Purpose <p>In the present study, we asked autistic and non-autistic children to recount personally experienced basic (fear, happiness, sadness) and self-conscious (guilt, pride, embarrassment) emotions. Using linguistic analyses, children’s recounts were examined in order to gain new insights into their understanding of these emotions.</p> Methods <p>Children were asked to recount personally experienced basic and self-conscious emotions. Discourse and content analyses were performed on their responses.</p> Results <p>For basic emotions, autistic and non-autistic children’s recounts did not differ in terms of number of responses provided following two requests, length of responses (i.e., number of verb and argument clauses in recounts), and the number of prompts needed to elicit basic emotion recounts. Moreover, children did not differ in the appropriateness of their recounts of basic emotions, although autistic children made fewer causal inferences about sadness than their non-autistic peers. In contrast, autistic children offered significantly fewer recounts of guilt and embarrassment and provided briefer recounts of guilt compared to their non-autistic peers. Autistic children also made fewer references to others in their recounts for guilt and embarrassment and were less likely to include social context information (i.e. references to others). Finally, autism symptomatology predicted appropriateness of content for guilt and pride, whereas vocabulary level predicted appropriateness of content for sadness and embarrassment.</p> Conclusion <p>Few linguistic differences were found in children’s recounting of basic emotions, whereas autistic children showed less appropriate content for self-conscious emotions. Implications of our findings in terms of possible avenues for intervention and training are given.</p>