The extracted definitions are presented in Appendix 1 (Table 1), and the resulting themes are shown diagrammatically as a metaphorical cultural concept soup (Fig. 2). In Fig. 2, diverse stakeholders, consciously or unconsciously, bring with them different cultural definitions to their work, and use what they think is the ‘cultural safety ladle’ to stir the soup. The thematic aromas that arise from the soup diffuse into social policy discourse to create ethereal meanings about which cultural concept–or group of concepts–are crucial to the success (or failure) of cultural reforms.
Definitional diversity
Our search yielded 42 definitions of cultural safety (Table 1) in social policy about First Nations Australians. The definitional diversity appeared to begin with the first published definition of cultural safety, namely Eckermann’s 1992 book ‘Binan Goonj: Bridging Cultures in Aboriginal Health’, as reprinted in 1994 [85]. Hence, it is the first entry in Table 1 (Row 1). A taxonomy was created for tracking the source of definitions AKE-dCSaf-AU (Ann Katrin-Eckermann, definition of cultural safety, Australia). The second definition is Robyn Williams’, published in the journal article, ‘Cultural Safety – What Does it Mean for Our Work Practice?’ (Williams, 1999): RW-dCSaf-AU (Table 1, Row 2).
Numerous modifications occur to Williams’ definition over the years, which is not the case for Eckermann’s definition. There are seven modifications to Williams’ definition (Table 1, Rows 3, 11, 15, 17, 19, 23, and 26). In our taxonomy, MBS-dCSaf(RW-dCSaf2)-AU, means that Maryann Bin-Sallik modified Robyn Williams’ definition. This occurs were Bin-Sallik extracted a paragraph from an unpublished paper of Williams’ [86] and reframed the paragraph (located in Appendix 3, p. 15) as Williams’ definition [87]. Other definitions of cultural safety (not Williams’ or Eckermann’s) were also modified by subsequent authors (Table 1, Rows 21, 24, 28, and 42).
The definitions also reflect diverse points and pathways within health (e.g., policy, hospital, nursing and midwifery, health workers, doctors, health equity, alcohol programs, health practitioners, women’s safety, general practice, suicide prevention), family and child safety (e.g., social work, education system, child care and young people, family violence), the mining industry, legal centres and legal aid, workplace health and safety, Australian trade and investment, Australian rules football, and program evaluation and libraries. This definitional diversity holds implications for ‘governance’ and ‘legitimation’ (Fig. 1) in the sense that definitions are vehicles of meaning for the governance actions driving organisational reforms that flow on to practitioner service delivery.
Developmentally dubious
Except for the AHPRA definitional development process (Table 1, Row 25), the processes for constructing definitions were opaque (dubious) in that no empirical, theoretical, or methodological processes are described for their construction (other than the fact that some draw-on earlier authors). There are no publicly available explanations of how definitions were developed, with whom they were developed, and whether any First Nations Australian community engagement occurred during the process of delineation. This lack of engagement and culturally informed process has implications for the ‘interpretive scheme’ concept of structuration (first level of Fig. 1) because community needs are not informing definitions and, thereby, guidance for practitioners to respond to community needs.
Problematic provenance
The cultural provenance of the definitions shows that many (n=14) (Table 1) are transformed from kawa whakaruruhau, a Māori First Nations concept from Aotearoa/New Zealand, and not from a concept of First Nations Australians (Table 1, Rows 1, 2, 3, 7, 11, 14, 15, 17, 19, 20, 23, 26, 27, and 30), such as the Wiradjuri concept of Nyaa-bi-nya: to examine, try, and evaluate [88]. Almost half of the definitions (n=20, Table 1, Rows 4, 5, 6, 8, 9, 10, 12, 13, 16, 18, 29, 31, 32, 33, 34, 35, 36, 37, 38, and 39) have unknown provenance. There are nine definitions (Table 1, Rows 21, 22, 24, 25, 28, 34, 39, 41, and 42) whose cultural origins are Australian and New Zealand (but not kawa whakaruruhau). This has implications for the ‘signification’ concept of structuration (Fig. 1) and social policy systems, because such ‘problematic provenance’ suggests the possibility of cultural appropriation [89] and signifies the devaluing of First Nations Australian ways of knowing.
Ostracised ontology
First Nations Australians are among the world’s oldest living cultures [90] and have local language terms for health and wellbeing. For example, there is Kaurna language: Yaitya Purruna/‘our own health and wellbeing’ [91]; Wiradjuri language: Waluwin/‘health and wellbeing’ [92]; Walpiri: wankaru/ ‘to promote and strengthen the life of Aboriginal people as a means of ensuring their survival and growth’ [93]; Aranda language: Kurruna Mwarre Ingkintja/‘good spirit men’s place’ [94], and Wirringa Baiya/‘women speak’ [95]. None of the definitions in Table 1 make reference to First Nations Australian languages to indicate that their meanings are based on translation of local worldviews. This ‘ostracised ontology’ has implications for the ‘domination’ concept of structuration and ‘legislation’ (Fig. 1, Level 1) because infusing legislation with cultural power should be based on the strengths of First Nations Australian cultures.
Validation vacillations
Validation involves stakeholder assessment of measurement tools to ensure they are culturally acceptable [26, 96]. First Nations Australians challenge the underlying Western cultural construct of many instruments [97–100]. None of the definitions (Table 1) come with information about their methodologies of cultural validation: what was the instrument used to validate their process and their definitions? Such ‘validation vacillations’ have implications for the ‘sanction’ concept of structuration (Fig. 1, Level 3) and the types of social services provided because who decides (sanctions) a service is culturally safe are First Nations Australians, who expect those services to reflect their cultural values.
Professional prose
Based on the individual (as opposed to organisational) authors the definitions are rendered in English and through the lenses of non-First Nations peoples (n=7, Table 1, Rows 1, 2, 14, 23, 30, 31, and 33) who were professionals with higher education qualifications, and who were employed by mainstream organisations. Similarly, other definitions were authored by First Nations experts (n=5, Table 1, Rows 9, 14, 26, 29, and 39). Even the First Nations authors are based in similar settings, though only a few authors are First Nations Australians (Maryann Bin-Sallik, Gregory Phillips, BJ Newton, Larissa Behrendt, and Sharon Gollan), but also with professional education and qualifications.
The result is where ‘professional prose’ structures the locus of power to rest with professionals rather than First Nations communities or service users. It is notable that almost all (except Row 30, Table 1) definitions developed by individual authors avoid referencing their own professional standing and power, despite the philosophical rampart of ‘reflexivity’. Other definitions reference professional/provider power (n=3, Table 1, Rows 21, 27, 30) and power imbalances/sharing/differentials (n=10, Table 1, Rows 8, 10, 16, 24, 25, 28, 34, 37, 40, and 41). This has implications for the ‘agency’ and ‘power’ concepts of structuration theory (Fig. 1, Level 3) because of the power of writing in creating meaning that influences practitioner attitudes and their practices.
Scholarly shenanigans
The definitions of cultural safety also suffer poor standards of attribution and citation (n=10, Table 1; Rows 2, 3, 12, 15, 17, 19, 23, 28, 40, and 41). Examples of ‘scholarly shenanigans’ are common. For instance, in an article about the importance of cultural safety to social work policy, Thompson and Duthie [101] quote Williams’ (1999) definition but attribute it to an article by Ramsden (1992), which does not contain any definition [102], and this shenanigan also occurs in an article by Elvidge and colleagues [27]. The definition AIHW-dCSaf (Table 1, Row 27) is incorrectly referenced to Papps and Ramsden [103], which is also incorrectly cited as the source of a so-called ‘definition’ (Table 1, Row 41).
There is one incorrect citation of Williams’ (1999) definition of cultural safety as Williams (2008) (Table 1, Row 19) and an internet search revealed n=124 instances where this incorrect reference was repeated [i.e., found by entering the incorrect reference: ‘Williams, R. (2008). Cultural safety: what does it mean for our work practice?’]. The definition by Williams (1999) is often attributed to Eckermann and colleagues (1994) (Table 1, Row 2, 23, and 26) [see for example 104]. The work is also misattributed in The National Best Practice Framework for Indigenous Cultural Competency in Australian Universities [43], and in academic articles [105, 106]. If scholars checked primary sources, they would have seen Eckermann’s definition (Table 1, Row 1) bears no resemblance to Williams’s definition (Table 1, Row 2). This has implications for the ‘modality’ and ‘facility’ concepts of structuration (Fig. 1, Level 2) because the inaccuracies are repeated in cultural safety discourse, a facility of meaning.
Meddlesome modifications
The definitions contain different ingredients (Table 1) with no explanation for their selection. In the health domain, these ‘meddlesome modifications’ of the concepts not only risk stultifying the cultural flavour of the soup, but also affect how definitions are interpreted. The modifications can be seen in the selection of key ingredients excised from each definition:
- recognised, assured, reflects, you/your culture, language, customs, attitudes, beliefs, and preferred ways (AKE-dCSaf, Row 1)
- determined, individuals, families, communities, practise, reflection, health practitioner, knowledge, skills, attitudes, behaviours, power differentials, safe, accessible, responsive, and racism (AHPRA-dCSaf, Row 25)
- defined, health consumer, individual, experience, care, ability, access services, and raise concerns (AIHW-dCSaf, Row 27)
- environment, diverse background, feels, valued, and accepted (LB-dCSaf, Row 29)
The meddlesome modifications hold implications for the ‘communication’ concept of structuration (Fig. 1) with the rationale that the selection of words (and who selects them) is significant for the communication of meaning to stakeholders and their organisations. Interestingly, the concepts of power (n=7), culture (n=12), reflexivity (n=6), and identity (n=20) are non-uniformly distributed and show selective word choices by authors.
Discombobulating discourse
The definitions in Table 1 contain a confusing of meanings. For example, ANMC-dCSaf states ‘regardless of race or ethnicity’ (Table 1, Row 4) whereas Eckermann’s definition implores ‘the need to be recognised within the healthcare system’ (Table 1, Row 1) and CATSINAM-dCSaf (Table 1, Row 16) states that it, ‘represents a key philosophical shift from providing care regardless of difference, to care that takes account of peoples’ unique needs’. Thus, while cultural safety may be a commonly used phrase, the ‘discombobulating discourse’ and definitional diversity within different policy points demonstrates the potential risk of divergent meanings through the ‘interpretive scheme’ concept of structuration and ‘practitioner-client interaction’ (Fig. 1, Levels 2 and 3). Then, when communicating with clients, practitioners may speak from a standpoint of either disregard or regard for race and justify both as correct choices by referencing the relevant definition.
Redundant reflexivity
All definitions (except the APHRA definition) occur without being reflexive to cultural diversity, power, and identity. The individual authors of organisational definitions (Table 1) are unknown, in that they do not identify themselves or their cultures, and in doing so demonstrate ‘redundant reflexivity’ (n=28; Table 1, Rows 4, 5, 6, 7, 8, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 22, 24, 25, 27, 28, 32, 34, 35, 36, 37, 38, 40, and 42). This invisibility of authors’ cultural identities prevents understanding of their cultural worldviews through which the definitions were developed.
There are also whole-of-organisation authors (n=23; Table 1, Rows 4, 5, 10, 11, 12, 13, 17, 19, 20, 21, 22, 24, 25, 27, 28, 32, 34, 35, 36, 37, 38, 40, and 42) with western colonial hierarchical governance structures that are at odds with First Nations community governance practices (Panaretto, Wenitong et al., 2014). There are some non-government organisations based on First Nations professional interests and advocacy (n=3; Table 1, Rows 7, 10, and 16), but where governance structures are still hierarchical based in colonial models. In contrast, four Aboriginal community controlled organisations offered definitions (VACCA in Row 6, VACCHO in Row 8, CAPSC in Row 17, and SNAICC in Row 18). None of the whole-of-organisation definitions are accompanied by reflexive statements about the philosophical implications of infusion their different governance structures with cultural safety.
Overall, the organisational authorship processes, and the resulting cultural safety definitions, lack grounding in the frequent call for reflexivity so often made by authors of cultural safety definitions. This has implications for the ‘norm’ concept of structuration (Fig. 1, Level 2) because it signifies a convention for reflexivity to be an optional, rather than essential, feature of cultural safety.
Fig. 2: Cultural concept soup with thematic results