Professional Documents
Culture Documents
Comprehension of Presupposition Triggers in Cantonese-Speaking Children With and Without Autism Spectrum Disorders
Comprehension of Presupposition Triggers in Cantonese-Speaking Children With and Without Autism Spectrum Disorders
Candice Chi-Hang Cheung, Yicheng Rong, Fei Chen, Man Tak Leung & Tempo
Po Yi Tang
To cite this article: Candice Chi-Hang Cheung, Yicheng Rong, Fei Chen, Man Tak Leung &
Tempo Po Yi Tang (2019): Comprehension of presupposition triggers in cantonese-speaking
children with and without autism spectrum disorders, Clinical Linguistics & Phonetics, DOI:
10.1080/02699206.2019.1673486
Introduction
Individuals with autism spectrum disorders (ASD) are known to have deficits in pragmatic
aspects of language or in the ability to use language to communicate effectively in a range of
social contexts (Baron-Cohen, 1988; Boucher, 2003; Eales, 1993; Lord & Paul, 1997; Martin &
McDonald, 2003; Ozonoff & Miller, 1996; Tager-Flusberg, 1981, 1996; Wilkinson, 1998). This
is evident from the huge body of literature showing that children with ASD have difficulties
with topic management (e.g. Dobbinson, Perkins, & Boucher, 1998; Hale & Tager-Flusberg,
2005; Rutter & Schopler, 1987), turn taking (e.g. Dobbinson et al., 1998; García-Pérez, Lee, &
Hobson, 2006), and conversational repair (e.g. Volden, 2004) in conversational contexts.
(1) The Prime Minister of Japan is planning to visit London next month.
While a recent study has shown that Cantonese-speaking children with ASD have a deficit in
comprehending presuppositions relative to typically developing (TD) children (Cheung et al.,
2017), it is worth mentioning that pragmatic phenomena that are closely connected to
presuppositions, such as scalar implicatures (Chevallier, Wilson, Happé, & Noveck, 2010;
Pijnacker, Hagoort, Buitelaar, Teunisse, & Geurts, 2009) and obligatory control (Janke &
Perovic, 2015, 2017), did not reveal a significant difference between children with ASD and
TD controls. While Cheung et al. (2017) is the first comprehensive study of autistic children’s
comprehension of four types of presupposition – namely, existential presuppositions, lexical
presuppositions, factive presuppositions, and structural presuppositions, one limitation of
their study is that it lumped together different types of presupposition trigger (e.g. change-of-
state verbs and iteratives were treated as a type of presupposition called lexical presupposi-
tions), making it impossible to investigate whether children with ASD show deficits in
comprehending a specific type of presupposition trigger, relative to TD controls.1
Moreover, it is unclear how children with ASD differ from TD children matched on
chronological age (CA) and TD children matched on language ability (LA) in terms of
their understanding of presupposition triggers. Furthermore, while there is an increasing
number of experimental studies on the properties of different presupposition triggers in
neurotypical adults (Amaral & Cummins, 2015; Jayez, Mongelli, Reboul, & van der Henst,
2015; Romoli, Khan, Sudo, & Snedeker, 2015; Schwarz, 2015; Tiemann, Kirsten, Beck,
Hertrich, & Rolke, 2015), studies investigating the acquisition of presuppositions in TD
children are limited and they mainly focus on a few presupposition triggers (Berger & Höhle,
2012; Dudley, Orita, Hacquard, & Lidz, 2015). In light of these research gaps, the present
study aimed to address two questions: (1) Do Cantonese-speaking children with ASD differ
from CA-matched and LA-matched TD children in terms of comprehension of seven types
of presupposition trigger – namely definite descriptions, factive predicates, change-of-state
1
The need to investigate whether children with ASD differ from TD controls with respect to each type of
presupposition trigger is in line with the growing number of experimental studies that have
compared presupposition triggers with one another in neurotypical adults in order to gain a better
understanding of the differences between triggers (Schwarz, 2015).
CLINICAL LINGUISTICS & PHONETICS 3
verbs, implicative verbs, iteratives, temporal clauses, and counterfactual conditionals? (2) Do
Cantonese-speaking children with ASD show a deficit in comprehending any specific types of
presupposition trigger relative to CA-matched and LA-matched TD children after control-
ling for the effects of chronological age, language ability, and non-verbal intelligence?
Presupposition triggers
In this section, we introduce the seven types of presupposition trigger investigated in our
study. We have already exemplified one type of definite description, using the Prime
Minister of Japan as an example (Strawson, 1950, 1952). Another type of definite descrip-
tion is possessives, such as Rachel’s son. These have the same property of presupposing the
existence of their referents, as shown in (2) (presupposition is marked by “≫”
throughout).
Factive predicates (e.g. know, regret, discover, find, odd, happy) presuppose the veracity of
the following complement clause (Karttunen, 2016; P. Kiparsky & Kiparsky, 1970; Scoville &
Gordon, 1980), as shown in (3). In other words, these predicates trigger the presupposition
that the content of the complement clause (i.e. John is a teacher in (3)) is true.
Change-of-state verbs (e.g. stop, begin, continue) are verbs that signal the beginning,
continuation, or end of an event or action (Abusch, 2002; Lorenz, 1992; Simons, 2001).
The presuppositions they trigger are based on their lexical meaning. For instance, in (4),
the change-of-state verb stop presupposes that the event depicted (i.e. that Dave argued
with Sally) had been taking place.
Implicative verbs (e.g. manage, forget) presuppose the sufficient and necessary condi-
tions for utterances containing them to hold true (Karttunen, 1971, 2016; Levinson, 1983).
For instance, the implicative verb manage presupposes try as its sufficient and necessary
condition because one cannot manage to do something without trying.
Iteratives (e.g. again, anymore) are adverbs that mark the repetitiveness of an event
(Karttunen, 2016; Levinson, 1983). For instance, in (6), the iterative again presupposes
that the event (i.e. that the student was late) has happened before.
4 C. C.-H. CHEUNG ET AL.
Temporal clauses are headed by after, before, since, and so on in English (Beaver &
Condoravdi, 2003; Heinämäki, 1974). As shown in (7), they presuppose the veracity of the
clause (i.e. Joe came) introduced by the temporal conjunction (i.e. after).
(8) If May had won the lottery, she would be in Paris now.
≫ May did not win the lottery.
As these descriptions indicate, each type of presupposition trigger involves a specific set
of lexical items or constructions. Since understanding the presuppositions clearly relies on
the hearer’s knowledge of the meaning of the presupposition triggers involved, and since
some types of presupposition trigger might be acquired earlier than others, we hypothesize
that children with ASD might show impairment in one or more specific types of
presupposition trigger rather than being impaired in all types of presupposition trigger
compared to their TD peers.
The participants’ language ability was assessed using the textual comprehension test
(TCT), a subtest of the Hong Kong Cantonese Oral Language Assessment Scale
(HKCOLAS; T’sou et al., 2006). Their non-verbal intelligence was assessed using
Raven’s Standard Progressive Matrices Test (Raven, 1981, 1989). Mann–Whitney U-tests
revealed no significant difference between the ASD and CA-matched TD groups in terms
of chronological age (U = 286.00, p = 0.633). However, compared to CA-matched TD
children, children with ASD scored significantly lower on the TCT (U = 155.00, p =
0.002). The ASD group did not differ significantly from the LA-matched TD group on the
raw mean score of TCT (U = 251.50, p = 0.503), but children with ASD were significantly
older than LA-matched TD children (U = 187.00, p = 0.045).
Table 1 summarizes the characteristics of the three groups of participants, namely
chronological age, age range, mean score for the textual comprehension test, and mean
score for non-verbal intelligence.
This study was approved by the Human Subjects Ethics Subcommittee at the Hong
Kong Polytechnic University. For all participants, a parent read and signed an informed
consent form prior to the start of the study. All participants were informed that they were
free to withdraw at any time for any reason. No financial compensation was offered.
2
The sentences used in the 56 test trials will be made available to any researchers upon request.
6 C. C.-H. CHEUNG ET AL.
and (10) are test trials that include a definite description (specifically, the possessives
Siu2fan1 go3 syu1baau1 ‘Siufan’s schoolbag’ and Daai6man4 go3 saimui6 ‘Daaiman’s
younger sister’) in the first utterance.3
Following Cheung et al. (2017), we prepared an answer sheet and asked the participants
to indicate whether the last utterance they heard was correct or not by circling either tick
(correct) or cross (incorrect) on the answer sheet for that particular stimulus. For instance,
in the example above, if the participants heard the sequence of utterances (9a, b), then
they had to circle the tick mark for their answer to be considered correct. If they heard the
sequence (10a, b), then they had to circle the cross to be considered correct.
To ensure that the test was effective for assessing children’s understanding of the seven
types of presupposition trigger, for each type of presupposition trigger, half of the test
trials contained the correct presupposition (9a, b) and half of the test trials contained the
incorrect presupposition (10a, b). For instance, for definite descriptions, four of the test
trials contained the correct presupposition and four contained the incorrect presupposi-
tion. Although Cantonese is a classifier language and does not have definite articles
(Cheng & Sybesma, 1999), in addition to possessives such as Siu2fan1 go3 syu1baau1
‘Siufan’s schoolbag’, definite expressions introduced by demonstratives such as go2 go3
gung1jyun4 ‘that garden’ are triggers of existential presuppositions. To minimize the effect
of the choice of definite description on children’s performance, we constructed four test
trials with possessives and four with definite expressions introduced by demonstratives.
Turning to factive predicates, we followed Cheung et al. (2017) in including both
factive (e.g. zi1dou6 ‘know’, hau6fui3 ‘regret’) and strong non-factive (e.g. ji5wai4 ‘falsely
think’, waan6soeng2 ‘imagine’) mental terms in the eight test trials – specifically, four of
each. Unlike factive mental terms, which presuppose the veracity of the complement
clause (see (11)), strong non-factive mental terms presuppose the falsity of the comple-
ment clause (see (12)).
3
All the examples contain the Cantonese romanizations in the first line, followed by word by word by
glosses in the second line, and English translations in the third line. The Cantonese romanizations
used in this paper follow the Linguistic Society of Hong Kong Cantonese Romanization Scheme.
CLINICAL LINGUISTICS & PHONETICS 7
As for change-of-state items, in addition to verbs that are also available in English such
as hoi1ci2 ‘start’, gaai3 ‘quit’, and ting4 ‘stop’, we included the Cantonese verbal particle
faan1, which indicates resumption of an activity or return to a state which has been
interrupted, as shown in (13) (Matthews & Yip, 2011).
For implicative verbs, we included ones whose counterparts are available in English
such as sing4gung1 ‘succeed’ and m4gei3dak1 ‘forget’, as shown in (14).
For iteratives, we included ones that are also available in English such as jau6si3 ‘again’
and mou5zoi3 ‘not anymore’, as in (15).
For temporal clauses, we included ones with temporal conjunctions such as zi1cin4
‘before’, zi1hau6 ‘after’, mou5noi6 ‘not long after’, and go2zan6 ‘at that time’ (Matthews &
Yip, 2011), as in (16).
Finally, Cantonese counterfactual conditionals behave like English ones in that they
could use the conjunction jyu4gwo2 ‘if’ to mark the if-clause (Matthews & Yip, 2011) and
in that the jyu4gwo2 ‘if’-clause is contrary to fact.
Table 2 summarises the number of test trials as well as the number of correct and
incorrect presuppositions for each type of presupposition trigger. All the test trials were
recorded as audio files.
Table 2. Description of the test trials for the seven types of presupposition trigger.
Types of presupposition trigger Correct presuppositions Incorrect presuppositions Total number of test trials
Definite descriptions 4 4 8
Factive predicates 4 4 8
Change-of-state verbs 4 4 8
Implicative verbs 4 4 8
Iteratives 4 4 8
Temporal clauses 4 4 8
Counterfactual conditionals 4 4 8
CLINICAL LINGUISTICS & PHONETICS 9
speech therapists in Hong Kong to assess children’s textual comprehension skills. Since
children must understand a wide range of vocabulary, phrases, and grammatical struc-
tures in order to understand the texts and answer the questions correctly, we considered
it a valid test for assessing the participants’ language ability.
4
As shown in (18) and (19), the participants were presented with a correct presupposition (18b) or an
incorrect presupposition (19b) in the second utterance. This design might overestimate the presup-
position understanding of the participants, as the second utterance could potentially highlight the
presuppositional meaning of the first utterance and make it more salient to the participants. However,
since one goal of the present study was to compare children with ASD with CA-matched TD and LA-
matched TD children regarding comprehension of the seven types of presupposition trigger, if the
design could indeed enhance the participants’ presupposition comprehension, we expect that it
would not affect our comparison among the three groups of children, as the enhanced presupposi-
tion comprehension should apply across the board.
10 C. C.-H. CHEUNG ET AL.
After listening to each test trial, the participants had to indicate whether the last
utterance they heard was correct or not by circling either tick (correct) or cross (incorrect)
on the answer sheet for that particular trial. In the example above, if participants heard the
sequence of utterances (18a, b), then they had to circle the tick mark for their answer to be
considered correct. If they heard the sequence (19a, b), then they had to circle the cross to
be considered correct.
All 56 test trials were pseudorandomized to avoid test trials with the same type of
presupposition trigger appearing in blocks. The order of presentation of the practice trials
and test trials was the same for all participants. To avoid causing fatigue to the partici-
pants, the test was divided into four sessions, each session consisting of three practice
trials and 14 test trials. Short breaks were given between sessions.
Results
Among the 56 test trials, 28 should be answered “yes” and 28 should be answered “no.” To
minimize the 50% chance correctness bias, signal detection criterion c was used as
a measure of response bias (Banks, 1970; Macmillan & Creelman, 1990; Snodgrass &
Corwin, 1988). No participant in the current study showed “yes” or “no” bias (c < – 1 for
“yes” bias; c < – 1 for “no” bias; Stanislaw & Todorov, 1999). Mann–Whitney U-tests
revealed that the scores for comprehending the seven types of presupposition trigger were
all significantly above chance in the ASD group and the two TD groups.5
Figure 1 shows the mean accuracy with which Cantonese-speaking children with ASD, CA-
matched TD children, and LA-matched TD children comprehended each type of presupposi-
tion trigger. A two-way repeated-measures analysis of variance (ANOVA), with Greenhouse-
Geisser corrections when appropriate, was conducted on the mean accuracy with trigger type
(seven types of presupposition trigger) as a within-subject factor, and group (three groups:
ASD, CA-matched TD, and LA-matched TD groups) as a between-subjects factor. The
analysis revealed significant main effects for both group (F(2, 68) = 5.52, p = 0.006) and
trigger type (F(6, 408) = 8.88, p < 0.001), while the interaction between group and trigger type
5
ASD group: definite descriptions: U = 27.00, p < 0.001; factive predicates: U = 40.50, p < 0.001; change-
of-state verbs: U = 13.50, p < 0.001; implicative verbs: U = 27.00, p < 0.001; iteratives: U = 40.50, p <
0.001; temporal clauses: U = 54.50, p < 0.001; and counterfactual conditionals: U = 40.50, p < 0.001.
CA-matched TD group: definite descriptions: U < 0.01, p < 0.001; factive predicates: U = < 0.001, p <
0.001; change-of-state verbs: U < 0.01, p < 0.001; implicative verbs: U < 0.001, p < 0.001; iteratives: U
< 0.01, p < 0.001; temporal clauses: U = 11.50, p < 0.001; and counterfactual conditionals: U < 0.01, p
< 0.001. LA-matched TD group: definite descriptions: U < 0.01, p < 0.001; factive predicates: U =
10.50, p < 0.001; change-of-state verbs: U < 0.01, p < 0.001; implicative verbs: U = 10.50, p < 0.001;
iteratives: U < 0.01, p < 0.001; temporal clauses: U = 31.50, p < 0.001; and counterfactual conditionals:
U < 0.01, p < 0.001.
CLINICAL LINGUISTICS & PHONETICS 11
Figure 1. Mean percentage accuracy on seven types of presupposition trigger for three comparison
groups: Cantonese-speaking children with ASD, CA-matched TD children, and LA-matched TD children.
Error bars: +/ – 1 standard error.
was not significant (F(12, 408) = 1.19, p = 0.305). The ANOVA results indicated that the three
groups of children performed differently in comprehending presupposition triggers. Tukey’s
HSD post hoc pairwise comparisons of the three groups indicated that the children with ASD
performed much worse than the CA-matched TD children in comprehending the seven types
of presupposition trigger (p = 0.004), but their performance did not differ from that of the LA-
matched TD children (p = 0.690). Furthermore, the absence of a group × trigger type
interaction implied that the relative difficulty of the seven types of presupposition trigger
did not differ among ASD, CA-matched TD, and LA-matched TD groups. Further post hoc
pairwise comparisons of the seven types of presupposition trigger showed that for all three
groups of children, comprehending temporal clauses was much more difficult than compre-
hending definite descriptions (p = 0.002), factive predicates (p = 0.001), change-of-state verbs
(p < 0.001), and iteratives (p < 0.001).
Spearman correlations were conducted to investigate whether children’s performance
correlated with their chronological age, language ability, and non-verbal intelligence. The
correlation results are presented in Table 3. For children with ASD, their chronological age
correlated with their comprehension of iteratives (r(27) = 0.47, p = 0.010) and temporal
clauses (r(27) = 0.44, p = 0.020), and their language ability correlated with their comprehen-
sion of factive predicates (r(27) = 0.59, p = 0.001), change-of-state verbs (r(27) = 0.60, p =
0.001), iteratives (r(27) = 0.55, p = 0.002), temporal clauses (r(27) = 0.43, p = 0.024), and
counterfactual conditionals (r(27) = 0.44, p = 0.019). As for the CA-matched TD and LA-
matched TD groups, mean accuracies on all seven types of presupposition trigger did not
significantly change as a function of chronological age or language ability, which might be
attributed to a ceiling performance around 90% in both groups of TD children.
Results of the above correlation analysis provided support for the influence of chron-
ological age, language ability, and/or non-verbal intelligence on children’s performance,
although in a different manner within each group. To examine the unique effect of group
12 C. C.-H. CHEUNG ET AL.
Table 3. Intercorrelations among chronological age, language ability, non-verbal intelligence, and
comprehension of the seven types of presupposition trigger by group.
ASD CA-matched TD LA-matched TD
CA LA NVIQ CA LA NVIQ CA LA NVIQ
Definite description 0.14 0.11 0.20 –0.01 0.27 0.25 0.30 0.17 –0.10
Factive predicate 0.17 0.59** 0.37 0.12 0.07 –0.39 0.33 0.16 0.06
Change-of-state verb 0.33 0.60** 0.31 –0.23 –0.35 0.34 –0.11 –0.12 0.61**
Implicative verb 0.07 0.31 0.05 –0.12 –0.17 –0.18 0.19 0.05 –0.07
Iterative 0.47* 0.55** 0.28 0.29 0.06 –0.43 –0.00 0.07 0.42
Temporal clause 0.44* 0.43* 0.33 0.30 0.19 0.11 0.35 0.22 0.24
Counterfactual conditional 0.30 0.44* 0.28 0.34 0.39 –0.16 0.36 0.24 0.13
CA = chronological age; LA = language ability; NVIQ = non-verbal intelligence.
* p < 0.05; ** p < 0.01.
(i.e. ASD group vs. CA-matched TD group; ASD group vs. LA-matched TD group) on
children’s understanding of the seven types of presupposition trigger, hierarchical multiple
regressions were further conducted. For the comparison between ASD and CA-matched TD
groups, chronological age and language ability were entered at the first step, because
a significant correlation between chronological age and language ability was found in the
ASD group (r(27) = 0.43, p = 0.024) as well as the CA-matched TD group (r(23) = 0.58, p =
0.004). Non-verbal intelligence was entered at the second step, and then group was entered
at the third step (see Table 4). To compare the performance between ASD and LA-matched
TD groups, chronological age was entered at the first step in the regression, language ability
was entered at the second step, non-verbal intelligence was entered at the third step, and
then group was entered at the last step (see Table 5). For all regressions, the dependent
variables were the participants’ scores on the seven types of presupposition trigger.
As shown in Table 4, regressions 1 through 7 examined the effect of group (ASD group
vs. CA-matched TD group) on comprehension of the seven types of presupposition
trigger, after controlling for the effects of chronological age, language ability, and non-
verbal intelligence. Results of the full model in regressions revealed that language ability
had a significant influence on the comprehension of some types of presupposition trigger,
namely factive predicates (t = 3.86, p < 0.001), change-of-state verbs (t = 2.57, p = 0.014),
iteratives (t = 3.28, p = 0.002), and counterfactual conditionals (t = 2.49, p = 0.017). After
the effects of chronological age, language ability, and non-verbal intelligence were con-
trolled for, 9.7% of the variance in the comprehension of temporal clauses was accounted
for by the participants’ group (F = 6.48, p = 0.014).
As shown in Table 5, regressions 8 through 14 examined whether group (ASD group
vs. LA-matched TD group) uniquely predicted the comprehension of the seven types of
presupposition trigger. Results of the full model in these regressions revealed
a significant influence of chronological age on the comprehension of temporal clauses
(t = 2.20, p = 0.033). Language ability had a significant influence on the comprehension
of factive predicates (t = 3.13, p = 0.003) and change-of-state verbs (t = 2.69, p = 0.010),
whereas non-verbal intelligence had a significant influence on the comprehension of
change-of-state verbs (t = 2.23, p = 0.031) and temporal clauses (t = 2.75, p = 0.009).
After the effects of chronological age, language ability, and non-verbal intelligence were
controlled for, 7.6% of the variance in the comprehension of temporal clauses was
accounted for by the participants’ group (F = 4.95, p = 0.031).
CLINICAL LINGUISTICS & PHONETICS 13
Table 4. Explaining comprehension of the seven types of presupposition trigger by group (ASD group
vs. CA-matched TD group).
Variables B SE R2 change F change p value
Regression 1: Definite description
Step 1: CA < 0.01 < 0.01 0.06 1.54 0.225
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.05 2.76 0.104
Step 3: Group < 0.01 0.04 0.03 2.71 0.105
Regression 2: Factive predicate
Step 1: CA < 0.01 < 0.01 0.33 11.71 < 0.001
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.02 1.40 0.243
Step 3: Group 0.13 0.06 0.01 0.74 0.394
Regression 3: Change-of-state verb
Step 1: CA < 0.01 < 0.01 0.14 3.96 0.026
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.03 1.69 0.200
Step 3: Group < 0.01 0.02 0.001 0.04 0.804
Regression 4: Implicative verb
Step 1: CA < 0.01 < 0.01 0.15 4.05 0.024
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.02 1.13 0.292
Step 3: Group 0.03 0.02 0.03 1.55 0.220
Regression 5: Iterative
Step 1: CA < 0.01 < 0.01 0.36 13.13 < 0.001
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.01 0.43 0.514
Step 3: Group 0.02 0.07 0.02 1.34 0.253
Regression 6: Temporal clause
Step 1: CA < 0.01 < 0.01 0.18 5.15 0.009
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.04 2.32 0.135
Step 3: Group 0.06 0.02 0.10 6.48 0.014
Regression 7: Counterfactual conditional
Step 1: CA < 0.01 < 0.01 0.25 7.86 0.001
LA < 0.01 < 0.01
Step 2: NVIQ < 0.01 < 0.01 0.03 1.75 0.193
Step 3: Group 0.03 0.02 0.04 2.30 0.137
In each regression table, B (raw beta) and SE (standard error of beta) are final statistics of the full model, including all three
steps, whereas R2 change, F change and p value are change statistics for the respective step as the last-entered step,
representing its additional contribution over and above the last model. CA = chronological age; LA = language ability;
NVIQ = non-verbal intelligence.
Discussion
The present study addressed two research questions: (1) Do Cantonese-speaking children
with ASD differ from CA-matched and LA-matched TD children in terms of comprehen-
sion of seven types of presupposition trigger? (2) Do Cantonese-speaking children with
ASD show a deficit in comprehending any specific types of presupposition trigger relative
to CA-matched and LA-matched TD children after controlling for the effects of chron-
ological age, language ability, and non-verbal intelligence?
The answer to the first question depends on which TD group (i.e. CA-matched or LA-
matched TD group) is compared with the ASD group. Specifically, we found that
Cantonese-speaking children with ASD performed worse than their CA-matched TD
peers in comprehending the seven types of presupposition trigger, but they performed
similarly to their LA-matched TD peers. This result is consistent with Cheung et al.’s
(2017) finding that Cantonese-speaking children with ASD generally performed worse
14 C. C.-H. CHEUNG ET AL.
Table 5. Explaining comprehension of the seven types of presupposition trigger by group (ASD group
vs. LA-matched TD group).
Variables B SE R2 change F change p Value
Regression 8: Definite description
Step 1: CA 0.001 0.002 0.021 1.010 0.320
Step 2: LA 0.001 0.001 0.023 1.101 0.300
Step 3: NVIQ 0.002 0.001 0.022 1.051 0.311
Step 4: Group 0.006 0.038 <0.001 0.022 0.882
Regression 9: Factive predicate
Step 1: CA <0.001 0.002 0.023 1.093 0.301
Step 2: LA 0.004 0.001 0.221 13.158 0.001
Step 3: NVIQ 0.002 0.001 0.041 2.549 0.118
Step 4: Group 0.024 0.040 0.006 0.367 0.548
Regression 10: Change-of-state verb
Step 1: CA <0.001 0.001 0.028 1.318 0.257
Step 2: LA 0.003 0.001 0.148 8.096 0.007
Step 3: NVIQ 0.002 0.001 0.091 5.431 0.024
Step 4: Group −0.032 0.032 0.017 1.014 0.320
Regression 11: Implicative verb
Step 1: CA 0.001 0.002 0.015 0.720 0.400
Step 2: LA 0.002 0.001 0.066 3.212 0.080
Step 3: NVIQ <0.001 0.001 0.002 0.098 0.756
Step 4: Group 0.038 0.045 0.015 0.694 0.409
Regression 12: Iterative
Step 1: CA 0.002 0.002 0.045 2.183 0.146
Step 2: LA 0.003 0.001 0.210 12.668 0.001
Step 3: NVIQ 0.002 0.001 0.039 2.405 0.128
Step 4: Group 0.074 0.037 0.061 4.035 0.051
Regression 13: Temporal clause
Step 1: CA 0.005 0.002 0.068 3.382 0.072
Step 2: LA 0.002 0.001 0.098 5.043 0.030
Step 3: NVIQ 0.004 0.001 0.100 5.995 0.018
Step 4: Group 0.101 0.046 0.076 4.954 0.031
Regression 14: Counterfactual
Step 1: CA 0.003 0.002 0.065 3.180 0.081
Step 2: LA 0.002 0.001 0.099 5.305 0.026
Step 3: NVIQ 0.002 0.001 0.058 3.276 0.077
Step 4: Group 0.036 0.042 0.013 0.727 0.399
In each regression table, B (raw beta) and SE (standard error of beta) are final statistics of the full model, including all three
steps, whereas R2 change, F change and p value are change statistics for the respective step as the last-entered step,
representing its additional contribution over and above the last model. CA = chronological age; LA = language ability;
NVIQ = non-verbal intelligence.
children with ASD did not differ significantly from that of the CA-matched and LA-
matched TD groups after controlling for the effects of chronological age, language
ability, and non-verbal intelligence. Our finding that children with ASD were impaired
in comprehending temporal clauses is in line with previous studies, which found that
children with ASD had difficulty interpreting temporal clauses (Boucher, 2001; Overweg,
Hartman, & Hendriks, 2018; Perkins, Dobbinson, Boucher, Bol, & Bloom, 2006) and
understanding temporal concepts such as past and future (Boucher, Pons, Lind, &
Williams, 2007; Maister & Plaisted-Grant, 2011; Peeters & Gillberg, 1999). Given that
correct interpretation of the presupposition triggered by a temporal clause relies on
understanding that the veracity of the clause introduced by the temporal conjunction
must be presupposed, if children with ASD are unable to grasp the presupposition due
to their impairment in temporal perception, they might wrongly consider (20b) as the
incorrect presupposition of (20a), and (21b) as the correct presupposition of (21a).
In other words, our study revealed that children with ASD were less accurate in judging
the veracity of the clause introduced by the temporal conjunction compared to the two
groups of TD children even after controlling for the effects of chronological age, language
ability, and non-verbal intelligence. Future studies should investigate the factors contri-
buting to this specific deficit in children with ASD, such as executive functioning and
theory of mind, which have been found to play a role in accounting for ASD children’s
interpretation of temporal order of events in complex sentences with the temporal
conjunctions before and after (Overweg et al., 2018).
Conclusions
The current study examined how 7- to 12-year-old Cantonese-speaking children with
ASD comprehend seven types of presupposition trigger – namely definite descriptions,
factive predicates, change-of-state verbs, implicative verbs, iteratives, temporal clauses,
and counterfactual conditionals – compared to CA-matched and LA-matched TD groups.
We found that children with ASD performed worse than the CA-matched TD group in
comprehending the seven types of presupposition trigger, but they performed similarly to
16 C. C.-H. CHEUNG ET AL.
the LA-matched TD group. We further found that Cantonese-speaking children with ASD
showed a deficit in comprehending the presupposition triggered by a temporal clause
relative to the two groups of TD children.
Acknowledgements
We thank all the children and parents for their participation in this study. We are grateful to LoveXpress
Foundation Limited and Speech Therapy Unit at the Hong Kong Polytechnic University for recruitment
of the participants. We would also like to thank Gordon Cheung, Harmony Choi, and Lydia Lai for their
help in constructing the stimuli for this study. We are greatly indebted to the anonymous reviewer for
the valuable comments and suggestions on an earlier version of this paper. We also thank the audience
at the research seminar organized by the Research Centre for Language, Cognition, and Neuroscience at
the Hong Kong Polytechnic University, where part of the material in this paper was presented. Special
thanks go to Gordon Cheung, Noel Chiu, Harmony Choi, Lydia Lai, Ally Lee, Catherine Wong,
Charlene Wong, and Stephanie Yuen for their help with data collection. We would also like to thank
Anne Mark for editorial assistance.
Declaration of interest
The authors report no conflicts of interest.
Funding
This study was supported by Central Research Grant [YBR7], awarded to the first author; the title of
the project is “Development of Mental State Verbs and Presuppositions in Cantonese-speaking
Children with and without Autism Spectrum Disorders”.
ORCID
Candice Chi-Hang Cheung http://orcid.org/0000-0002-1502-5724
References
Abusch, D. (2002). Lexical alternatives as a source of pragmatic presuppositions. In B. Jackson (Ed.),
Proceedings of semantics and linguistic theory 12 (pp. 1–19). Ithaca, NY: Cornell University, CLC
Publications.
Amaral, P., & Cummins, C. (2015). A cross-linguistic study on information backgrounding and
presupposition projection. In F. Schwarz (Ed.), Experimental perspectives on presuppositions (pp.
157–172). Cham, Switzerland: Springer International Publishing.
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th
ed., text revision). Washington, DC: Author.
Banks, W. P. (1970). Signal detection theory and human memory. Psychological Bulletin, 74, 81–99.
Baron-Cohen, S. (1988). Social and pragmatic deficits in autism: Cognitive or affective? Journal of
Autism and Developmental Disorders, 18, 379–402.
Beaver, D., & Condoravdi, C. (2003). A uniform analysis of ‘before’ and ‘after’. In R. B. Young &
Y. Zhou (Eds.), Proceedings of semantics and linguistic theory 13 (pp. 37–54). Ithaca, NY: Cornell
University, CLC Publications.
Berger, F., & Höhle, B. (2012). Restrictions on addition: Children’s interpretation of the focus
particles auch ‘also’ and nur ‘only’ in German. Journal of Child Language, 39, 383–410.
CLINICAL LINGUISTICS & PHONETICS 17
Boucher, J. (2001). ‘Lost in a sea of time’: Time-parsing and autism. In C. Hoerl & T. McCormack (Eds.),
Time and memory: Issues in philosophy and psychology (pp. 111–135). Oxford, UK: Clarendon Press.
Boucher, J. (2003). Language development in autism. International Journal of Pediatric
Otorhinolaryngology, 67(Suppl.), S159–S163.
Boucher, J., Pons, F., Lind, S., & Williams, D. (2007). Temporal cognition in children with autistic
spectrum disorders: Tests of diachronic thinking. Journal of Autism and Developmental Disorders,
37, 1413–1429.
Cheng, L. L.-S., & Sybesma, R. (1999). Bare and not-so-bare nouns and the structure of NP.
Linguistic Inquiry, 30, 509–542.
Cheung, C. C. H., Politzer-Ahles, S., Hwang, H., Chui, R. L. Y., Leung, M. T., & Tang, T. P. Y.
(2017). Comprehension of presuppositions in school-age Cantonese-speaking children with and
without autism spectrum disorders. Clinical Linguistics & Phonetics, 31, 557–572.
Chevallier, C., Wilson, D., Happé, F., & Noveck, I. (2010). Scalar inferences in autism spectrum
disorders. Journal of Autism and Developmental Disorders, 40, 1104–1117.
Dobbinson, S., Perkins, M. R., & Boucher, J. (1998). Structural patterns in conversations with
a woman who has autism. Journal of Communication Disorders, 31, 113–134.
Dudley, R., Orita, N., Hacquard, V., & Lidz, J. (2015). Three-year-olds’ understanding of know and
think. In F. Schwarz (Ed.), Experimental perspectives on presuppositions (pp. 241–262). Cham,
Switzerland: Springer International Publishing.
Eales, M. J. (1993). Pragmatic impairments in adults with childhood diagnoses of autism or
developmental receptive language disorder. Journal of Autism and Developmental Disorders, 23,
593–617.
García-Pérez, R. M., Lee, A., & Hobson, R. P. (2006). On intersubjective engagement in autism:
A controlled study of nonverbal aspects of conversation. Journal of Autism and Developmental
Disorders, 37, 1310–1322.
Hale, C. M., & Tager-Flusberg, H. (2005). Brief report: The relationship between discourse deficits
and autism symptomatology. Journal of Autism and Developmental Disorders, 35, 519–524.
Heinämäki, O. (1974). Semantics of English temporal connectives. (Unpublished doctoral disserta-
tion). Indiana University, Bloomington.
Janke, V., & Perovic, A. (2015). Intact grammar in HFA? Evidence from control and binding.
Lingua, 164, 68–86.
Janke, V., & Perovic, A. (2017). Contrasting complement control, temporal adjunct control and
controlled verbal gerund subjects in ASD: The role of contextual cues in reference assignment.
Frontiers in Psychology, 8, 448.
Jayez, J., Mongelli, V., Reboul, A., & van der Henst, J.-B. (2015). Weak and strong triggers. In
F. Schwarz (Ed.), Experimental perspectives on presuppositions (pp. 173–193). Cham, Switzerland:
Springer International Publishing.
Karttunen, L. (1971). Implicative verbs. Language, 47, 340–358.
Karttunen, L. (1974). Presupposition and linguistic context. Theoretical Linguistics, 1, 181–193.
Karttunen, L. (2016). Presupposition: What went wrong? In M. Moroney, C. R. Little, J. Collard, &
D. Burgdorf (Eds.), Proceedings of semantics and linguistic theory 26 (pp. 705–731). Ithaca, NY:
Cornell University, CLC Publications.
Kiparsky, P., & Kiparsky, C. (1970). Fact. In M. Bierwisch & K. E. Heidolph (Eds.), Progress in
linguistics (pp. 143–173). The Hague, Netherlands: Mouton.
Levinson, S. C. (1983). Pragmatics. Cambridge, UK: Cambridge University Press.
Lord, C., & Paul, R. (1997). Language and communication in autism. In D. J. Cohen &
F. R. Volkmar (Eds.), Handbook of autism and pervasive developmental disorders (2nd ed., pp.
195–225). New York, NY: John Wiley.
Lorenz, S. (1992). On the role of reasoning about change in the projection of presuppositions. Report
234, IWBS, IBM Germany.
Macmillan, N. A., & Creelman, C. D. (1990). Response bias: Characteristics of detection theory,
threshold theory, and “nonparametric” indexes. Psychological Bulletin, 107, 401–413.
Maister, L., & Plaisted-Grant, K. C. (2011). Time perception and its relationship to memory in
autism spectrum conditions. Developmental Science, 14, 1311–1322.
18 C. C.-H. CHEUNG ET AL.
Martin, I., & McDonald, S. (2003). Weak coherence, no theory of mind, or executive dysfunction?
Solving the puzzle of pragmatic language disorders. Brain and Language, 85, 451–466.
Matthews, S., & Yip, V. (2011). Cantonese: A comprehensive grammar. New York, NY: Routledge.
Overweg, J., Hartman, C. A., & Hendriks, P. (2018). Temporarily out of order: Temporal perspec-
tive taking in language in children with autism spectrum disorder. Frontiers in Psychology, 9,
1663.
Ozonoff, S., & Miller, J. N. (1996). An exploration of right-hemisphere contributions to the
pragmatic impairments of autism. Brain and Language, 52, 411–434.
Peeters, T., & Gillberg, C. (1999). Autism: Medical and educational aspects (2nd ed.). London, UK:
Whurr.
Perkins, M., Dobbinson, S., Boucher, J., Bol, S., & Bloom, P. (2006). Lexical knowledge and lexical
use in autism. Journal of Autism and Developmental Disorders, 36, 795–805.
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, 607–618.
Raven, J. (1981). Manual for Raven’s Progressive Matrices and Vocabulary Scales. Research supple-
ment no. 1: The 1979 British standardisation of the Standard Progressive Matrices and Mill Hill
Vocabulary Scales, together with comparative data from earlier studies in the UK, US, Canada,
Germany, and Ireland. Oxford, UK/San Antonio, TX: Oxford Psychologists Press/The
Psychological Corporation.
Raven, J. (1989). The raven progressive matrices: A review of national norming studies and ethnic and
socioeconomic variation within the United States. Journal of Educational Measurement, 26, 1–16.
Romoli, J., Khan, M., Sudo, Y., & Snedeker, J. (2015). Resolving temporary referential ambiguity
using presupposed content. In F. Schwarz (Ed.), Experimental perspectives on presuppositions (pp.
67–108). Cham, Switzerland: Springer International Publishing.
Rutter, M. R., & Schopler, E. (1987). Autism and pervasive developmental disorders: Concepts and
diagnostic issues. Journal of Autism and Developmental Disorders, 17, 159–186.
Schwarz, F. (2015). Symmetry and incrementality in conditionals. In F. Schwarz (Ed.), Experimental
perspectives on presuppositions (pp. 195–213). Cham, Switzerland: Springer International
Publishing.
Scoville, R. P., & Gordon, A. M. (1980). Children’s understanding of factive presuppositions: An
experiment and a review. Journal of Child Language, 7, 381–399.
Simons, M. (2001). On the conversational basis of some presuppositions. In R. Hastings, B. Jackson,
& Z. Zvolensky (Eds.), Proceedings of semantics and linguistic theory 11 (pp. 431–448). Ithaca,
NY: Cornell University, CLC Publications.
Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to
dementia and amnesia. Journal of Experimental Psychology: General, 117, 34–50.
Stalnaker, R. (1973). Presuppositions. Journal of Philosophical Logic, 2, 447–457.
Stalnaker, R. (1974). Pragmatic presuppositions. In M. Munitz & P. Unger (Eds.), Semantics and
philosophy (pp. 197–214). New York, NY: New York University Press.
Stalnaker, R. (1998). On the representation of context. Journal of Logic, Language and Information,
7, 3–19.
Stalnaker, R. (2002). Common ground. Linguistics & Philosophy, 25, 701–721.
Stanislaw, H., & Todorov, N. (1999). Calculation of signal detection theory measures. Behavior
Research Methods, Instruments, & Computers, 31, 137–149.
Strawson, P. F. (1950). On referring. Mind, 59, 320–344.
Strawson, P. F. (1952). Introduction to logical theory. London, UK: Methuen.
T’sou, B. K. Y., Lee, T., Tung, P., Chan, A., Man, Y., & To, C. K. S. (2006). Hong Kong Cantonese
oral language assessment scale. Hong Kong, China: Language Information Sciences Research
Centre, City University of Hong Kong.
Tager-Flusberg, H. (1981). On the nature of linguistic functioning in early infantile autism. Journal
of Autism and Developmental Disorders, 11, 45–56.
Tager-Flusberg, H. (1996). Brief report: Current theory and research on language and communica-
tion in autism. Journal of Autism and Developmental Disorders, 26, 169–172.
CLINICAL LINGUISTICS & PHONETICS 19
Tiemann, S., Kirsten, M., Beck, S., Hertrich, I., & Rolke, B. (2015). Presupposition processing and
accommodation: An experiment on wieder (‘again’) and consequences for other triggers. In
F. Schwarz (Ed.), Experimental perspectives on presuppositions (pp. 39–65). Cham, Switzerland:
Springer International Publishing.
Volden, J. (2004). Conversational repair in speakers with autism spectrum disorder. International
Journal of Language & Communication Disorders, 39, 171–189.
Wilkinson, K. M. (1998). Profiles of language and communication skills in autism. Mental
Retardation and Developmental Disabilities Research Reviews, 4, 73–79.