al. 2011).

All of these topics fall under Experimental Pragmatics and make for a rich

database. At the root of most of these studies is the divide between linguistically

encoded meanings and intended meanings. This distinction drives experimental

endeavors, reveals robust effects that any theory would need to account for and,

importantly, establishes a forum for resolving important debates in linguistic-‐

pragmatics.

References

Ackerman, B.P. (1981). Performative bias in children’s interpretations of

ambiguous referential communications. Child Development, 52, 1224-‐1230.

Ackerman, B.P., Szymanski, J., & Silver, D. (1990). Children’s use of common ground

in interpreting ambiguous referential utterances, Developmental Psychology,

26, 234-‐245.

Almor, A., Arunachalam, S., & Strickland, B. (2007). When the creampuff beat the

boxer: working memory, cost, and function in reading metaphoric reference.

Metaphor and Symbol, 22(2), 169–193.

Bambini, V., & Resta, D. (2012). Metaphor and Experimental Pragmatics: When

Theory Meets Empirical Investigation. humana.Mente, 23, 37–60.

Barner, D., Brooks, N., & Bale, A. (2011). Accessing the unsaid: The role of scalar

alternatives in children’s pragmatic inference. Cognition, 118(1), 84.

Baron-‐Cohen, S., Wheelwright, S., Skinner, R., Martin, J., and Clubley, E. (2001). ‘The

autism spectrum quotient (AQ): Evidence from Asperger syndrome/high

functioning autism, males and females, scientists and mathematicians’,

Journal of Autism and Developmental Disorders, 31, 5-‐17.

Bašnáková J., Weber, K., Petersson, K.M., van Berkum, J., & Hagoort, P. (in press).

Beyond the Language Given: The Neural Correlates of Inferring Speaker

Meaning. Cerebral Cortex.

Blasko, D. G., & Connine, C. M. (1993). Effects of familiarity and aptness on

metaphor processing. Journal of experimental psychology. Learning, memory,

and cognition, 19(2), 295–308.

Bott, L. & Noveck, I. A. (2004). ‘Some utterances are underinformative: The onset

and time course of scalar inferences’ Journal of Memory and Language, 51(3),

437-‐457.

Bott, L., Bailey, T.M., & Grodner, D. (2012). ‘Distinguishing speed from accuracy in

scalar implicatures’, Journal of Memory and Language, 66, 123-‐142.

Braine, M., & Rumain, B. (1981). Children’s comprehension of “or”: Evidence for a

sequence of competencies. Journal of Experimental Child Psychology, 31, 46-‐70.

Breheny, R., Katsos, N., Williams, J. (2006). ‘Are scalar implicatures generated by

default?’, Cognition, 100, 434-‐463.

Breheny, R., Ferguson, H. J., & Katsos, N. (2013). Taking the epistemic step: Toward

a model of on-‐line access to conversational implicatures. Cognition, 126(3),

423-‐440.

Brennan S.E., & Clark H.H. (1996). ‘Conceptual pacts and lexical choice in

conversation’, Journal of Experimental Psychology: Learning, Memory, and

Cognition, 22, 1482-‐1493.

Brennan, S. E., & Hanna, J. E. (2009). Partner-‐Specific Adaptation in Dialog. Topics

in Cognitive Science, 1(2), 274-‐291.

Brown-‐Schmidt, S. (2009). Partner-‐specific interpretation of maintained

referential precedents during interactive dialog. Journal of memory and

language, 61(2), 171-‐190.

Carston, R. (2002). Thoughts and utterances. Oxford: Blackwell.

Carston, R. (2010a). Lexical pragmatics, ad hoc concepts and metaphor: from a

relevance theory perspective. Italian Journal of Linguistics, 22(1), 153–180.

Carston, R. (2010b). Metaphor: Ad Hoc Concepts, Literal Meaning and Mental

Images. Proceedings of the Aristotelian Society, 110(3), 295–321.

Chevallier, C., Noveck, I., Bott, L., Lanza, V., Nazir, T., & Sperber, D. (2008). Making

disjunctions exclusive. Quarterly Journal of Experimental Psychology, 61(11),

1741-‐1760

Chevallier, C., Noveck, I., Happé, F., & Wilson, D. (2011). What's in a voice? Prosody

as a test case for the Theory of Mind account of autism. Neuropsychologia,

49(3) 507-‐517.

Chierchia, G. (2001). Scalar implicatures, polarity phenomena, and the

syntax/pragmatics interface, in A. Belleti (ed.) Structures and beyond. Oxford:

Oxford University Press.

Chierchia, G., Fox, D., & Spector, B. (to appear). The grammatical view of scalar

implicatures and the relationship between semantics and pragmatics. in P.

Portner, C. Maienborn et K. von Heusinger (Eds.) An International Handbook

of Natural Language Meaning. Mouton de Gruyter.

Clark, E. V. (1990). On the pragmatics of contrast. Journal of Child Language,

17(02), 417-‐431.

Clark, H. H., & Gerrig, R. J. (1984). On the Pretense Theory of Irony, 113(1), 121–

126.

Davies, C., & Katsos, N. (2010). Over-‐informative children:

Production/comprehension asymmetry or tolerance to pragmatic violations?.

Lingua, 120(8), 1956-‐1972.

De Neys, W., & Schaeken, W. (2007). When people are more logical under cognitive

load: Dual task impact on scalar implicature, Experimental Psychology, 54,

128-‐133.

Filik, R., & Moxey, L. (2010). The on-‐line processing of written irony. Cognition,

116(3), 421–436.

Flavell, J.H., Speer, J.R., Green, F.L., August, D.L., & Whitehurst, G.J. (1981). The

development of comprehension monitoring and knowledge about

communication, Monographs of the Society for Research in Chile Development,

46(5), 1-‐65.

Gerrig, R. J., & Healy, A. F. (1983). Dual Processes in Metaphor Understanding�:

Comprehension and Appreciation. Cognition, 9(4).

Geurts, B. (2010). Quantity implicatures. Cambridge University Press.

Gibbs, R. W. (1986). On the psycholinguistics of sarcasm. Journal of Experimental

Psychology: General, 115(1), 3–15.

Gibbs, R. W. (1994). The poetics of mind: Figurative thought, language, and

understanding. Cambridge: Cambridge University Press.

Gibbs, R. W. (1999). Interpreting what speakers say and implicate. Brain and

Language, 68(3), 466–485.

Gibbs, R. W. (2002). A new look at literal meaning in understanding what is said

and implicated. Journal of Pragmatics, 34(4), 457–486.

Gibbs, R. J., & Gerrig, R. (1989). How context makes metaphor comprehension seem

“special”. Metaphor and Symbol, 4(3), 145–158.

Gibbs Jr, R. W., & Moise, J. F. (1997). Pragmatics in understanding what is said.

Cognition, 62(1), 51-‐74.

Gildea, P., & Glucksberg, S. (1983). On understanding metaphor: The role of

context. Journal of Verbal Learning and Verbal Behavior, 22(5), 577–590.

Giora, R. (1997). Understanding figurative and literal language: The graded

salience hypothesis. Cognitive Linguistics, 8(3), 183–206.

Giora, R. (2002). Literal vs. figurative language: Different or equal? Journal of

Pragmatics, 34(4), 487–506.

Giora, R. (2003). On our mind: Salience, context, and figurative language. New York:

Oxford University Press.

Giora, R., & Fein, O. (1999). Irony: Context and salience. Metaphor and Symbol,

14(4), 241–257.

Giora, R., Fein, O., Kaufman, R., Eisenberg, D., & Erez, S. (2009). Does an “ironic

situation” favor an ironic interpretation? In G. Brône & J. Vandaele (Eds.),

Cognitive poetics: Goals, gains and gaps (pp. 383–399). Berlin/New York:

Mouton de Gruyter.

Grice, H. P. (1975). Logic and conversation. In S. Davis (Ed.), Pragmatics: A reader

(p. 305-‐315). New York: Oxford University Press.

Grice, P. H. (1989). Studies in the Way of Words. Cambridge, Massachusetts:

Harvard University Press.

Grodner, D., Klein, N. M., Carbary, K. M., & Tanenhaus, M. K. (2010). ‘‘‘Some’’, and

possibly all, scalar inferences are not delayed: Evidence for immediate

pragmatic enrichment’, Cognition, 116, 42-‐55.

Guasti, M. T., Chierchia, G., Crain, S., Foppolo, F., Gualmini, A., & Meroni, L. (2005).

‘Why Children and Adults Sometimes (But Not Always) Compute

Implicatures’, Language and Cognitive Processes, 20, 667-‐696.

Happé, F. G. (1993). Communicative competence and theory of mind in autism: a

test of relevance theory. Cognition, 48(2), 101–119.

Huang, Y., & Snedeker, J. (2009). Semantic meaning and pragmatic interpretation in

5-‐year-‐olds: Evidence from real-‐time spoken language comprehension,

Developmental Psychology, 45(6), 1723-‐1739.

Inhoff, A. W., Lima, S. D., & Carroll, P. J. (1984). Contextual effects on metaphor

comprehension in reading. Memory & Cognition, 12(6), 558–567.

Ironsmith, M., & Whitehurst, G. (1978). The development of listeners abilities in

communication: How children deal with ambiguous information, Child

Development, 49, 348-‐352.

Janus, R., & Bever, T. (1985). Processing of metaphoric language: An investigation

of the three-‐stage model of metaphor comprehension. Journal of

Psycholinguistic Research, 14(5), 473–487.

Jorgensen, J., Miller, G., & Sperber, D. (1984). Test of the mention theory of irony.

Journal of Experimental Psychology. General, 113(1), 112–120.

Katsos, N., & Bishop, D. V. (2011). Pragmatic tolerance: Implications for the

acquisition of informativeness and implicature. Cognition, 120(1), 67-‐81.

Keysar, B., Barr, D. J., Balin, J. A., & Brauner, J. S. (2000). Taking perspective in

conversation: The role of mutual knowledge in comprehension. Psychological

Science, 11(1), 32-‐38.

Kronmüller, E., & Barr, D. J. (2007). Perspective-‐free pragmatics: Broken

precedents and the recovery-‐from-‐preemption hypothesis. Journal of Memory

and Language, 56(3), 436-‐455.

Krönmuller, E., Morisseau, T. & Noveck, I. A. (in press). Show me the pragmatic

contribution: A developmental investigation of referential communication.

Journal of Child Language.

Kutas, M., & Hillyard, S. A. (1984). Brain potentials during reading reflect word

expectancy and semantic association. Nature 307, 161-163.

Lai, V. T., Curran, T., & Menn, L. (2009). Comprehending conventional and novel

metaphors: an ERP study. Brain research, 1284, 145–55.

Levinson, S.C. (2000). Presumptive meanings: The theory of generalized

conversational implicature. Cambridge, MA: MIT Press.

Marr, D. (1982). Vision: A computational investigation into the human

representation and processing of visual information. New York: W.H. Freeman.

Metzing, C., & Brennan, S. E. (2003). When conceptual pacts are broken: partner-‐

specific effects on the comprehension of referring expressions. Journal of

Memory and Language, 49, 201–213.

Nieuwland, M. S., Ditman, T., & Kuperberg, G. R. (2010). ‘On the incrementality of

pragmatic processing: An ERP investigation of informativeness and

pragmatic abilities’, Journal of Memory and Language, 63(3), 324-‐346.

Noveck, I. A. (2001). When children are more logical than adults: Investigations of

scalar implicature, Cognition, 78, 165-‐188.

Noveck, I. A., Bianco, M., & Castry, A. (2001). The costs and benefits of metaphor,

Metaphor and Symbol, 16(1and2), 109-‐121.

Noveck, I. A., Guelminger, R., Georgieff, N., & Labruyere, N. (2007). What autism can

reveal about every… not sentences. Journal of Semantics, 24(1), 73-‐90.

Noveck, I. A., & Posada, A. (2003). Characterizing the time course of an implicature:

An evoked potentials study, Brain and Language, 85(2), 203-‐210.

Noveck, I.A. & Reboul, A. (2008). Experimental Pragmatics: A Gricean turn in the

study of language, Trends in Cognitive Sciences, 12, 425-‐431.

Noveck, I. A., & Sperber, D. (2007). The why and how of experimental pragmatics:

The case of ‘scalar inferences’, in N. Burton-‐Roberts (Ed.) Advances in

Pragmatics. Basingstoke: Palgrave.

Noveck, I.A. & Spotorno, N. Narrowing. In L. Goldstein (Ed.), Brevity. Oxford: OUP.

Ortony, A., Schallert, D. L., Reynolds, R. E., & Antos, S. J. (1978). Interpreting

metaphors and idioms: Some effects of context on comprehension. Journal of

Verbal Learning and Verbal Behavior, 17(4), 465–477.

Paris, S. (1973). Comprehension of language connectives and propositional logical

relationships. Journal of Experimental Child Psychology, 16, 278-‐291.

Peleg, O., Giora, R., & Fein, O. (2001). Salience and Context Effects: Two Are Better

Than One. Metaphor and Symbol, 16(3-‐4), 173–192.

Pijnacker, J., Hagoort, P., Buitelaar, J., Teunisse, J. P., & Geurts, B. (2009). Pragmatic

inferences in high-‐functioning adults with autism and Asperger syndrome.

Journal of autism and developmental disorders, 39(4), 607-‐618.

Pouscoulous, N., Noveck, I. A., Politzer, G., & Bastide, A. (2007). A developmental

investigation of processing costs in implicature production, Language

Acquisition, 14, 347-‐375.

Pynte, J., Besson, M., Robichon, F. H., & Poli, J. (1996). The time-‐course of metaphor

comprehension: an event-‐related potential study. Brain and language, 55(3),

293–316.

Rubio Fernández, P. (2007). Suppression in metaphor interpretation: Differences

between meaning selection and meaning construction. Journal of Semantics,

24(4), 345–371.

Schumacher, P. B. (in press). The hepatitis called ...: Electrophysiological Evidence

for Enriched Composition. In Meibauer Jörg & Markus Steinbach (eds.).

Experimental Pragmatics/Semantics. Amsterdam/Philadelphia: John

Benjamins.

Searle, J. R. (1979). Expression and meaning. Studies in the theory of speech acts.

Cambridge: Cambridge University Press.

Shintel, H., & Keysar, B. (2009). Less is more: A minimalist account of joint action in

communication. Topics in Cognitive Science, 1(2), 260-‐273.

Sperber, D. & Noveck, I.A. (2004) Introduction. In I.A. Noveck & D. Sperber (Eds.).

Experimental Pragmatics. London: Palgrave Macmillan.

Sperber, D., & Wilson, D. (1981). Irony and the use-‐mention distinction. In P. Cole

(Ed.), Radical pragmatics (pp. 295–318). New York: Academic Press Elsevier.

Sperber, D., & Wilson, D. (1986). Relevance: Communication and Cognition. Oxford:

Blackwell; Wiley-‐Blackwell; 2nd Edition edition 1995.

Sperber, D., & Wilson, D. (2008). A deflationary account of metaphors. In R. W.

Gibbs (Ed.), The Cambridge handbook of metaphor and thought (pp. 84–105).

New York: Cambridge Univesity Press.

Spotorno, N., Koun, E., Prado, J., Van Der Henst, J.-‐B., & Noveck, I. A. (2012). Neural

evidence that utterance-‐processing entails mentalizing: The case of irony.

NeuroImage, 63(1), 25–39.

Sternberg, R. J. (1979). Developmental patterns in the encoding and combination of

logical connectives. Journal of Experimental Child Psychology, 28(3), 469-‐498.

Van Der Henst, J. B., & Sperber, D. (2004). Testing the cognitive and the

communicative principles of relevance. In Noveck, I. & Sperber, D. (Eds).

Experimental pragmatics. London: Palgrave Macmillan

Wilson, D. (2009). Irony and metarepresentation. UCLWPL, 21(June), 183–226.

Wilson, D., & Carston, R. (2007). A unitary approach to lexical pragmatics:

relevance, inference and ad hoc concepts. In N. Burton-‐Roberts (Ed.),

Pragmatics (pp. 230–259). Basingstoke and New York: Palgrave Macmillan.

Wilson, D., & Sperber, D. (2012a). Explaining Irony. In D. Wilson & D. Sperber

(Eds.), Meaning and Relevance. Cambridge: Cambridge University Press.

Wilson, D., & Sperber, D. (2012b). Meaning and Relevance. Cambridge: Cambridge

University Press.

Documents

Experimental Pragmatics