Normal redefined: Exploring decontextualization of lorises (Nycticebus & Xanthonycticebus spp.) on social media platforms

1 Introduction

Prolific and unsustainable trade of wildlife is a major threat to biodiversity across the globe (Maxwell et al., 2016; Cardoso et al., 2021), and emerging and established platforms on the internet have become conduits for the illegal trade of numerous species (Milner-Gulland, 2018). With the advent of web 2.0 technologies, many social media sites have arisen. Social media platforms (SMPs) have been identified to facilitate and directly stimulate the illegal wildlife trade and promote the keeping of non-human animals as pets (Nekaris et al., 2013; IFAW, 2018). The emerging discipline of critical animal media study has established that our mistreatment of nonhuman animals is due to public consent, and we need to understand why the public still supports this content in the media and SMPs (Almiron et al, 2018). Here, we examine illegal wildlife trade within the context of SMPs by examining online imagery of a group of legally protected species.

There is a growing trend in the media of displaying nonhuman organisms within an anthropogenic framework (Malamud, 2010). Displaying exotic pets in the media has been shown to promote positive perceptions of their exploitation and increase their demand, directly fuelling the wildlife trade in exotic animals (Sollund, 2011; Moloney et al., 2021). The proliferation of images and videos of animals in anthropogenic settings has been shown to be detrimental to the conservation of nonhuman animals (Schroepfer et al., 2011). For example, when primates were displayed in both natural and anthropogenic settings it was shown that images of primates in anthropogenic settings, and with humans, distorted public perceptions of their abundance and suitability as pets (Ross et al., 2011; Schroepfer et al., 2011; Leighty et al., 2015). Similarly, even when images of exploited chimpanzees (Pan spp.) are paired with overt explanations of the issues of such treatment, children remember the exploitation as a normal part of our interactions with chimpanzees (Bettinger et al., 2010). These findings have significant implications for primates that are popular on SMPs, such as slow and pygmy lorises (Nekaris et al., 2013; Nekaris et al., 2015).

Decontextualization is a term originally from psychology (Etheredge et al, 2021). Most often, the term is used to describe ‘decontextualized language’, which relates to language learning in children, where new word definitions are learned outside of the here-and-now. The communicator typically assumes that the receiver knows as little as possible and is therefore explicit in how they explain the meaning (Gee, 2014). The receiver has to create meaning from the words alone. Decontextualized language is used to discuss abstract concepts as well as entities that are not present. The ability to create and use decontextualized language is a crucial skill in attaining literacy (Peterson and Mccabe, 1994; Chatpongcharoen et al., 2021) and is linked to proficiency with academic language later in development (Uccelli et al., 2018). Decontextualized images have been used within an environmental communication setting precisely for their flexibility of interpretation. The ambiguity of generic images has been used to support communication in everything from educational documentaries to greenwashing, the “intentional action to mislead or deceive consumers with false claims on organization environmental posture” (Andreoli et al., 2017; Hansen, 2017).

Within our context, Malamud (2010) refers to the removal of wildlife from its proper ecological context as ‘decontextualization’ and for our work we will use his definition. However, we build upon their work by using a new methodological approach and research context. While Malamud explored decontextualization with qualitative methods within the context of traditional media, we assess the presence of decontextualization with a quantitative approach within the setting of social media. Nekaris et al., 2015 created a foundation for how to measure decontextualization quantitatively when they examined videos of lorises (Nycticebus spp., Xanthonycticebus spp.) on social media sites to determine whether uploaders violated the ‘five freedoms’ of animal welfare. Four of the conditions defined in their work not only quantified welfare but also measured decontextualization, (human/non-conspecific contact, presence of daylight, signs of stress or ill health, and unnatural conditions). Thus, we use these four established conditions, and an additional condition described below, to assess decontextualization across social SMPs.

Slow lorises are small nocturnal primates from Asia that live a strictly arboreal lifestyle and have a diet specialised for exudates and live prey (Streicher et al, 2012; Nekaris and Starr, 2015). Due to these highly specialised dietary and habitat needs, slow lorises have not thrived in even the best captive settings, where a diet of no to extremely low amounts of fruit is recommended (Fuller et al., 2013; Cabana et al., 2019). Slow lorises possess many characteristic features that are considered ‘cute’, such as a large head, round face, and large neotenous eyes all of which induce caregiving behaviour in humans and are known as ‘baby schema’ (Glocker et al., 2009; Estren, 2012). It may be this cuteness that leads to the high demand for individuals in the illegal wildlife pet trade. Indeed, all species of slow loris are listed by the IUCN Red List as either Vulnerable, Endangered, or Critically Endangered. Extensive wildlife trade led CITES to ban the commercial trade of lorises in 2007 (Nekaris and Nijman, 2007) yet they remain heavily exploited in illegal trade for pets (Nijman et al., 2017), traditional medicines and use as photo props by tourists (Osterberg and Nekaris, 2015). Additionally, pet and photo prop slow lorises frequently display signs of stress including arm displays, stress faces, loss of hair and excessive venom production when in close proximity to humans (Nekaris et al., 2015). Being the only venomous primates, their teeth are also extracted to prevent their venomous bite and facial abscesses may be a sign of this in online imagery. A combined lack of success in captive breeding programmes coupled with restricted legal trade of slow lorises indicate that a majority of slow lorises on SMPs are illegally obtained from the wild (Nekaris et al., 2013; Musing et al., 2015). A 2015 study by Nekaris et al., examined 100 YouTube videos of slow lorises to determine whether they violated the ‘five freedoms’ of animal welfare and they additionally assessed whether the conditions contributed to likes or views. They found that approximately 31% of videos contained all five violations and unnatural environmental conditions (91%) and daylight (87%) were the most prominent conditions. Additionally, viewers liked videos more when slow lorises were displayed in daylight or when they showed signs of stress or ill health.

For the first time, we compare imagery of a threatened species across three SMPs, YouTube, TikTok, and Giphy, which have differing outreach/entertainment aims. In our case study, we examine decontextualization of lorises on each platform by breaking it down into individual conditions, based upon the categories from Nekaris et al., 2015. The main demographic difference between the SMPs appears to be the age of users, with a vast majority of TikTok and Giphy’s users being between 16-24 though they also reach into the 25-44 age range, while YouTube appears to cover nearly all age ranges relatively equally (Insider Intelligence, 2022; Similar Web, 2022). Our null hypothesis predicts that decontextualization will not differ across our three platforms. Although YouTube, TikTok, and Giphy have different purposes and formats we expect that decontextualization scores would be variable generally based on the content length and purpose. For example, Giphy only allows for brief content that is often just a few frames and the GIFs from the platform are mainly used as a tool to emote on other SMPs. For this content, we anticipate that most context cues will not be available to viewers, and the content will be the most decontextualized. In contrast, YouTube videos are longer, providing more opportunity for context, and can be entertaining or educational in nature. We therefore predict that there will be fewer elements of decontextualization (see Supplementary Material). We also expect that TikTok will fall somewhere in the middle as a short form video sharing platform that can host entertainment and educational content. Based upon the prior literature, we predict that lorises in ill health and that are featured in daylight will have the highest numbers of views and likes amongst the conditions of decontextualization (Nekaris et al., 2015). We consider our results in the context of the social distortion hypothesis of Ross et al., 2011.

2 Methods

We conducted surveys of the top most viewed 100 relevant videos of slow lorises across three sites: YouTube (n=40), TikTok (n=40), and Giphy (n=20)- for descriptions and functionality see Supplementary Table 1. We selected YouTube because it is one of the most popular SMPs and we chose to survey Giphy because it is integrated into some of the most popular SMPs: Facebook, Instagram, Snapchat, and TikTok (Statista, 2021). Only 20 videos could be collected from Giphy due to the prevalence of a smaller range of slow loris images. Lastly, we surveyed TikTok because of its growing relevance amongst popular SMPs (Statista, 2021). Across all sites, we searched the phrase “slow loris” and only noted videos with lorises. The first thirty posts were coded by two independent researchers who were trained in the method. As the two coders had a 100% match in their coding, we did not see the need to run a reliability test, and the first author continued with the remaining coding. Each video was analysed in situ and the five main conditions for decontextualization were recorded, including human/non-conspecific contact, daylight, signs of stress or ill health, unnatural conditions (Nekaris et al., 2015) as well as a novel measurement of anthropogenic context (Table 1). We recorded the five conditions for each video as either present (1) or absent (0) and combined these measurements to create a decontextualization score (DS) for each video, with 0 indicating no presence of the five conditions up to 5 indicating the presence of all conditions (See Table 1 for coding definitions). Additionally, the number of views and likes were recorded for the posts on YouTube and TikTok to measure viewer engagement. These data were not able to be recorded for the Giphy images as this measurement is not provided by that SMP.

TABLE 1

Table 1 Description of each of the five conditions for decontextualization that we coded in 100 slow loris videos across three social media platforms.

2.1 Data analysis

We ran a Spearman non-parametric test to determine the correlation between decontextualization score and the number of likes/views. We then ran Generalised Linear Mixed Models (GLMMs) to understand the influence of each decontextualization condition (i.e., anthropogenic context, daylight, human non conspecific contact, signs of stress or ill health, unnatural conditions) separately on the number of likes and views on TikTok and YouTube videos. We used the SMP and user ID as random effects in the analysis. We used the length of time that the video had been available (as a log of days) as offset in the analysis. We ran all analyses with R v 4.1.0. We ran GLMs via the “glmmTMB” function in the “glmmTMB” package as this function includes several fit families that are suitable to deal with count distributions (Brooks et al., 2017). We tested the different families present in the package and selected the model based on the QQ plot residuals and residual vs predicted plot from the package “DHARMa”. We calculated the pseudo R-squared via the “pR2” function from the package “pscl”. We considered p = 0.05 as level of significance.

To visualise how decontextualization factors were combined within videos we ran Non-Metric Dimensional Scaling (NMDS) with presence/absence of each element for each video. We used the “metaMDS” function in the package “vegan” to run the NMDS. The NMDS ordination was plotted using the ggplot package with decontextualization factors indicated with labels and arrows and images indicated with points.

We then ran statistical analyses to determine if the elements of the decontextualization score varied between SMPs and which elements were more characteristic. We tested differences in the first two dimensions of the NMDS via Kruskal-Wallis tests and, if significant, we ran pairwise comparisons using Wilcoxon rank sum test and Bonferroni-Holm correction.

3 Results

Imagery ranged from ecologically correct portrayals to videos completely devoid of natural behaviours or settings as well as animated material (Figures 1, 2). Figure 1 provides examples of decontextualized imagery found on the sampled SMPs and illustrates what decontextualization score (DS) the videos would receive. These scores are based upon the number of abnormal conditions present in the loris imagery.

FIGURE 1

Figure 1 Examples of levels of decontextualization- copyrighted images that we did not have the rights to are illustrated. Each image is shown with a corresponding decontextualization score (DS) ranging from zero, indicating no elements of decontextualization to five, indicating the presence of all elements of decontextualization. (A) Represents an ecologically accurate image of a slow loris. (B) The inner ring shows a slightly decontextualized image for a conservation campaign logo. (C) The slow loris from an animated television show portrayed as a house pet represents the furthest level of decontextualization. (D) A wild slow loris is captured in a natural setting for a study and is pictured with a researcher (K. A. I. Nekaris). (E) An illustration of the pop singer Rihanna with a slow loris representing the viral photo of her with a photo prop loris. (F) A slow loris being handled for research showing a stress response of clinging. (G) A pet loris from a viral ‘tickle’ video in a defence pose.(H) A slow loris at a zoo eating a grasshopper with some vegetation around. (I) A slow loris from a viral ‘rice ball’ video holding unnatural food.

FIGURE 2

Figure 2 Radar chart representing the prevalence of the five decontextualization conditions for videos analysed on YouTube (N=40), TikTok (N=40), and Giphy (N=20), illustrating differences in decontextualization between the three SMPs.

Of the 100 videos analysed across the three platforms, the average decontextualization score was found to be 4.04 (YouTube- averaged 4.1, TikTok- averaged 3.75, Giphy- averaged 4.5) and 77% of all videos had four to five conditions of decontextualization present. The average length of the videos were as follows: YouTube- averaged 327.6 seconds, TikTok- averaged 27.2 seconds, and Giphy- averaged 2.3 seconds. Despite this considerable difference in duration, average decontextualization scores across the three SMPs showed no clear differences, with all three displaying high decontextualization scores and high prevalence of decontextualization conditions (Figure 2). Only two of the 100 videos had no elements of decontextualization: a YouTube video uploaded by Nat Geo WILD and a clip from the same video on Giphy and TikTok had no videos of fully contextualized lorises. Across all platforms, three videos had only one condition, five videos had two conditions, 13 had three conditions, 33 had four conditions, and 44 contained all five conditions of decontextualization. The most prevalent unnatural conditions across all videos were daylight (93%) and ‘anthropogenic context’ (91%), followed by signs of stress or ill health (76%), human/non-conspecific contact (73%), and ‘unnatural conditions’ (71%). ‘Anthropogenic context’ was highly common across all platforms (Giphy- 95%, YouTube- 90%, TikTok- 90%) and daylight was similarly prevalent across all SMPs (Giphy- 95%, YouTube- 93%, TikTok- 93%). ‘Unnatural conditions’ were highest on Giphy (95%) followed by TikTok (78%) and in approximately half of YouTube videos (53%). Signs of stress or ill health were highest on Giphy (95%), then YouTube (83%), and lowest on TikTok (60%). Human/non-conspecific contact was most common in YouTube videos (93%), followed by Giphy (70%), and then TikTok (55%) (Figure 2).

YouTube videos averaged 1,215,813 views and 8,898 likes and TikTok averaged 586,874 views and 190,098 likes. We found a significant correlation between decontextualization scores and number of views, but not between decontextualization scores and number of likes (Figure 3). When conditions were analysed separately, the number of views on TikTok and YouTube videos were significantly higher when videos presented lorises with signs of stress or ill health and in unnatural conditions (Table 2; Figure 4). The number of likes on TikTok and YouTube videos were significantly higher when lorises in videos were in an anthropogenic context.

FIGURE 3

Figure 3 Spearman correlation plots representing the non-significant correlation between number of likes of loris videos on three social media platforms and decontextualization score (A), and the significant positive correlation between number of views and decontextualization score (B).

TABLE 2

Table 2 Results of the Generalised Linear Models to understand the influence of each parameter used to calculate the decontextualization score on the number of likes and views on TikTok and YouTube videos. Reference categories for predictors: “Yes”.

FIGURE 4

Figure 4 Violin plots with jittered data points and quartiles (1^st, median, 3^rd) representing the predicted model values for significant predictors influencing the number of likes and views on TikTok and YouTube videos. Videos with anthropogenic context had more likes (A), videos with signs of stress or ill health (B) and unnatural conditions (C) had more views.

All elements of the decontextualization score were important in defining differences between videos, with human/non-conspecific contact and unnatural conditions having the highest r-squared (NMDS: stress = 0.08; Table 3; Figure 5). Videos on TikTok and YouTube had more diverse elements of the decontextualization score than videos on Giphy (Figure 5). NMDS1, mainly characterised by higher frequency of daylight and lower frequency of human/non-conspecific contact or signs of stress or ill health, was significantly different between SMPs (Kruskal-Wallis: χ 2 = 19.25, p<0.001). Pairwise comparisons using Wilcoxon rank sum test and Bonferroni-Holm correction revealed that YouTube had a lower NMDS1 than Giphy (p<0.001) and TikTok (p<0.001), while no differences were found between Giphy and TikTok (p=0.477). NMDS2, mainly characterised by unnatural conditions and anthropogenic context, was significantly different between SMPs (Kruskal-Wallis: χ 2 = 12.89, p=0.002), with YouTube having a lower NMDS2 that Giphy (p- 0.004) and TikTok (p=0.005), while no differences were found between Giphy and TikTok (p=0.447).

TABLE 3

Table 3 First and second axes of the Non-Metric Dimensional Scaling and relative importance of each parameter.

FIGURE 5

Figure 5 Non-Metric Dimensional Scaling representing the elements of the decontextualization score in videos from Giphy, TikTok and YouTube.

4 Discussion

We found that decontextualization occurred at high levels on all three platforms we monitored. Our results also support previous research that showed that such decontextualization levels are linked to higher levels of viewer interaction through views and likes (Nekaris et al., 2015; Kitson and Nekaris, 2020). In 2020, Kitson and Nekaris conducted a study in Turkey focused on two-shot images, posted to the popular SMP Instagram, of tourists with photo-prop lorises. The study found that, regardless of the attractiveness of a photo (lighting, focus, rule of thirds, and visibility), images of people with lorises gained significantly more likes than general photos. This lack of context for imagery hinders understanding and leaves interpretation open to the viewer (Peterson and Mccabe, 1994; Hansen, 2017). Due to the flexible interpretation of natural imagery, decontextualized images have infamously been used to support greenwashing campaigns (Hansen, 2017). Prior research has established that displaying images of exotic animals in the media, especially in anthropogenic settings has negative consequences for their conservation and increases their desirability as a pet (Schroepfer et al., 2011; Sollund, 2011). The exploitation of slow lorises despite their protected status is prolific and stimulated by the advertising that happens on SMPs (Nekaris et al., 2013; Musing et al., 2015; Osterberg and Nekaris, 2015; Nijman et al., 2017; IFAW, 2018). Although some videos of slow lorises can potentially increase public understanding about certain aspects of slow loris conservation, a proper representation of the natural state and behaviour of slow lorises is crucial for creating positive effects.

The 100 videos of lorises that we analysed overwhelmingly displayed individuals in inappropriate settings. Decontextualization scores were not affected by the SMP’s content type or general length as we predicted. Instead, our results support the null hypothesis that there is no difference and indeed all platforms showed equivalently high average decontextualization. Furthermore, it was not just private content producers decontextualizing lorises. Wild ecology was lost even in content produced by educational organisations such as Animal Planet (DS=4), Nat Geo Wild (DS=4), Seeker (DS=3), the Brookfield Zoo (DS=4), and the San Diego Zoo (DS=5). The videos across these five different educational institutions included all five conditions of decontextualization. The ubiquity of decontextualized content shown in these data have the potential to contribute to the normalisation of an image of a loris outside its ecological context, normalising its presence in the hands of humans (Malamud, 2010; Moloney et al., 2021).

There was a clear preference for videos that displayed lorises in anthropogenic contexts. This is illustrated by the fact that, on YouTube and TikTok, videos showing lorises in anthropogenic contexts had more likes than videos with natural settings. Moreover, videos with signs of stress or ill health and unnatural conditions also had more views than videos portraying healthy lorises in natural conditions. As we anticipated, these results are in line with previous findings on YouTube where portrayals of lorises in ill health and unnatural conditions had higher numbers of views and likes (Nekaris et al., 2015). Our findings are correlational, making further study necessary to fully understand the variables responsible for any causational relationship. Regardless, the further that popular images move away from ecological reality and the more human-centric they become, the easier it is to see animals as objects and accept images of them as unhealthy pets as ‘normal’ (Malamud, 2010; Moloney et al., 2021).

Figures that have historically helped to push the normalisation of lorises in contact with people were celebrities (Nekaris et al., 2013). Giphy bears the signs of this through animated GIFs of the pop singer Rihanna with a slow loris on her head (Figure 1). This is a direct reference to when the singer took a selfie with a slow loris that was illegally used as a photo prop on the street in Thailand. Famous individuals such as Ariana Grande, Haley Williams, Ricky Gervais, and more have actively directed their audiences to decontextualized loris content on YouTube in the past, such as the infamous tickle video, where a large portion of the viewers declared their desires and intentions to get one as a pet (Nekaris et al., 2013). Since its popularisation, it appears that decontextualized loris content now has the power to create fame and influence. Just over half of the most relevant videos on TikTok were uploaded by a single user, @yojenka, and displayed slow lorises as house pets. This content creator was quite influential, with over 200,000 followers and upwards of five million likes on her videos, manufacturing a type of celebrity with this decontextualized content and implicitly pushing the message that this unnatural condition is normal. Although most of the content on TikTok was produced by a single user, analysis of the conditions of decontextualization showed that all videos on the platform had diverse elements of decontextualization. This illustrates that the decontextualized content is what is truly popular, not just particular users and this tracks with previous findings (Nekaris et al., 2013; Kitson and Nekaris, 2020). Banning a single user will not stem the problem as creators can simply make another account or their place in the spotlight will be taken by a creator who is producing the same content. Mitigation of this problem needs to come from a higher level of content monitoring and reporting functionality from SMPs to combat this pervasive content. A case in point here is the 2009 ‘tickling slow loris’ video mentioned above. Although campaigning meant that the original uploader deleted it, 1000s of copies remain, and were seen on all of the platforms in this study. There is a potential that restricting or banning content will ultimately lead people to seek it out (Jansen and Martin, 2015; Ohlheiser, 2020). Educational campaigns are needed to place slow lorises within accurate context to enhance ecological knowledge (Bergman et al., 2022), due to the striking lack of properly contextualized content even from educational institutions.

This decontextualization was brought to an extreme on 31 May 2021 when an animated television show entitled Housebroken was released by the Fox Broadcasting Company. The show follows the lives of anthropomorphised household pets; the cast of characters includes domestic animals such as dogs, cats, as well as a guinea pig, a tortoise, a goldfish, and a pygmy loris. The pygmy loris is named Tchotchke, defined by the Cambridge English Dictionary as a small decorative object, and he is the only non-speaking character. He communicates to the other animals only through body language and the use of a drink umbrella, which is part of his character design (for a representation similar to the character see Figure 1). This is a clear reference to the viral video of a pet pygmy loris, shown holding a drink umbrella (Nekaris et al., 2013). It has been demonstrated that YouTube content portraying prosimians as pets leads to viewers desiring them as pets and even proclaiming to take steps to acquire one (Nekaris et al., 2013; Clarke et al., 2019). This television programme may reinforce the social acceptability of owning lorises as pets and may contribute to the view that slow lorises are good pets as it portrays the character as quiet, tame, and friendly. The fact that the only exotic animal included was a pygmy loris highlights the extent that decontextualization can normalise pet keeping. This normalisation originates from the social proof phenomenon in which people copy the actions of others because they presume that an action is acceptable if other people are doing it. On SMPs and in the media, widespread imagery with views, likes, and acclaim heavily influence the perceptions of viewers (Schnuerch and Gibbons, 2015; Fong et al., 2020).

Slow and pygmy lorises represent a model group for understanding decontextualization due to their prolific exploitation (Nekaris et al., 2015; Osterberg and Nekaris, 2015; Nijman et al., 2017; Kitson and Nekaris, 2020). Our findings indicate a need to explore how other species are decontextualized on SMPs. There are numerous species that are similarly portrayed as pets and shown in non-natural and often anthropogenic contexts on SMPs- such as meerkats (Suricata suricatta), kinkajous (Potos flavus), ring-tailed lemurs (Lemur catta), squirrel monkeys (Saimiri spp.), and capuchins (Cebus spp. & Sapajus spp.) (Vázquez et al., 2016). In particular, sloths (Bradypus spp. & Choloepus spp.) face equivalent threats to slow lorises and have a high demand in the illegal pet trade and photo prop industry (Moreno and Plese, 2006; Carder et al., 2018). Sloths are nocturnal, slow-moving mammals that have a large round face with a permanent smile that makes them appear ‘cute’ in the same way as the slow loris (Glocker et al., 2009; Estren, 2012). Sloths also endure clipping of their claws and their teeth to make them easier to handle (Carder et al., 2018), which is a parallel to the tooth extractions in the slow loris trade (Nekaris et al., 2010). Images of sloths are prominent in the media (Lenzi et al., 2020) and they are seemingly decontextualized in the same manner as the slow loris, which was particularly heightened after celebrity endorsement from actress Kristen Bell when she appeared on the Ellen Show, filmed in the USA, in 2012 describing how a live sloth was used as entertainment at her birthday party. This indicates a pattern of decontextualization and validation that should be explored in further research on additional taxa. Future studies should also focus on how the diverse conditions of decontextualization impact public attitudes towards the animals on display, their trade as pets, and the viewer’s perceptions of what behaviours are natural for species impacted by decontextualization (c.f., Ross et al., 2011; Schroepfer et al., 2011).

These data show that overall, YouTube, TikTok, and Giphy need more properly contextualized content that shows lorises in natural scenes, specifically in nocturnal and ecologically appropriate settings (e.g., in the dark interacting with natural substrates). Content producers should also focus on only portraying lorises in contact with conspecifics and portraying healthy lorises that are not exhibiting stress-related behaviours or ill health. Where Giphy needs sweeping content reforms for decontextualized conditions, YouTube needs particular focus on anthropogenic context, daylight, human/non-conspecific contact, and signs of stress or ill health. TikTok need to focus changes mostly on anthropogenic context, daylight, and unnatural conditions. With regard to content moderation, SMPs should attempt to make connections with established loris conservation organizations to create guidelines for filtering out inappropriate loris content.

Our findings indicate that there is a need for an examination of the animal-based content present on SMPs. Social media has great potential to connect users with wildlife in ways that promote appropriate concern and understanding (Pimentel, 2022). Within a zoo context, images of animals shown with keepers have been demonstrated to increase viewers’ willingness to donate to conservation; yet our findings showed zoos also contributed to the decontextualization of animal imagery (Spooner and Stride, 2021). Conservation and educational organisations should examine the ways that they portray non-human animals in their posts on SMPs. They should not only follow IUCN guidelines to avoid circulating images of humans in close proximity to primates (Waters et al., 2021), but conservation organisations should also be aware of the presence of decontextualization conditions in their content so that their messages are not misconstrued. Decontextualization is an issue that needs to be considered and addressed by the conservation community at large and research regarding why the public supports decontextualized animal imagery is critical for the future of conservation efforts.

This study was limited in the number of SMPs that were included, and the volume of data analysed. Future studies of decontextualization should include, and directly sample from, other popular SMPs such as Instagram and Facebook. It would have been valuable to get a sample of all the content relevant to slow lorises instead of the most popular or relevant content. Future research may additionally focus on how slow and pygmy lorises are specifically portrayed by science education and zoological organisations as these may be seen by audiences to be endorsing or approving decontextualizated posts through their own social media presence.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

Ethical review and approval was not required for the animal study because this study was based completely online and focused on analysing content created for web 2.0 sites by the public as well as private institutions. We examined how videos portrayed a specific endangered animal. We did not perform the study with live animal subjects.

Author contributions

KN and LQ contributed to conception and design of the study. MC performed the statistical analyses, created most figures, and wrote up results with LQ. LQ wrote the initial manuscript draft and KN and KF reviewed and edited the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

Funding

This work was part of the Little Fireface Project’s wildlife trade monitoring, funding by People’s Trust for Endangered Species and Cleveland Zoo and Zoo Society.

Acknowledgments

Special thanks to Vincent Nijman and Abigail Hines for giving feedback on the content of this article. We thank Wawan Tarniwan and Andrew Walmsley for photographs of wild slow lorises.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcosc.2023.1067355/full#supplementary-material

References

Almiron N., Cole N., Freeman M. (2018). Critical animal and media studies: Expanding the understanding of oppression in communication research. Eur. J. Commun. 33, 367–380. doi: 10.1177/0267323118763937

CrossRef Full Text | Google Scholar

Andreoli T. P., Crespo A., Minciotti S. (2017). What has been (short) written about greenwashing: A bibliometric research and a critical analysis of the articles found regarding this theme. Rev. Gestao Soc 11, 54–72. doi: 10.24857/rgsa.v11i2.1294

CrossRef Full Text | Google Scholar

Bergman J. N., Buxton R. T., Lin H.-Y., Lenda M., Attinello K., Hajdasz A. C., et al. (2022). Evaluating the benefits and risks of social media for wildlife conservation. FACETS 7, 360–397. doi: 10.1139/facets2021-0112

CrossRef Full Text | Google Scholar

Bettinger T. L., Kuhar C. W., Lehnhardt K., Cox D., Cress D. (2010). Discovering the unexpected: Lessons learned from evaluating conservation education programs in Africa. Am. J. Primatol. 72, 445–449. doi: 10.1002/ajp.20735

PubMed Abstract | CrossRef Full Text | Google Scholar

Brooks M. E., Kristensen K., van Benthem K. J., Magnusson A., Berg C. W., Nielsen A., et al. (2017). glmmTMB balances speed and flexibility among packages for zero‐inflated Generalized Linear Mixed Modeling. R J. 9, 378–400. doi: 10.3929/ethz-b-000240890

CrossRef Full Text | Google Scholar

Cabana F., Clayton J. B., Nekaris K. A. I., Wirdateti W., Knights D., Seedorf H. (2019). Nutrient-based diet modifications impact on the gut microbiome of the javan slow loris (Nycticebus javanicus). Sci. Rep. 9, 1–11. doi: 10.1038/s41598-019-40911-0

PubMed Abstract | CrossRef Full Text | Google Scholar

Carder G., Plese T., Machado F. C., Paterson S., Matthews N., McAnea L., et al. (2018). The impact of ‘selfie’ tourism on the behaviour and welfare of brown-throated three-toed sloths. Animals 8, 216. doi: 10.3390/ani8110216

PubMed Abstract | CrossRef Full Text | Google Scholar

Cardoso P., Amponsah-Mensah K., Barreiros J. P., Bouhuys J., Cheung H., Davies, et al. (2021). Scientists’ warning to humanity on illegal or unsustainable wildlife trade. Biol. Conserv. 263, 109341. doi: 10.1016/J.BIOCON.2021.109341

CrossRef Full Text | Google Scholar

Chatpongcharoen P., Campera M., Laithong P., Gibson N. L., Nekaris K. A. I. (2021). Naturalising diet to reduce stereotypic behaviours in slow lorises rescued from wildlife trade. Appl. Anim. Behav. Sci. 242, 105413. doi: 10.1016/j.applanim.2021.105413

CrossRef Full Text | Google Scholar

Clarke T. A., Reuter K. E., LaFleur M., Schaefer M. S. (2019). A viral video and pet lemurs on twitter. PloS One . 14 (1), e0208577. doi: 10.1371/journal.pone.0208577

PubMed Abstract | CrossRef Full Text | Google Scholar

Estren M. J. (2012). The neoteny barrier: Seeking respect for the non-cute. J. Anim. Ethics 2, 6–11. doi: 10.5406/janimalethics.2.1.0006

CrossRef Full Text | Google Scholar

Etheredge R. I., Schartl M., Jordan A. (2021). Decontextualized learning for interpretable hierarchical representations of visual patterns. Patterns 2, 100193. doi: 10.1016/j.patter.2020.100193

PubMed Abstract | CrossRef Full Text | Google Scholar

Fitch-Snyder H., Schulze H., Larson L. (2001). “Management of lorises in captivity,” in A husbandry manual for Asian lorisines (Nycticebus & loris ssp.) (San Diego, CA: Zoological Society of San Diego).

Google Scholar

Fong K. C. H., Au C. H., Lam E. T. H., Chiu D. K. (2020). Social network services for academic libraries: A study based on social capital and social proof. J. Acad. Librariansh. 46, 102091. doi: 10.1016/j.acalib.2019.102091

CrossRef Full Text | Google Scholar

Fuller G., Kuhar C. W., Dennis P. M., Lukas K. E. (2013). A survey of husbandry practices for lorisid primates in north American zoos and related facilities. Zoo Biol. 32, 88–100. doi: 10.1002/zoo.21049

PubMed Abstract | CrossRef Full Text | Google Scholar

Gee J. P. (2014). Decontextualized language: A problem, not a solution. Int. Multiling. Res. J. 8, 9–23. doi: 10.1080/19313152.2014.852424

CrossRef Full Text | Google Scholar

Glocker M. L., Langleben D. D., Ruparel K., Loughead J. W., Gur R. C., Sachser N. (2009). Baby schema in infant faces induces cuteness perception and motivation for caretaking in adults. Ethol. 115, 257–263. doi: 10.1111/j.1439-0310.2008.01603.x

CrossRef Full Text | Google Scholar

Hansen A. (2017). “Using visual images to show environmental problems,” in The routledge handbook of ecolinguistics. Eds. Fill A. F., Penz H. (New York: Routledge), 179–195.

Google Scholar

IFAW (2018). Disrupt: wildlife cybercrime-uncovering the scale of online wildlife trade (London: IFAW).

Google Scholar

Insider Intelligence (2022) TikTok is eating into YouTube’s dominance–here’s what marketers should know. Available at: https://www.insiderintelligence.com/content/tiktok-youtube-marketers (Accessed January 15, 2023).

Google Scholar

Jansen S., Martin B. (2015). The Streisand effect and censorship backfire. Int. J. Commun. 9, 656–671.

Google Scholar

Kitson H., Nekaris K. A. I. (2020). “Slow lorises (Nycticebus spp.) as photo props on Instagram,” In Evolution, ecology and conservation of lorises and pottos Eds. Nekaris K. A. I., Burrows A. M. (Cambridge: Cambridge University Press) 374–380.

Google Scholar

Leighty K. A., Valuska A. J., Grand A. P., Bettinger T. L., Mellen J. D., Ross S. R., et al. (2015). Impact of visual context on public perceptions of non-human primate performers. PloS One 10, e0118487. doi: 10.1371/journal.pone.0118487

PubMed Abstract | CrossRef Full Text | Google Scholar

Lenzi C., Speiran S., Grasso C. (2020). “Let me take a selfie”: Implications of social media for public perceptions of wild animals. Soc Anim. 1, 1–20. doi: 10.1163/15685306-BJA10023

CrossRef Full Text | Google Scholar

Malamud R. (2010). Animals on film: The ethics of the human gaze. Spring 83, 1–26.

Google Scholar

Maxwell S. L., Fuller R. A., Brooks T. M., Watson J. E. M. (2016). Biodiversity: The ravages of guns, nets and bulldozers. Nature 536, 143–145. doi: 10.1038/536143a

PubMed Abstract | CrossRef Full Text | Google Scholar

Milner-Gulland E. J. (2018). Documenting and tackling the illegal wildlife trade: Change and continuity over 40 years. Oryx 52, 597–598. doi: 10.1017/S0030605318001047

CrossRef Full Text | Google Scholar

Moloney G. K., Tuke J., Dal Grande E., Nielsen T., Chaber A. L. (2021). Is YouTube promoting the exotic pet trade? analysis of the global public perception of popular YouTube videos featuring threatened exotic animals. PloS One 16, e0235451. doi: 10.1371/journal.pone.0235451.t003

PubMed Abstract | CrossRef Full Text | Google Scholar

Moreno S., Plese T. (2006). The illegal traffic in sloths and threats to their survival in Colombia. Edentata 2006, 10–18. doi: 10.1896/1413-4411.7.1.10

CrossRef Full Text | Google Scholar

Musing L., Suzuki K., Nekaris K. A. I. (2015). Crossing international borders: The trade of slow lorises Nycticebus spp. as pets in Japan. Asian Primates J. 5, 12–24.

Google Scholar

Nekaris K. A. I., Campbell N., Coggins T. G., Rode E. J., Nijman V. (2013). Tickled to death: analysing public perceptions of ‘cute’ videos of threatened species (slow lorises–Nycticebus spp.) on web 2.0 sites. PloS One 8, e69215. doi: 10.1371/journal.pone.0069215

PubMed Abstract | CrossRef Full Text | Google Scholar

Nekaris K. A. I., Musing L., Gil Vazquez A., Donati G. (2015). Is tickling torture? assessing welfare towards slow lorises (Nycticebus spp.) within web 2.0 videos. Folia Primatol 86, 534–551. doi: 10.1159/000444231

CrossRef Full Text | Google Scholar

Nekaris K. A. I., Nijman V. (2007). CITES proposal highlights rarity of Asian nocturnal primates (Lorisidae: Nycticebus). Folia Primatol 78, 211–214. doi: 10.1159/000102316

CrossRef Full Text | Google Scholar

Nekaris K. A. I., Shepherd C. R., Starr C. R., Nijman V. (2010). Exploring cultural drivers for wildlife trade via an ethnoprimatological approach: A case study of slender and slow lorises (Loris and nycticebus) in south and southeast Asia. Am. J. Primatol. 72, 877–886. doi: 10.1002/ajp.20842

PubMed Abstract | CrossRef Full Text | Google Scholar

Nekaris K. A. I., Starr C. R. (2015). Conservation and ecology of the neglected slow loris: priorities and prospects. Endanger. Species Res. 28, 87–95. doi: 10.3354/esr00674

CrossRef Full Text | Google Scholar

Nijman V., Spaan D., Rode-Margono J., Nekaris K. A. I. (2017). Changes in the primate trade in Indonesian wildlife markets over a 25-year period: Fewer apes and langurs, more macaques, and slow lorises. Am. J. Primatol 79, e22517. doi: 10.1002/ajp.22517

CrossRef Full Text | Google Scholar

Ohlheiser A. (2020)Twitter’s ban almost doubled attention for biden story - the social media company’s attempt to stop misinformation from spreading brought the Streisand effect into action. In: MIT Technology review. Available at: https://www.technologyreview.com/2020/10/16/1010644/twitter-ban-hunter-biden-emails-backfires/ (Accessed January 15, 2023).

Google Scholar

Osterberg P., Nekaris K. A. I. (2015). The use of animals as photo props to attract tourists in Thailand: a case study of the slow loris Nycticebus spp. TRAFFIC Bull. 27, 13–18.

Google Scholar

Peterson C., Mccabe A. (1994). A social interactionist account of developing decontextualized narrative skill. Dev. Psychol. 30, 937–948. doi: 10.1037/0012-1649.30.6.937

CrossRef Full Text | Google Scholar

Pimentel D. (2022). Saving species in a snap: On the feasibility and efficacy of augmented reality-based wildlife interactions for conservation. Nat. Conserv. 66, 126151. doi: 10.1016/j.jnc.2022.126151

CrossRef Full Text | Google Scholar

Ross S. R., Vreeman V. M., Lonsdorf E. V. (2011). Specific image characteristics influence attitudes about chimpanzee conservation and use as pets. PloS One 6, e22050. doi: 10.1371/journal.pone.0022050

PubMed Abstract | CrossRef Full Text | Google Scholar

Schnuerch R., Gibbons H. (2015). Social proof in the human brain: electrophysiological signatures of agreement and disagreement with the majority. J. Psychophysiol. 52, 1328–1342. doi: 10.1111/psyp.12461

CrossRef Full Text | Google Scholar

Schroepfer K. K., Rosati A. G., Chartrand T., Hare B. (2011). Use of “entertainment” chimpanzees in commercials distorts public perception regarding their conservation status. PloS One 6, e26048. doi: 10.1371/journal.pone.0026048

PubMed Abstract | CrossRef Full Text | Google Scholar

Similar Web (2022) (Giphy.com). Available at: https://www.similarweb.com/website/giphy.com/#overview (Accessed January 15, 2023).

Google Scholar

Sollund R. (2011). Expressions of speciesism: the effects of keeping companion animals on animal abuse, animal trafficking and species decline. Crime Law Soc Change 55, 437–451. doi: 10.1007/s10611-011-9296-3

CrossRef Full Text | Google Scholar

Spooner S. L., Stride J. R. (2021). Animal-human two-shot images: Their out-of-context interpretation and the implications for zoo and conservation settings. Zoo Biol. 40, 563–574. doi: 10.1002/zoo.21636

PubMed Abstract | CrossRef Full Text | Google Scholar

Statista (2021) Most used social media (Statista.com). Available at: https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/ (Accessed January 02, 2022).

Google Scholar

Streicher U., Wilson A., Collins R. L., Nekaris K. A. I. (2012). “Exudates and animal prey characterize slow loris (Nycticebus pygmaeus, n. coucang and n. javanicus) diet in captivity and after release into the wild,” in Leaping ahead. Eds. Masters J., Gamba M., Génin F. (New York, NY: Springer), 165–172. doi: 10.1007/978-1-4614-4511-1_19

CrossRef Full Text | Google Scholar

Uccelli P., Demir-Lira Ö.E., Rowe M. L., Levine S., Goldin-Meadow S. (2018). Children’s early decontextualized talk predicts academic language proficiency in midadolescence. Child Dev. 90, 1650–1663. doi: 10.1111/cdev.13034

PubMed Abstract | CrossRef Full Text | Google Scholar

Vázquez A. G., Musing L., Nekaris K. A. I., Martínez I. J. (2016). El efecto de los vídeos de YouTube en la percepción de especies exóticas como mascotas potenciales. Chronica Naturae 6, 13–23.

Google Scholar

Waters S., Setchell J. M., Maréchal L., Oram F., Cheyne S. M. (2021). Best practice guidelines for responsible images of non-human primates. In: The IUCN primate specialist group section for human-primate interactions. Available at: http://eprints.lincoln.ac.uk/id/eprint/43949/1/HPI-Imagery-Guidelines.pdf (Accessed January 02, 2022).

Google Scholar