Introduction

Along with the rise of digital museums and the use of communication technology by museums, the display and use of cultural artefact visuals are becoming more and more common. Not only are artefact digital pictures gradually being made available to the public as copyright allows through interactive devices (Grammenos et al., 2011), online activities (UNESCO, 2020), museum collection databases and their services (Seifert et al., 2017), but also visual elements of cultural artefacts are digitally edited and reused. These visuals may “hide” in everyday environments such as urban branding (Law, 2023), mascots, news, cultural products, and advertisements of goods and services, among many others.

Cultural artefacts have been utilised for more than just education or decoration. From the perspective of Critical Heritage Studies (CHS), they also play an important role in representing, expressing, and negotiating culturally relevant identities in the present (Walsh, 1992; Hall, 1999). Some studies have shown that for this purpose neither authorised nor individual interpretations of heritage were as neutral as they seem and that the meaning of heritage in the present will be made, used, and negotiated by different agencies (Hooper-Greenhill, 1992, 2000; Graham et al., 2000; Harvey, 2001; Smith, 2006; Bonacchi, 2022). As a representative of history and culture, heritage may serve as an authorised resource for modern nation-states. Smith (2006) proposes the heuristic notion of the Authorised Heritage Discourse (AHD) to conceptualise an authoritative expression and the making of heritage to facilitate power analysis into heritage studies. Through the official definition and granting of cultural heritage, the AHD emphasises the professionalism and national significance of heritage protection and tends to view members of the public as being passive receivers of heritage education or marketing. To construct a harmonious national heritage discourse in the present, the AHD selectively remembers and forgets the past, leading to possible dissonances among individuals and communities (Yan, 2015; Dellios, 2022). In such a discourse, cultural artefacts are actively incorporated into national narratives. However, the use and engagement of heritage for identity-related purposes is not only officially hosted but may also occur with private and everyday heritage-making (Jones, 1997; Smith, 2006, 2021). The popular and personal engagement matters as much as professionalism in heritage-making, whether it is reproducing or challenging the AHD and existing power relations (Smith, 2006, 2021; Mason and Baveystock, 2009; Wang, 2012; Ku, 2018; Hu, 2023).

Compared to heritage-making within obvious interpretive contexts, such as those provided by museums or heritage sites, in the digital age, cultural artefacts may only leave a fragmented visual exposure. For instance, we may only have a passing acquaintance with any particular image or may quickly swipe through a collection of pictures shared on a museum’s social media that are not explicitly involved with storytelling or emotional investment. However, it does not necessarily mean that heritage-making is not occurring in these exposures to pictures (Waterton, 2009). When digital artefact visuals or other heritage items occur decontextualised in everyday life, are they viewed as simply “objective” images or are they still playing a part in national representation (for both officials and individuals)? This is the key underlying question prompting this research.

This research critically explores viewers’ responses to “objective” artefact digital visuals according to their contemporary national identity. We borrow “in-group/out-group” concepts in Social Identity Theory (SIT) to conceptualise viewers’ categorisation and meaning-making of subjective “domestic (in-group)” and “foreign (out-group)” digital artefact visuals, and design the lab or online experiments based on large-scale materials to go beyond museum or heritage sites contexts. The “in-group/out-group artefacts” concept helps academically divide people’s subjective meaning-making from archaeological professional knowledge (i.e., defining “national artefacts” from one’s own way, not from excavation sites or documentary evidence) and ensures the de-essentialisation and openness regarding contemporary identities like nation, ethnicity, and culture. We proposed two main research questions:

RQ1. How do people understand and classify the images of in-group/out-group artefacts regarding their contemporary national identity?

RQ2. Are there differences in their evaluations of self-defined in-group/out-group artefact visuals?

We expected RQ1 to explore whether members of the public can link the digital artefact visuals to national representations and what this linkage may be. RQ2 considered whether these impressions of “in-group/out-group artefacts” (if they exist) banally express national identities, such as a preference for in-group artefact appreciation or value evaluations. We also added two extra sub-questions, exploring the extent to which this in-group/out-group categorisation on artefact visuals influences intergroup attitudes towards living people, and relates to one’s characteristics, such as the degree of cultural identity, ethnic essentialism, and cultural exposure:

RQ3. After viewing typical in-group/out-group artefact visuals, do their interpersonal attitudes towards living in-group/out-group people differ?

RQ4. Whether the above responses relate to one’s degree of in-group cultural identity, ethnic essentialism, and cultural exposure?

This topic makes China a suitable context. In China, official heritage discourses function similarly to the international AHD to a certain extent (Yan, 2015). In the digital age, with top-down directives like The Internet + Chinese Civilisation (Bollo and Zhang, 2017) and Smart Museums (Wang CF, 2020), many museums have created online accessible collection databases. Additionally, Internet cultural products serving cultural tourism mean that edited heritage visuals are highly accessible. Chinese young people are one of the mainstream audiences for these heritage disseminations. According to a report by the Douyin platform (2024), a higher proportion of users born after 2000 are interested in museum-related online content compared to users in other age groups.

As Fig. 1 shows, we conducted two mixed-method studies following a design of (QUAL + QUAN) → QUAN (Schoonenboom and Johnson, 2017; Creswell and Creswell, 2022). The material consisted of 1000 pottery, bronze, and porcelain digital pictures collected from online collection databases based in China and abroad. We chose artefacts categorised as European in the databases as possible representatives of the out-group to highlight the differences in the eyes of non-experts. The three artefact material types were selected to ensure the quantity and the historical periods they cover. The first study focused on the in-group/out-group categorisation of artefact pictures of 60 Chinese university students in response to RQ1. We collected their descriptions of in-group/out-group artefacts after categorisation experiments and interviewed them about their definition of “Chinese cultural artefacts”. We coded and analysed how they defined in-group artefacts, and the descriptions were then tagged as 22 features at the pixel-level, object-level, and semantic-level (Dalrymple et al., 2019). Based on Study 1, Study 2 screened 30 subjective typical in-group artefacts, untypical artefacts, and typical out-group artefacts, respectively, in response to RQ2 (Supplementary Table A1). Typical in-group and out-group artefacts were distinctive and conformed to the “stereotypes”, while the untypical pictures consisted of a portion with hard-to-distinguish features and another portion with clearly “dissonant” evidence, e.g., porcelains with realistically painted figures. We set the three corresponding cultural priming conditions (Hong et al., 1997, 2000, 2003; Wong and Hong, 2005) and an extra control group, collecting 256 valid online subjects’ evaluations of the artefact visuals and their representing culture without any contextual information or comparison. To address RQ3, participants’ implicit attitudes toward “living” individuals—measured through in-group and out-group face pictures—were assessed, given that untypical artefact visuals may elicit cognitive dissonance or influence implicit interpersonal attitudes (Festinger, 1957; Tunbridge and Ashworth, 1996; Liu, 2020). The quantitative differences in the data between these groups were compared. Both studies collected the subjects’ three quantitative features: Chinese cultural identity, ethnic essentialism, and exposure to in-group/out-group cultures to explore RQ4. The correlations between the subjects’ responses and individual features were statistically analysed.

Fig. 1
Fig. 1
Full size image

The research procedure and methods.

The outcomes mainly demonstrate that the reproduced artefact visuals of everyday life, despite carrying no apparent narrative purpose, may nonetheless be both framed by the AHD and, in doing so, may be understood as banal expressions of a nation among the young participants. This helps rethink the social consequences of possible stereotyping in unreflective digital reproduction of cultural heritage visuals in museum practices. Making museum collections digitally visible and accessible is one of the most fundamental digital heritage tasks. However, this research implies that existing national narratives and power structures within heritage representation may be unwittingly perpetuated in decontextualised digital heritage visuals. At least at this point, the accessibility of digital visuals may not be an innovation but an expansion of convention.

Literature review

In this section, our task is to establish theoretical connections between people’s categorisation/stereotypes of digital visuals, national identity, and digital heritage, with an awareness of potential conflicts among these concepts in different disciplines or epistemologies. In heritage studies, this tension exists between the AHD and the agency of communities and visitors with sub-national identities in negotiating the meaning of heritage (Smith, 2006). Maintaining a critical stance in stereotyping research requires avoiding the framing of stereotypes as either pathological or politically neutral (Pickering, 1995, 2001). Briefly, we advocate incorporating the audience’s subjectivities into the negotiation of heritage meaning with the nationalist and professional AHD, while critically relating the “stereotypes” in their meaning-making to mainstream national narratives and the AHD ideas in a less mechanical way.

The study of everyday encounters with national cultural or heritage symbols has a history. Anderson (2006[1983]) proposed the nation as an “imagined community” and identified the role of media such as newspapers and museums in the formation of nations. Billig (1995) developed the concept of “banal nationalism” to describe the “unwaved flags” of ordinary life. Although nationhood is not the first theme in most people’s lives, the flags, language, news, etc. construct “borders” in a banalised, or everyday, way to distinguish “us” from “them”. The increasing digital cultural artefact images coming out of museums may also be seen as an extension of national flags consumed for imagined communities. For instance, Waterton (2009, p. 49) pointed out that the images of unmanned landscapes in tourist brochures of England convey “a single logic anchored to the AHD”, making heritage distant, elitist, and national. Grincheva and Stainforth (2024) discussed how digital heritage in galleries, libraries, archives, and museums is integrated into a national/international framework, influencing and shaping real geopolitics in virtual spaces. In comparison to seeing nationalism as a radical performance, we use banal nationalism as a conceptual tool to explore how the representations of nationhood in digital artefact pictures are “taken for granted and unreflectively accepted” (Ichijo, 2017, p. 261).

The critical analysis of heritage visuals as national representations is in parallel with studies on audience agencies. Bourdieu et al. (1996, 1979, 1990) believe that the distinction between one’s legibility of the artwork reproduces class structure and cultural capital (Fyfe, 2004). Merriman (2000[1991]) and Dicks (2008, 2016) extend Bourdieu’s ideas to cultural consumption beyond middle-class aesthetic tastes, exploring the role of historical museums in awakening a sense of place and collective memory among different members of the public. The agency of people was given more attention, and the (re)production or resistance of nationhood and existing social structures in everyday life and mass culture, such as in memes and video games, is also identified (Edensor, 2020, 2002; Fox and Miller-Idriss, 2008; Antonsich, 2016; Bos, 2018; Wiggins, 2019). For instance, Mason and Baveystock (2009) studied the “ICONS of England” online project, a platform where members of the public can nominate and vote for the symbol of England’s national heritage. The nominations showed a mixture of official heritage and popular cultural forms. Overall, they reflect the diversity of how people use heritage or other cultural contents to represent social identity, as well as the compatibility or tension between them and professional discourses and national narratives.

Identifying “our artefacts” also relates to social psychological concepts and measurements, such as in-group/out-group categorisation (Castano et al., 2002) and in-group favouritism (Brewer, 1979; Fischer and Derham, 2016; Strickland-Hughes et al., 2020). These concepts and measurements are mainly based on the SIT (Tajfel et al., 1971; Tajfel, 1974, 1979, 1982; Turner, 1975; Hogg and Abrams, 1998; Brown, 2000). Briefly, the SIT theorises that people may simplify the characteristics of humans, generate categories, and form corresponding group stereotypes. After determining the group categorisation to which they belong, individuals may exaggerate the in-group/out-group differences and produce behaviours like in-group favouritism and out-group derogation. The objects of favouritism include not only people’s face and voice materials (Young et al., 2010; Perrachione et al., 2010) but also historical texts and evaluations (Huang and Liu, 2018; Iordanou et al., 2020). Some aesthetic studies on artworks (Mastandrea et al., 2021; Darda, 2022) have discussed the role of the unconscious categorisation of “us” and “them” in the artwork viewing experience, but they focused more on the aesthetic evaluation of paintings rather than the cultural meanings of cultural artefacts for individuals.

As Billig (1995) argued, the SIT tends to overemphasise people’s motivation to distinguish groups and ignore the particular meaning of social identity and the way it becomes inhabited. Similarly, the dilemma of stereotyping research is that we must avoid pathologising stereotyping as the prejudice of a minority that can be solved by providing “the truth”, while also avoiding naturalising it as a common cognitive strategy, thereby overlooking the fact that stereotyping participates in cultural practices and processes (Pickering 1995, 2001). Pickering distinguished between categories and stereotypes. Categories are tools for constructing mental maps and help everyday social relations and interactions, while stereotypes have special attributes related to ideological views, values, and issues of power (Pickering 2001, p. 2–3). By mentioning this, we are not intending to distinguish between which responses of participants to “our artefacts” are categorisation and which are stereotypes, but rather to provide a conceptual space for discussing the extent to which the “in-group/out-group categorisation”—a concept that can be empirically measured—becomes a stereotype reproducing the AHD, banal nationalism or other ideological views. The research is expected to promote critical thinking on stereotyping in culturally relevant digital media production (Degand, 2019).

Methods

Study 1: The in-group/out-group categorisation of artefact pictures

Stimuli

The material consisted of 1000 pictures of pottery, bronze, and porcelain artefacts categorised as from the Chinese and European regions in online databases. We began to construct the material by identifying shape-types of artefacts with Chinese teaching materials of artefacts (Li, 2005) and the Beazley Archive Pottery DatabaseFootnote 1 (Supplementary Appendix A). Starting with shapes is because the shape (or the extended functional meaning) of an artefact is intuitive and basic when named. In the British Museum’s database, for instance, many artefacts are named only by their shape type (e.g., oil-flask). Some shape names have the same meaning in the two cultural contexts, such as Ping = bottle/vase, Wan = bowl, while some shapes are archaeologically named and unique to that culture. We then searched these shapes from online collection databases of 38 Chinese museums for “Chinese artefacts”, and the British Museum and the Metropolitan Museum of Art for “European artefacts” (using Europe as a keyword). When there was not enough of a particular shape type, we included a small number of pictures from auction houses or taken by visitors. Within each shape, we consciously avoided repetition in appearance and chronology to manually balance other factors such as decoration, colour, and date. Inevitably, some specific artefact shape types were only found with specific ages and appearances. The material consisted of 170 potteries, 160 bronzes, 200 porcelains as “Chinese representations”, 170 potteries, 100 bronzes, and 200 porcelains as “European representations”. The classification criteria of “Chinese/European” here were simple and relative, as which region an artefact belongs to is somehow an open question, and we were only interested in how viewers processed them. Ironically, to study the in-group/out-group classification of audiences, we ourselves have to make similar classifications when collecting experimental materials. Therefore, we cannot escape accusations of potential biases. This point will be reflected upon at the end of the article. After standardised processing (Supplementary Appendix A), the pictures were divided randomly into 5 large groups of 200 pictures each by shape name. Within the large groups, they were further divided randomly into 4 blocks of 50 pictures each.

Participants

A total of 60 subjects (M = 23.50, SD = 2.18 years old) were recruited through the university online forum on a voluntary basis, including 29 males and 31 females. It required that participants were not colour-blind or colour-impaired, had normal or corrected visual acuity, and were not from an academic background in cultural artefact or museum studies. All subjects were right-handed. They received a modest payment upon completion of the experiment.

Procedure

As seen in Supplementary Appendix A, subjects were first given a guideline, telling them that they would see a series of digital pictures of global artefacts on the monitor in a randomised order, and the data would only be used for research purposes and would be anonymised. They were invited to decide whether the artefact in the picture belonged to “Chinese tangible cultural heritage”. If the answer was yes, the subject could press the F key at any time after the picture appeared, the J key if the opposite was true, or the space bar if they were unsure. They were told that they did not have to seek knowledgeable correctness and could freely decide based on experiences or first impressions, and they were encouraged to show preferences whenever possible. Each subject randomly evaluated 200 pictures (into 4 blocks) from any of the 5 large groups. The order of presentation between and within groups was randomised via the Latin square matrix. The experiment was designed and performed on Tobii Studio – version 3.4.8.

After the 4 blocks, the subjects completed a questionnaire. First, they provided 5–10 descriptions of the artefacts they classified as yes and no, respectively. The Chinese cultural identity scale (Cleveland and Bartikowski, 2018), ethnic essentialism scale (No et al., 2008; Long, 2019)Footnote 2 and cultural exposure questionnaire (Wang C, 2020; Xu, 2018) were afterwards measured via scales or questions (Supplementary Appendix A). It is noted that the discussions of “identity”, “essentialism”, and “cultural exposure” in this paper are still general rather than these measured variables, unless we specifically mention statistical analysis and the scales. The subjects then participated in an interview on the following open-ended questions: How do you understand the Chinese historical and cultural tangible heritage/artefacts? (Probe: If you have to make a definition, what would it be?) Did you have a clear classification criterion in the experiment? (Probe: Which materials did you think were or were not clearly Chinese artefacts? Which type was more difficult for you to judge? What choices did you make in this case?)Footnote 3.

Study 2: The evaluations of self-defined in-group/out-group artefact pictures

Stimuli

We selected 30 pictures that were consistently categorised as in-group/out-group with the lowest mean standardised response time (RT, see Supplementary Table B1) in Study 1 as the material for the typical in-group/out-group artefact priming group, respectively, and 30 that were classified as in-group in half of the subjects and had the highest mean standardised RT as the untypical artefact priming group (Supplementary Table A1).

Participants

This study adopted a between-subjects experimental design containing 4 conditions (control/typical in-group/untypical/typical out-group). The recommended subject size of 232 was calculated by Gpower 3.1 (Faul et al., 2007)Footnote 4. We collected 256 subjects’ valid data (M = 23.8, SD = 3.47 years old), with the number of subjects in the 4 groups being 64/66/62/64. Among them, 148 were recruited from the university forum, and the others came from registered users of the NAODAO online research platform (Chen et al., 2022)Footnote 5; one hundred and ten subjects were male and 144 were female, and 2 subjects preferred not to tell the gender; ten of them were left-handed. All subjects received a modest payment upon completion of the experiment.

Procedure

The main experimental programme is based on PsychoPy – version 2021.2.3. The procedure was run on the NAODAO platform. All subjects gave informed consent before starting the online experiment. They were randomly divided into one of the 4 groups by the platform. Except for the control group, which did not include artefact pictures, the process was:

  1. i.

    Pre-test. Subjects first filled in demographics. Considering that traditional cultural images may enhance cultural identity (Chang, 2020), the identity and essentialism scales data were also collected in advance.

  2. ii.

    Artefact evaluation. Subjects were only told to evaluate 30 artefact pictures without being told any additional information. Each picture was randomly presented, accompanied by 4 sequenced evaluations examining the subjects’ liking, perception of aesthetic value, economic value and symbolic valueFootnote 6. Some questions were adapted from Su et al. (2020). Subjects rated their agreement with the description by clicking the mouse on a 7-level scale (including decimals).

  3. iii.

    Cultural imagination and evaluation. Subjects were asked to freely associate how they would describe the cultural connotations behind these artefacts in one minute. Then they completed 9 questions in randomised order on how much they agreed with the declarative sentences about the culture they had imagined. These sentences implied three factors: perceived cultural proximity, perceived cultural continuity, and emotions when imagining culture (see Supplementary Table A2 for the validation factor analysis).

  4. iv.

    Implicit Association Test (IAT). The implicit intergroup attitudes towards “living” humans were measured by D-values in the IAT. The D-value means the extent to which subjects associated positive words with in-group face pictures relative to negative words. Higher D suggested higher in-group favouritism towards human faces. The stimuli (Supplementary Appendix F) and process followed Wang et al. (2008), Zhao (2013), Liang et al. (2018), Robin Scaife et al.Footnote 7, the project implicitFootnote 8 and Greenwald et al. (2003). The subjective typicality of the artefact and face pictures, and cultural exposure data were collected at last.

Data analysis

The corpus data was coded by three raters, validated by Krippendorff’s Alpha inter-rater reliability coefficient (Krippendorff, 2019) using K-Alpha Calculator (Marzi et al., 2024) in Supplementary Appendix B. The quantitative data were analysed by variance (ANOVA), non-parametric tests, and correlations using SPSS – version 25, which is shown in detail in the Results. The R – version 4.1.1 was used for visualisation.

Results

Coding and analysis of the self-definition of in-group artefacts

The responses of the 60 subjects in Study 1 to the question “How do you understand the Chinese historical and cultural tangible heritage/artefacts?” were coded and categorised as Table 1. Forty subjects thought Chinese artefacts are something valuable to “us”, including historical (S49: “Something that has been used in history and then preserved for research. We can study it to better understand our history and culture.”), symbolic (S50: “It is a part that clearly distinguishes itself from Western or other nations.”), aesthetic (S26: “For example, some porcelains that everyone senses pretty…”) and education values (S57: “Kids these days don’t really like to see these things, and I feel they should know about them through a variety of mediums.”). Thirty-nine and Thirty-one subjects defined Chinese artefacts through the scope of time and space, respectively, using expressions such as “thousands of years”, “olden times”, “China (used as a region)”, or “central plains”. Thirty subjects emphasised the connection between objects and ancient people, for instance, considering in-group artefacts to be something handmade and left behind by previous generations. A small number of subjects judged based on their personal experiences, or understood in-group cultural artefacts from an essentialist perspective.

Table 1 The subjects’ understanding of “in-group artefacts” (including duplicate codes).
Full size table

Coding and analysis of the descriptions of in-group/out-group artefact images

In the post-test questionnaire, participants provided descriptions of the cultural artefacts they classified as in-group and out-group. A total of 363 descriptions for in-group artefacts and 325 for their counterpart were collected for coding. Taking reference from Dalrymple et al. (2019), the coding started by identifying three first-level themes (pixel, object, and semantic) and then sub-themes, ultimately creating a system in Fig. 2a involving 22 sub-tags (Table 2):

  1. i.

    Pixel-level features (P1–P6) refer to the characteristics of colour, structure, and texture complexity. These features were purely visual stimuli without being recognised as specific objects and can be objectively described (e.g., green, dim, intricate patterns).

  2. ii.

    Object-level features (O1–O6), mainly involving the recognition of themes, materials and shapes in the artefacts. Participants were able to describe these elements as objects, and the descriptions generally did not involve subjective feelings (e.g., auspicious clouds, foreigners, handles).

  3. iii.

    Semantic-level features (S1-S10), referring to subjective or perceptual description. The words may carry personal sentiment or certain knowledge (e.g., simple and elegant, strange, religious worship)Footnote 9.

Fig. 2: Statistics of artefact visual features.
Fig. 2: Statistics of artefact visual features.
Full size image

a Coding system for the description of artefact visual features. b Comparing the frequency of the coded tags for in-group versus out-group artefact descriptions. The distribution of the number of tagged points was arranged from top to bottom according to the difference between the number of out-group points and that of in-group points, from highest to lowest. c The relationship between the entropy and the number of people categorising a picture as in-group.

Table 2 The coding for characterisation of in-group/out-group cultural artefact pictures (including duplicate codes).
Full size table

The differences in frequency and content of 1231 tagged references between in-group and out-group artefacts were compared in Fig. 2b. The object-level and the semantic-level took up a larger portion of the weight at both ends. Also, we noted that the pixel-level and object-level features are relatively measurable. As a complement, the quantitative relationship between the computable visual features of a picture and its classification behaviour was analysed. For each pictureFootnote 10, we computed three behavioural metrics and a series of pixel-level and object-level visual features. The behavioural metrics included possibility (the number of participants categorised as Chinese), categorisation entropy (the inconsistency of categorisation, see Fig. 2c), and average standardised RT for each picture, measuring how each artefact was categorised (Supplementary Table B1). Based on the pixel- and object-level sub-themes, we tried to find the potential calculating or labelling methods for verification (following the work of Li and Chen, 2009; Zhang, 2017; Zhang et al., 2020; Iigaya et al., 2021; Lv et al., 2018). The detailed methodology is listed in Supplementary Tables C1 and C2.

Combining the subjects’ description coding and the statistical analysis of the correlation between the visual features and the behavioural variables (Supplementary Appendix D), we inferred the ambiguous figures of in-group/out-group artefacts in the eyes of Chinese young people. The main pixel-level elements of subjective impressions about “Chinese” artefacts include white or cold tones, relatively simple colour schemes or textures, low brightness and shadow contrast, low saturation, and symmetrical structure, while the opposite was more in line with the impression of the out-group counterpart. However, the absolute values of the correlation coefficients between pixel-level features and behavioural data were low, and there were some repetitive descriptions in both in-group/out-group artefacts, suggesting that the pixel-level features were not distinctive.

On the object-level, the impressions of in-group artefact visuals include ice crack texture, figures and cultural scenes (e.g., appreciating flowers in a garden), and painting style and content (flowers, birds, or landscapes) in in-group culture. Loong and auspicious clouds were the more frequently mentioned mythical elements. For out-group artefacts, motifs such as triangles, foreign people, and mythical elements like angels and gods were more impressive among participants. In terms of the material, subjects’ descriptors of in-group artefacts mainly included bronze (54 times) and porcelain (72 times). Descriptions of the material objects of out-group ones mostly referred to gold-rimmed decorations (12 times), and they seldom mentioned the material of the artefact’s whole body. Statistically, bronze was the material with the highest probability of classification as in-group, the highest uniformity and the shortest RT (Kruskal–Wallis H-test, 139.417, P < 0.001; 81.505, P < 0.001; 31.398, P < 0.001)Footnote 11. For in-group artefact shapes, subjects could relatively easily name academic vessel types, but they were mainly limited to the widely known names such as Ding. Additionally, some subjects had impressions of the decorative details (a frequently mentioned feature was “three-legged”). As for the out-group artefact shapes, subjects used simple and daily life words like “plate” and “bottle”. They also noticed some detailed shape features like “sharp”, “upturned”, and other local shape features. Moreover, statistical analysis showed that artefact shapes with a higher width-height ratio may be more in line with stereotypes of in-group artefacts (Supplementary Appendix D).

According to the coded results, on the semantic layer, participants had perceptual awareness of some objective features. For example, some subjects thought that the colours of in-group artefacts were “introverted” and “atmospheric”, the shapes were “conventional” and “harmonious and beautiful”, while the colours of out-group artefacts were “exaggerated and bold” and “high-profile”, and the shapes were “humorous” and “strange”. Subjects not only made judgements about the art style, such as “Chinese paintings are freehand and Western paintings are realistic”, but also categorised them as combining specific cultural knowledge, such as the “Roman soldiers with shields”, “Indian Buddhist”. They would use self-reference, such as if its function can be seen (e.g., seeing “don’t know how to use” ones as out-group), or speculating using personal experiences or feelings, such as “I just feel like we (Chinese) are more frugal and like to repair things (S27)”, “I just imagined some (Chinese) TV series I have seen and put that thing imaginatively, and if I felt harmonious, I would choose yes (S11).” Furthermore, subjects paid more attention to temporal descriptions for Chinese artefacts (using words such as “aged”, “old-fashioned”) but geographic descriptions for foreign ones like “ancient Egypt” and “Africa”, although they may not be all actually “correct”. This also indicates that the subjects followed the principle of “Chinese/non-Chinese” rather than “Chinese/European” in categorising.

Since digital artefact pictures did not tell a story about which “country” they belonged to, where did the “stereotypes” come from? We further coded and analysed subjects’ responses to the probe interview questions. Eighteen subjects mentioned history books or history education (S17:” I used to learn some of that in history courses”), and 14 subjects mentioned museums or archaeological sites (S13:” I recalled that I had vaguely seen these in museums…”). Four subjects mentioned news or documentary (S35: “Mainly from the websites of the mobile phones, press and news”). Six subjects mentioned films, TV series or advertisements (S60:” In some films and TV shows, I saw features from our ancient palaces…”). Two people mentioned items from their life scenarios (S09:” Back in my parents’ day, the house would be decorated with those big porcelains…”). Together, authorised knowledge and popular culture both help link artefact images to cultural identity.

Quantitative differences in the in-group/out-group artefact visual evaluations

In Study 2, we compared the subjects’ responses towards typical in-group artefacts, untypical artefacts, and typical out-group artefacts (30 screened in each group). Briefly, untypical artefacts tended to be recognised as in-group in the between-subject experimental design, and in-group and out-group artefacts caused differences in almost all evaluations (Fig. 3a). However, the effect of artefact visual contact on interpersonal attitudes towards “living” people was limited.

Fig. 3: The quantitative differences under priming conditions.
Fig. 3: The quantitative differences under priming conditions.
Full size image

a Means and standard deviations of subjects’ scores for different artefact picture groups. b Means and standard deviations in implicit intergroup attitudes towards “living” people after exposure to artefact visuals across priming groups.

The typicality of “Chinese cultural artefacts”

The one-sample Wilcoxon signed-rank test on the 7-point scale evaluations of the typicality of in-group artefactsFootnote 12 (for detailed tests and P-values see Supplementary Appendix E) found that the scores of 30 typical in-group cultural artefact pictures were significantly higher than the median of 4. Among the 30 typical out-group artefact pictures, 27 had significantly lower scores than 4. Surprisingly, the non-typical artefacts that subjects could not agree on in-group/out-group in Study 1 were relatively easily received as in-group in Study 2. Twenty-seven untypical artefacts had significantly higher results than the median of 4.

Artefact evaluations

There was a significant difference in artefact liking (Shapiro–Wilk test, P = 0.279; F(2,189) = 12.727, P < 0.001, \(\eta\)P2 = 0.119). Subjects’ liking of typical out-group artefacts was significantly lower than for untypical ones (P = 0.003, 95% CI [−0.736, −0.152]) and typical in-group ones (P < 0.001, 95% CI [−1.020, −0.444]). Regarding the evaluation of the aesthetic value (Shapiro–Wilk test, P = 0.228; F(2,189) = 14.633, P < 0.001, \(\eta\)P2 = 0.134), multiple comparisons revealed that the rating of aesthetic value followed the order of typical in-group > untypical > typical out-group (P < 0.05). There was also a significant difference in the evaluation of the economic value (Shapiro–Wilk test, P = 0.010; Kruskal–Wallis H-test, 19.740, P < 0.001) and symbolic value (Shapiro–Wilk test, P = 0.011; Kruskal–Wallis H-test, 12.639, P = 0.002). The subjects considered the economic and symbolic values of typical in-group artefacts to be significantly higher than those of untypical artefacts (Mann–Whitney U-test, P = 0.008; P = 0.022) and typical out-group ones (Mann–Whitney U-test, P < 0.001; P < 0.001). However, the evaluations of out-group pictures were still at a neutral level.

Cultural evaluations

A significant difference was shown in cultural proximity (Shapiro–Wilk test, P = 0.005; Kruskal–Wallis H-test, 43.592, P < 0.001), cultural continuity (Shapiro–Wilk test, P < 0.001; Kruskal–Wallis H-test, 9.887, P = 0.007) and emotions when imagining the culture (Shapiro–Wilk test, P < 0.001; Kruskal–Wallis H-test, 37.695, P < 0.001). The three variables for typical out-group pictures were significantly lower than those for untypical pictures (Mann–Whitney U-test, P < 0.001; P = 0.050; P < 0.001) and typical in-group pictures (Mann–Whitney U-test, P < 0.001; P = 0.002; P < 0.001).

Implicit attitudes towards “living” faces after artefact visual exposure

As seen in Fig. 3b, although the mean value of in-group favouritism towards human face pictures (D-value) was slightly higher in the untypical artefact group, there was no statistically significant difference in the ANOVA (Shapiro–Wilk test, P = 0.139; F(3,252) = 1.112, P = 0.345, \(\eta\)P2 = 0.013).

Quantitative relations to in-group cultural identity, ethnic essentialism, and cultural exposure

In-group identity and essentialism

The correlation analysis found that the higher the subjects’ score on the Chinese cultural identity scale, the lower the RT for the classification of typical in-group cultural artefacts in Study 1Footnote 13 (r = −0.352, P = 0.006). Higher RT may indicate that subjects are more cautious about categorising and prefer to make accurate distinctions (Blascovich et al., 1997). Furthermore, subjects with a high level of cultural identity scale had a higher degree of liking for artefacts (r = 0.149, P = 0.039) and symbolic value assessment (r = 0.169, P = 0.019), with a higher sense of intimacy (r = 0.241, P = 0.001), continuity (r = 0.297, P < 0.001) and emotion (r = 0.229, P = 0.001) about the culture behind the objects after viewing. However, subjects with a high score on the essentialism scale had relatively low perceptions of the economic value (r = −0.153, P = 0.034) and symbolic value (r = −0.178, P = 0.014). This is difficult to explain because the subjects were confronted with three different artefact groups. Since our study only used scales to measure identification and essentialism, and they may have more complex performances (Desmond and Emirbayer, 2009), this finding needs to be analysed with more qualitative instruments in future studies.

Exposure to in-group and out-group cultures

The categorisation RT of typical out-group artefacts was positively correlated with the duration of reading domestic news (r = 0.328, P = 0.011) and negatively correlated with the degree of exposure to people from European (r = −0.286, P = 0.027). In addition, the duration of consuming in-group cultural products (in the questionnaire, it refers to films, TV series, animations, music, and literature) was positively correlated with RT spent categorising untypical cultural artefacts (r = 0.411, P = 0.001). The D-value negatively correlated with the frequency of in-group cultural product consumption (r = −0.182, P = 0.004). There was also a positive correlation between the consumption of cultural products in European areas and the artefacts’ economic value evaluations (frequency: r = 0.170, P = 0.018; duration: r = 0.147, P = 0.042). One’s contact with people from Europe was positively correlated with the liking for artefacts (r = 0.157, P = 0.029), cultural intimacy (r = 0.158, P = 0.029), and emotions when imagining culture (r = 0.223, P = 0.002), and negatively correlated with the essentialism scale score (r = −0.148, P = 0.018).

Discussion

This research mainly found that: (i) The subjects had certain abilities to recognise in-group cultural artefacts but not all understood them through “professional knowledge”. (ii) The image of in-group/out-group artefacts was linked to a mixture of objective and perceptual descriptions, from colours, shapes, and specific objects to personal feelings and experiences. (iii) The categorisation may produce unconscious expressions of in-group favouritism of artefacts but have a limited effect on, or work independently from, their attitudes towards “living” in-group/out-group human face pictures. (iv) The above behaviours were partially statistically associated with one’s in-group cultural identity, ethnic essentialism, and cultural exposure. Together, they demonstrate the potential for pure digital heritage visuals to be read for national representations by lay audiences in the absence of an obvious narrative, providing a deeper understanding of how heritage meaning-making continues to be framed by banal national identity expressions in the digital visual age.

First, the digital artefact visuals can carry national identity representations without external professional explanations. In response to RQ1, “What is a Chinese artefact?”, by letting the subjects freely share their answers, we found that the objective indicators, such as temporal and geographical scope, did not enjoy absolute priority. Subjects overall also preferred to understand the in-group cultural heritage regarding diverse public values. As shown in the results, this value is collectivist for “us”, a concept discussed in research on banal nationalism (Mihelj et al., 2009; Antonsich, 2016). Some emphasised the inheritance from the ancients, revealing a linear view of history. This is how the AHD tends to convey—something “intended for those who ‘belong’—a society which is imagined as, in broad terms, culturally homogeneous and unified” (Hall, 1999, p. 6). Under these self-reported definitions, certain features of artefact digital pictures were subjectively linked to national representations among subjects. Visual objects (such as bronze material and myth for in-group artefacts and human figures for out-group ones) and perceptual semantics (such as speculations on date, geographic, function, and style) in artefacts seemed to play a greater role in distinguishing in-group/out-group impressions. There were some other reasons including using descriptors of humans such as “humorous” to describe artefacts, a tendency to classify more bronzes as in-group, and categorising “unseen before” or “do not see their function” artefacts as out-group.

Second, the in-group favouritism on artefacts showed a banal expression of national identity and did not explicitly manifest as out-group hostility. Subjects overall showed in-group favouritism in evaluating the artefact visuals. In Study 2, responding to RQ2, visually in-group artefacts were scored higher in a series of value evaluations. Additionally, the subjects’ intimacy, positive emotion, and continuous feeling about the imagined culture after viewing typical out-group artefacts were relatively lower, although these ratings were still at a neutral level. This contrasts interestingly with some previous studies on audiences’ evaluations of cross-cultural travelling destinations or television works (Kastenholz, 2010; Lu et al., 2019), which showed that cultural distance did not cause an inevitable decrease in appreciation. In-group favouritism is more likely to show up when heritage was not narratively processed.

Many reasons may make an artefact image difficult to classify into an in-group/out-group in Study 1 (Supplementary Table A1). They might be prominent in both “Eastern” and “Western” characteristics, or not align with any cultural instincts, or the quality of the digital pictures themselves may not be high enough in resolution. However, this group of atypical artefact pictures was overall considered more “in-group” by subjects in Study 2. Some subjects in Study 1 may explain how they dealt with their cognitive dissonance, e.g., S38: “Maybe being a Chinese myself, I tried to list as many as I can that I thought to look good in the definition of traditional Chinese culture”; S10: “(For these difficult to judge) I still chose to favour China at last, because I felt like it might be an outcome of some cultural collisions.” This also led to a non-significant difference in the evaluation of atypical artefacts and typical in-groups. Exceptions appeared in aesthetic evaluation, where atypical artefacts are significantly in the middle, as well as symbolic and economic value, where atypical artefacts and out-group artefacts are rated more closely together. This may be because the symbolic and economic value of an artefact was underestimated when it had no obvious or conflicting cultural symbols. The deeper motivation may remain to be further verified through interviews.

When checking whether the above responses are statistically related to one’s degrees of in-group cultural identity, ethnic essentialism, or cultural exposure (RQ4), analysis showed that subjects with higher levels of in-group identification scales accepted typical in-group cultural artefacts more quickly. Castano et al. (2002) found that people who identified highly with the in-group had stricter entry criteria for the in-group and therefore categorised targets as out-group whenever possible. However, in this research, higher identifiers had lower RT to in-group artefacts, while there was no significant linear correlation for untypical and out-group artefacts. It shows that the in-group/out-group categorisation on artefacts may not be motivated by the over-exclusion. Also, people’s attitudes toward the in-group/out-group human faces did not differ significantly based on experimental groupings (RQ3). This raises the question of how, and to what extent, the perception of cultural artefacts and interpersonal interactions shape one another, or remain separate domains.

Third, personal and everyday meaning-making engages with the process of linking artefact visuals and national representations. The linkage between artefacts and nations may come from digital or non-digital cultural exposure, including textbooks, museums, news, films, etc. Among our screened materials for Study 2 (Supplementary Table A1), simple colour tones and shapes could activate in-group stereotypes, whereas the activation of out-group artefact images required stronger and more accurate signals, such as foreign human figures and realistic painting styles. From the perspective of SIT, this may be due to asymmetrical cultural exposure (Perrachione et al., 2010). Additionally, we would like to regard everyday in-group artefact visuals as fuzzy images of cultural memory and “unwaved flags”, a visible and external store distinguished from individual memories (Assmann, 2011), which continues in the digital space.

The statistical studies also found that subjects’ responses to in-group and out-group artefact visuals were related to exposure to the two-sided cultures. In-group news exposure may make subjects more cautious when confronted with out-group artefact categorisation, while cultural products may make subjects more cautious when confronted with untypical artefacts. Cross-cultural encounters may improve the evaluations of artefact visuals. It reflects that the daily consumption of news, cultural products or travels may form an experience of national representations that engages with the perception of the artefact visuals. This may be due to the (potentially conflicting) representations and meanings about heritage embedded in everyday cultural exposures (Longhurst et al., 2004; Waterton, 2009; Ichijo, 2017; Bonacchi, 2022).

Finally, to what extent does this in-group/out-group categorisation relate to a stereotype responding to certain ideological views or social values? This categorisation simplifies the visual information of cultural artefacts on multiple levels, from the pixel to the semantic levels. This dynamic of answers creates tension or harmony with the AHD. It may challenge the legitimacy of “official expertise”, but it may also rationalise the gap between popular and official heritage-making and be subordinate to national narratives. Specifically, our results show that people’s understandings of national culture were not simple imitations of professional discourses (Miller-Idriss, 2006). Subjects may use subjective, emotional, and perceptual criteria in the in-group/out-group categorisation, and popular culture, such as TV series and advertisements, can also engage with their judgements. However, the subjects’ self-definitions of “in-group artefacts” were still largely dominated by the AHD, and there is still a large gap in cultural capital between experts and lay audiences. While audiences’ private understanding can sometimes be seen as a challenge to the AHD, it also implicitly acknowledges and legitimates the professional heritage knowledge and its national symbolic values. For instance, some subjects unwittingly mentioned in the interviews that they were “not good at history”, and museums and history education were the basis for some subjects’ judgements. Once viewers consider themselves “uncultured” (Lu, 2014) and believe that heritage has no direct relevance to their current private lives, it is difficult for them to participate in heritage-making in the present. More importantly, simplified national representations in the artefacts cause potential difficulties in using heritage to represent sub-national social identities. If the national representations of artefacts are anchored in a single logic, sub-national identities and their associations with heritage visuals will be more easily overlooked.

This poses a challenge to the technological optimism in achieving digital accessibility in heritage work. Scanning museum collections into digital visuals and making them available to a wider audience or licensing them for more creative or commercial uses have become commonplace. However, choosing what to digitise, or the differences in messages an audience may get from a digital image, is an issue about social inclusion and diversity. The results show that audiences may categorise, interpret, or use them within the framework of the AHD and national identity-making. In doing so, the multiple features of heritage visuals are simplified, professional knowledge is not properly conveyed, and borders between groups are unconsciously shaped and reinforced. Inadvertently, the heritage digitisation is likely to be leveraged to realise a digital expansion of a tradition rather than a progressive vision it may aim to make (Witcomb, 2003; Henning, 2006); audiences will continue to think of heritage as something old, professional, national symbolic, and irrelevant to their sub-national memories, emotions, and participation in the present. This is not to deny the contributions of any existing heritage digitisation to cultural diversity, but to recognise a gap between simply digitising heritage and the vision it has been endowed with. Although digital heritage does become a political resource for constructing normative national identities, the subjects’ descriptions and feelings of in-group artefact visuals were also somewhat personal, spontaneous, and positive. They have the potential to be used to generate an understanding of diversity and belonging and empower individual/community agencies (Castelló and Mihelj, 2018). However, it cannot be done without going beyond a technical understanding of digital heritage, and being careful about “the play of ideological forces set in motion by processes of stereotyping” (Pickering, 1995, p. 692).

Limitations and directions for future research

There are some deficiencies in this research. First, this is a study on a specific group in a certain cultural context (mainly university students in China), limiting the generalisability of the findings. It lacks consideration of how other socio-cultural backgrounds and individual identities, including age, class, gender, ethnicity, educational background, and geographic location, may influence perceptions of in-group/out-group artefacts. Moreover, focusing only on the dimensions of national-level categorisation may result in the neglect of other social classifications (Desmond and Emirbayer, 2009; Smith, 2006). In future work, it is important to consider the diversity of what “in-group” means in different contexts, rather than only focusing on differences between nations.

Second, we as researchers also did “in-group/out-group categorisation” in our experimental design and analysis, which has cultural limitations. Although we aimed to construct an ideal psychological experimental material of “Chinese/European” artefacts that is perfectly balanced in date, material, shape, colour, and pattern, it was hard to reach this goal, and we may also be influenced by essentialist ideas. Our consideration of how a “diverse” and “balanced” experimental material of in-group/out-group artefacts might be limited by our cultural framework. For instance, our accumulated knowledge of the diversity of Chinese bronzes is greater than that of others. We also cannot distinguish which artefact features are indeed archaeologically unique to the Chinese/European cultural region and which are socially constructed. If a research team from a different cultural background follows our proposed methodology, it is possible that their selection of artefact visuals and coding of data would be very different. Therefore, engaging with different cultural contexts regarding both research materials, subjects, and even researchers themselves will yield more insights into this mixed-methods exploration on national identity perceptions.

Third, conflicts in some concepts between qualitative and quantitative research remain to be discussed. For instance, we used the term “identity” in a fragmented way because it can carry different meanings in studies with positivist and interpretive traditions. In the former, identity can be measured through indicators like differential treatment of in-groups/out-groups or a Likert scale. However, in interpretive studies, it is more revealed by in-depth interviews, textual analyses, etc. The same applies to concepts including “essentialism” and “cultural exposure”, where we acknowledge that this article has to oscillate between a concept as a measurable variable and the same concept in historical and discursive contexts. This results in CHS and SIT combining in a less-than-satisfactory manner, and leads to a reflection on the quantitative approach in our mixed-method design. Whilst expanding the sample size of artefacts could be one possible future of this research, before moving on to future research, we need to stop and think about whether the measurement-based approach of this paper is an alternative, a complement, or even a “mistake”—to interpretive heritage studies, which focus upon contextualised samples of heritage, communities and their cultural backgrounds.

Conclusion

This paper explores mixed methods to study Chinese young people’s (mainly university students) preferences for in-group/out-group categorisation and related evaluations of the cultural artefact digital pictures, illustrating that digital heritage visuals may be banally read by lay audiences as national representations in the absence of external interpretation. The results show a relationship between young people’s visual engagement with digital artefacts and contemporary national identity construction under the potential influence of an authorised heritage discourse. This raises reflections on the social consequences of everyday heritage digitisation practices, i.e., to what extent the heritage visualisation and accessibility have promoted innovation or just maintained traditional narratives and social relations.