ChapterPDF Available

Pragmatic reasoning in autism



The main points to be developed in this chapter are the following: • Pragmatic competence is not a homogeneous thing, but manifests itself in the ability to deal with a broad variety of tasks, including disambiguation, establishing coherence, interpreting metaphors, and much more. • Pragmatic deficiencies observed in people with autism are not equally severe across all these sub-tasks. • Contrary to the widely held view that mind reading is essential to pragmatics, it is unclear whether mind reading is always involved in pragmatic reasoning, nor is it even clear how important it really is. Hence, as things currently stand, it is only a hypothesis that pragmatic deficits in autism are largely due to mind-reading deficits. • At least some of the data that have been claimed to provide support for this hypothesis are better explained in terms of executive function deficits, which are frequently associated with ASD.
Pragmatic reasoning in autism
Bart Geurts
Mikhail Kissine
Bob van Tiel
A quarter of a century ago, Happ´e’s review of the communicative deficits
associated with Autism Spectrum Disorder (ASD) opened as follows:
Interacting with a bright and verbal autistic child can be an eye-
opening experience: one discovers one is talking in metaphors! A
request to “Stick you coat down over there” is met by a serious
request for glue. Ask if she will “give you a hand,” and she will
answer that she needs to keep both hands and cannot cut one off
to give to you. Tell him that his sister is “crying her eyes out”
and he will look anxiously on the floor for her eyeballs. Ask her to
read a passage “out loud” and she will obligingly shout through
to the end of it. These overliteral interpretations are made in all
earnestness, and they tell us something important about autism
and about the way in which we normally communicate. (Happ´e,
1995, p. 271)
Apart from usefully reminding us that much of our everyday talk is non-
literal, this passage vividly illustrates the pragmatic impairments that are
thought to typify ASD. As a more recent source puts it:
[...] interpretation of an utterance demands an ability to go be-
yond linguistically given meaning by using and connecting rel-
evant contextual information, with which individuals with ASD
have great difficulties. (Kim, Paul, Tager-Flusberg, & Lord, 2014,
p. 247)
To appear in Thinking, reasoning and decision making in autism,
edited by Kinga Morsanyi and Ruth Byrne, Psychology Press.
Difficulties handling metaphor, irony, sarcasm, and other non-literal uses of
language have been central to the definition of autism, from the very first
descriptions to the most up-to-date nosological criteria. Kanner (1944), for
example, mentioned excessive “literalness” as one of the core features of the
twenty children with autism he described, and Asperger (1991) included lack
of understanding of jokes among the characteristics of “autistic affect”. With
remarkable consistency, the latest edition of the Diagnostic and Statistical
Manual of the American Psychiatric Association (DSM-V; American Psy-
chiatric Association, 2013, 48) includes the following characteristics among
those required for an ASD diagnosis:
[d]ifficulties understanding what is not explicitly stated (e.g. mak-
ing inferences) and nonliteral or ambiguous meanings of language
(e.g. idioms, humor, metaphors, multiple meanings that depend
on the context for interpretation).
Of course, the autism spectrum is characterized by an enormous heterogene-
ity of clinical profiles, and this is especially true for linguistic abilities; in
around 50% of children with ASD expressive language emerges with an im-
portant delay, and around 30% never achieve verbal communication (e.g. Kim
et al., 2014). However, pragmatic deficits seem to impact utterance interpre-
tation in autism irrespective of language level and persist even in individuals
whose language skills are typical otherwise. In this chapter we survey these
difficulties and explore some ways of explaining them.
One line of explanation which has been particularly influential links the
pragmatic deficits in autism with the reduced ability to attribute mental
states to oneself and others, an ability that is variously known as “mind
reading”, “theory of mind”, “mentalizing”, and “folk psychology” (we will
use the first term). In autistic people this ability tends to be compromised to
a greater or lesser extent, and this has often been taken to explain why their
pragmatic abilities are compromised, as well (cf. Chapter 3). This view hinges
on the assumption that mind reading is essentially involved in pragmatic in-
terpretation, a strong version of which is held by Sperber and Wilson (2002),
who go so far as to state that “pragmatic interpretation is ultimately an ex-
ercise in mind-reading” (p. 3) If this claim was correct, then we should expect
pragmatic abilities in ASD to be impaired across the board, but that is not
what we find. Instead, while some pragmatic skills are reduced in autism, oth-
ers seem to be unaffected (Deliens, Papastamou, Ruytenbeek, Geelhand de
Merxem, & Kissine, 2018; Kissine, 2012, 2016).
The main points to be developed in this chapter are the following:
Pragmatic competence is not a homogeneous thing, but manifests itself
in the ability to deal with a broad variety of tasks, including disam-
biguation, establishing coherence, interpreting metaphors, and much
Pragmatic deficiencies observed in people with autism are not equally
severe across all these sub-tasks.
Contrary to Sperber and Wilson’s claim, which is still widely held,
it is unclear whether mind reading is always involved in pragmatic
reasoning, nor is it even clear how important it really is. Hence, as
things currently stand, it is only a hypothesis that pragmatic deficits
in autism are largely due to mind-reading deficits.
At least some of the data that have been claimed to provide support
for this hypothesis are better explained in terms of executive function
deficits, which are frequently associated with ASD.
One last remark before we get going. Although communication perforce in-
volves speakers and hearers, pragmatic theories usually take the hearer’s
point of view. Of course, pragmatic difficulties in autism also surface on the
production side, most notably in the management of conversation dynamics
and social interaction. However, most of the empirical research adopts the
hearer’s perspective and focuses on utterance interpretation, and therefore
that is also the perspective adopted here.
What is pragmatic reasoning?
To answer this question, we first have to establish what pragmatics is, but
since that is a contentious topic (which is putting it very mildly), we will
proceed mainly by way of examples. It is relatively uncontentious to say that
pragmatics concerns language use, i.e. communication, and that context is
key to pragmatic reasoning, where “context” is a broad notion that includes
not only the situation in which communication takes place, but also the local
time, the previous discourse, linguistic and world knowledge shared between
interlocutors, and so on. As a first stab at characterizing pragmatic reasoning,
we might say that it supports the production and interpretation of language
by integrating contextual information. The following examples will serve to
make this more concrete.
Referential expressions
In each of the following examples, the purpose of the italicised expressions is
to refer to some entity or other:
(1) Fred is staring at the moon again.
(2) Wilma went after the thief, but he got away.
(3) There’s no dignity anymore and I think that’s very important. (Mae
In (1), the most likely target of the name “Fred” is some male individual
named “Fred”, but it depends on the common ground between speaker and
hearer which Fred “Fred” refers to. The same holds for “the moon”, which
probably refers to Earth’s natural satellite, but might in principle refer to any
other moon that is currently salient. In (2), “the thief” refers to some salient
thief, while “he” refers either to the aforementioned thief or, less likely, to
another salient individual. Finally, though it could be in (3) that the speaker
was pointing at an object while uttering the pronoun “that”, it is more likely
that she meant to refer either to the quality of dignity or to the fact that
there is no dignity anymore.
Nearly every sentence uttered in everyday life contains one or more refer-
ential expressions, and in order to interpret a given utterance, the hearer will
have to decide what these expressions refer to. As is shown by the examples
in (1)-(3), the target of a referential expression is constrained but not deter-
mined by its form; context always plays a role, and pragmatic reasoning is
always needed to interpret referential expressions.
Each of the italicised expressions in (1)-(3) is indeterminate in the sense
that, in principle, it could refer to any of a range of potential targets. But
interpretative indeterminacies are not only caused by referential expressions;
they also arise through various types of ambiguity. For example:
(4) The defendant drew a gun.
(5) One morning I shot an elephant in my pajamas. How he got in my
pajamas, I don’t know. (Groucho Marx)
Due to fact that the verb “draw”, like the vast majority of words, is lexically
ambiguous, (4) could mean either that the defendant pulled out a gun or
produced a picture of a gun. In (5), the source of the ambiguity is syntactic:
“in my pajamas” may be an adjunct either to “elephant” or to “I shot an
elephant”. Together with referential expressions, lexical and syntactic am-
biguities infuse an utterance with considerable amounts of indeterminacy,
which hearers have to resolve by means of pragmatic reasoning.
Non-literal meaning
Although there may be bags weighing 1,000 kilograms, (6) is most likely
understood as a hyperbolical statement meaning that the bag in question is
very heavy:
(6) That bag weighs a ton.
Likewise, though in the end it always depends on the context, the most
obvious interpretation of (7) is not that Trump is an orifice, but rather that
he is stupid, irritating, or contemptible, or all of the above, in which case the
speaker’s statement is a metaphor:
(7) Trump is an arse.
non-literal meanings come in many varieties, of which irony is perhaps the
most remarkable, because it usually involves not just a shift, but a reversal
of meaning. For example, if it has been raining for days, and I say:
(8) The weather is great.
then my utterance may well be understood as conveying that the weather is
awful, so the conveyed meaning is the opposite from the literal meaning.
Conversational implicatures
The foregoing cases involved various hurdles that must be cleared in order
to determine the meaning of an utterance. But the import of an utterance
generally goes beyond its meaning:
(9) Can you close the window?
(10) A: Did you do all of your homework?
B: I did some of it.
An utterance of (9) would normally imply that the speaker is making a
request, while he seems to be asking a question. Similarly, B’s answer in (10)
may imply that he didn’t do all of his homework, but this is not part of
the literal meaning of B’s utterance. In this case, the pragmatic reasoning
that derives the conclusion that B didn’t do all of her homework might go
as follows:
i. B has said that he did some of his homework.
ii. Being cooperative, he wouldn’t have said this if he did all of his home-
work, for in that case, it would have been more cooperative to say “I
did all of it”.
iii. Therefore, A is entitled to infer that B didn’t do all of his homework.
Such pragmatic inferences are known as “conversational implicatures”. Rather
than being part of the meaning of the speaker’s utterance, they are deriv-
able, and supposed to be derived, from it, and they critically depend on the
assumption that the speaker is cooperative (Grice, 1975).
This quick survey should suffice to show that pragmatic reasoning is a het-
erogeneous phenomenon. It comes in many flavours that vary along several
dimensions, of which we will discuss two. The most important distinction
to be drawn is between forms of reasoning that help to pin down the con-
textual meaning of an utterance, on the one hand, and forms of reasoning
that derive further inferences from that meaning, on the other. This distinc-
tion presupposes that interpretation is a two-stage process. During the first,
“pre-propositional” stage, the utterance meaning is assembled, on the basis of
which further inferences are derived during the second, “post-propositional”
stage. It should be noted that this two-stage model, which is widely accepted
in the field of pragmatics (e.g. Recanati, 2004), need not be viewed as a pro-
cessing model, no matter how tempting it may seem to do so. For example,
the fact that conversational implicatures are post-propositional inferences
does not imply that they cannot be computed until the speaker has com-
pleted his utterance (Geurts, 2010, pp. 72–74). Nor need it be the case that
post-propositional pragmatics is intrinsically harder than pre-propositional
pragmatics, as has often been suggested. Though it may be true that, as a
matter of statistical fact, resolving pronouns tends to be easier than inter-
preting metaphors, for example, there is no evidence that the first type of
task is necessarily simpler than the second.
A second way of classifying forms of pragmatic reasoning is in terms of
additivity, as we will call it. Some forms of pragmatic reasoning merely add
further information to what is given already. Ambiguity resolution is a case
in point. Given that (4) can mean either that the defendant pulled out a
gun or produced a picture of a gun, resolving the ambiguity results in a
more informative interpretation which doesn’t lose any of the information in
the ambiguous input. By contrast, if (8) is interpreted as meaning that the
weather is awful, then this implies a reversal of the standard meaning; so the
pragmatic reasoning underlying irony is non-additive.
pre-propositional post-propositional
additive reference (exs. 1-3) quantity implicatures (ex. 10)
ambiguity resolution (exs. 4-5) indirect speech acts (ex. 9)
non-additive metaphor (exs. 7, 11) irony (ex. 8)
hyperbole (ex. 6)
Table 1: Pragmatic phenomena classified along two dimensions. Caveat:
this classification is not entirely theory-neutral. For example, according to
some analyses, indirect speech acts are non-additive.
Table 1 summarizes the foregoing survey by classifying pragmatic phe-
nomena along the two dimensions we have used. Note that, although pre-
propositional forms of reasoning tend to be additive while post-propositional
forms need not be, the two distinctions do not coincide. In particular, whereas
the derivation of non-literal meanings is generally non-additive, it need not
be post-propositional. For example, in (7) the word “arse” is interpreted
metaphorically, and therefore non-additively, but since the contextual mean-
ing of the sentence is affected, the reasoning involved must be pre-propositional.
Whether pre- or post-propositional, additive or non-additive, pragmatic
reasoning is a form of abduction, or “inference to the best explanation”
(Geurts, 2010; Harman, 1965; Lipton, 2004). Making sense of an utterance
is like diagnosing a disease or a malfunctioning toaster. An utterance is an
action with a purpose, and in order to make sense of it, the hearer needs to
find out what that purpose is and how it might be served by the utterance.
This diagnostic procedure relies on the assumption that, up to a point at
least, speakers aim to optimize utterances for their purposes. To illustrate,
consider example (4) again:
(4) The defendant drew a gun.
Suppose this is said, in court, as part of a report on a robbery. Then it is
virtually certain that this utterance serves its purpose best on the assumption
that, in this context, the meaning of (4) is that the defendant pulled a gun.
This holds more or less regardless of what that purpose is. Similarly, in
order to save the premise that (11) is meaningful at all, it is most likely
that “commit suicide” is not to be construed in its literal sense, and that a
non-literal interpretation serves the utterance’s purpose best:
(11) I’m sorry I’m late: my car committed suicide on the motorway.
Since pragmatic reasoning is abductive, it is non-monotonic and lacks the ab-
solute certainty characteristic of deductive reasoning: in principle, any con-
clusion established by means of pragmatic reasoning may be overruled by
further evidence.
As we saw in the introduction, there is a popular notion that pragmatic
reasoning is essentially an exercise in mind reading. To a large extent, this
idea owes its popularity to the broader view that communication is a matter
of expressing intentions, opinions, desires, etc., on the part of speaker, and
grasping the expressed mental states, on the part of the hearer. Although this
view has deep historical roots, in modern times it is especially associated with
Grice (1957). On a Gricean account, if Barney promises Betty:
(12) I’ll walk the dog.
he thereby conveys the intention to walk the dog, and communication will
have succeeded only if Betty comes to understand that Barney has the in-
tention to walk the dog. It has been shown, notably by Bach and Harnish
(1979), that this style of analysis is applicable to a sizeable class of utterance
types, although it is conceded even by its proponents that it doesn’t apply
to all speech acts. In particular, ceremonial speech acts like, “I christen this
ship the USS Stormy Daniels”, don’t seem to require the speaker to be in
any particular mental state (Kissine, 2013b; Strawson, 1964).
Be that as it may, even if many utterances can be construed as expressing
mental states, it doesn’t follow that that is their primary purpose. After all,
although utterances typically convey information about the speaker’s gen-
der, age, and language skills, it is clearly not their primary purpose to do so.
Furthermore, although the Gricean approach may be the most popular (espe-
cially outside the field of pragmatics proper), it is not the only game in town.
For example, it is uncontroversial that promises create social commitments
for speakers: no matter what Barney’s intentions are, if he promises Betty
to walk the dog, he thereby comes under obligation to walk the dog, and
obligations are social facts. Similar observations can be made about ques-
tions, assertions, and other speech act types, and they raise the possibility
that, in general, the primary purpose of a speech act is to undertake social
commitments rather than to express psychological states (Brandom, 1994;
Clark, 2006; Geurts, 2018, 2019). This is not the place to argue for this or
any other account of speech acts. The point we want to make here is merely
that, contrary to what is often claimed or implied, it does not go without
saying that the Gricean account is correct. That is to say, it is debatable
whether pragmatic interpretation is essentially an exercise in mind reading
(Kissine, 2012, 2013a, 2016; Thompson, 2014).
Could it be, then, that some varieties of pragmatic reasoning necessarily
involve mind reading? One such variety might be metaphor, as Happ´e (1995)
and many others have suggested. Prima facie, at least, metaphors seem harder
to process than lexical ambiguities, for example, and since mind reading is
often considered to be hard, it might be surmised that processing metaphors
requires mind reading, and that this is the reason why people with autism
appear to experience difficulties with metaphorical expressions. However, this
line of thinking is problematic in at least two ways. One is that it is by no
means clear that mind reading must be hard (Geurts & Rubio-Fern´andez,
2015). The other is that, as far as we can see, there is no a priori reason for
supposing that understanding metaphors requires mind reading.
Let us belabour this point a bit by way of example (11). If someone
tells you that her car “committed suicide”, then you will infer that her car
broke down. How? Well, to begin with, if the expression “commit suicide”
is taken literally, then the speaker’s statement doesn’t make sense. Since
it is quite common for expressions to be used non-literally, hearers have a
general strategy for dealing with this kind of situation, which is to look
for more plausible meanings that are saliently associated with the literal
meaning. So you ask yourself what concepts are associated with suicide that
would make sense in this particular context, and realizing that suicide entails
the end of proper functioning, you infer that, in the context of (11), the
meaning of “commit suicide” is “break down”. This chain of inferences does
not refer to the speaker’s mental state at any point. True, you might go
on to conclude that this meaning was intended by the speaker, but that
doesn’t seem to be a prerequisite for interpreting her utterance. Hence, a
theoretical analysis of metaphor doesn’t require the require the assumption
that an utterance expresses a communicative intention that must be grasped
by the hearer in order for communication to succeed. More generally, as things
currently stand, it is an open question how much mind reading is involved in
In the last paragraph we went through a somewhat protracted argument
to the effect that, in the context of (11), “commit suicide” probably means
something like “break down”. This type of argument is quite common in
pragmatics, though opinions diverge on how it is to be formalized. It bears
emphasising, however, that the principal purpose of this style of analysis is
to show why it is reasonable for the hearer to adopt this interpretation rather
than another. There is nothing in the analysis to imply that, in order to get to
this interpretation, hearers have to go through the same chain of reasoning.
The key point here is that, as a rule, pragmatic theories are not processing
theories. A pragmatic theory may impose constraints on possible processing
models, but such constraints rarely follow straightforwardly from the theory
(Geurts & Rubio-Fern´andez, 2015).
Giving this issue its due would take us way beyond the scope of the
present chapter, but there is one phenomenon that should be mentioned in
this connection: pragmatic inferences have a way of becoming conventional.
For example, the original sense of “arse” was anatomical, and the pejorative
sense developed out of it. In the meantime the latter has conventionalized,
and is now probably more common than the former, but nevertheless the
pejorative sense remains recognizably metaphorical. In all likelihood, this
will have consequences for how the word is processed, but such consequences
do not fall into the province of pragmatic theory.
Pragmatic abilities in ASD
Contrary to what is suggested by the quote at the beginning of this chapter,
two decades of increasingly sophisticated experimental studies have shown
that pragmatic impairments in autism are neither global nor uniform. In the
remainder of this chapter we review a representative sample of these studies.
Referential expressions
Let us start with two instances of additive pre-propositional reasoning. Stud-
ies on reference processing in ASD are relatively scarce, and mostly focus on
production (see Malkin, Abbot-Smith, & Williams, 2018, for a recent meta-
analysis). A robust result of this literature is that individuals with ASD may
use fewer referential expressions, such as pronouns, which can make it hard
for their addressees to understand what they want to say (Arnold, Bennetto,
& Diehl, 2009; Colle, Baron-Cohen, Wheelwright, & van der Lely, 2008; de
Villiers, 2011). However, while the production of referential expressions is not
always optimally tuned to the hearer’s perspective, there is no evidence that
ASD is generally associated with a reduced ability to understand referen-
tial expressions. For example, using act-out tasks, Hobson, Lee, and Hobson
(2010) showed that adolescents with ASD correctly understand personal pro-
nouns; likewise, a story retelling task yielded no indication that children with
ASD have problems with reference resolution (Novogrodsky, 2013).
Another variety of pragmatic reasoning that is additive and pre-propositional
is involved in the resolution of lexical ambiguity. Difficulties in integrating
global linguistic context to resolve linguistic ambiguity have been amongst
the main arguments for Weak Coherence accounts of autism, according to
which cognitive processing in ASD is atypically focused on local properties
of the stimulus, at the expense of the global context (Frith, 1989; Happ´e
& Frith, 2006). If this is correct, one should expect individuals with ASD
to experience difficulties in using global, contextual information to cancel a
common meaning of an ambiguous word in favour of a less common one. The
bulk of evidence here comes from homograph tasks, in which participants
have to read aloud sentences that favour either more frequent or rarer pro-
nunciations of words with distinct meanings that happen to share the same
spelling. For example, whereas the sentence in (13) favours the more common
pronunciation of lead as [li:d], in (14) it is the less common [lEd] that makes
more sense.
(13) Mary wanted to take the dog for a walk, so she went to the cupboard
and took the lead.
(14) The scrap metal man first took the copper and iron and then he took
the lead. (from Jolliffe & Baron-Cohen, 1999)
Several studies using this paradigm found that children and adults with ASD
have difficulties inhibiting frequent but contextually inappropriate pronunci-
ations of homographs (Frith & Snowling, 1983; Jolliffe & Baron-Cohen, 1999).
However, what homograph tasks tap is not the pragmatic competence to use
context in order to select the appropriate lexical meaning, but rather the ca-
pacity to inhibit the most salient pronunciation of the homograph and switch
to the less frequent one. Such inhibition and flexibility skills belong to the do-
main of executive functions, viz. of cognitive skills required for (short-term)
management and control of behaviour. Since there is evidence that executive
dysfunction, especially with respect to flexibility, is part of the ASD cog-
nitive profile, it makes more sense to explain findings with the homograph
paradigm along these lines (Hill, 2004; Ozonoff, South, & Provencal, 2005;
Russell & Hughes, 1994; Zelazo, Jacques, Burack, & Frye, 2002).
This view is supported by studies using finer-grained tasks that assessed
the activation of contextually appropriate meanings without relying on pro-
nunciation found no specific impairments in autism. For example, Brock,
Norbury, Einav, and Nation (2008) used the Visual World paradigm to com-
pare anticipatory fixations on representations of a target word (e.g. “ham-
ster”) as compared to a phonological competitor (e.g. “hammer”). When the
verb primed the target, as in (15), the rate of anticipatory fixations on the
competitor was lower than after verbs whose meaning was neutral in this
respect, as in (16).
(15) John stroked the hamster quietly.
(16) John chose the hamster reluctantly.
Crucially, Brock et al. (2008) did not find a specific effect of ASD diagnosis,
even though language level did play a role. These results, which are consistent
with other studies (Hahn, Snedeker, & Rabagliati, 2015; Norbury, 2005a),
strongly suggest that autism is not, or at least not uniformly, characterized
by a reduced ability to recruit contextual information for the purpose of
Impaired comprehension of metaphors is one of the official diagnostic crite-
ria for ASD (DSM-V; American Psychiatric Association, 2013, 48), and this
diagnostic was supported already by one of the first studies on pragmatics
and ASD, which found that both children and adults with ASD had difficul-
ties understanding metaphors (Chapter 5; Happ´e, 1993). In Happ´e’s Exp. 1,
participants had to complete sentences by selecting one out of six words.
Only one of these words was appropriate, but it required a metaphorical
interpretation of the sentence, as illustrated by the word “volcano” in (17):
(17) Father was very very cross. He really was a .. .
Happ´e found that participants with ASD were significantly less likely to
select the appropriate word than participants without ASD. This difference
was weaker in sentences like (18) that made the analogical relation explicit.
(18) Father was very very cross. He was like a . . .
Happ´e reported a further finding that she argued to be relevant for the nexus
between pragmatics and mind reading. She measured to what extent partici-
pants were able to reason about other people’s beliefs by using various tasks
that required participants to attribute false beliefs (e.g. “Sally wrongly be-
lieves the marble is in the box”), and found that participants with ASD who
passed these mind-reading tasks performed equally well on the metaphor task
as participants without ASD, and significantly better than participants with
ASD who failed at these mind-reading tasks. Based on these findings, Happ´e
concluded that pragmatic skills and mind-reading skills are closely related, a
conclusion that was widely accepted in the literature.
However, Norbury (2005b) present data that undermine Happ´e’s conclu-
sions. Using a metaphor task similar to Happ´e’s, she tested mind-reading
abilities in a similar way. In addition, she measured the semantic knowl-
edge of participants using the Test of Word Knowledge, which measures how
well participants understand the meanings of words by asking them, e.g. to
provide synonyms and explain lexical ambiguities (Wiig & Secord, 1992).
Norbury found that semantic knowledge, rather than mind-reading abilities,
modulates performance on the metaphor task: once differences in semantic
knowledge are controlled for, individuals with ASD understand metaphors
just as well as individuals without ASD (cf. also Kalandadze, Norbury, Nær-
land, & Næss, 2018; Kasirer & Mashal, 2014). These data argue against the
idea that autism is characterized by problems with metaphor comprehension,
as well as against the hypothesis that pragmatic competence always requires
mind reading.
Chouinard and Cummine (2016) report data suggesting that individu-
als with ASD may process metaphors differently from individuals without
ASD, even if they ultimately arrive at the same interpretation. In this study,
participants had to indicate whether metaphorical sentences like (19) were
literally true or false:
(19) Some hearts are ice.
Participants with and without ASD were slower to respond that metaphorical
sentences were literally false when compared to non-metaphorical sentences,
which suggests that both groups of participants derived the metaphorical in-
terpretation of (18). However, this slowdown was greater for participants with
ASD, who were also more likely to make errors on metaphorical sentences
than on non-metaphorical controls. These results suggest that participants
with ASD had problems inhibiting the unintended interpretation of these
Indirect speech acts
Indirect speech acts are clear cases of post-propositional inferences. It has
often been claimed in the literature that individuals with ASD experience
difficulties interpreting (20) or (21) as requests to close the door:
(20) Could you close the door?
(21) I’d like you to close the door.
As the syntactic structures of (20) and (21) are prototypically associated with
questions and assertions, respectively, it was thought that individuals with
ASD should have difficulties overriding these structural cues in order to de-
rive contextually appropriate interpretations. However, there is no clear-cut
experimental evidence to support this hypothesis. Paul and Cohen (1985) re-
port that, in a spontaneous drawing task, adults with ASD sometimes failed
to follow an experimenter’s suggestion to colour a part of the drawing blue
(instead of red), when this request was conveyed indirectly, e.g. as “Why
colour this house red?” or “Doesn’t it need blue?” (Participants did com-
ply with suggestions like “Why not colour this blue?” or “It needs blue.”)
These results were interpreted as indicating a deficit in pragmatic process-
ing, but the failure to comply with such convoluted requests hardly counts
as evidence for a deficient understanding of indirect requests. When Kissine,
De Brabanter, and Leybaert (2012) analysed an extensive video corpus of
low-functioning children with ASD, they found that these children are as
likely to comply with direct requests, like (22), as with indirect ones, such as
(22) Pour the milk.
(23) You forgot the water in your bag. [Intended meaning: “Go and fetch
the water from your bag.”]
Though it seems that children and adults with ASD are able to comply
with indirect requests, one may still wonder whether their pragmatic under-
standing of such requests is standard in all respects. MacKay and Shaw (2005)
presented children diagnosed with Asperger syndrome or high-functioning
autism with illustrated short stories, in some of which a protagonist made
an indirect request, such as (24):
(24) These crisps look lovely.
Although children with ASD correctly understood the directive meaning of
such indirect requests equally often as a typically-developing comparison
group, they often failed to correctly explain why the speaker used the in-
direct form. Difficulties pertaining to the understanding of the reasons why
the speaker opted for this or that form are certainly telling about these chil-
dren’s interactional deficits. This issue, however, is orthogonal to the assess-
ment of the capacity to obtain pragmatic interpretations in ASD. A plausible
implication of MacKay and Shaw’s (2005) results is that the correct interpre-
tation of at least some indirect speech acts can be achieved without making
assumptions about the speaker’s beliefs and intentions; this is in line with
the experimental evidence on metaphor in ASD.
This line of thought is supported by two further studies which strongly
suggest that in spite of well-documented difficulties in mental state attri-
bution, individuals with ASD are capable of deriving indirectly conveyed
meanings. Kissine et al. (2015) had children with ASD play with several
Mr. Potato Heads, in the presence of two experimenters: one who played
with the child while the other was ostensibly withdrawn from the interac-
tion, her back to the child, and reading a magazine. In the first context,
when the first experimenter uttered (25), children accurately responded to
her utterance by looking for a hat to put on a Mr. Potato Head.
(25) Oh! He has no hat.
But when the second speaker produced the same utterance later on, this time
meant as a comment on the magazine she was reading, almost none of the
children in the ASD group seemed to interpret her utterance as a request
or a suggestion to put a hat on the other Mr. Potato Head they were play-
ing with at that moment. Hence this study showed that children with ASD
are flexible about the interpretations they assign to an utterance, and that
their interpretations are not determined solely by linguistic form. In Deliens,
Papastamou, et al. (2018, Exp. 1), adults with ASD were provided with a
touchscreen displaying a grid with geometrical figures and yes and no but-
tons. Participants were instructed to follow pre-recorded audio instructions.
Some instructions, such as (26), could only be complied with by moving a
shape on the grid; some others, such as (27), unambiguously were questions,
to be answered by touching the yes or the no button.
(26) Move the green triangle on the right of the red square.
(27) Is the green triangle on the right of the red square?
Other instructions, however, were ambiguous and could be interpreted either
as questions (to which the answer was “yes”) or requests:
(28) Can you move the green triangle on the right of the red square?
(29) Is it possible to move the green triangle on the right of the red square?
Participants with ASD responded by moving a shape in the grid equally often
as neuro-typical controls. Analyses of reaction times and eye-movements did
not reveal any significant group differences, either. Importantly, replicating
an earlier study with neurotypical participants (Ruytenbeek, Ostaschenko,
& Kissine, 2017), these results showed that when participants opted for a
“request” interpretation of an ambiguous instruction, they did not respond
more slowly than to unambiguous, imperative requests, and did not look at
the yes and no buttons (which would have been indicative of hesitation
between a “question” and a “request” response).
Scalar implicatures
Like indirect requests, the interpretation of scalar expressions, such as “some”
and “or”, involves additive and post-propositional pragmatic reasoning (Geurts,
2010; van Tiel, 2014). It is commonly assumed that the literal meaning of
“some” and “or” can be paraphrased as “at least some and possibly all” and
“or and possibly and”. Someone who utters (30), however, may imply that
not all dogs are mammals, and someone who utters (31) that zebras do not
have black and white stripes. These scalar inferences are usually explained
as a variety of conversational implicature. Thus someone who says (30) could
have been more informative, hence more cooperative, by saying “All dogs are
mammals.” Why didn’t she? Presumably because she believes not all dogs
are mammals. Thus, (30) and (31) are true when interpreted literally, but
false if their respective scalar inferences are derived.
(30) Some dogs are mammals.
(31) Zebras have black or white stripes.
Pijnacker, Hagoort, Buitelaar, Teunisse, and Geurts (2009) were the first
to test whether individuals with ASD were less likely to derive scalar in-
ferences than individuals without ASD. In their study, participants had to
indicate whether they considered sentences such as (30) and (31) to be true
or false. Participants who derived the scalar inferences were predicted to an-
swer “false”, while participants who interpreted the sentences literally were
predicted to answer “true”. Pijnacker et al. found no difference in the rates
of “false” responses between participants with and without ASD. Moreover,
within the group of participants with ASD, Pijnacker et al. observed a sig-
nificant effect of language competence, such that individuals with greater
language competence were more likely to derive scalar inferences. Both of
these findings were confirmed in later studies on children and adolescents
with ASD (Chevallier, Wilson, Happ´e, & Noveck, 2010; Su & Su, 2015; van
Tiel & Kissine, in press).
As noted earlier, one of the hallmarks of pragmatic inferences is that
they are defeasible, i.e. they can be overruled by further evidence. Thus,
while (32) will normally imply that not all of the boxes contain strawberries,
this inference need not be warranted if it is given that the speaker didn’t
inspect all the boxes, because, in that case, the reason the speaker didn’t say
“All of the boxes have strawberries” is that she lacks the requisite knowledge:
(32) Some of the boxes have strawberries.
Hochstein, Bale, and Barner (2018) showed that, indeed, whereas neurotyp-
ical participants vary their interpretation of sentences like (32) based on the
speaker’s knowledge, participants with ASD interpreted them with a scalar
implicature irrespective of whether the speaker knew what was in all of the
boxes. This finding suggests that mind-reading deficits may affect the prag-
matic understanding of scalar expressions in individuals with ASD. However,
an alternative explanation would be that individuals with ASD have prob-
lems inhibiting the scalar inference, which may be relatively salient, much
like individuals with ASD have problems inhibiting the most salient pronun-
ciation of homographs, as we have seen.
In any case, the observation that individuals with and without ASD are
equally likely to derive scalar inferences should not be construed as imply-
ing that they are equally adept at reasoning about the speaker’s reasons for
being underinformative. Thus, van Tiel and Kissine (in press) investigated
whether the probability of deriving various underinformativity-based infer-
ences varies with the degree to which one has autistic traits, as measured
by the autism spectrum quotient questionnaire (Baron-Cohen, Wheelwright,
Skinner, Martin, & Clubley, 2001). In line with previous research, van Tiel
and Kissine found no effect of autistic traits on the probability of deriving
scalar inferences; however, they did find that participants with more autis-
tic traits were less likely to derive so-called distributivity inferences, which
belong to the same pragmatic family as scalar inferences (Geurts, 2010). Dis-
tributivity inferences are licensed when “or” is embedded under a universal
quantifier, as in (33).
(33) Each of the shapes is red or green.
Someone who utters this sentence may imply that there are both red shapes
and green shapes. The pragmatic reasoning that underlies these distributivity
inferences may be spelled out as follows:
i. The speaker has said that each of the shapes is red or green.
ii. Being cooperative, he wouldn’t have said this if all of the shapes are
red and none are green, for in that case, it would have been more
cooperative to say “Each of the shapes is red.”
iii. He also wouldn’t said this if all of the shapes are green and none are
red, for in that case, it would have been more cooperative to say “Each
of the shapes is green.”
iv. Therefore, the hearer is entitled to infer that there are both red shapes
and green shapes.
The observation that the probability of deriving distributivity inferences but
not scalar inferences varies with the degree to which one has autistic traits
shows that the pragmatic skills of people with ASD may vary even within
the same class of pragmatic inferences.
Irony may be one of the most complex pragmatic phenomena, as it is defined
by a conflict between literal and conveyed meaning, in some respect or other.
In some cases at least, reversing the literal interpretation and getting to the
ironic meaning requires that the speaker’s intentions be taken into account.
For instance, a speaker who volunteers (34) as a comment on a long and
confused conference presentation, may be mistaken, insincere, or ironic:
(34) This talk was great.
If mistaken, the speaker says something she wrongly believes to be true;
if insincere, she says something she takes to be false without wanting her
hearers to realize that; and if she is being ironic, she says something she
takes to be false, and does want her hearers to realize. The most reliable
route to rule out the first two possibilities in the favour of the third is to
assume that the speaker cannot possibly have liked the talk, and, further-
more, that she believes that this is also obvious to her addressee (see, e.g.
Bryant, 2012; Deliens, Antoniou, Clin, Ostaschenko, & Kissine, 2018). It
doesn’t come a surprise, therefore, that experimental studies indicate that
the capacity to understand irony strongly correlates with mind-reading com-
petence (e.g. Akimoto, Miyazawa, & Muramoto, 2012; Filippova & Asting-
ton, 2008; Spotorno & Noveck, 2014). These findings make it plausible that
the robustly attested mind-reading deficits in ASD should cause difficulties
in irony comprehension. Given the central role mind reading plays in irony
comprehension, one should expect these difficulties to be more severe rela-
tive to the other types of pragmatic phenomena surveyed in the foregoing,
including metaphor (Andr´es-Roqueta & Katsos, 2017; Kissine, 2012, 2013a,
Some studies report difficulties in understanding irony in adults with
ASD, as tapped by story completion or interpretation tasks (Happ´e, 1993;
Kaland et al., 2002; Martin & McDonald, 2004), but this pattern is not en-
tirely consistent with other studies. Somewhat surprisingly, in a recent meta-
analysis of non-literal language in autism, Kalandadze et al. (2018) found
that, overall, group differences between individuals with ASD and neurotyp-
icals are stronger in the comprehension of metaphor than of irony. A closer
look reveals, however, that irony comprehension in autism may have been
over-estimated due to methodological factors. Since these factors have im-
portant theoretical implications, it is worth looking in some detail at stud-
ies that report intact irony comprehension in ASD. In Chevallier, Noveck,
Happ´e, and Wilson (2011), ironic items were uttered with a marked prosody,
while non-ironic items were uttered in a neutral tone of voice. Participants
had then to choose between two options: a literal one and an ironic one, the
latter being explicitly incongruent with the target literal meaning.
(35) Context: Glenn tells Phil that he decided to come by plane rather than
by train.
Ben says: How clever of you! [target]
Option 1: Ben really thinks that Glenn’s decision was right because
the trains are always late. [admiration]
Option 2: Ben actually thinks that Glenn is silly because the plane
takes longer than the train. [irony]
Chevallier et al.’s participants with ASD seemed to have no trouble correctly
choosing the ironic option in this task. Wang, Lee, Sigman, and Dapretto
(2006), Colich et al. (2012) and Pexman et al. (2011) also report above-chance
performance in ASD in a forced-choice discrimination task in which ironic
items were associated with contextual incongruence and marked prosody.1
In all these cases, participants could rely on prosodic cues and evident
incongruence with the literal meaning to help them choose between two in-
terpretations. However, in real life cues to irony are not nearly as systematic,
and therefore the interpreter’s task is much harder. Deliens, Antoniou, et al.
(2018) showed that, in a forced-choice task, participants correctly classified
marked prosodic contours and facial expressions as ironic or non-ironic. How-
ever, the same cues caused poor performance in an act-out task, in which
participants had to figure out, based on an utterance, which of two avail-
able objects the speaker really wanted. In each trial of this task, participants
watched a short video with two protagonists sitting in front of two objects,
e.g. a cup of tea and a glass of milk. One protagonist offered one of these two
objects to the second one, asking “Would you like a cup of tea?”, for example.
1Interestingly, in Wang et al. (2006) and Colich et al. (2012) neural activation patterns
differed between ASD and neurotypical groups.
After that the second protagonist reacted in a positive or negative way, ut-
tering the target sentence, e.g. “Yes, you know how much I love tea” or “No,
I don’t like tea.” In some trials, this target sentence was ironic. Participants
had to select the item that the second protagonist really wanted.
Crucially, in this design, ironic stimuli were not systematically associated
with distinctive prosody, contextual incongruence, or facial expression, as
sometimes only one or two of these cues were available. For example, in some
but not all trials the first protagonist provided background information about
the second protagonist’s preferences, e.g. by saying “I know that you like
tea for your breakfast.” Moreover, only some ironic targets were associated
with a distinctively ironic intonation and/or facial expression, and non-ironic
stimuli were also associated with marked but non-ironic prosody or facial
expressions. In brief, this design did not associate irony with a distinctive
pattern of stimuli, and participants with ASD performed significantly worse
than neurotypical controls (Deliens, Papastamou, et al., 2018, Exp. 2).2It
appears, therefore, that in adults with ASD the capacity to detect irony
requires special circumstances that are usually not given in everyday life.
Therefore, it seems fair to say that the positive results reported in earlier
studies overestimate the pragmatic competence of adults with ASD.
Engaging in everyday communication continually requires pragmatic reason-
ing in order to deal with various types of reference, ambiguity, non-literal
meaning, implicatures, indirect speech acts, and so on. It is well established
that people with ASD exhibit pragmatic deficiencies, but contrary to popular
opinion it is not the case that, in ASD, pragmatic competence is impaired
across the board. Instead, the experimental record suggests a motley pattern
of local deficiences, which fail to align with any of the divisions standardly
made by pragmatic theories. For instance, although it has been claimed that
the understanding of metaphor is uniformly impaired in ASD, this idea is not
borne out by experimental evidence, and the same holds for the claim that
people with ASD always have problems with the post-propositional reasoning
required for understanding indirect speech acts or irony.
2These were the same participants who, in Deliens, Papastamou, et al. (2018, Exp. 1),
produced the same rates of indirect requests as the comparison group.
It is well established that people with ASD tend to have problems adopt-
ing other people’s perspective. Therefore, the existence of intact pragmatic
processing in ASD speaks against the notion that all pragmatic processing re-
quires mind reading. This conclusion is confirmed by theoretical analysis, for
as things stand it is an open question when and to what extent communica-
tion relies on mind reading. As a consequence, there is currently no support
for the view that pragmatic reasoning which does require mind reading is
uniformly impaired in ASD, simply because we do not know yet which forms
of pragmatic reasoning require mind reading, and which don’t. Be that as it
may, it is unlikely that mind-reading deficits are the only causal factor under-
lying impaired pragmatic reasoning in ASD, if only because it is in the nature
of pragmatic reasoning, as a form of inference to the best explanation, that
it requires executive abilities such as inhibition and flexibility, which tend to
be compromised in people with ASD. Therefore, we should expect pragmatic
reasoning in ASD to be impaired for this reason, too, and as we have seen
at several points in this chapter, there is experimental evidence confirming
that expectation.
Pragmatic theories make various distinctions between types of commu-
nicative skills and sub-skills, and there is a certain amount of consensus
on what the key distinctions are. This chapter started out from typological
distinctions that we take to be fairly uncontroversial, and used them as a
backdrop for discussing pragmatic impairments associated with ASD. This
was the most obvious procedure to adopt, but it has led us to the conclu-
sion that this may not be the best way of approaching pragmatic reasoning
in ASD. Instead, it seems more advisable to focus on the types of process-
ing involved in pragmatic reasoning, and consider in each case if and how
it might be compromised in ASD. Typologies offered by pragmatic theories
might still be useful as a heuristic point of departure for such an approach,
but mental processes are not bound to respect any typology of pragmatic
reasoning, understood as a set of communicative skills.
The authors are grateful to Ruth Byrne, Kinga Morsanyi, and each other, for
many helpful comments on earlier versions of this chapter or parts thereof.
Akimoto, Y., Miyazawa, S., & Muramoto, T. (2012). Comprehension pro-
cesses of verbal irony: the effects of salience, egocentric context, and
allocentric theory of mind. Metaphor and symbol,27 , 217–242.
American Psychiatric Association. (2013). Diagnostic and statistical manual
of mental disorders (5th ed.). Washington, DC: Author.
Andr´es-Roqueta, C., & Katsos, N. (2017). The contribution of grammar,
vocabulary and theory of mind in pragmatic language competence in
children with autistic spectrum disorders. Frontiers in psychology ,8,
Arnold, J. E., Bennetto, L., & Diehl, J. J. (2009). Reference production in
young speakers with and without autism: Effects of discourse status
and processing constraints. Cognition,110 , 131–146.
Asperger, H. (1991). ‘Autistic psychopathy’ in childhood. In U. Frith (Ed.),
Autism and Asperger syndrome (pp. 37–92). Cambridge, United King-
dom: University Press.
Bach, K., & Harnish, R. M. (1979). Linguistic communication and speech
acts. Cambridge, MA: MIT Press.
Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & 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 ,
Brandom, R. (1994). Making it explicit. Cambridge, MA: Harvard University
Brock, J., Norbury, C. F., Einav, S., & Nation, K. (2008). Do individuals with
autism process words in context? Evidence from language-mediated
eye-movements. Cognition,108 , 896–904.
Bryant, G. A. (2012). Is verbal irony special? Language and linguistics
compass ,6, 673–685.
Chevallier, C., Noveck, I., Happ´e, F. G. E., & Wilson, D. (2011). What’s
in a voice? Prosody as a test case for the theory of mind account of
autism. Neuropsychologia,49 , 507–517.
Chevallier, C., Wilson, D., Happ´e, F. G. E., & Noveck, I. (2010). Scalar
inferences in autism spectrum disorders. Journal of autism and devel-
opmental disorders,40 , 1104–1017.
Chouinard, B., & Cummine, J. (2016). All the world’s a stage: Evaluation of
two stages of metaphor comprehension in people with autism spectrum
disorder. Research in Autism Spectrum Disorders,23 , 107–121.
Clark, H. H. (2006). Social actions, social commitments. In N. Enfield &
S. C. Levinson (Eds.), Roots of human sociality: culture, cognition and
interaction (pp. 126–150). Oxford, United Kingdom: Berg.
Colich, N. L., Wang, A.-T., Rudie, J. D., Hernandez, L. M., Bookheimer,
S. Y., & Dapretto, M. (2012). Atypical neural processing of ironic
and sincere remarks in children and adolescent with autism spectrum
disorders. Metaphor and symbol,27 , 70–92.
Colle, L., Baron-Cohen, S., Wheelwright, S., & van der Lely, H. K. J. (2008).
Narrative discourse in adults with high-functioning autism or Asperger
syndrome. Journal of autism and developmental disorders,38 , 28–40.
Deliens, G., Antoniou, K., Clin, E., Ostaschenko, E., & Kissine, M. (2018).
Context, facial expression and prosody in irony processing. Journal of
memory and language,99 , 35–48.
Deliens, G., Papastamou, F., Ruytenbeek, N., Geelhand de Merxem, P., &
Kissine, M. (2018). Selective pragmatic impairment in autism spectrum
disorder: indirect requests vs irony. Journal of autism and developmen-
tal disorders.
de Villiers, J. (2011). ‘I saw the yellowish going south’: narrative discourse
in autism spectrum disorder. Belgian journal of linguistics,25 , 3-29.
Filippova, E., & Astington, J. W. (2008). Further development in social rea-
soning revealed in discourse irony understanding. Child development ,
79 , 126–138.
Frith, U. (1989). Autism: explaining the enigma. Oxford, United Kingdom:
Frith, U., & Snowling, M. (1983). Reading for meaning and reading for
sound in autistic and dyslexic children. British journal of developmental
psychology,1, 329–342.
Geurts, B. (2010). Quantity implicatures. Cambridge, United Kingdom:
University Press.
Geurts, B. (2018). Making sense of self talk. Review of philosophy and
psychology,9, 271–285.
Geurts, B. (2019). Communication as commitment sharing: speech acts,
implicatures, common ground. Theoretical linguistics .
Geurts, B., & Rubio-Fern´andez, P. (2015). Pragmatics and processing. Ratio,
28 , 446–469.
Grice, H. P. (1957). Meaning. Philosophical review,66 , 377–388.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan
(Eds.), Syntax and semantics 3: Speech acts (pp. 41–58). New York,
NY: Academic Press.
Hahn, N., Snedeker, J., & Rabagliati, H. (2015). Rapid linguistic ambiguity
resolution in young children with autism spectrum disorder: Eye track-
ing evidence for the limits of weak central coherence. Autism Research,
8, 717–726.
Happ´e, F. G. E. (1993). Communicative competence and theory of mind in
autism: a test of relevance theory. Cognition,48 (2), 101-119.
Happ´e, F. G. E. (1995). Understanding minds and metaphors: insights from
the study of figurative language in autism. Metaphor and symbolic
activity,10 , 275–295.
Happ´e, F. G. E., & Frith, U. (2006). The weak coherence account: detail-
focused cognitive style in autism spectrum disorders. Journal of autism
and developmental disorders,36 , 5–25.
Harman, G. H. (1965). The inference to the best explanation. Philosophical
review,74 , 88–95.
Hill, E. (2004). Executive dysfunction in autism. Trends in cognitive sciences,
8, 26–32.
Hobson, R. P., Lee, A., & Hobson, J. A. (2010, 01). Personal pronouns and
communicative engagement in autism. Journal of autism and develop-
mental disorders,40 , 653–664.
Hochstein, L., Bale, A., & Barner, D. (2018). Scalar implicature in absence of
epistemic reasoning? The case of autism spectrum disorder. Language
learning and development,14 , 224–240.
Jolliffe, T., & Baron-Cohen, S. (1999). A test of central coherence theory:
linguistic processing in high-functioning adults with autism or Asperger
syndrome: is local coherence impaired. Cognition,71 , 149–185.
Kaland, N., Moller-Nielsen, A., Callesen, K., Mortensen, E. L., Gottlieb, D.,
& Smith, L. (2002). A new ‘advanced’ test of theory of mind: evidence
from children and adolescents with Asperger syndrome. Journal of
child psychology and psychiatry and allied disciplines,43 , 517–528.
Kalandadze, T., Norbury, C., Nærland, T., & Næss, K.-A. B. (2018). Fig-
urative language comprehension in individuals with autism spectrum
disorder: a meta-analytic review. Autism,22 , 99-117.
Kanner, L. (1944). Early infantile autism. The Journal of pediatrics ,25 ,
Kasirer, A., & Mashal, N. (2014). Verbal creativity in autism: comprehen-
sion and generation of metaphoric language in high-functioning autism
spectrum disorder and typical development. Frontiers in human neu-
roscience,8, 615.
Kim, S. H., Paul, R., Tager-Flusberg, H., & Lord, C. (2014). Language
and communication in autism. In Handbook of autism and pervasive
developmental disorders (4th ed.) (pp. 230–262). Hoboken, NJ: Wiley.
Kissine, M. (2012). Pragmatics, cognitive flexibility and autism spectrum
disorders. Mind and language,27 , 1–28.
Kissine, M. (2013a). From utterances to speech acts. Cambridge, United
Kingdom: University Press.
Kissine, M. (2013b). Speech act classifications. In M. Sbis`a & K. Turner
(Eds.), Pragmatics of speech actions (pp. 173–202). Berlin, Germany:
De Gruyter.
Kissine, M. (2016). Pragmatics as metacognitive control. Frontiers in psy-
chology,6, 2057.
Kissine, M., Cano-Chervel, J., Carlier, S., De Brabanter, P., Ducenne, L.,
Pairon, M.-C., . . . Leybaert, J. (2015). Children with autism un-
derstand indirect speech acts: evidence from a semi-structured act-out
task. PLoS ONE ,10 , e0142191.
Kissine, M., De Brabanter, P., & Leybaert, J. (2012). The interpretation of
requests in children with autism: the effect of sentence-type. Autism,
16 , 523–532.
Lipton, P. (2004). Inference to the best explanation. London, United King-
dom: Routledge.
MacKay, G., & Shaw, A. (2005). A comparative study of figurative language
in children with autistic spectrum disorders. Child language teaching
and therapy,20 , 13–32.
Malkin, L., Abbot-Smith, K., & Williams, D. (2018). Is verbal reference
impaired in autism spectrum disorder? A systematic review. Autism
and developmental language impairments,3, 1–24.
Martin, I., & McDonald, S. (2004). An exploration of causes of non-literal
language problems in individuals with Asperger syndrome. Journal of
autism and developmental disorders,34 , 311–328.
Norbury, C. F. (2005a). Barking up the wrong tree? Lexical ambiguity
resolution in children with language impairments and autistic spectrum
disorders. Journal of experimental child psychology ,90 , 141–171.
Norbury, C. F. (2005b). The relationship between theory of mind and
metaphor: evidence from children with language impairement and
autistic spectrum disorder. British journal of developmental psychol-
ogy,23 , 383–399.
Novogrodsky, R. (2013). Subject pronoun use by children with autism spec-
trum disorders (ASD). Clinical linguistics and phonetics,27 , 85–93.
Ozonoff, S., South, M., & Provencal, S. (2005). Executive functions. In
F. R. Volkmar, R. Paul, A. Klin, & D. Cohen (Eds.), Handbook of
autism and pervasive developmental disorders (3rd ed.) (pp. 606–627).
Hoboken, NJ: Wiley.
Paul, R., & Cohen, D. J. (1985). Comprehension of indirect requests in adults
with autistic disorders and mental retardation. Journal of speech and
hearing research,28 , 475–479.
Pexman, P. M., Rostad, K. R., McMorris, C. A., Climie, E. A., Stowkowy, J.,
& Glenwright, M. R. (2011). Processing of ironic language in children
with high-functioning autism spectrum disorder. Journal of autism and
developmental disorders,41 , 1097–1112.
Pijnacker, J., Hagoort, P., Buitelaar, J., Teunisse, J.-P., & Geurts, B. (2009).
Pragmatic inferences in high-functioning adults with autism and As-
perger syndrome. Journal of autism and developmental disorders,39 ,
Recanati, F. (2004). Literal meaning. Cambridge, United Kingdom: Univer-
sity Press.
Russell, J., & Hughes, C. (1994). Evidence for executive dysfunction in
autism. Neuropsychologia,32 , 477–492.
Ruytenbeek, N., Ostaschenko, E., & Kissine, M. (2017). Indirect request pro-
cessing, sentence-types and illocutionary forces. Journal of pragmatics,
119 , 46–62.
Sperber, D., & Wilson, D. (2002). Pragmatics, modularity and mind-reading.
Mind and language,17 , 3–23.
Spotorno, N., & Noveck, I. A. (2014). When is irony effortful? Journal of
experimental psychology. General,143 , 1649–1665.
Strawson, P. F. (1964). Intention and convention in speech acts. Philosophical
review,73 , 439-460.
Su, Y., & Su, L.-Y. (2015). Interpretation of logical words in Mandarin-
speaking children with autism spectrum disorders: uncovering knowl-
edge of semantics and pragmatics. Journal of autism and developmental
disorders,45 , 1938–1950.
Thompson, J. R. (2014). Meaning and mindreading. Mind and language,
29 , 167–200.
van Tiel, B., & Kissine, M. (in press). Quantity-based reasoning in the
broader autism phenotype: a web-based study.To appear in: Applied
van Tiel, B. (2014). Quantity matters (Unpublished doctoral dissertation).
University of Nijmegen.
Wang, A. T., Lee, S. S., Sigman, M., & Dapretto, M. (2006). Neural basis of
irony comprehension in children with autism: the role of prosody and
context. Brain ,129 , 932–943.
Wiig, E. H., & Secord, W. (1992). Test of word knowledge. London, United
Kingdom: The Psychological Corporation.
Zelazo, P. D., Jacques, S., Burack, J. A., & Frye, D. (2002). The rela-
tion between theory of mind and rule use: evidence from persons with
autism-spectrum disorders. Infant and child development,11 , 171–195.
... Accumulating evidence suggests that children with autism spectrum disorder (ASD) experience difficulties understanding non-literal language (of which humor is one subtype), including irony [28][29][30], metaphors [31][32][33], and idioms [32,34,35]. The evidence also suggests they lean towards literal interpretations of idioms instead of their lexicalized idiomatic meaning [32,36] and show difficulties in understanding and enjoying jokes, especially those that require considering people's (false) mental states [37]. ...
Full-text available
Semantic jokes involve resolving an incongruity emerging from wordplay or from a violation of world knowledge. The research has shown individuals with autism spectrum disorder (ASD) demonstrate a lower performance on humor tasks involving social situations; however, less is known about their semantic joke comprehension. This study examines semantic joke comprehension among adolescents with ASD and its possible relationship to vocabulary size, theory of mind (ToM), and idiom comprehension. Thirty-two adolescents with ASD and 32 typically developed (TD) peers participated. Semantic joke comprehension was assessed via a multiple-choice questionnaire and time-limited computer program. Vocabulary, ToM abilities, and idiom comprehension were also tested. The results reveal that adolescents with ASD are as fast in processing semantic jokes as their age- and vocabulary-matched TD peers, but less accurate. Age and idiom comprehension significantly contributed to semantic joke comprehension among both groups. As semantic joke comprehension is based on incongruity resolution, the greater difficulties in comprehension among the adolescents with ASD may have been due to deficits in simultaneously retaining two alternative interpretations and selecting the relevant one (and not due reduced ToM abilities). Similar to the TD group, semantic joke comprehension among the ASD group appeared to be more developed with age.
... Both the understanding for non-literal language and the TB task might be applications of advanced pragmatics understanding. However, they might not necessarily be seen as a unified phenomenon nor be subject to a same development [45]. ...
Full-text available
The true belief (TB) control condition of the classical location-change task asks children to ascribe a veridical belief to an agent to predict her action (analog to the false belief (FB) condition to test Theory of Mind (ToM) abilities). Studies that administered TB tasks to a broad age range of children yielded surprising findings of a U-shaped performance curve in this seemingly trivial task. Children before age four perform competently in the TB condition. Children who begin to solve the FB condition at age four, however, fail the TB condition and only from around age 10, children succeed again. New evidence suggests that the decline in performance around age four reflects pragmatic confusions caused by the triviality of the task rather than real competence deficits in ToM. Based on these results, it can be hypothesized that the recovery of performance at the end of the U-shaped curve reflects underlying developments in children's growing pragmatic awareness. The aim of the current set of studies, therefore, was to test whether the developmental change at the end of the U-shaped performance curve can be explained by changes in children's pragmatic understanding and by more general underlying developmental changes in recursive ToM or recursive thinking in general. Results from Study 1 (N = 81, 6-10 years) suggest that children's recursive ToM, but not their advanced pragmatic understanding or general recursive thinking abilities predict their TB performance. However, this relationship could not be replicated in Study 2 (N = 87, 6-10 years) and Study 3 (N = 64, 6-10 years) in which neither recursive ToM nor advanced pragmatic understanding or recursive thinking explained children's performance in the TB task. The studies therefore remain inconclusive regarding explanations for the end of the U-shaped performance curve. Future research needs to investigate potential pragmatic and general cognitive foundations of this developmental change more thoroughly.
... 10 We are here ignoring the entry "2" for the age of one of the participants. 11 Note that the circumstance that rational reconstructions in terms of abductive reasoning like this can be carried out does not mean that they play a role for the underlying psychological processes, or that they could not have become conventionalized in time (for discussion, see Geurts, Kissine, & van Tiel, 2020). ...
Full-text available
It is widely held that there are important differences between indicative conditionals (e.g. "If the authors are linguists, they have written a linguistics paper") and subjunctive conditionals (e.g. "If the authors had been linguists, they would have written a linguistics paper"). A central difference is that indicatives and subjunctives convey different stances towards the truth of their antecedents. Indicatives (often) convey neutrality: for example, about whether the authors in question are linguists. Subjunctives (often) convey the falsity of the antecedent: for example, that the authors in question are not linguists. This paper tests prominent accounts of how these different stances are conveyed: whether by presupposition or conversational implicature. Experiment 1 tests the presupposition account by investigating whether the stances project-remain constant-when embedded under operators like negations, possibility modals, and interrogatives, a key characteristic of presuppositions. Experiment 2 tests the conversational-implicature account by investigating whether the stances can be cancelled without producing a contradiction, a key characteristic of implicatures. The results provide evidence that both stances-neutrality about the antecedent in indicatives and the falsity of the antecedent in subjunctives-are conveyed by conversational implicatures.
Pragmatik dil becerilerin değerlendirilmesi, özellikle dil ve sosyal gelişim alanında bozuklukları olan çocukların değerlendirilmesinde merkeze alınması gereken bir konudur. Sosyal, dilsel, bilişsel ve kültürel etkilerin pragmatik dil üzerindeki karmaşık etkileşimi nedeniyle, değerlendirmeye yönelik norm bağımlı bir yaklaşım ele almak tek başına yeterli değildir. Pragmatik dil becerileri çevre, ortam ve bağlamdan etkilenen becerilerdir ve bu becerilerin yapılandırılmış ve standardize yöntemlerle değerlendirmek güçleşmektedir. Dolayısıyla, pragmatik dil becerilerinin değerlendirilmesinde standardize testlerin yanı sıra, çocukları farklı ortam ve bağlamlarda, farklı kişilerle etkileşim kurarken gözlemleme fırsatı olan ebeveyn, öğretmen ve uzmanların gözlemlerini de kullanmak önemlidir. Bu nedenle ebeveyn ve uzmanlar tarafından yanıtlanan kontrol listelerinin de etkili bir şekilde kullanılması gerekli görülmektedir. Bununla birlikte çocukların pragmatik dil becerilerini doğal gözlem ve anlatı analizi yöntemleriyle de değerlendirmek mümkündür ve uzmana çocuğun pragmatik dil performansı hakkında daha detaylı bilgi sağlamaktadır. Bu bilgiler ışığında bu derlemenin amacı pragmatik dil becerilerinin değerlendirilmesinde kullanılan değerlendirme yöntem ve araçlarını literatür doğrultusunda incelemektir.
Full-text available
One of the main distinctive features of social interaction in our species is that we use language to coordinate our future activities, and in many cases far ahead. Non-human primates don't do this, as a consequence of which their interactions remain comparatively simple and short-range. This paper offers the first in-depth discussion of this key distinctive feature of human communication. It is argued that its evolution was conditional on two developments: an increase of responsiveness during the communicative exchange and the emergence of normative behaviours in the follow-up. Responsiveness was required to coordinate future interactions, but wasn't enough for coordinating interactions beyond the immediate future, which required normativity, to boot.
Full-text available
Pragmatic difficulties are considered a hallmark of autism spectrum conditions (ASC), but remain poorly understood. We discuss and evaluate existing hypotheses regarding the literalism of ASC individuals, that is, their tendency for literal interpretations of non-literal communicative intentions. We present evidence that reveals a developmental stage at which neurotypical children also have a tendency for literalism and suggest an explanation for such behaviour that links it to other behavioural, rule-following, patterns typical of that age. We discuss evidence showing that strict adherence to rules is also widespread in ASC, and suggest that literalism might be linked to such rule-following behaviour.
Full-text available
How to use language properly and acquire the capacity for language use has become the focus of linguists and philosophers for centuries. Therefore, pragmatic competence underlying language use arouses enormous interests of language acquisition practitioners. This study reveals the core properties of various models or theories of pragmatic competence, such as the communicative componential models, the form-function mapping proposal of the functionalist, the tripartite cognitive model, and the current integrated model of pragmatic competence. The common core includes (but not limited to) integration of thought and communication, one uniform pragmatic mechanism, dynamic form-function mapping, and complementarity between grammatical and pragmatic competences. With the findings as a departure, a brief outline for further investigation of pragmatic competence is proposed finally, including pathological and neurobiological examination of pragmatic competence.
Full-text available
There is a tendency across the Autistic Spectrum Conditions (ASC) to understand non-literal uses of language in a literal way. Different accounts for such a literalist bias have been proposed. Three of them can be considered ‘classical’ by now: the Executive Dysfunction theory, the Theory of Mind theory and the Weak Central Coherence theory. Currently, there is another hypothesis gaining traction, namely the Structural Language hypothesis. In this paper, we critically analyse these four accounts, with a focus on the Structural Language hypothesis, which holds that issues in figurative meaning comprehension relate not to ASC core traits but to structural language problems. We argue that the notion of ‘structural language’ is vague, and also that many studies taken to support the hypothesis have not actually tested the literalist bias. We conclude by suggesting interesting areas for further research.
Full-text available
Figurative language processing impairments in autism have been widely demonstrated, and have been considered a defining feature of autism. Studies in this area often consider different types of figurative language together, and less attention has been paid to identifying the factors that might contribute to difficulties in processing specific types of figurative language. Here we present a preregistered systematic review and meta-analysis of studies assessing the comprehension of idioms and proverbs in autism spectrum disorder (ASD) as compared to typically developing (TD) individuals. Idioms are widely used multi-word figurative expressions, which are understood by using contextual information. Proverbs are a related type of fixed, figurative, formulaic expression in a sentential form, typically linked with wisdom. Idioms and proverbs represent forms of figurative language which are more conventionalized and frequently opaquer than metaphors, pointing to a unique way that they are processed in conversational contexts. Our analysis encompassed a total of 11 studies from 10 papers (involving 235 autistic and 224 TD individuals), which met our inclusion criteria (the ASD and TD groups were matched on both chronological age and intelligence). The analysis of accuracy data revealed a group difference favouring the TD over the ASD group, with a medium effect size, and no indication of a publication bias. Participants’ age was unrelated to the magnitude of group differences, but there was a trend for smaller group differences in the case of participants with higher (verbal) intelligence. We discuss these findings with reference to theories related to the nature of figurative language impairments in autism.
Full-text available
This presentation shares strategies and resources to help learners with ASD develop effective communication skills needed for success in community settings.
Full-text available
The main tenet of this paper is that human communication is first and foremost a matter of negotiating commitments, rather than one of conveying intentions, beliefs, and other mental states. Every speech act causes the speaker to become committed to the hearer to act on a propositional content. Hence, commitments are relations between speakers, hearers, and propositions. Their purpose is to enable speakers and hearers to coordinate their actions: communication is coordinated action for action coordination. To illustrate the potential of the approach, commitment-based analyses are offered for a representative sample of speech act types, conversational implicatures, as well as for common ground.
Full-text available
We conducted a web-based study investigating whether the probability of deriving four types of pragmatic inferences depends on the degree to which one has traits associated with the autism spectrum, as measured by the autism spectrum quotient test (Baron-Cohen, Wheelwright, Skinner, Martin, & Clubley, 2001). In line with previous research, we show that, independently of their autism spectrum quotient, participants are likely to derive those pragmatic inferences that can be derived by reasoning solely about alternatives that the speaker could have used. However, if the derivation of the pragmatic inference draws upon more complex counterfactual reasoning about what the speaker could have said, the probability that it is derived decreases significantly with one’s autism quotient. We discuss the consequences for theories of pragmatics in autism and for linguistic theorizing in general.
Full-text available
Autism Spectrum Disorder (ASD) is often described as being characterised by a uniform pragmatic impairment. However, recent evidence suggests that some areas of pragmatic functioning are preserved. This study seeks to determine to which extent context-based derivation of non-linguistically encoded meaning is functional in ASD. We compare the performance of 24 adults with ASD, and matched neuro-typical adults in two act-out pragmatic tasks. The first task examines generation of indirect request interpretations, and the second the comprehension of irony. Intact contextual comprehension of indirect requests contrasts with marked difficulties in understanding irony. These results suggest that preserved pragmatics in ASD is limited to egocentric processing of context, which does not rely on assumptions about the speaker’s mental states.
Full-text available
Background and aims Pragmatic language is a key difficulty in autism spectrum disorder. One such pragmatic skill is verbal reference, which allows the current entity of shared interest between speakers to be identified and thus enables fluid conversation. The aim of this review was to determine the extent to which studies have found that verbal reference is impaired in autism spectrum disorder. We organise the review in terms of the methodology used and the modality (production versus comprehension) in which proficiency with verbal reference was assessed. Evidence for the potential cognitive underpinnings of these skills is also reviewed. Main contribution and methods To our knowledge, this is the first systematic review of verbal reference in autism spectrum disorder. PsychINFO and Web of Science were systematically screened using the combination of search terms outlined in this paper. Twenty-four studies met our inclusion criteria. Twenty-two of these examined production, whereby the methodology ranged from elicited conversation through to elicited narrative, the ‘director’ task and other referential communication paradigms. Three studies examined reference interpretation. (One study investigated both production and appropriacy judgement). Four studies examined the relationship between appropriate usage of verbal reference and formal language (lexico-syntactic ability). Two studies investigated whether reference production related to Theory of Mind or Executive Functioning. Conclusion and implications Across a range of elicited production tasks, the predominant finding was that children and adults with autism spectrum disorder demonstrate a deficit in the production of appropriate verbal reference in comparison not only to typically developing groups, but also to groups with Developmental Language Disorder or Down syndrome. In contrast, the studies of reference interpretation which compared performance to typical control groups all found no between-group differences in this regard. To understand this cross-modality discrepancy, we need studies with the same sample of individuals, whereby the task requirements for comprehension and production are as closely matched as possible. The field also requires the development of experimental manipulations which allow us to pinpoint precisely if and how each comprehension and/or production task requires mentalising and/or various components of executive functioning. Only through such detailed and controlled experimental work would it be possible to determine the precise location of impairments in verbal reference in autism spectrum disorder. A better understanding of this would contribute to the development of interventions.
Full-text available
People talk not only to others but also to themselves. The self talk we engage in may be overt or covert, and is associated with a variety of higher mental functions, including reasoning, problem solving, planning and plan execution, attention, and motivation. When talking to herself, a speaker takes devices from her mother tongue, originally designed for interpersonal communication, and employs them to communicate with herself. But what could it even mean to communicate with oneself? To answer that question, we need a theory of communication that explains how the same linguistic devices may be used to communicate with others and oneself. On the received view, which defines communication as information exchange, self talk appears to be an anomaly, for it is hard to see the point of exchanging information with oneself. However, if communication is analysed as a way of negotiating commitments between speaker and hearer, then communication may be useful even when speaker and hearer coincide. Thus a commitment-based approach allows us to make sense of self talk as well as social talk.
We investigated “scalar implicature” in adolescents and children with autism spectrum disorder (ASD) to test whether theory of mind deficits associated with autism affect pragmatic inferences in language. We tested scalar implicature computation in adolescents with ASD (12–18 years) and asked whether they reason about mental states when computing inferences. Like previous studies, we found the adolescents with ASD computed implicatures to the same degree as neurotypical adults. However, we also found that this ability may not rely on epistemic reasoning. In a test of epistemic reasoning (which probed so-called “ignorance implicature”) we found that adolescents with ASD were able to make epistemic inferences required by Gricean models of scalar implicature when they were explicitly required by the task to do so. However, in a second task, which asked whether subjects spontaneously reason about mental states in the service of scalar implicature when not explicitly asked to, we found that adolescents with ASD did not engage in epistemic reasoning, leading them to compute scalar implicatures in contexts in which they were not justified. Based on these data, we argue that epistemic reasoning may not be a core, constitutive component of scalar implicature.
According to the literalist view of speech acts, morpho-syntactic sentence types are associated directly at the semantic level with an illocutionary force. By contrast, according to contextualist theories illocutionary force emerges from contexts of use. To date, however, there is little experimental evidence relevant to this debate. We propose two experimental, eye-tracking studies to test two predictions of the literalist view: First, unlike for the highly conventionalised Can you? forms, whenever a non-conventionalised construction such as Is it possible to? is interpreted as a request, its question interpretation should also be activated. Second, the directive interpretation of modal You must declaratives should activate the statement interpretation and, therefore, be costlier than that of imperatives. In Study 1, we show, first, that, in contexts where both the non-directive and directive interpretation of indirect requests are available, the latter are processed as fast as[ 9 _ T D $ D I F F ] that of the corresponding imperatives, independently of the conventionalisation degree of the indirect request at hand. Second, eye fixation data show that the comprehension of indirect requests does not activate their direct meaning. Study 2 shows that modal You must declaratives are understood as imperatives and do not activate a statement interpretation; this supports the view that obligation modal requests are as direct as imperative requests.