Can an Emoji Be Considered as Defamation? A Legal Analysis of Burrows v Houda [2020]

This article considers the Australian case of Burrows v Houda 2020 NSWDC 485 and the English case of Lord McAlpine v Bercow 2013 EWHC 1342 (QB). Both cases considered the question of whether emojis could be considered to be defamatory and answered the question in the affirmative. This article also explores whether the South African courts will follow the lead of the Australian and English courts and concludes that emojis also have the potential to be considered defamatory in our law.


Introduction
In August 2020 an Australian court heard the novel matter of Burrows v Houda (hereafter the Burrows case) and was called upon to decide whether the use of the ʺzipper-mouth emojiʺ could be considered as defamatory. 1 The court ruled on 27 th August 2020 that the use of this emoji was capable of giving rise to a defamatory meaning. This is merely a preliminary ruling in the matter and the matter is currently ongoing.
An emoji is ʺa small digital image or icon used to express an idea, emotion, etc., in electronic communicationsʺ. 2 Pelletier states that: [e]mojis are often included alongside text to portray an emotion or otherwise add to readersʹ understanding of the text. The most basic emoji is a yellow smiley face, but emojis span from face emotions, like happy, sad, and angry, to animals and ordinary objects, like a cat or a briefcase. With such options, tweets are largely only limited by individual usersʹ creativity. 3 Emojis can be included in text messages, emails and messages, and posts on social media sites, and blogging or other sites. Emojis can also be thought of as playing ʺa role in social media analogous to non-verbal behaviour in offline speechʺ 4 and they attempt to ʺsmooth out the rough edges of digital life.ʺ 5 Some writers have even gone so far as to state that ʺ[o]ver time, it will feel increasingly ʹweirdʹ and ʹcreepyʹ to write online messages without emojis.ʺ 6 McMahon and Kirley postulate that while emojis are widely perceived as ʺcute or benign adjuncts to online communicationsʺ and are used to ʺhumanize truncated digital messages by conveying humor (sic) emotion, and sociability, emoji (can) perform a far more sinister role.ʺ 7 This sinister role includes its use in hate speech, cyber-bullying, witness intimidation, online grooming of children for sexual abuse, and other crimes. 8 There are currently almost 3 300 different emojis in existence 9 and July 17 has been designated World Emoji Day. It is estimated that approximately 700 million emojis are used daily in Facebook posts and that over 900 million emojis are sent as standalone messages (i.e. without any accompanying text) on Facebook Messenger. 10 Over 10 billion emojis are used daily and 95% of internet users have used an emoji, with studies showing that the use of emojis in a social media post can increase engagement with that post: ranging from an increase on Twitter of 25.4% and increasing the amount of likes on Facebook by 57% and the number of comments on posts by 33%. 11 Additionally, emojis are also effective in digital marketing, as commercial emails which include an emoji in their email subject lines, have a higher unique open rate than emails that did not contain emojis in their subject lines.
With these figures, it is safe to state that the use of emojis is here to stay. As such, our courts will have to increasingly engage with the impact of these emojis in different areas of the law. Courts in the USA have had to decide whether the use of emojis can be considered as criminal intimidation, 12 and online sexual grooming of children by paedophiles. 13 In this article we explore whether emojis can be considered as defamatory.
This note will first discuss the general law relating to defamation in Australia, as a backdrop to an exploration of the Burrows case. Next, it will briefly discuss the English case of Lord McAlpine of West Green v Bercow (hereafter referred to as the Bercow case), before hypothetically discussing whether an emoji could be held to be defamatory in the context of South African law.

General background to Australian defamation law
The law of defamation in Australia is governed largely by a combination of common law and the Defamation Act. 14 The common law is inherited mainly from the English law of defamation, and it has been developed by the Australian courts over many decades. Each of the states and territories of Australia has also modified the common law by introducing statutes which apply to that state or territory. However, each of these individual statutes is 10 SouthFloridaReporter.com 2020 https://southfloridareporter.com/%F0%9F%98% 84-95-of-internet-users-have-used-an-emoji-over-10-billion-emojis-are-sent-daily/. 11 SouthFloridaReporter.com 2020 https://southfloridareporter.com/%F0%9F%98% 84-95-of-internet-users-have-used-an-emoji-over-10-billion-emojis-are-sent-daily/. In order for a defamation action to be founded in Australia, the material must be published to a person who is capable of understanding it to be defamatory. This person must be an individual other than the plaintiff and the publication must be intentional or a probable result arising from the defendant's conduct. As in South Africa, a person who republishes a defamatory statement can also be held liable for it. 16 There may also be instances of multiple publishers, in that the author, the printer, the publisher and the distributor may all be held liable for defamatory material published by them.
The publisher of the defamatory material will be held liable even if he acted with reasonable care or there was no intention on his part to cause the plaintiff any damages. 17 Australian law has deemed the owner of a website that publishes defamatory material to be the publisher of that material. 18 The Australian courts have recently decided on many matters which are relevant to internet and social media usage. In Trkulja v Yahoo! Inc. LLC the court ruled that search engines (such as Google) could be deemed to be publishers and can be sued for defamation for defamatory material appearing on their search engines. 19 They would, however, be able to use a defence of innocent dissemination to escape liability in appropriate cases. It is interesting to note that while a South African court may order an apology in defamation matters, an Australian court may not make this order, and the usual Australian remedy is a damages award. 25 The Australian courts have dealt with many more facets of internet and social media defamation than the South African courts. One should, however, bear in mind the warning of legal experts in Australia that:  26 Pelley 2019 https://www.afr.com/companies/media-and-marketing/australia-thedefamation-capital-of-the-world-20190904-p52nuh. 27 Douglas 2020 https://theconversation.com/australias-outdated-defamation-lawsare-changing-but-theres-no-revolution-yet-143532. interest" defence based on the English law, and amendments to the capping of damages.
Australia is also investigating legislative changes to allow social media companies to be held liable for defamatory comments made by third parties on their platforms. While the common law already allows for this, as established in the Voller case above, that judgment is currently being appealed.
The cross-border global nature of social media networking sites raises interesting Public Interest litigation issues especially around the issues of holding social media companies liable in situations where the social media company is not based in the country of origin of the litigation. Most social media companies are based in the United States of America, whose laws provide that ʺno provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.ʺ 28 This provision has resulted in some US companies simply ignoring Australian judgments or obtaining judgments from American courts, meaning that the company does not have to comply with the Australian judgment. 29 In the absence of global agreement on cross-border jurisdictional issues, the cross-border enforceability of judgements against social media companies will always be a thorn in the side of litigants.

Facts
The facts in this matter are not very clear as the court did not canvass the facts thoroughly. 30 This is probably because this is merely a preliminary ruling on the ability of the comments to be regarded as defamation and not a complete consideration of all the elements of the defamation action. These preliminary rulings allow for the cost of proceedings to be kept down, as they encourage parties to settle the matter prior to a full hearing of it. https://theconversation.com/a-push-to-make-social-mediacompanies-liable-in-defamation-is-great-for-newspapers-and-lawyers-but-not-you-127513. This strategy has also been pursued by Google in other jurisdictions for e.g. Canada in the cases of Google Inc. v Equustek Solutions Inc. 2017 SCC 34 and Google LLC v Equustek Solutions Inc. Case No 5:17-cv-04207-EJD. 30 Burrows v Houda 2020 NSWDC 485. P SINGH PER / PELJ 2021 (24) 7 The plaintiff, Zali Burrows, is a fairly well-known Australian lawyer. 31 The defendant, Adam Houda, is also a fairly well-known Australian criminal lawyer. The Herald, an Australian newspaper, had published an article in which it reported on a judge's criticism of the plaintiff's conduct in a legal matter before him and his suggestion that her conduct be referred to the Law Society for potential disciplinary action. Houda is being sued by the plaintiff in respect of two Twitter posts made in regard to links to this Herald article on 27 May 2020 and 28 July 2019.
It is unclear from the judgment what the actual post on 27 May 2020 was. From a reading of the case and a perusal of the attachments submitted to the court, labelled as "MCO A" and "MCO B", it is submitted that the 27 May 2020 post was a link to the Herald newspaper article. This post of the defendant received ʺretweets, ʹlikesʹ and, in particular, a reply which asked: ʹJuly 2019 story. But what has happened to her since?ʹ The defendant's response is the emoji commonly referred to as ʹzipper-mouth faceʹ". 32 Image 1 taken from Burrows judgment.
Three comments were also made by third parties as responses to Houdaʹs post. The first reply made the statement ʺʹJudge Wilson recommended Ms Burrows' clients be banned for life by ASIC and prosecuted for signing affidavits they knew to be falseʹ, followed by a series of hashtags and links to other Twitter users. This is followed by the words 'tick-tock' and an emoji showing a clockʺ. 33 This post also contained two document stubs. 31 Whitbourn 2020 https://www.smh.com.au/national/judge-considers-zipper-mouthemoji-in-sydney-lawyers-defamation-fight-20200828-p55q7x.html.
P SINGH PER / PELJ 2021 (24)  8 The second reply consisted of the third party utilising the ʺretweet with commentʺ option on Twitter to republish Houdaʹs post. The additional comment of the third party consisted of the use of three emojis. These emojis were ʺcollisionʺ, ʺface with tears of joyʺ, and ʺghostʺ.
Image 2 taken from Burrows judgement.
The third reply from a third party also ʺretweeted with commentʺ Houdaʹs post and added the words ʺOhmigod bro!!!!!ʺ. 34 Image 3 taken from Burrows judgement.
The 28 July 2020 tweet was a response by a third party to Houdaʹs tweet, in which this user stated ʺOutright crims -signing false affidavits FFSʺ. 35 The court noted that the defendant did not challenge the inclusion of the posts from third parties as part of the defamation claim against himself.

Judgment
Gibson DCJ stated that she would approach the topic of whether an emoji could be considered to be defamatory with care, as this was the first time that an Australian court had been asked to rule on the issue of whether an emoji could be considered to be defamatory.
Gibson DCJ then proceeded to investigate the meaning an ordinary and reasonable Twitter user would assign to the emoji symbols. In doing this she consulted an internet dictionary known as Emojipedia. [a]lthough there has been some academic criticism of judges for referring to sources such as Wikipedia…the nature of modern communications makes consultation of internet dictionaries, such as Emojipedia, a necessary step for the trier of fact who seeks to determine what the ordinary reasonable Twitter reader would make of the use of these symbols. 36 Gibson DCJ then proceeded to discuss the difference between an emoticon and an emoji. She states: An 'emoticon' is a portmanteau term (from 'emotional icon') for pictures made from punctuation marks, letters and numbers to create an image displaying a sentiment and predates the internet as signs….can be created with a keyboard…[a]n emoji is a more recent invention, consisting of pictographs of faces, objects and symbols… [t]his results in the definition: 'a small digital image or icon used to express an idea, emotion, etc., in electronic communications.' Emoji are used in addition to other meaning-conferring tools, such as hashtags and buttons for 'like' or 'retweet'. 37 The court then considered the English judgment in the Bercow case, in which the court was called upon to decide on the meaning of an ʺinnocent faceʺ emoticon, which took the form of words rather than symbols. 38 This case will be discussed below.
Gibson DCJ in the Burrows judgment then proceeded to determine whether it was appropriate for a judge to decide on the meaning of an emoji without the assistance of expert evidence or jury input. She did not consider the use of an expert to be necessary in this case, for the following reasons: firstly, neither of the parties had suggested that an expert should be called and as such she did not wish to impose this on them; and secondly, Australia already had rulings regarding the meaning of emoji and other non-verbal tools (e.g. the like button and hashtags) in other areas of law not involving defamation and it had not been a requirement in those cases that an expert be called.
She also expressed the opinion that emojis have largely replaced emoticons and that the ʺʹzipper-mouth faceʹ emoji has a meaning to denote ʹa secretʹ or ʹstop talkingʹ, in circumstances where a person impliedly knows the answer but is forbidden or reluctant to answer.ʺ 39 Dibson DCJ then considered the emojis used by the third parties who replied to Houdaʹs tweet. She concluded that the words ʺtick-tockʺ accompanied by the picture of the clock implied that the clock was ticking for the plaintiff. She then considered the tweets of the second reply, being the ʺcollision emoji, the face with tears of joy emoji and the ghost emojiʺ. She noted their meanings as per Emojipedia, but stated that she would not be bound to these strictly, as the meanings conveyed by a publication are a matter of a more extensive impression.
She described the emojis meaning as per Emojipedia as follows: 'Collision'...may be used to illustrate a clash but is commonly used to represent something is excellent or exciting in some way…'Ghost' which indicates something fun or goofy…the raised arms have resulted in its use to indicate 'an excited Yay!' 40 The Court did not define what ʺface with tears of joysʺ meant. It is submitted that this was probably because this emoji is very popular, and its meaning is both well-known and self-explanatory.
The plaintiff submitted to the court that Houdaʹs tweets and the subsequent responses to these tweets created a suggestion that she was facing a potential legal battle after a judge had made critical comments in respect of her competency as a lawyer and/or that the plaintiff had committed an act of misconduct during a court case and that the judge had recommended that she be referred to a lawyers' professional body for possible disciplinary sanction and/or that the plaintiff was a criminal who signs false affidavits and/or that the ʺplaintiff's conduct as a lawyer during court proceedings over which Judge Wilson presided was so poor that the judge recommended that her clients be banned for life by ASIC and prosecuted for signing affidavits that they knew to be false.ʺ 41 The defendant's counsel submitted that none of the defendant's tweets gave rise to a suggestion that the plaintiff was being, or was likely to be, disciplined by any professional body for her conduct. Although the defendant had shared the ʺstubʺ (i.e. the link) to the Herald article, he made no additional comment with the exception of three dots (…) and a tag labelled ʺ#auspolʺ. They argued that neither the dots nor the hashtag was sufficient to give rise to any of the implications raised by the plaintiff. In addition, they argued that the ʺzipper-mouth faceʺ ʺconveys nothing other than the defendant cannot reply.ʺ 42 The plaintiff did not agree with this argument, pointing out that the ʺzippermouth faceʺ was made as a response to an enquiry in regard to the alleged disciplinary inquiry. In addition, they argued that the defendant had posted 40 Burrows v Houda 2020 NSWDC 485 para 34.

42
Burrows v Houda 2020 NSWDC 485 para 37. P SINGH PER / PELJ 2021 (24)  11 the stub to the article almost a year after the article had originally been posted by the Herald. The plaintiffs submitted: [t]hat this 'zipper-mouth face' is worth a thousand words -the emoji implies that there has been a finding damaging to the plaintiff, but the defendant is not at liberty to disclose the result, and instead must hint at it by posting the newspaper story from the previous year and using the 'zipper-mouth face', so the reader can guess the rest…this Delphic response is the equivalent of shouting 'fire' in a crowded theatre, giving rise to a defamatory meaning which is further inflamed by the three following comments. 43 The court ruled that the defendant's comments, when read in context, were capable of being understood as stating that the plaintiff was facing disciplinary action by a professional body. The defendant had posted a stub that was almost a year old, and when asked about the outcome of this article he posted the ʺzipper-mouth faceʺ in reply to that question. The court held that this ʺis a case where ʹjoining the dotsʹ…to achieve the meaning is a particularly likely exercise when carried out on a social media site, where the exchange of such information is more likely than a serious publication to contain hints of a sensational nature.ʺ 44 The court then proceeded to consider the plaintiff's submissions that the defendant's comments suggested that she was a criminal who signs false affidavits and/or that her poor conduct as a lawyer led to the judge recommending that her clients be banned by ASIC and be prosecuted for knowingly signing false affidavits. 45 The court held that the tweets in question were reasonably capable of conveying these suggestions and stated: I am satisfied that the ordinary reasonable social media reader would infer that, while the clients 'signed' the false affidavits, the plaintiff, who was a solicitor reported to be in trouble with the judge, would also reasonably have been seen to be in trouble for her role in the preparation of the offending affidavits and/or their presentation to the court. This is underlined by the words 'tick tock' and use of the 'clock face three oʹ clock' emoji, which infer that the plaintiff's time (in terms of being dealt with for her wrongdoing) was up. The third and fourth posts add further emoji and comment to the defendant's post when they retweet it. They reinforce that the plaintiff's conduct is extremely serious and the subject of professional sanctions… I am satisfied that, in circumstances where the tweet clearly identifies that there is to be a prosecution for false swearing of affidavits, the ordinary reasonable social media reader would infer that one of those likely to be prosecuted would be the plaintiff, particularly given the 'tick tock' and the three excited emoji in the third reply. This

43
Burrows v Houda 2020 NSWDC 485 paras 39-40. 44 Burrows v Houda 2020 NSWDC 485 para 41. 45 Burrows v Houda 2020 NSWDC 485 paras 43-50. Bercow denied that her comment was defamatory and stated that she was simply asking a question in her tweet. She further accepted that her question implied that Lord McAlpine was trending (on Twitter), but denied that this was defamatory in any way as she did not suggest a reason as to why Lord McAlpine was trending. 52 The sole issue in this case was to determine whether the tweet could be considered as defamatory.
The court held that: It is common ground between the parties that the words 'innocent face' are to be read like a stage direction, or an emoticon (a type of symbol commonly used in a text message or email). Readers are to imagine that they can see the Defendantʹs face as she asks the question in the Tweet. The words direct the reader to imagine that the expression on her face is one of innocence, that is an expression which purports to indicate (sincerely, on the Defendantʹs case, but insincerely or ironically on the Claimantʹs case) that she does not know the answer to her question. 53 The court then proceeded to examine the meaning that the ordinary reasonable Twitter follower of Bercow would place on her tweet. The court concluded that the ʺinnocent faceʺ emoticon was not an innocent question, but rather that her tweet would be viewed in the light of the circumstances surrounding the tweet. 54 At the point in time when she had tweeted, the plaintiff's name was trending as there was a flurry of articles in the press in which an unnamed person was accused of being a paedophile. Thus any Twitter user who read her post would interpret her question not as an innocent query, but rather as an insinuation that the plaintiff was the unnamed person and that he had been publicly exposed for conduct for which he was guilty and that she believed this to be righteous. 55 The court held that English law accepted that a question could be understood to convey a defamatory meaning, 56 and stated:  (24)  14 simply wants to know the answer to a factual question… I find that the Tweet meant, in its natural and ordinary defamatory meaning, that the Claimant was a paedophile who was guilty of sexually abusing boys living in care…If I were wrong about that, I would find that the Tweet bore an innuendo meaning to the same effect. 57 Bercow finally settled the matter with Lord McAlpine for the sum of £15 000.

South Africa
Currently our South African courts have not yet had to hear any matter involving emojis, nor have they been called upon to decide whether an emoji could be considered defamatory. I will briefly canvass the South African law of defamation before hypothesising the potential reaction of our courts to the factual scenario as presented in Burrows v Houda.
Defamation is the unlawful, intentional publication of words or conduct which have the effect of injuring another personʹs reputation. 58 The injury to the person's reputation is judged objectively, in that a reasonable member of society must think lesser of the person due to the defamatory statement or conduct. 59 In order to establish a defamation action, a plaintiff must prove ʺ(a) the wrongful and (b) intentional (c) publication of (d) a defamatory statement (e) concerning the plaintiffʺ. 60 Once the plaintiff proves the publication of a defamatory statement which refers to him/her, then the wrongfulness and fault element are presumed to have been proven in favour of the plaintiff, and should the defendant wish to rebut this presumption, then he/she must raise a defence to exclude the presence of wrongfulness or intention. 61 The truthfulness of the statement/conduct is not a prerequisite for the statement/conduct to be held to be defamatory, and the defendant can be held liable for defamation even for the publication of truthful information. 62

Publication
In order to prove publication, the plaintiff must show that the defamatory statement or conduct was communicated to a third party. not be intentional, provided it is foreseeable. 63 Publication occurs only if the third party understands the publication to be defamatory. However, it is not necessary that he understand the material to be defamatory at the time he first hears or reads the words. If he later discovers the meaning of the material to be defamatory, then the requirement of publication will have been satisfied. 64 In Burrows the publication occurred via a publication of the tweet on Twitter.
In terms of South African law, publication would not be an issue. Once a tweet is published on Twitter it is generally available for any viewer to see, regardless of whether such a viewer has a Twitter account or not. The user does have an option to limit the audience to his post. However, it is submitted that even should a user utilise this option and limit his audience, the tweet would still have been published to third parties.
A person who repeats the defamatory statement of another is also regarded as a publisher of that material. In Burrows, the postings were made by Houda and other third parties. Houda did not deny ownership of his Twitter profile, nor did he raise the fact that some of the postings were made by third parties and not himself. 65 Should these issues arise in a South African court they would not be novel as our courts have already had to deal with issues of the ownership of online profiles and anonymous third-party postings in Dutch Reformed Church Vergesig v Sooknunan. 66 In this case the defendant, Sooknunan, denied that he was the owner of the Facebook page on which the defamatory statements were published and argued that even if he were the owner of the page, he was not responsible for the posts of other anonymous third parties who might have posted defamatory material on the page. The court adopted a pragmatic approach in establishing the ownership of disputed profiles on Facebook. 67 Should a dispute arise over the ownership of a Twitter profile in South Africa, it is submitted that the court would in practice consider the content and pictures on the profile to determine ownership. Sooknunan also argued that he should not be responsible for the posts of third parties on the page. The ʺcourt dealt with this issue by holding that the creator of a Facebook profile is responsible for posts made on the profile as his or her role is akin to that of a publisher who has made a forum available for the posting of unlawful content.ʺ 68 Thus, in South Africa a defendant in Houdaʹs position, could be held liable for the defamatory postings of third parties on his Twitter profile.

Defamatory statement
The plaintiff also has to establish that the defendant published statements or conduct that could be regarded as defamatory. This publication must both actually harm the plaintiffʹs good name or reputation and be objectively unreasonable. 69 In order to determine if the words or conduct complained of are defamatory in nature, a court must explore the meaning of the words or conduct. This meaning could firstly be its primary meaning, which is the ordinary meaning that a reasonable reader or listener would ascribe to the words or conduct. A court must not only look at what the words or conduct expressly state, but also at any implications that arise from it. 70 Secondly, in appropriate circumstances the court must consider whether the words or conduct may have a secondary meaning. A secondary meaning is usually words or conduct that may appear prima facie to be innocent, but are in fact defamatory due to an innuendo, or a special meaning, which is understandable to the reader/listener. In addition, the words may also be defamatory in their primary meaning, but carry an additional sting due to the innuendo. 71 In a situation where words or conduct are capable of multiple meanings, the court will attempt to determine, on a balance of probabilities, the meaning that the ordinary reader or listener would ascribe to the words. 72 Once the court has determined the primary and/or secondary meaning of the words or conduct, it will attempt to determine if those words or conduct 68 Singh 2014 Obiter 626. 69 Neethling, Potgieter and Visser Neethling's Law of Personality 135. This is judged in an objective manner and means that the injury to a person's reputation must have factually caused a reasonable member of society to think lesser of him; i.e. that his reputation has been lowered in their estimation. can be held to be defamatory. In order to do so the court will determine whether the words or conduct lower the plaintiff in the estimation of rightthinking persons generally. 73 In more esoteric defamation matters which concern words or conduct that might be defamatory to certain subcultures or sectors of society only, our courts will consider whether the words or conduct are defamatory to ʺsubstantial and respectableʺ sections of South African societyprovided their views are not contrary to good morals. 74 A plaintiff will usually be considered to be lowered in the estimation of rightthinking persons generally if the words or conduct reflect upon his moral character 75 or if they subject him to hatred, 76 contempt 77 or ridicule. 78 Could an emoji be considered as defamatory in South Africa? It is submitted that the issue of defamatory matter will be the crux of the law in reaching this conclusion. Any words or conduct can be considered as defamatory in South Africa, provided that the publication of this statement or conduct lowers the plaintiff's reputation. Our courts have already recognised that a digitally altered photo could be considered to be defamatory. 79 It is thus submitted that provided the plaintiff can establish that the publication of the emoji caused his reputation to be lowered, there would be no obstacle to our court's extending our current legal principles to an act of defamation perpetrated via the medium of an emoji.
Below, I canvass some of the issues I believe could arise if a defamation matter involving an emoji arises in a South African court.
The popularity of emojis in modern conversations cannot be denied. While the meaning behind the most popular emojis can be said to be widely understood, the same cannot be said about the meanings behind less popular emojis. An example is the ʺsmiling face with open mouth and tightly shut eyesʺ emoji, which in one study had 54% of respondents state that it was a positive emojiwhereas 44% of respondents labelled it as a negative emoji. If such a matter were to arise in a South African court, it is submitted that the court would attempt to assign to the emoji the meaning which an ordinary reasonable reader of an emoji on that medium would assign to it. This might involve the court's entering an exploration of whether the average Twitter user differs from the average Facebook or WhatsApp user. There have been many overseas studies which attempt to determine the meaning that an average user assigns to emojis. However, it could be argued that these studies are of limited use as they did not contain South African study subjects, who may hold quite different views of the emojis being studied.
One should also bear in mind that different countries and cultures may assign different meanings to the same emoji, and subtleties and different nuances in cultural interpretations may lead to miscommunications if the parties or the persons to whom the potentially defamatory emoji is communicated share different cultures or countries of origin. This would be especially problematic in South Africanot only when South African citizens communicate with people from other countries, but also in communications among South Africans, as we have such a rich and diverse citizenship.
Additionally, it should be noted that the same emoji can bear a very different coded meaning to members of subcultures. For example, a crown emoji can be thought of as a royal symbol to normal users, whereas that same symbol in the sex-trafficking subculture means that the lady who is being trafficked is under the control of a "pimp". 81 It is submitted that the points raised above should not deter our courts from considering the defamatory aspect of an emoji. Our courts should use existing principles and adapt these to online usage customs and to customs among subcultures. Our current principles have been developed over a period of time and are solidly rooted in our Roman-Dutch law and constitutional principles, and can be easily adapted to new problems that are encountered with modern technology.
Most users are also unaware that the senders and recipients of emojis do not always see the exact same emoji on their devicesespecially if each of them uses different devices or operating software/platforms. 82 As some emoji images are considered to be intellectual property, some of them are depicted differently on different platforms. Let us take a very commonly used emoji as an examplethe face with rolling eyes emoji. On an Apple device 81 Goldman 2019  this emoji will appear with the eyeballs mid centre at the top of the eyes with a straight mouth. On a device utilising Google as the platform, this very same emoji will appear with larger eyeballs that are still on the top of the eye, but which are now slightly to the left of the eye instead of the centre of the eye. The mouth is also different and is now not straight, but rather is downturned. This same emoji on a device utilising the Samsung software platform is also very different with much larger and more detailed eyes that are placed differently from the eye locations of the same emoji in both the Apple and the Google devices. The Samsung device emoji also has eyebrows and a mouth that are completely different from those of the same emoji as depicted in the other two devicesmost importantly, its mouth is a smiling one. 83 It is thus possible for the same emoji to elicit different reactions and meanings between senders and recipients who are using different devices. While this is usually insignificant, it can sometimes lead to miscommunication and unintended consequences. 84 If such a scenario should arise in South Africa, it is submitted that our courts would have to carefully consider the fault requirement for defamation and canvass any potential defences that the defendant could raise due to possible contamination of the emoji selected by the defendant (sender) and received by the recipient due to their having different devices or using different software platforms.

Reference to plaintiff
This element is not usually a problematic one as the plaintiff merely has to prove that the ordinary reasonable reader or listener would associate the defamatory words or conduct with them. 85 In the Burrows case, Burrows would have no difficulty in South Africa in showing that the posts complained of referred to her, as the ordinary reader of the posts could clearly conclude from the context that Burrows was the person to whom the posts referred.

Possible defences to defamation in South Africa
There is no closed list of defences which a defendant may raise to a defamation claim in South Africa. The most commonly raised defences to defamation claims in South Africa may be divided into two categories - firstly those which attempt to rebut the wrongfulness element of the defamation, and secondly those which attempt to rebut the fault element of the defamation.
A defendant who is attempting to rebut wrongfulness could raise any of the following defences: Truth for the public benefit, 86 fair comment 87 or qualified privilege. 88 In addition, if the defendant is a media entity, it may also raise the defence that the publication was reasonable. 89 A defendant who is attempting to rebut the fault element may raise any of the following defences to show that he lacked the intention to defame the plaintiff: mistake (of both law and fact), 90 jest 91 or rixa (provocation). 92

Possible South African court awards
While most plaintiffs consider an award of damages to be the only or the main outcome of a defamation action, our courts also have the ability to award other orders.
Our courts have approved interdicts to remove offensive posts from social media sites. 93 Our courts have also held that they have the ability to order the defendant to tender an apology to the plaintiff 94 and that an apology tendered in the same medium in which the plaintiff was defamed could completely clear the plaintiff's name. 95 It is thus submitted that should a defendant find himself in Houdaʹs position in South Africa, he should consider the option of retracting his tweet and apologising for the tweet on Twitter. Even if such an action does not completely exonerate him, the court will take this retraction and apology into account when considering an award of damages.

Conclusion
It is submitted that our courts in South Africa would be willing to hold that an emoji could be considered as defamatory in our law. In order to reach this 86 Johnson  150, 195, 197, 199, 202, 203. 95 Isparta v Richter 2013 6 SA 529 (GP) para 41.