How do autistic children interpret 'some' when 'all' is true?
With only 'agree' or 'disagree' available, autistic children rejected misleading 'some' statements at least as often as other children, but with a middle option they chose the extremes while typically developing peers chose 'agree a bit'.
Source
The Understanding of Scalar Implicatures in Children With Autism Spectrum Disorder: Dichotomized Responses to Violations of Informativeness
Study at a glance
- Design
- Case-control — Three groups (ASD, IQ-matched, age-matched typically developing) judged optimal, logically false and underinformative statements, first with two response options (Experiment 1) then with three (Experiment 2), in fixed order.
- N
- N=66 · 66 Flemish-speaking 10-year-olds: 22 with ASD plus an IQ-matched and an age-matched group; the same children did both experiments. The extracted text lacks the Experiment 1 participant and materials section, so group matching details are not available.
- Population
- Flemish-speaking children around 10 years old with and without autism spectrum disorder
- Outcome
- Agreement judgements to underinformative (e.g., 'some' when 'all' is true), optimal and false statements
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The researchers tested 66 ten-year-old Flemish-speaking children: 22 with autism, a group matched on IQ, and a group matched on age. Children judged a character's statements about pictures, some true and fully informative, some false, and some true but underinformative, such as saying 'some' when 'all' applied. In Experiment 1 they could only agree or disagree; in Experiment 2 they could also 'agree a bit'.
What they found
In the two-option task, all groups handled true and false statements equally, and the autistic group rejected underinformative statements more than the others in the regression analysis, though not in the non-parametric analysis. In the three-option task, typically developing children preferred 'agree a bit' for underinformative statements, showing tolerance, whereas autistic children more often totally agreed or totally disagreed. The IQ-matched group showed the same split pattern as the autistic group, suggesting general intellectual or verbal ability, not autism alone, is involved.
The limits
What it doesn't show
Groups were small and the key contrasts were not always consistent across the two statistical approaches used. Because the IQ-matched group behaved like the autistic group, the study cannot attribute the pattern specifically to autism, and the authors did not separate verbal from non-verbal IQ. All children did the binary task before the ternary one, so rule-switching could be a confound, and the extracted paper text omits the Experiment 1 participant and materials section, so matching details cannot be checked here.
Key terms
- Scalar implicature
- The inference that a weaker term like 'some' implies 'not all', because the speaker could have said the stronger term.
- Underinformative statement
- A statement that is literally true but weaker than it could be, such as 'some elephants have trunks'.
- Logical vs pragmatic interpretation
- Logical reads 'some' as 'at least some, possibly all'; pragmatic reads it as 'some but not all'.
- Pragmatic tolerance
- Recognising that a statement is underinformative while not treating it as flatly false, shown by choosing 'agree a bit'.
- Ternary judgment task
- A response format with three options (totally agree, agree a bit, totally disagree) instead of two.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
In the three-option task, what did typically developing children prefer for underinformative statements?
Common questions
Why does the response format matter so much?
With only two options, children must treat a slightly misleading statement as right or wrong, hiding whether they see it as merely imperfect; a middle option reveals that tolerance.
Does this show that autism causes difficulty with 'some'?
Not on its own, because children matched on IQ without autism showed the same pattern, pointing to general cognitive or verbal abilities.
What is a non-scalar item?
An ad hoc informativeness contrast not based on a word scale like some/all, used to compare with classic scalar terms.
More on Autism