Staying on topic or shifting away? Interests drive conversation topic in autistic and neurotypical children

Intense interests are a defining feature of autism, yet their causal role in pragmatic language use remains poorly specified. We investigated whether intense interests directly shape conversational topic management, a core component of pragmatics, in two pre-registered experimental studies with 7–11-year-old children.

Across both studies, we used a novel conversation elicitation paradigm that allowed experimental control within an otherwise naturalistic interaction. The experimenter embedded pre-planned conversational probes (e.g., “We like trips to castles”), and children’s responses were coded for appropriate topic acknowledgement and elaboration. During the Zoom-based conversation, children were also exposed to visual distractors via screenshare. Critically, these distractors—or the ongoing conversational topic—were manipulated to be either neutral or related to the child’s high-interest domain.

In both studies, high-interest content systematically altered children’s topic management. In Study 1 (N = 132), children were less likely to maintain the conversational topic in the presence of high-interest distractors compared to neutral distractors. In Study 2 (N = 72), children were less likely to shift topic when a shift was pragmatically warranted (e.g., to comment on a warning sign) if the conversation involved a high-interest topic rather than a neutral one.

Diagnostic status modulated these effects in different ways across studies. In Study 1, the disruptive effect of high-interest distractors was significantly greater for autistic children. In Study 2, there was no interaction between diagnostic status and interest condition.

To move beyond categorical diagnostic comparisons, we conducted regression analyses treating interest intensity and autistic traits as continuous dimensions across all participants. In both studies, interest intensity explained substantial unique variance in topic management, with large effect sizes, even when controlling for autistic traits, cognitive flexibility, and core language abilities.

These findings provide the first direct experimental evidence that intense interests exert a causal influence on conversational pragmatics. We argue that interest intensity is a key mechanistic variable for explaining individual differences in topic management across and within diagnostic groups and should be incorporated into theories of conversation and dialogue.