A crucial aspect of Open Science concerns opening science to society through ‘open engagement of societal actors and open dialogue with other knowledge systems’ (UNESCO, 2023, p 1), connecting Open Science to achieving equity, diversity, and inclusion in science (Bertram et al., 2023). Thus, the movement of Open Science can be discussed in relation to epistemic injustice: opening science to society relates to the status of people in society as epistemic agents, and their role as knowers, interpreters, and providers of information (Fricker, 2007). By recognising non-researchers as epistemic agents through practices like citizen science and co-produced research, Open Science emphasises the importance of respecting community-held knowledge and language (UNESCO, 2023, p 19, 32; Bertram et al., 2023). Language is used to search for and synthesise existing literature, precisely define technical terms, develop research questions and hypotheses, and present findings and interpretations (Schmalz et al., 2025). As such, language does not merely describe the objects of study, it actively constructs them (Botha et al., 2023; Schmalz et al., 2025). Narrative choices embedded in scholarly writing can thus either reinforce existing power hierarchies or, in line with the aim of Open Science, contribute to more inclusive and participatory forms of knowledge production.

We investigate the importance of language in scholarly writing by focusing on research on autism. Milton and Bracher (2013) criticize autism research both for failing to fully engage autistic lived experiences in scientific work and for framing autism using problematic narratives. They argue that, while autistic individuals often do express their perspectives, these perspectives are frequently misunderstood, dismissed, or filtered through non-autistic interpretations. This highlights how language can shape not only how autism is described, but how it is understood throughout the research process. These issues are evident already when it comes to the theoretical framing of autism. For instance, the ‘medical model’ has often been used in scientific discourse despite having been criticized for suppressing autistic voices and pathologizing their communication styles (Milton and Bracher, 2013). However, when applied uncritically, alternative models can also overlook the diversity of perspectives within autistic communities. Such biases extend into empirical research methods, where using standardized methods that rely on language may exclude individuals who are non-speaking, situationally mute, or have co-occurring learning disabilities, leading to samples which are typically more verbal and more educated (Botha et al., 2023). Finally, in the dissemination phase, language used in articles is considered important, as evidenced by the debate around using person-first (“person with autism”) versus identity-first (“autistic person”) language. Despite there being no consensus on this matter within the scientific community, autistic communities often advocate for identity-first language as a reflection of political and cultural identity. In her work, Botha et al. (2023) stress the importance of listening to what autistic communities are saying as they are the ones being represented. When the subjects of research are not meaningfully included in shaping how they are described, there is a risk of increasing epistemic injustice, not only through exclusion from research processes, but also through the imposition of external narratives which may not align with lived experience (Milton and Bracher, 2013).

We contribute to these debates and to diversifying the largely Anglophone focus of existing body of work in both Open Science and autism by focusing our investigation on how autism is framed within the Croatian scientific literature. Through its technical infrastructure, supporting Open Science, primarily through the University Computing Centre (SRCE) that provides access to over 540 journals (platform HRČAK) and a diverse range of digital objects (platform DABAR), Croatia has established a great technical foundation which makes it easier for both scientists to open their science and the public to reach it (Bolkovac et al., 2025). Thus, literature on autism is accessible to a wider range of professionals and individuals. Our aim is not to evaluate the scientific validity of existing autism research in Croatia, but to qualitatively analyse the ethical and epistemological implications of how such research is communicated. To do so, we will conduct a content analysis of a sample of Croatian scientific publications on autism. Data extraction will be performed manually by multiple coders using a systematic protocol to evaluate linguistic patterns across the texts based on two criteria:1) on the use of person-first versus identity-first language when referring to autistic people and 2) on the use of pathologizing language when describing autism. We will present the results of our analysis and discuss them in light of debates within autism research that aim at raising awareness of community-preferred language and engagement with the perspectives of communities that are being studied.