Invisible Animals: Exploring Public Discourses to Understand the Contemporary Status of Donkeys in Britain

ABSTRACT Established representations of donkeys in western literature and popular culture have often been negative, portraying the animals as stupid, inept, and bad tempered. To understand whether such representations are reflected in contemporary understandings of donkeys, we constructed a digitized body of texts (a corpus) from contemporary (public-facing) news articles and (public-produced) social media posts about donkeys, which we analyzed using quantitative and qualitative language analysis techniques drawn from corpus linguistics and discourse analysis. We supplemented this with focus groups conducted with members of the public to gather insights and reflections on key patterns of representation in two key news texts identified in the linguistic analysis. This combined approach, which is novel in its application to animal welfare topics, revealed that donkeys are most commonly represented as figures of entertainment, ridicule, and as victims of hardship and suffering (i.e., animals in need). We argue that such representations can sideline, obscure and “invisibilize” the real animals, leading to persistent misunderstandings and false conceptions about donkeys, which are further perpetuated through language use and can be difficult to disrupt and change. These findings have important implications for animal welfare, education programs, and public communication about donkeys. Finally, we highlight areas where incorrect, unhelpful, and potentially damaging representations can be challenged by those seeking to improve the contemporary status of donkeys.

Like many other animals, donkeys are frequently projected as signs and symbols in human society: imagined according to popular ideas and normative orderings of who they are, how they (should) behave, what relationship they (should) have with humans, and what role(s) they fulfill or ought to perform.As animal studies scholars have long maintained, the conceptual and material "placing" of animals is inextricably linked to how they are perceived and utilized in societyboth historically and in modern times (Philo & Wilbert, 2000;Urbanik, 2012;Wilcox & Rutherford, 2018;Wolch & Emel, 1998).To investigate how donkeys are understood and represented in contemporary British culture, including the various roles they occupy in different settings (production, working, therapy, and companion animals), we examined their representations in the language of news, social media, and focus-group discussions.Commissioned by The Donkey Sanctuary, a UK-based global charity dedicated to transforming the lives of donkeys and mules around the world, this research was designed to provide insights into the contemporary status of donkeys in Britain.
This paper argues that dominant representations of donkeys as figures of entertainment and ridicule, or as victims of hardship and suffering, can sideline, obscure, or "invisibilize" the real animal, leading to misunderstandings and false conceptions about donkeys, which are further perpetuated through repetition in language use.These findings have important implications for animal welfare, communication and advocacy for organizations involved in sharing public information, challenging negative stereotypes, and informing and promoting positive and accurate representations of donkeys.

Background Literature
Despite their vast contribution to human civilization (Mitchell, 2018), there has been limited engagement with the topic of donkeys in the humanities and social sciences. 1 A systematic review of journal articles containing the word "donkey" between 1896 and 2018 found that the majority were about donkey biology and physiology (McClean & Gonzalez, 2018). 2  Mitchell (2018) provides an important account of the historical uses of donkeys from an archeological perspective, but the most comprehensive cultural analysis comes from Jill Bough (2011a) in her book "Donkey."Here she discusses the variety of social, cultural, and religious meanings that have been attributed to donkeys, including in art and literature.She argues that donkeys have been "much maligned" (2011a, unpaginated) throughout history and "have many symbolic burdens to overcome before they can be understood in their own right" (Bough, 2011b, p. 61).This piece of work offered an important framing for our study on contemporary discourses about donkeys.In this section, we bring together what is already known about the cultural status of donkeys from various sources, which document social histories and portrayals of donkeys in popular and folk culture, before examining the role of language in shaping social perceptions of animals.
"Beast of Burden" -A Social History of Donkeys in Britain Donkeys have lived in Britain for nearly 2,000 years, following their introduction to England by the Romans in AD 43, though it was not until the eleventh century that their presence became more widespread (Mitchell, 2018).Since then, donkeys have been used to plow fields, carry people, transport goods, and support military efforts.During the eighteenth and nineteenth centuries, donkeys became an essential part of Britain's expanding industrial activities, used to deliver coal to houses and factories and carry miners to and from work in northern England.Despite their enormous contribution to society, the treatment of donkeys has been far from favorable.Historical evidence from the thirteenth century suggests that donkeys were poorly perceived and harshly treated at this time: The ass 3 … [is] kindly heavy and slow, and unlusty, dull and witless and forgetful … For he is put to travail overnight and is beaten with staves, and sticked and pricked with pricks, and his mouth is wrung with a barnacle, and is led hither and thither, and withdrawn from leas and pasture … Bartholomew Anglicus, Medieval Lore; 13th century (cited in Bough, 2011a) The perception of donkeys as "witless" or foolish persisted into the eighteenth century, when shamed husbands were strapped to a cart drawn by a donkey and paraded around town as part of public-ridicule punishments (George, 1994;cited in Bough, 2011a).During the nineteenth century, donkey neglect and abuse were rife.Determined to highlight the ubiquitous cruelty to domestic animals in Britain, MP and animal rights campaigner Richard Martin led an historic court case against a donkey owner in 1822, insisting that the donkey appear in court so that the animal's wounds could be witnessed.The owner was subsequently convicted of animal cruelty, leading to Parliament passing the first legislation in the world to prevent cruelty to animals (Act to Prevent the Cruel and Improper Treatment of Cattle 1822 -"Martin's Act").Despite this, Blackwood's Edinburgh Magazine described the typical treatment of donkeys in 1840 as "unfed, homeless, vagrant, unpitied, untended, kicked, lashed, spurred, tormented, troubled, thumped and thrashed" (cited in Bough, 2011a). 4Martin's Act 1822 did however mark a turning point for donkeys and other animals in Britain as there followed a steady rise in animal protection legislation and the establishment of animal welfare charities (e.g., The Royal Society for the Prevention of Cruelty to Animals (RSPCA), founded in 1824; RSPCA, 2022).The Donkey Sanctuary, established in 1969, is currently one of the largest such charities in the world, generating over £45 million in public donations and welcoming over 390,000 visitors to its UK sites annually (The Donkey Sanctuary, 2022).Public donations fund welfare advice, veterinary interventions, rescue and rehoming schemes, and the provision of lifelong care to over 4,000 donkeys in sanctuaries across Europe.Such substantial public support indicates that donkeys have charity appeal among parts of the British public today.

Popular and Folk Culture Representations
To the best of our knowledge, there have not been any systematic studies on contemporary representations of donkeys in popular (Western) culture and media, though the limited literature does offer some insight into how donkeys have been perceived and treated historically (Baker, 1993;Bough, 2011a;Kean, 1998).Bough (2011a) discusses the many ways in which donkeys have been (mis)understood (and stereotyped) throughout Western history.Early characterizations include fables from Ancient Greece and Rome, which have been "deeply influential in depictions of donkeys as servile, stubborn and stupid" (Bough, 2011a, p. 130).Though later Christian representations portrayed donkeys as "symbol[s] of patience, humility, suffering and service" (Bough, 2011a, p. 130), contemporary iterations of donkey characters in popular literature and film, on the whole, portray the donkey as stubborn, stupid, and dull.For instance, Shakespeare's donkey in A Midsummer Night's Dream (c.1600), Bottom, a fool with the head of an ass, is portrayed as an "unintelligent yokel," "a rude mechanical" with "uninspiring qualities" (Bough, 2011a, p. 134), while Eeyore in Winnie the Pooh (1926) is melancholic, alone, quiet, and introspective.Some alternative characterizations can be found, such as the animated character Shrek (Adamson et al., 2004;Jenson & Adamson, 2001;Miller, 2007;Mitchell, 2010;Trousdale, 2007), but the vast majority of contemporary representations of donkeys are not particularly positive.The perceived qualities of stubbornness and stupidity have seeped into contemporary political discourse and imagery.For instance, in the US political context, the donkey is used to symbolize the Democrats, with certain satirical cartoon renderings inviting associations with the "idiotic," and for over a century both the donkey and the elephant (symbol for the Republican party, though largely uncontroversial) have "maintained their potential for unflattering characterisations of the party" (Baker, 1993, pp. 92-94).Meanwhile in the UK, the "Led by Donkeys" campaign group (Wollaston, 2019) gained significant following for its political campaigns criticizing Brexit and the UK Government. 5  Representations of animal characters in literature, film and television will influence and inform contemporary understandings of real animals (DeMello, 2012;Ferguson, 2000).While nuances in donkey and ass characterizations do exist, such representations are, on the whole, not particularly favorable.

Investigating the Representation of Animals in Language
It is well established that public attitudes can be shaped by those with discursive influence (e.g., the media, government, and public representatives) (Driscoll, 1992;Stibbe, 2001;Van Dijk, 2005).As Stibbe (2001, p. 147) states: "social constructions exist, not in the minds of individual people but within the constant interaction and exchange of information in a society."Language "reflects, and also shapes, how we see the world … [it] teaches us cultural values and norms" (DeMello, 2012, p. 284).The numerous studies on the linguistic representation of nonhuman animals (e.g., Cook & Sealey, 2017;Goatly, 2006;McClaughlin, 2019;Sealey, 2018;Stibbe, 2001Stibbe, , 2005) ) shed light on how language and discourse can influence the ways in which societies perceive (and ultimately treat) animals.Language can have a significant impact on the lived experience of animals, in some cases turning "biological diversity into a political hierarchy" (Heuberger, 2017, p. 340), which gives rise to speciesist ideologies and, in turn, actions that favor or exploit certain animals.For example, intensive farming discourse influences industry practice and treatment of pigs by "constructing" them as machines, objects, and commodities (Stibbe, 2003).Drawing on the pig-as-machine metaphor, pigs are represented in terms of "performance," "sow durability," "sow breakdown," and as having "salvage value" (Stibbe, 2003).Conversely, Molloy (2011) finds that where newspaper reports of livestock "escapees" named animals individually using intertextual references (e.g., "Free Wooly" to name a sheep using a play on the film Free Willy; "Butch" and "Sundance" the Tamworth pigs; and a pig described as "Babe like"), they were ultimately spared slaughter.Where similar reports named groups of animals in food terms ("Traffic Ham: Porkers Close Motorway after Lorry Crash," Daily Mirror, 2009, p. 31), the animals were ultimately killed (Molloy, 2011).
For these reasons, it is important to reflect on how anthropomorphic and zoomorphic representations might influence public ideas about (and treatment of) the actual animal.Whilst anthropomorphism has been promoted in certain arenas (for discussions, see Daston & Mitman, 2005;Horowitz & Bekoff, 2007;Mitchell et al., 1997), critical animal studies scholars and ecolinguists advise caution with the use of anthropomorphic and zoomorphic representations (e.g., Heuberger, 2017;Serpell, 2002), with some highlighting the adverse effects on animal welfare (Bradshaw & Casey, 2007;Larsen et al., 2018;Mota-Rojas et al., 2021).Such representations have the power to justify and reproduce certain assumptions about (and negative perceptions of) animals, and these, in turn, can structure and influence social attitudes and behaviors toward them (Gibbs, 1999;Goatly, 2006), leading to anthropocentric and speciesist positions.
A review of zoomorphisms (i.e., attributing animal qualities, traits, and characteristics to humans in language) produced by university students highlights negative connotations associated with donkeys (Sommer & Sommer, 2011).It found that 98% and 94% of students rated the zoomorphic use of "ass" and "donkey," respectively, as "uncomplimentary" (Sommer & Sommer, 2011).Likewise, a study of metaphors relating to (lack of) intelligence and stupidity found that donkeys made up the largest animal group in the dataset (Allan, 2008), and an historical analysis of English metaphors for animals (using the Historical Thesaurus of English) found "donkey" to be associated with the terms "stupidity," "foolish person," and "contempt" (Paterson, 2016).This is not restricted to the English language context: the word "donkey" has negative connotations in languages beyond English, as Table 1 shows. 6 While there have been very few cultural analyses of donkeys specifically, it is clear that generalizations about donkeys are prevalent in language.To investigate contemporary understandings of donkeys, we examined their representations in the language of popular discourses, which we further explored with focus groups.As we go on to demonstrate, certain representations of donkeys are maintained in contemporary discourses, which has implications for their status and for those working to improve their welfare.

Ethics Approval
Research was carried out under the research policy and guidelines of The Donkey Sanctuary and received ethical approval from the executive team therein (Project code: DDCL19).All data were stored securely and in compliance with current data protection legislation (GDPR; EU Reg 2016/679).
Table 1.International meanings of the word "donkey" (Source: Various language forums a ).

Approach
We adopted qualitative and quantitative approaches to investigate the discursive representations of donkeys in Britain, drawing on corpus-assisted discourse studies (CADS) (Ancarno, 2020;Gillings et al., 2023;Partington et al., 2013; see also Baker, 2006) and focus groups to gather insights on contemporary understandings of donkeys.CADS combines corpus linguistics, a largely (but not exclusively) quantitative approach to language analysis, which uses specialist statistical software 7 to identify salient patterns in language use (Hoffmann et al., 2008), with discourse analysis (Widdowson, 2007), a largely qualitative approach, which involves examining the expression and interpretation of meaning in texts.Critical approaches to discourse analysis (Van Dijk, 2005) examine ideologies, stereotypes, and power relations in language, which are of key importance in this study.Speciesist or anthropomorphic language exemplify the kinds of ideologies, stereotypes, and power relations in which critical discourse analysis is typically interested.CADS helps to ensure against "cherry picking" (Gillings et al., 2023), where expected and/or convenient patterns of representation are chosen to demonstrate a preconceived argument, whilst allowing for less dominant and perhaps unexpected patterns in the discourse to emerge.We discussed key findings from the CADS analysis with focus groups, which are a well-established tool for gathering insights from a group of people who are guided to discuss a specific topic through open-ended questions (Barbour & Kitzinger, 1999;Bloor et al., 2001).

Data and Methods
Data for this study comprise real-world texts about donkeys and feedback from focusgroup participants.We built an approximately 1-million word "corpus" (a digital database of language texts) of English language texts about donkeys published between 2015 and 2020; 54% of the corpus comprises popular representations of donkeys, while 46% comprises professional/informed representations of donkeys (Table 2).The texts were drawn from three discourse genres: British online and print news media (see Table 3); social media posts gathered from the Twitter platform (tweets) over a 24-hour period 8 ; and "informed" 9 texts produced by The Donkey Sanctuary and published on its website (every webpage and linked PDF published in this domain was included).These genres reflect a range of beliefs, attitudes, and values at a national scale.News media is known to both reflect and shape the views of readers (Kuhn, 2007;Louw, 2005;Richardson, 2017).Twitter has been used by pollsters and governments alike to gauge public opinion on specific matters (e.g., Bovet et al., 2018;Tavoschi et al., 2020).The purpose of the paper was not to elevate one type of discourse above another, since they each have different purposes, strengths, and limitations (see McClaughlin et al., 2022) for findings relating to the similarities and contrasts between the discourses). 10We first conducted a CADS analysis on the corpus of informed and popular texts about donkeys, identifying salient grammatical, lexical, and thematic patterns, which guided our selection of appropriate stimulus materials from the popular discourses (tweets, key texts, and their related images and videos) for focus-group discussion (see McClaughlin, 2019 for a full discussion of the salient results from the CADS analysis).Though these patterns facilitated discussion, participants were not limited in the range of topics they could discuss, and their perspectives added depth to our understanding of the quantitative results.We ran three focus groups of five participants lasting between 60 and 90 minutes (see Table 4 for demographic information and Table 5 for a breakdown of the focus groups).These were conducted online using the Zoom video conferencing platform and were audio recorded before being transcribed for analysis. 11Informed consent was gathered from participants, and during transcription, names were changed to preserve anonymity: for example, P1A signifies Participant A from focus group 1).
Transcriptions were coded using NVivo (v.12) software.We initially identified a broad suite of issues through "topic coding" (Richards, 2005) and reflected on these in relation to the literature, relating our findings back to the broader themes emerging from a corpus analysis of both the informed and popular discourses about donkeys, including "negative stereotypes," "roles and functions," "capacities and cognition," "sentience and subjectivity," "care and concern," "harm and cruelty," "objects of entertainment," and "animals in need."These themes were generally approached from different perspectives across the informed and popular discourses (see McClaughlin, 2019).The primary focus of this paper is the popular discourses and public responses to them.We discuss in detail two news stories that were highlighted in the corpus analysis findings with focus group participants: the first related to welfare of animals in television entertainment; the second related to working animals beyond Britain.In the following section, we report the results from the focus groups and the salient patterns identified in the popular discourses (news and Twitter) on which the focus groups were based.

Results
Donkeys prominently featured in popular discourses as figures of entertainment and ridicule, (i.e., "joke" animals).In fact, 85% of tweets and 63% of news texts that we originally gathered contained non-literal references to "donkey(s)" (the word), which denoted fictional characters, or reflected figurative, metaphorical, and zoomorphic references.These texts used the word "donkey(s)" zoomorphically: that is, to refer to (mostly negative) human traits and characteristics, both within a sentence or as a hashtag ("#donkey").Non-literal references to donkeys were often directed at particular individuals on Twitter, referring to (human) characteristics of ignorance and stupidity (e.g., "@ … … … … Just because you're so unintelligent … doesn't mean you have to hate others … #donkey").Often these tweets were used in conjunction with profane and/or discriminatory language.Though these non-literal references to donkeys were not retained for full corpus linguistic analysis (and do not form part of the final dataset shown in Table 2), we noted that almost all took the form of insults, appearing in the contexts of politics, particularly in reference to inept politicians (e.g., example a, Table 6); football, particularly in reference to clumsy and inefficient players (e.g., example b, Table 6), and people's physical appearanceusually womenwho are perceived as or claimed to be unattractive (e.g., example c, Table 6).This particular usage is a trope of US rightwing commentary about Democratic Party politician Alexandria Ocasio-Cortez, whose party emblem is the donkey.
It is clear that non-literal references to donkey-naming terms reflect the established negative representations of donkeys identified in the literature.In turn, these representations can be seen in the literal references to donkeys included in our corpus, which also frequently painted the real animals as figures of ridicule and entertainment.A considerable number of tweets featured jokes about bestiality with donkeys.Donkey bestiality is almost exclusively treated as a light-hearted topic in our corpus, and often used to shame or ridicule the people (alleged to be) involved in the act.Importantly, the same comedic response is not found in discussions of bestiality with other species (see DeMello, 2012, for examples).We also found that donkeys occasionally featured in fictional contexts in order to ridicule people.For instance, in an imagining of "Brexit Britain Festival" on Twitter, donkey rides are proposed to be replaced by "Tories giving piggy backs": Ideas for the Brexit Britain festival.Maypole: see Theresa [May] do the Polka Donkey rides: Tories giving piggy backs Coconut shy: they will be shy, no trade deals mean we can't import them Welly boot toss: aim for the economics experts Don't forget to visit the adequate food tent!! Zoomorphic and semi-zoomorphic references to donkeys can, however, reveal something of how real donkeys are perceived.In the above tweet, humans (politicians) are put in the place of donkeys in a demeaning position of servitude, and this representation is reflected Table 6.Examples of the word "donkey" used as an insult.Source: Twitter.
Reference Example a "@realDonaldTrump You absolute donkey" b "@BBCSport He's a donkey" c "@AOC good ole donkey teeth has nothing better to do" in one of the most widely reported (and tweeted) media stories in the corpus.It concerns the television show "Gordon, Gino and Fred: Christmas Road Trip," which aired in the UK in 2019 (henceforth "CRT").The show featured three celebrity chefs riding a donkey cart in Morocco after their van purportedly broke down (Figure 1).CRT elicited many critical comments and complaints from viewers (e.g., "It is clearly unacceptable for a small animal like this to pull four adult men on a cart.It's quite disconcerting that no one involved questioned this" [Gaffney, quoted in Bryant, 2020]).
A tweet containing a clip of the show also provoked furious and empathic responses from Twitter users, as shown in extracts d through f (Table 7).The phrase "poor donkey" was salient in the corpus as a result of public evaluations of this particular story, and examining the wider context of this phrase revealed concern over the physical effort required to pull the cart (extracts g-j, Table 7).On the other hand, some commentators on Twitter used humor to ridicule the celebrities in CRT by calling them "the three asses," although this can inadvertently re-entrench negative ideas of donkeys as "joke" animals.This demonstrates that it is possible to inadvertently reinforce stereotypical and/or zoomorphic representations in attempting to be a "voice" for donkeys.
We invited our focus-group participants to respond to a video clip and still (Figure 1) from the show.In line with the response on Twitter, our participants expressed concerns over the donkey's physical condition (comments k-m, Table 8).Similarly to the Twitter user in extract f (Table 7, above), who felt it necessary to protect themselves and stop watching, our participants described the negative emotional effect that witnessing this depiction of suffering had on them (comments n-p, Table 8).
The affective language of these responses indicates that the roles of donkeys as figures of entertainment and ridicule do not translate well from metaphorical (zoomorphic) or figurative representations to demeaning the real animals for "light" entertainment.Focus-group participants recognized that donkeys play a valuable role as working animals beyond the UK context, but for them, CRT "completely ignore[s] the situation of these animals" (P2D), where life for humans and donkeys in Morocco was not accurately represented.One said, "it's fine to show that the locals use donkeys in their day-to-day lives […] but you don't need to extend it in this way … " (P3B).Participants felt that the use of donkeys in television entertainment was "wrong" (P1E) and "unacceptable" (P3B) and that it showed "a blatant disregard for animal welfare" (P2D).The strong public reaction to this show, which made headlines in UK newspapers and sparked a debate on Twitter, indicates that many are uncomfortable with this "comedy donkey" trope.
The consideration of working animals outside the UK context was a feature of both our news and Twitter corpora, where donkeys were frequently discussed in relation to the labor services they provide to humans, highlighted by words like "make," "pull," "ride," "carry," and "work" (Figure 2). 12.In relation to this, we identified wide-ranging concern over international donkey welfare in our British news and Twitter corpora.Two of the main themes that emerged (in a statistically salient way) were "care and concern" and "harm and cruelty."The most prevalent words (nouns, verbs, and collocates 13 ) from these two themes were: "suffer," "crippled," "exhausted," "mistreated," "slave," "cruelty," and "injured."Donkeys are commonly presented as "animals in need" (of human help or care) in these domains, and this was echoed in our focus-group discussions, where participants expressed concern over the treatment and welfare of the donkeys.One prominent news story headlined "Mali donkey horror" (henceforth "MDH"), published by the news website express.co.uk (Hopps, 2019), captured both of these themes.In its discussion of working donkeys in a waste disposal site in Bamako, Mali (Figure 3), MDH uses  descriptive language to portray a vivid scene to its readers: "It is the workplace from hell.A stinking, sprawling mountain of rubbish as far as the eye can see, dotted with exhausted donkeys delivering back-breaking loads of waste beneath a blazing sun" (Hopps, 2019).The British public engaged with this story on social media in a significant way.Twitter users expressed concern over rubbish tip "work" leaving donkeys "exhausted" (both words were statistically significant in the tweets).
When we asked our focus group participants to reflect on the language and imagery of the MDH article, their initial responses were those of shock.They offered comments such as "Oh my god" (P1C); "oh! [gasp]" (P1B); "how horrible" (P1D); and "horrendous" (P1C) whilst reading the article.When asked what prompted their responses, participants explained that they were moved by the physical conditions the donkeys were working in: "they don't have enough to eat so therefore they'd eat anything" (P1D); "they're going to go and pick up plastic and god knows what to eat" (P1D); "these poor donkeys choking in the air" (P1C).Though donkey hardship is foregrounded in this article, we found that human social actors were concealed in language used to describe donkey hardship both in the stories themselves and in quoted responses to them as this excerpt from the MDH article shows (Hopps, 2019, emphasis added): They are wheezing heavily, literally gasping for breath, their bodies shuddering and their legs wobbling like jelly.They are struck several times to "encourage" them.
The above quotation does not mention who the donkeys are struck by; the human involvement is simply implied.In fact, wherever donkey hardship is mentioned in MDH, humans are not present in the immediate context.This is achieved either by way of agentless passive grammatical constructions, which allow for the agent of the verb to be omitted (e.g., "they are struck"), or through metonymy, where the human social actors are substituted for objects ("Pulling burdens some four times their body weight, they have ill-made harnesses that rub with every step, causing seeping wounds to fester," Hopps, 2019, emphasis added).This serves to distance the owner (and their role) from the plight of donkeys in this context.It is possible that this is an intentional choice by the writer due to the uncomfortable truth that human workers are also suffering in this context (see Watson et al.'s [2020] discussion of "shared suffering").Alternatively, the suffering of donkeys may not be considered the result of intentional cruelty (i.e., "deviant violence," DeMello, 2012, p. 242), but rather the result of wider environmental and socioeconomic challenges in lower-middle income countries (LMICs).Our focus group participants seemed to confirm this, insofar as they felt the owners were not necessarily to blame in this context, where working conditions are poor, wages are meager, and human health is regularly and significantly compromised.
These and other salient popular media stories identified in the corpus analysis suggest that there are different attitudes toward donkey ownership in Europe and LMICS.It may be that extratextual political factors come into play here.In another prominent news text shared on Twitter, originally headlined "Blackmuled: Costa del Cruelty Scandal" (Parker, 2016), the donkey can be seen as a conduit for political comment.The article, which discusses the cruel treatment of donkeys in Spain (with a British man in the role of rescuer), was published at the height of the Brexit debates.Here it appears that the topic of animal cruelty has been exploited for political influence, which if true, illustrates how anthropocentric representations can sideline the plight of the animals involved.In these ways, donkeys and related issues are veiled and obscured in popular discourses in favor of human emotions, politics, and entertainment.

Discussion
Popular representations have real life consequences for the animals involved, including their relationships with humans and the roles they occupy in human lives (DeMello, 2012;Molloy, 2011). 14According to Stibbe (2001, p. 148), "[i]deologies, embedded and disseminated through discourse, influence the individual mental representations of members of a society, which in turn influence their actions."It follows, then, that speciesist ideologies can be perpetuated through participation in and exposure to discourse, which in turn can have a direct impact on the welfare of animals.This is problematic for those seeking to elevate the status of donkeys and improve their welfare.
As a result of the lack of attention given to donkeys in policies and planning relating to international development and disaster-risk reduction (Clancy et al., 2021), they have been described as "invisible workers" (Valette, 2015) and "invisible livestock" (The Brooke, 2019).Our findings highlight that donkeys have been discursively obscured in ways that lead these animals (and their very real contribution to society) to be overshadowed by speciesist ideologies embedded in the language used to represent them.In particular, accounts of the living and welfare standards of donkeys in popular discourses can obscure or invisibilize the animals by (i) enabling humans to explore their own emotional response to hardship and suffering of donkeys, and (ii) providing a platform for people to criticize other people (the animal is secondary in this discourse).
Donkeys are discursively represented in multiple and complex ways.A major pattern of representation in the discourses we examined places donkeys as emblems of suffering.
Here, the CRT and MDH stories exemplify the role of donkeys as objects of pity.We identified a tendency for people to explore their own emotional responses to harm and suffering: that is, their own feelings of discomfort, above those of the animals involved.The emotional responses of the animals are not overtly present in discussions surrounding these texts.Depictions of donkeys abroad in the media present an incomplete and over-simplified reality, which reinforces negative evaluations of the reality of human and animal relationships in a way that is reminiscent of "poverty porn" (i.e., media misrepresentation and vilification of social welfare and people in poverty for entertainment; see Mooney, 2011, p. 4).
Anthropomorphic and zoomorphic representations have the power to justify and reproduce certain assumptions about animals, and these, in turn, can structure and influence social attitudes and behaviors toward them (Gibbs, 1999;Goatly, 2006).In particular, zoomorphic language can distance people from the real animal, conceptually and emotionally (Goatly, 2006).This is consistent with Heuberger (2017, p. 342), who argues that animal metaphors used to criticize humans also "reinforce the idea that these animals actually have the corresponding negative traits, which decreases their status," but he asserts that "even positive metaphors do not necessarily bring about an amelioration of the status of animals, but often result in reduction and inequality" (Heuberger, 2017).In the popular discourses we examined, placing donkeys in the role of joke animals is generally accepted, providing the representation is zoomorphic and relates to figurative or metaphorical donkeys, or to real donkeys who are not visible.This representation is not accepted when the joke is transferred into real-life contexts in a way that is visible, however.Demeaning working donkeys for "light" entertainment is particularly controversial for members of the British public represented in our data; the incompatibility of the two roles is jarring, perhaps because it leads to the realization that even light-hearted non-literal representations can contribute to negative stereotypes.Strongly negative responses about donkey suffering, both from social media users and our focus-group participants, indicate that issues of harm and cruelty are a major concern, which is especially centered on the physical impact of labor.
The dominant representations of donkeys as figures of entertainment and ridicule, and as vulnerable "animals in need" may result from the historical use of donkeys in Britain and contemporary understandings of the way donkeys are utilized abroad.Some social commentators argue that shame and guilt from the past is held in collective memory.For instance, Moore (2007) contends that donkeys have a revered status in British society as a direct result of societal guilt for historical poor treatment.Equally, the prevalence of certain symbolic representations and stereotypes of donkeys might reflect the lack of real-life encounters and interactions with them.As DeMello (2012, p. 283) points out: " … for many people, the real relationships that humans once had with animals have been largely supplanted by symbolic representations."This is an interesting theory and would need to be explored in further research on public perceptions of donkeys in different contexts.
It is important to note that our findings represent a "snapshot" of contemporary popular opinion and, as such, some voices and perspectives are not represented. 15To some extent, this was addressed by seeking perspectives about donkeys from members of the public in a series of focus groups, though the majority of participants were over 55 years old, female, well-educated, white, and either had some pre-existing knowledge of and/or interest in donkeys (see Table 4). 16We factored this into our analysis, acknowledging that our participant group represented an "engaged" audience.This lent itself to in-depth discussions on issues such as anthropomorphism and human responsibility toward animals.
Our results highlight that when people are actively engaged with real-life issues 17 (as our focus-group participants were), the language appears to shift, and the sentience and subjectivity of donkeys is more openly acknowledged.However, this does not necessarily mean the animal is recognized for their own intrinsic value.We found that even sympathetic language centered on human emotional responses to the situationwhen donkeys are presented as animals in needtheir sentience and subjectivity is not necessarily prioritized in an overt way.Unfortunately, collective moral responsibility concerning the discursive representations of donkeys (and animals more generally) was rarely recognized or understood in the discourses we examined.
We also wish to highlight the important role of the media and social media in shaping public perceptions of animals: inaccurate and sensationalist portrayals of animals may sell stories, but they can have negative consequences for animal welfare (Molloy, 2011).Both news and social media users have a responsibility to consider the effects of anthropocentric language use and address or challenge harmful representations.Social media can be a polarizing platform for discourse about donkeys, which contributes to the perpetuation of myths and misunderstandings.It is important to foreground donkeys and issues relating to them whilst challenging anthropomorphic and anthropocentric discourse (including zoomorphisms and obscuring language), which reinforce negative, inaccurate ideas of donkeys.Though discourse is just one part of the wider picture, which includes popular representations in literature, film, and television, there is little doubt that such representations present barriers to human understanding of the real animals.

Conclusion
According to their representations in the public-facing and public-produced discourses we examined, donkeys continue to occupy a lowly status in popular texts and are frequently subject to common misconceptions and negative stereotypes in language.In contrast, the response of focus-group participants (who were engaged with donkey welfare issues) demonstrated that simply engaging people is in itself an important first step regardless of their level of understanding or interest in donkeys.Unhelpful and potentially damaging representations can be effectively challenged by those seeking to improve the contemporary status and understanding of donkeys.
We recommend that donkeys are discursively framed for their intrinsic value and not simply for the roles they perform in human lives.In order to achieve this, collective responsibility must be taken for promoting welfare-positive discourse, whilst being mindful of the potential for such discourse to be counterproductive.So-called "weak" forms of anthropocentrism are expected as humans can only ever experience the natural world from a human perspective (see Heuberger, 2017); expressing empathy for donkeys experiencing hardship is an acceptable (and admirable) position.Strongly anthropocentric ideologies must be addressed, however.These ideologies ignore and/ or conceal the animals' emotional experiences of hardship as well as those responsible for it.A shift in the way that donkeys are understood and represented needs to take place, and we outline here three main ways in which this might be achieved through language.First, by using donkey-first language, which considers donkeys' emotional needs and their responses to their experiences.Second, by avoiding anthropocentric language choices in evaluations of human traits and behaviors (zoomorphic language) or obscuring animals in accounts of their own hardship.This also means offering (linguistically) accurate representations of the complex relationships between humans and donkeys, including where power resides, and where social and structural inequality might have a bearing on animal suffering (Watson et al., 2020).Finally, we recommend the avoidance and active rejection of anthropocentric language that ridicules donkeys, particularly in relation to cruelty.Those who have elected to take responsibility for donkey welfare can (and should continue to) provide "counter discourses" (Macgilchrist, 2007;Martin, 2004) to challenge the harmful discourses that invisibilize donkeys (and other animals).Beyond this, we propose that there is a wider social responsibility to do the same in order to conventionalize the kinds of constructive language use that we recommend here.Through the promotion of critical, reflective language, which acknowledges the sentience of donkeys, the status of the donkey may be transformed, both in Britain and beyond.

Figure 1 .
Figure 1.The Christmas Road Trip story in The Mirror Online, January 12, 2020.

Table 2 .
Descriptive statistics for the donkey discourse corpus.

Table 3 .
Breakdown of news publications in the news corpus.

Table 7 .
Excerpts from public commentary about the television show "Christmas Road Trip" (Source: Twitter).

Table 8 .
Excerpts from focus-group participants commenting on the television show "Christmas Road Trip" (Source: primary research by authors).