The purpose of this study is to investigate the correlation and the causality direction between mastering linguistic tests and second-order reasoning tasks in the autistic population. Linguistic tests include two types of tasks: 1. Mastery of multiple embeddings or recursion and 2. Mastery of dialogue particles. Second-order tests include standard false-belief tasks.

The following research questions are to be investigated:

1. Do particular developments in syntax such as recursion comprehension affect particular kinds of thinking such as second-order Theory of Mind reasoning in autistic children? More concretely: Does children's ability to comprehend the linguistic structure involving or consisting of recursive structures predict performance on second-order false-belief tests, beyond development in general language abilities?
2. Does knowledge of dialogue particles predict performance on second-order false-belief tests, beyond development in language abilities and working memory? In order to investigate these questions, a between-subjects empirical design includes data collection and data analysis in two steps: one measuring the strength of correlations between variables, the other establishing causality effect via training.

The study is divided into four stages: pilot, pre-testing, intervention, post-testing.

1. The central goal of the pilot testing is to ensure reliability of tasks in Danish: vocabulary familiarity and how children understand what they are asked about. Second-order false-belief tasks and recursion tasks will be given to children in the pilot stage.

2. Pre-testing stage is the stage where all the variables are measured (dependent, independent and confounding). Selection of pretests reflects considerations about confounding variables and inclusion criteria.

   Confounding variables include: verbal comprehension (from Wechsler Intelligence Scale for Children IV), working memory (from Wechsler Intelligence Scale for Children IV), comprehension of grammar and syntactic structures (from Clinical Evaluation of Language Fundamentals-4), pragmatic profile (from Clinical Evaluation of Language Fundamentals-4).

   Independent variables include above-mentioned linguistic tests. Dependent variable is second-order false-belief tests.

3. At the intervention stage training of recursion comprehension for a randomized half of sample size will be provided as four individual sessions,15 minutes each. Training of recursion comprehension will include cooperative dialogue and feedback on the sentences and small dialogues analogous to the tasks in the pretests.

4. After the training is completed, all the sample subjects will be given second-order false-belief and recursion tasks again at the post-testing stage.

The goal is to recruit 60 subjects in total. For the training part 30 subjects will be assigned by randomization to the intervention group and 30 subjects- to the control group. The recruiting process and participation will not require disclosure of any personal data. Parental informed, written consent for their children to take part in the study is a prerequisite for every child's participation. Parents will also receive debriefing information in written form upon the completion of the empirical part of the study.

Participants are ensured of full anonymity. All data will be anonymized by using subject's identification, and not names. All personal record of any kind, including video and tape recordings of testing, will be kept according to the Danish Privacy Law regulations.

Every child will be tested individually by a trained psychologist in a quiet comfortable room away from the classroom. In case testing procedure will be in any way unpleasant or excessively demanding for some participants, the session will be cancelled, finished earlier or the number of tasks will be revised.

Inclusion Criteria:

1. Autism Spectrum Disorder is already established and confirmed. Participants will be recruited by contacting special schools directly, where the established diagnosis is the admission criteria.
2. Monolinguals, Danish native speakers.
3. Absence of significant impairments in language development.
4. Intelligence quotient (IQ): more than 85 (i.e. within the normal range).

Exclusion Criteria:

1. Medical treatment affecting performance (antidepressants, antipsychotic and stimulants).
2. Subjects that have been tested with Wechsler Intelligence Scale test during the last 1.5 year.
3. Comorbidity (difficult with Attention Deficit Hyperactivity Disorder, but other psychiatric conditions and brain damage condition of any sort should be excluded).