This paper reports the first case in which a linguist served as an expert witness in Hong Kong, a former British colony that has operated as a special administrative region of the People’s Republic of China (PRC) since 1997. The dispute was on the meaning of the political slogan ‘Liberate Hong Kong, Revolution of Our Times’, which was widely adopted during the 2019–2020 protests. The keywords ‘liberate’ and ‘revolution’ are smoking gun evidence for the prosecution in a large cluster of cases that involve sedition law and national security offences. Section I of the paper provides background information about a case the author was involved in, which was concerned with whether the slogan was seditious. Section II describes the analysis conducted, which concludes that the slogan as a whole refers to a need to rectify a problem and to return to the original, a more desirable state of affairs for Hong Kong, without specifying what problem there is and what the desirable state of affairs looks like. Section III highlights some critical issues in the analysis, discussing challenges faced and ethical questions for the expert witness. Section IV is a postscript that briefly describes the outcome of the case.

1 The Case

In 2019, massive protests broke out in Hong Kong to oppose a proposed extradition bill [50] that would allow fugitives to be extradited to Taiwan and mainland China. The fear, fury, and frustration that the proposed bill generated needs to be understood in the context of another incident that took place a few years prior: five booksellers connected with a dissident bookstore in Hong Kong disappeared one by one in Hong Kong and Thailand, with there being no immigration record of them exiting the territory, and reappeared in mainland China confessing to various crimes on national television [31]. There was concern that the proposed bill did not have sufficient human rights safeguards [4], and would speed up the erosion of freedoms guaranteed by One Country, Two Systems. One Country, Two Systems is a constitutional principle whereby under China’s rule, the former British colony would enjoy a high degree of autonomy and retain its own systems and ways of life. The Basic Law, Hong Kong’s constitutional document, gives effect to this principle. Under Article 8 of the Basic Law, Hong Kong retains its common law legal system, though the power of interpreting the Basic Law is vested in the Standing Committee of the National People’s Congress (NPCSC).

After more than a quarter of its population stormed the streets, including a rally attended by two million people, the extradition bill was withdrawn. However, the protests persisted as people were angry about the use of force by the police and demanded amnesty for arrestees. An even more expansive law was then imposed on Hong Kong, bypassing its legislature altogether. The National Security Law (NSL), enacted in June 2020, has successfully quashed the protests and dramatically increased the risks of dissent [7].

More than 10,000 people have been arrested in the Anti-Extradition Law Amendment Bill Movement (hereafter anti-ELAB). Many are now being put on trial, on charges ranging from illegal assembly and rioting, to language crimes such as incitement and sedition. All these cases, including the one reported in this paper, have been tried without a jury. One of the most contentious slogans used in the protests is 光復香港 時代革命 [gwong1 fuk6 hoeng1 gong2 si4 doi6 gaak3 ming6] (Liberate Hong Kong, Revolution of Our Times).

The first court case that dealt with the slogan was HKSAR v. Tong Ying Kit [21], which was also the first NSL case tried in Hong Kong. With a flag bearing the slogan hoisted at his back, a motorcyclist collided with a police checkline, injuring three policemen. The defendant was charged with incitement of secession and committing terrorist activities. The slogan informs the determination of these charges, both of which are NSL offices (Article 20 and 24 respectively) and carry a maximum penalty of life imprisonment. The prosecution called a historian to give his expert opinion on the slogan, while the defence called a professor of politics and public administration and a professor of journalism and communication to do the same. Based on historical uses of some of the keywords in the slogan and on his analysis of the contemporary context, the historian opines that the meaning of the slogan is “to cause the consequence of separating the territory of residence from the State sovereignty; in the context of Hong Kong’s political language, these words were raised necessarily for the objective of separating the HKSAR from the PRC” [21: 115]. Drawing from social scientific methods, the defence experts demonstrate that keywords of the slogan have been recontextualised in the last few years and the slogan has taken on a range of different meanings.

In another case that was tried at about the same time, HKSAR v Tam Tak Chi [20], the meaning of the same slogan was also contested. Even though this case involves the same slogan, the charges are based on colonial-era sedition law instead of NSL. The historian who testified in Tong Ying Kit submitted his opinion for the prosecution, while I was called by the defence as an expert witness, along with the professor of politics and public administration who also served in the case above. To my knowledge, this is the first time in a Hong Kong courtroom where a linguist (other than handwriting experts) served as an expert witness. Using linguistic methods, my analysis leads to the same conclusion as the defence experts in Tong Ying Kit: that the slogan has divergent and indeterminate meanings.

Tam Tak Chi is a DJ and a social activist who faced 14 charges based on his participation in and speeches he made during various public meetings between January and July of 2020. He was charged with multiple counts of the offence of uttering seditious words, contrary to Sect. 10(1) (b) of the Crimes Ordinance, Cap. 2000. Section 10(5) defines seditious words as “words having a seditious intention”. One key question for the court was whether the slogan is seditious. Since the slogan was widely adopted during the protests, and many arrestees are still pending trial, the determination of its meaning could affect the fate of many more defendants.

2 The Analysis

My role as an expert was to explain to the court, from a linguistics perspective, how the meaning of certain phrases is to be ascertained, and more specifically, what the meaning of the slogan is. I was not asked to determine what the defendant meant when he uttered the slogan in different occasions. I was also invited to comment on a report produced by the prosecution’s expert witness.

The analysis presented below does not replicate my expert report but has been rewritten for the academic audience. In the following analysis, I have adopted Shuy’s inverted pyramid [44] by moving from larger to smaller context. At the top of his inverted pyramid is speech events, followed by schemas, agendas, speech acts, conversational strategies, with grammar and lexicon at the bottom. The ingenious design of the inverted pyramid is that it corrects the temptation to take individual words as smoking gun evidence without regard for how they are used in context.

Although the analysis below follows the inverted pyramid approach, it takes place at a more macroscopic level than that of Shuy’s illustrative case [44], because my task here is not to analyse a particular utterance, but to locate the ordinary meaning of a phrase used by many people at a particular historical juncture. This approach may be particularly valuable in the current case, as keywords in the slogan—‘liberate’ and ‘revolution’—are the perfect smoking gun. For example, veteran Hong Kong politician Maria Tam has said that 光復 (liberate) in Chinese implies separatism and therefore contravenes the Basic Law [57].

2.1 Speech Events

Speech events, à la Dell Hymes, are the basic analytical unit of a verbal interaction. Speech events are governed by social norms, which frame people’s expectations about what happens in the interaction. Before I get to the nature of the speech events in which the slogan was used in Hong Kong, I used Google Trends to get a broad sense of interest in the slogan over time. Figure 1 shows that the slogan was first used as search term in 2016, and became a popular search term in June 2019.

Fig. 1
figure 1

Frequency of the political slogan appearing as a search term on Google (Hong Kong) between 2004 and May 2021, as indicated on Google Trends

Full size image

While the plot does not tell us anything about the meaning of the slogan, it provides some guidance on locating speech events that matter. The plot triangulates journalist reports that the slogan first appeared in February 2016, when Edward Leung adopted it in his election campaign in the Legislative Council by-election. He chanted the slogan on stage and printed it on his campaign flyers. Leung was a member of Hong Kong Indigenous, a localist political party.

The slogan, authored by Edward Leung’s campaign team, contains four compound words in Chinese combined in a unique manner that has not appeared before. It is almost certain that its subsequent usage in Hong Kong was an intertextual reference rather than a novel creation arrived at independently. The slogan would later appear in other election campaigns in Hong Kong. It was adopted by the localist party Youngspiration in the Legislative Council election of September 2016, whose representatives vowed to “continue revolution of our times that liberates Hong Kong” (my translation) [27]. It was also used as a campaign slogan by some candidates in the district council election of 2019. Despite the clear intertextual linkage between Edward Leung’s original slogan and these subsequent speech events, it is debatable whether these later speakers were just animatingFootnote 1 the original slogan, or whether they may be authors who recontextualised and reappropriated the slogan for new purposes.

Regardless of what Leung intended his slogan to mean in 2016, the slogan did not generate widespread interest at the time, as the Google Trends graph attests. It gained huge popularity during the anti-ELAB protests, appearing on posters, graffiti, social media posts, and other protest-related promotional materials. It was also chanted by tens of thousands of protestors on the street. Again, despite the intertextual relationship with Edward Leung’s election slogan, we cannot simply assume that everyone who used the slogan in 2019 shared all of Leung’s ideologies. Even if people were aware of the intertextuality, they might not know the intent of the original speaker.

To explain how the same slogan may be appropriated and recontextualised, we can look at Donald Trump’s election slogan in 2016: ‘Make America Great Again’. If we see someone wearing a red MAGA cap today, we infer that they are a Trump supporter. However, the same slogan had been used by republican Ronald Reagan (‘Let's Make America Great Again’) in his successful 1980 presidential campaign and by democrat Bill Clinton in his 1992 presidential campaign. Its earliest recorded use traces back to 1940 but few people are aware of it today. The fact that the same slogan was adopted by democrats and republicans shows that political slogans are often vaguely formulated, such that it is capable of expressing different ideologies and political agendas. Change in meaning does not have to come from wilful appropriation—there are also plenty of examples where popular usage of a word or a phrase is based on a misinterpretation of the original (such as the popular expression ‘the road less travelled’, taken from Robert Frost’s Poem, “The Road Not Taken”, which does not imply that his choice entails any bravery or fearlessness). Once slogans enter the public imaginary, they are often reappropriated for different purposes. The same is true for songs. One example in Hong Kong is the song Under the Vast Sky by the band Beyond, which has been used in both anti-government protests and pro-establishment rallies [30]. Its lyrics talks about the determination to pursue one’s dreams. We can convey a diverse range of meanings through words that are not our own, especially when they are vaguely formulated and resonate with us emotionally.

In analysing speech events, we need to consider the power of genre, which affects how people interpret words that they hear. Political speech tends to be hyperbolic and metaphorical; that is to say that speaker meaning often diverges from the literal meaning of their words. This is why in Watts v. United States (1969), the US Supreme Court notes that “The language of the political arena … is often vituperative, abusive, and inexact.” Literal meaning is not necessarily the most frequently occurring meaning (when someone says ‘I see’, they usually refer to their understanding rather than visual perception). Words used in political contexts cannot be taken too literally, as many of them are also metaphorical expressions. For example, national anthems that portray an armed struggle may be used as political songs, but in context they are rarely literally a call to arms every time they are sung.

Just as the literal meaning is not necessarily the most frequently occurring meaning, the most frequently occurring meaning may not be the intended meaning adopted by a particular speaker at a particular time. The meaning of the slogan in each occurrence can only be determined contextually, paying attention to contextual cues such as tone of voice, facial expressions and gestures, speaker identity, audience characteristics and reaction, time, location, and shared knowledge.

In sum, the slogan in question has been used in two prominent types of speech events in Hong Kong: election and social movement. We may ask whether the same slogan diverges in meaning in these different speech events. While there is clearly an intertextual relationship among its instances of use, only closer examination could reveal whether the slogan has been recontextualised and has transformed in meaning.

2.2 Schemas

Shuy defines schemas as how participants think about what is being talked about [44]. This may sometimes be gauged through analysing the verbal and non-verbal context of an utterance. I was also able to draw from ethnographic and journalistic interviews that directly probed into the thoughts of speakers, evidence that may not be available had the relevant speech events not drawn wide attention.

A natural starting point is how the slogan was framed in its first appearance. In his 2016 campaign speech, Edward Leung describes ‘Revolution of Our Times’ (時代革命) as a revolution for Hong Kongers to take back their freedom and to take back a Hong Kong that belongs to them. He said this was an era that belonged to people who believe in freedom, embrace freedom, and would use blood and sweat to earn freedom. Although his slogan does not articulate specific political demands, the rest of his speech suggests that this ‘revolution’ has to do with regaining freedom and autonomy. He did not specify the means through which he would achieve this goal, though he did mention in press interviews that his resistance to oppress had “no bottom line” [33]. He did advocate for Hong Kong independence earlier in his political life, but declared in 2016 when entering the election that he would support the Basic Law and the HKSAR.

It is noteworthy that Edward Leung participated in a series of local movements that also bore the name of ‘liberation’ in 2015. These movements, such as Liberate Sheung Shui and Liberate Sha Tin, targeted various social problems that local districts in Hong Kong were facing. In fact, in the same campaign speech and same promotional materials, Leung referred to ‘Liberate Hong Kong’ (光復香港) as well as those other ‘liberation’ movements without any contextual cues that signal any divergence in intended meaning (see further investigation of the use of the term ‘liberate’ in contemporary Hong Kong under Grammar and Lexicon below). It is reasonable to assume that the meaning of ‘liberate’ is not different in Liberate Hong Kong and Liberate Sheung Shui, unless there is conflicting evidence suggesting otherwise.

Consistent with this interpretation, the co-author of the slogan said in an interview that ‘liberation’ referred to “previous ‘reclaim’ demonstrations targeting parallel traders at the border”, but the slogan expanded the scope of ‘reclaim’ to mean “restoring the lost glamour of Hong Kong”. It also used ‘revolution’ to mean letting young people lead social changes [57]. There is a lack of specificity as to the nature of the glamour that was lost, and what social changes should happen.

Although we do not have to take his words at face value, an important context for this slogan is that it was adopted in a campaign during a lawful election. Regardless of Leung’s political ideology and long-term agenda, which are not the subject of this investigation, his immediate goal at the time of the campaign was to win votes. He compared votes with bullets, and invited his audience to use a non-violent means to create change—by casting their votes for him (my translation: “we do not have guns or cannons, but the votes in our hand are a revolution of the political map”). Leung lost the election and ended up serving a six-year jail sentence for rioting and assaulting a police officer. He was still in prison when the anti-ELAB protests happened.

The 2019 district council election took place in November, when the protests were still at a peak. After an election candidate made a Facebook post containing the slogan, an election officer—who was charged with vetting candidates for their allegiance to the city and to its constitution—asked her what she meant [17]. She replied (my translation): “The high degree of autonomy, fifty-years no change, and other rights and freedoms, as promised in the Basic Law, are receding quickly. The slogan expresses hope that Hong Kong can reinstate the core values that contributed to its pride and its success. It also expresses hope that the Central Government will keep its Basic Law promises, reform the political system, and allow Hong Kong to move forward.” [17] The officer was satisfied with her answer and she proceeded to win the election. The fact that the officer queried about what she used the slogan to mean, instead of directly deeming the slogan to be subversive, shows that the slogan has an indeterminate meaning. The fact that her answer was accepted suggests that her interpretation was a reasonable one. Her interpretation conveys a desire for preserving the rights and freedoms that Hong Kong has enjoyed, under the current One Country, Two Systems political framework.

Another candidate in the same election also explained his understanding of the slogan [56]: “He thinks that every Hong Konger understands the slogan in their own way. For him, liberating Hong Kong is to restore the system of a high degree of autonomy as promised in the Sino-British Joint Declaration, unlike the current system where the mainland takes the lead on everything. Revolution of our times is the citizens’ radical but non-violent way of resistance, to regain the power to govern ourselves, to replace antiquated systems that should not be kept in 2019…” [56; my translation]. Again, although there is a clear yearning for more political autonomy in Hong Kong, given that such autonomy is actually guaranteed by law, it is hard to argue that his interpretation suggests any intention to topple the current political regime.

In the context of social movement, journalistic interviews held during the anti-ELAB protests in 2019 provide some clues about how protesters understood the slogan. In its television show Hong Kong Connection, RTHK interviewed some protesters to see what they thought the slogan meant. One of the interviewees said, “Liberating Hong Kong refers to Hong Kongers having our say and our rights back.” Another said, “Revolution is about reforming or breaking through one’s way of thinking.”

The views expressed by interviewees in all these reports show that what they understand to be ‘revolution’ (革命) is not so different from ‘transformation’ (變革) or ‘reform’ (改革), and they embraced the slogan to mean that they want to maintain their ways of life, their rights, and their freedoms. These journalistic interviews corroborate with results of an onsite survey conducted between June and August 2019 involving 12,231 respondents. Lee et al. (2019) found a lack of uniformity in the protestors’ views, but on the whole 88.2 percent of the respondents find the slogan representative [24]. The researchers conclude that the slogan acts like an empty signifier “with no particularly defined meanings” [24: 21] when compared with other slogans with extraordinarily high recognition:

In fact, protesters of different age groups and political affiliations held markedly different views regarding to what extent these two slogans could represent the movement. Among the protesters who were aged 30 or below, 85.1 percent thought that “Hongkongers, add oil”Footnote 2 could represent the protests, while 95.4 percent of the older protesters thought so. Meanwhile, younger protesters had stronger identification with “liberate Hong Kong, revolution of our times”: 93.5 percent thought this slogan could represent the protests, while 77.5 percent older protesters thought so. This difference in identification between “Hongkongers, add oil” and “liberate Hong Kong, revolution of our times” was also found in protesters with different political orientations. Significantly more “moderate democrats” than “localists” thought “Hongkongers, add oil” could represent the protests. Yet the “localists” identified with “liberate Hong Kong, revolution of our times” more than the “moderate democrats.” [24: 22]

Although the slogan was embraced more by younger protesters than older protesters, and more by localists than moderate democrats, the fact that most protesters identified with it means that its meaning is vague enough to appeal to people with a wide spectrum of political views.

2.3 Agendas

Shuy uses the term ‘agenda’ to refer to what participants contribute to what is being talked about [44]. His framework suggests assessing agendas through topics they introduce to a conversation and responses they provide to topics introduced by others. This is the level where Shuy analyses criminal intentionality in speech crimes cases. Agendas are not easy to determine in the current case, as the slogan was often used in a standalone manner rather than as part of a conversation, and uniformity of intention was unlikely as it was chanted millions of times by tens of thousands of people.

Generally speaking, election candidates vow to contribute to their goal by participating in democratic politics. Regardless of the language used, election campaigns often contain promises or desires to claiming or reclaiming a seat, or to gain majority control as a party. Trump, for example, has vowed to ‘win back’ the White House in 2024. The term ‘liberate’ is also used in election slogans [54] in contemporary Taiwan (as in Liberate Yi Lan). Similarly, in the election context, ‘revolution’ does not always refer to overthrowing a regime. In the US, democratic candidate Bernie Sanders ran a presidential campaign with the theme of “Our Revolution”. His 2016 book, which carries the same title, discusses his ambitions on climate change, free college tuition, and income inequality, etc. [40]. The synopsis starts with the sentences “This is your country. Take it back.”, which conveys a sense of reclaiming which is not dissimilar to the slogan in question. Change is the essence of politics; clearly, not all changes are seditious or subversive. Leung’s use of the slogan seems to have a similar rhetorical effect. He has explicitly mentioned using votes as a means of revolution; he may well have other agendas but there is no clear discourse evidence for them in his election speech and campaign materials.

Let us now turn to the anti-ELAB protests. In a social movement, protestors participate in rallies to draw attention to their cause and put pressure on their political leaders to create change. Slogans are a way of articulating their purpose and of demonstrating unity. Political slogans do not always represent agendas accurately or literally: one good example is ‘Defund the Police’. Protestors chanting the slogan rarely believe that no policing should exist.

The slogan in the current case often appears as a standalone text on posters; when uttered, it is often chanted with other slogans. The 2019 protests started with a clear demand: withdraw the extradition bill. As the social movement escalated, by the end of June the most prominent slogan had become: ‘Five Demands, Not One Less’. The five demands include: full withdrawal of the extradition bill, commissioning an inquiry into alleged police brutality, retracting the classification of protesters as rioters, amnesty for arrested protesters, and universal suffrage. These demands are entirely compatible with the existing political structure of Hong Kong. This is not to say that no protester at the time had a secessionist agenda. There were small groups of people who waved a Hong Kong Independence flag and were shouting related slogans during the protests. The question here is whether the slogan ‘Liberate Hong Kong, Revolution of Our Times’, which a large majority of protestors [see 24]) thought was representative of the movement, was perceived ordinarily as carrying a secessionist or a broader meaning.

As Lee et al. (2019) work suggests [24], the slogan represents a higher level of radicalisation than ‘Hong Kongers, Add Oil’. The timing of its appearance in the protests may be helpful in deciphering its significance. According to a police report in the current case, the slogan was first heard in the protest on 21 July 2019. That was also the day when hundreds of thugs attacked commuters and protesters with steel rods and rattan canes in the Yuen Long subway station with little interference from the police. This incident shook Hong Kong. The slogan’s explosion of popularity after the incident could be taken to reflect the rage that people felt. People wanted freedom from fear, and they wanted change. The slogan is sufficiently broad to capture the emotion.

What further complicates the analysis of agendas in the context of the social movement is the participant roles that protestors played. Most protestors chanting the slogan were animators rather than authors, in Goffman’s terms [9]. Slogans are often chanted in a dialogic mannerFootnote 3 (known as call and response in musical performance and in some oral traditions) during Hong Kong protests: someone in the crowd (who may or may not have a microphone and may or may not have any official leadership role) takes the lead to shout the first half of the slogan, and the rest of the crowd responds by providing the second half to complete it. Table 1 shows some common adjacency pairs during the 2019 protests:

Table 1 Common adjacency pairs during the 2019 protests in Hong Kong
Full size table

The great majority of speakers in these adjacency pairs were not involved in formulating the slogans. It is exactly because the slogans have become formulaic that such a ritualised turn-taking pattern is possible. Although the interlocutors in all these examples may be considered animators, the initiator (A) in the table has some freedom to decide which first pair parts they will conjure up; their options are however limited by the repertoire of slogans the crowd (B) has. The collective animation forms a powerful spectacle as large crowds appear to be speaking with one voice. Iterative completions of adjacency pairs help to construct and reinforce group identity.

In sum, we cannot safely determine people’s agendas by their use of political slogans, especially in the contexts of election and social movement. What complicates the interpretation further is the performative use of slogans in a social movement as a demonstration of unity. Slogans used in the anti-ELAB protests could be used as a way of channelling frustration and anger without actual consensus about an agenda.

2.4 Speech Acts

Shuy, following Austin, understands speech acts as how participants convey their contributions [1, 44]. Speech acts are ways of getting things done with language, such as admitting, offering, and reporting. One key aspect of the speech act theory is the conceptual differentiation between the illocution and the perlocution: the illocution reflects speaker intention, and the perlocution is the addressee uptake. The speaker is not in full control of the perlocution.

Let us consider the most prominent slogan in the anti-ELAB protests: ‘Five Demands, Not One Less’. It is a commissive act that contains a pledge or a threat: if the five demands are not met, the protests will continue. Even though there is no explicit addressee, the slogan is directed at the government, because only the government can fulfil the demands made. The implicit threat simultaneously acknowledges the authority of the government to right the wrong. Another popular slogan in the protests was ‘Hong Kongers, Add Oil’, which is a directive that commands and inspires the addressees—other protesters—to continue with their effort.

Returning to the slogan in question: ‘Liberate Hong Kong, Revolution of Our Times’. The explicit act of the slogan appears to be commissive: it declares a desire or intention to ‘liberate Hong Kong’ by supporting or participating in ‘the revolution of our times’. Who is the target addressee for this slogan? In the election context, the slogan indicates a commitment or a promise by the candidates to their voters. If the perlocutionary effect had worked as intended, voters might become convinced and go and cast a vote for the candidate. In the mass rallies, the addressee is less clear—the intended addressees are likely to include the government and international media, as well as the other participants of the protests. Unlike the five demands, this slogan has no specificity about what it means to liberate Hong Kong or what revolution should take place. The commissive act articulates a determination to free Hong Kong from its current problems, with the means to achieve the goal left unsaid.

The explicit speech act is not always the intended act, as speech acts can be performed implicitly and indirectly (e.g., ‘Can you pass the salt?’ is not a question that probes the addressee’s ability to pass the salt but a request for them to do so). As discussed above, hyperbole is the hallmark of political speech. The implicit act expressed by the slogan in the social movement context may well be the expression of hopefulness and encouragement to other participants, as well as iteration of determination to the government and to international media. However, as discussed above, since participants’ agendas are diverse and conflicting, they might have intended different effects from the same speech.

2.5 Conversational Strategies

Conversational strategies are about how participants try to influence each other using various communication techniques. Hyperbole, as mentioned above, is a well-known technique in political speech. Another strategy, commonly seen in election campaigns and political slogans, involves the strategic use of linguistic indeterminacy.

The slogan in question is linguistically indeterminate in two ways: it is vague and general. Vagueness refers to categories having fuzzy boundaries: for example, there is no clear boundary between when ‘red’ ends and when ‘orange’ begins. Generality is the opposite of specificity. The general term ‘mother’ includes specific references to mothers and mothers-in-law [29].

The slogan is general in that there are different types of “liberation” and “revolution”; collocation could potentially help disambiguate, but in the case of this slogan the collocative words are not particularly helpful. ‘Liberate’ and ‘revolution’ are also vague, in that there are peripheral members of the categories that constitute borderline cases. There is no clear boundary between an advancement in digital technologies and a digital revolution, just as not everyone agrees that abortion rights constitute a kind of liberation.

Under-specification can be useful for mobilisation. The 2019 protests were known for having no formal leadership or any hierarchical structure. Slogans were used to unify disparate groups supporting a range of tactics and with varying political beliefs and levels of radicalization—those slogans were about creating and sustaining solidarity above all, and therefore their vagueness and open-endedness were functional. As a protest slogan, its vague and general framing boosted its popularity, because protesters from a wide political spectrum can interpret it in ways that suit their own ideology.

More importantly, the legal distinction of whether a word or a phrase is seditious does not map neatly onto semantic boundaries. The words ‘liberation’ and ‘revolution’, whether in Chinese or English, do not consist of polysemous senses differentiated by whether seditious intentions exist. Such intentionality simply cannot be ascertained by the literal meaning of the slogan.

2.6 Grammar and Lexicon

The Chinese slogan consists of four compound words: 光復 (liberate), 香港 (Hong Kong), 時代 (era), 革命 (revolution). The terms 光復 (liberate) and 革命(revolution) are the most controversial. A three-prong approach was used to study each keyword. First, I conducted a quantitative survey on Google Ngram and Google Trends, which sketch the frequency of usage of these terms in a time series, allowing for a glimpse of the connection between these terms and socio-political events at the time. Secondly, dictionary definitions are researched and compared. Finally, I performed a qualitative analysis of the recent use of these terms to gauge their semantic range from authentic usage.

Google Ngram is based on roughly 5 million, or 4%, of all books ever published. The Chinese subcorpus on Ngram Viewer contains 13 billion Chinese words. There are obvious limitations with the Ngram data, considering that it consists of only books (and therefore may not contain the most recent usage), and that its Chinese subcorpus cannot differentiate geographical differences in usage (such as that between Hong Kong and other Chinese speaking regions). On the other hand, Google Trends has data, between 2004 and present, of frequency of the keywords used as a search term. Because of the restriction of its operation, Google has relatively limited data from mainland China; however, the system does have data of search frequency specifically for Hong Kong.

2.6.1 光復 (liberate)

Let us start with the term光復 (liberate).

Google Ngram as captured in Fig. 2 above shows that after a huge spike in usage in the 1940s (corresponding with the Japanese invasion and the Chinese civil war), the term has enjoyed limited but consistent usage since.

Fig. 2
figure 2

Frequency of the word 光復 (liberate) appearing as a keyword in books published between 1800 and May 2021, as sampled by Google Books

Full size image

Referring to Fig. 3 above, as a search term in Hong Kong on Google, 光復 (liberate) saw some interest in 2012 and 2015, and a much more significant spike in 2019 when the social movement broke out.

Fig. 3
figure 3

Frequency of the word 光復 (liberate) appearing as a search term on Google (Hong Kong) between 2004 and May 2021, as indicated on Google Trends

Full size image

Despite the limitations of these datasets, these quantitative data show that the usage and interest of these keywords are not stable and unchanging over time but correspond with sociopolitical events of the time. The timing of sudden increases in usage or interest helps to corroborate the qualitative analysis on how the keywords go through recontextualization. The Google Trends graph also helps ensure that my qualitative analysis does not miss out important contexts of usage.

I then looked up the term in eight contemporary Chinese dictionaries,Footnote 4 but in the interest of space the results will not be reproduced here. Perhaps the most obvious observation from these dictionary entries is their similarity. It is however important to note that dictionaries cannot be understood as aggregated sources of independent data because dictionary-makers routinely refer to one another’s work in the editorial process. The dictionaries provide a consistent understanding of the term, suggesting that its base meaning is restore/reclaim (恢復/收復). The term may be used to refer to the reclamation of land/sovereignty (收復失去的土地), just as it could be used to talk about more abstract kinds of reclamation, such as restoring the original state (恢復本來的情況) and restoring light in darkness (在黑暗中重見光明).

Next I examined the range of meanings 光復 (liberate) can carry by examining its recent usage in Hong Kong. My survey of usage used the Internet as a database (through the mediation of Google and Yahoo as search engines), using the keywords as search terms. The purpose of the qualitative analysis is not to establish the relative frequency of meanings ascribed to these terms, but to investigate the range of meanings that are commonly carried by these terms in their contemporary usage in Hong Kong.

One limitation of using the Internet as a database of authentic language usage is that the searchable content consists almost exclusively of written language, and that while it covers a wide range of genres such as newspaper reports, government press release, discussion forums and personal blogs, it does not cover genres that are used offline (such as conversations that happen in elevators, gossiping in coffee shops, or exchanges between vendors and clients). However, since my investigation here focuses on semantic range, this limitation means that the actual range can only be larger than what could be sampled from the Internet and therefore does not invalidate the findings.

The qualitative analysis shows that ‘liberate’ (光復) has been used in the sociopolitical context in Hong Kong, with a meaning that has little to do with secessionism. Recent popular use of the term “liberate”(光復) in Hong Kong began on 15 September 2012, in a social movement that the protesters called Liberate Sheung Shui. This was a social movement that targeted parallel traders from mainland China who obstructed roads and whose activities pushed up rent and caused inflation, affecting the livelihood of local residents in the Sheung Shui neighborhood. This is how the Legislative Council at the time summarised the protest (my translation):

In September 2012, some citizens initiated the “Liberate Sheung Shui Station” movement online, protesting against parallel trading activities that affect residents’ living environment in the Northern District. The government announced 6 anti-parallel trading measures on 18th September 2012, and declared that it will step up efforts in combating such activities to protect the daily life of Northern residents from disturbances. [28: 1]

This portrayal is consistent with the newspaper reports at the time. The Immigration Department and the Police jointly mounted an anti-illegal employment operation codenamed “Windsand” [22] and raided blackspots for parallel goods trading. The government response was praised by a senior representative from the mainland Chinese government [see 38]. In short, the government responded to these ‘liberation’ activities by acknowledging and rectifying the social problems that they targeted. Sing Tao Daily even ran a story about the Hong Kong government’s new policies against parallel trading [46], with a headline that describes how the government is ‘liberating’ the district (my translation): “Six Measures by the Hong Kong Government to ‘Liberate the Northern District’”.

Liberate Sheung Shui was not a singular incident. There was another surge of movements in 2015, including Liberate Tuen Mun, Liberate Yuen Long, and Liberate Sha Tin, focusing on parallel trading activities that disturbed local residents in different neighbourhoods in Hong Kong. Even after the anti-ELAB protest broke out in 2019, there were local protests with the ‘liberation’ theme that continued to focus on parallel trading. A poster for a Liberate Sheung Shui event on 13 July 2019 contains these goals (my translation): Kick Out Parallel Trading, Restore Quiet Neighbourhood [16]. Interestingly, an organiser of one of the protests in 2019 expressed hope that people with alternative intentions do not show up to “cause chaos”, showing awareness that people who stand behind a ‘liberation’ slogan may have divergent agendas and intentions [13].

Apart from parallel trading activities, there were also ‘liberation’ protests that focused on other social issues. Liberate Tung Chung happened in 2018, where protesters targeted unlicensed tour guides bringing “unknown” tourist groups to the Tung Chung neighbourhood via the newly opened Hong Kong–Zhuhai–Macau Bridge [12], overcrowding public transportation and causing hygiene problems. The police issued a no-objection notice to the protest.

Another notable ‘liberation’ movement in Hong Kong is Liberate Tuen Mun Park, first held in July and September 2019. This movement was a protest against middle-aged women ‘dama’ (literally ‘big mama’) singers from mainland China, who regularly appeared in the Tuen Mun park to sing and dance in allegedly indecent clothing to earn tips from elderly men. Their noise and allegedly erotic manner disturbed local residents. It is notable that the protesters were given a go-ahead by the police in advance. As a response to the protest, the District Council shut down designated public areas for performers in the park [15]. These movements also triggered healthy public discussions about proper use of public space [14]. According to media reports, Liberate Tuen Mun Park activities persisted in 2020 [36]. These cases show that the term 光復(liberate) is not tied to a particular social issue but can be used flexibly to talk about the rectification of different social issues.

The term has also been used in the election context. In an interview in 2015, the then district counsellor who represented the North District used Liberate the Northern District as a way of articulating his goal of winning more seats for his party in a lawful election [19]. Here the problem that he wanted to rectify was that his party used to be but was no longer successful in the North District elections, and he planned to win back more seats for his party by efforts such as community service, revitalising his party through renovating its headquarters, establishing a creative media unit and systematizing fundraising. It is of interest to note that his use of ‘Liberate the Northern District’ diverges in meaning from the same phrase used by Singtao daily mentioned above, as the speakers have different ideas about the problem that North District should be ‘liberated’ from.

In sum, the basic meaning of 光復 (liberate) in these contexts is ‘restore’, which broadly refers to rectifying a problem and returning to the original, more desirable state of affairs (such as a peaceful and orderly Sheung Shui). Many ‘liberation’ movements were held in various neighbourhoods in Hong Kong between 2012 and 2020, including Sheung Shui, Sha Tin, Yuen Long, Tung Chung, Hung Hom/To Kwa Wan, and Tuen Mun Park. These movements focused on various social issues ranging from parallel trading, unlicensed tourism, to noise problems. Placed in these contexts, the phrase Liberate Hong Kong simply means liberating of the Hong Kong society from an unspecified set of social and political problems.

In their unpublished report produced for the court, the prosecution expert posits that 光復 (liberate) means “reclaiming a demised country or retaking a lost territory”. This restrictive interpretation is at odds with my findings. As evident in both dictionary definitions and documented usage, his interpretation cannot be the only possible sense of the term. The obvious solution is to adopt a broader and more basic meaning, as stated above: to rectify a problem and to return to the original, a more desirable state of affairs. If 光復 (liberate) was perceived to have the restrictive sense that the prosecution expert posited, it would be surprising that the Hong Kong government did nothing between 2012 and 2020 to counter such subversion and separatism, but has even taken action to address and alleviate issues raised in these movements.

Finally, it is also of interest to note that the term 光復 (liberate) in the slogan has been differently translated into English as ‘liberate’, ‘restore’, ‘reclaim’ and ‘revive’ [see e.g., 59], which also adds weight to analysis that the term has been interpreted to have a wide semantic range.

2.6.2 革命(revolution)

With the term 革命 (revolution), I will again begin with a quick survey of its usage pattern. Google Ngram data, as captured in Fig. 4, show that the term 革命 (revolution) is more frequently used than 光復 (liberate), which is not surprising. Its frequency of usage peaked in between 1940 and 1980s, with smaller peaks in the 1920s and around 1910.

Fig. 4
figure 4

Frequency of the word 革命 (revolution) appearing as a keyword in books published between 1800 and May 2021, as sampled by Google Books

Full size image

As Fig. 5 shows, the interest in the term as a Google Search Term peaked in 2011 for Hong Kong, during Arab Spring that broke out in Egypt, Tunisia and beyond. The other recent peaks happened in October 2014, June 2017, and August 2019. While the October 2014 peak corresponds with the ‘Umbrella Revolution’ and the August 2019 corresponds with the anti-ELAB protest, it is less clear what prompted the June 2017 peak. There is of course the 1 July rally in Hong Kong every year, but the 2017 rally had one of the smallest attendance in recent years. It could be related to the Toilet Revolution (廁所革命) that President Xi Jinping proposed in 2015 and pushed forward again in June 2017 [58].

Fig. 5
figure 5

Frequency of the word 革命 (revolution) appearing as a search term on Google (Hong Kong) between 2004 and May 2021, as indicated on Google Trends

Full size image

Dictionary definitions of the term 革命 (revolution) are again rather consistent. It denotes major change. Such change may be societal, political, economic, or cultural. It can also be used metaphorically to describe a kind of revolutionary consciousness or aptitude.

In their report, the prosecution expert posits that from ancient to contemporary times, the term革命 (revolution) has not been applied differently: that revolution “without exception causes the end of the current political regime, replacing it with a new regime established by the revolutionaries.” (para. 47; my translation). This restrictive and singular definition is incompatible with some of the usage I have identified. It ignores the power of collocation in shaping the meaning of the term. 革命 (revolution) is usually used in the political context, but can also be used in social, economic, and intellectual contexts, such as the Scientific Revolution, the Industrial Revolution, and the sexual revolution in the 60s in the US. Since the term basically denotes major change, the kind of change it refers to depends on the attributive adjective that modifies it.

Contrary to the assertion that any revolution aims to overthrow the government, there are modern examples in Chinese usage where revolutionary movements are patriotic, such as the Poetry Revolution in 1899 and the Literary Revolution in 1917. Like every other revolution, it entails a major change—in this case from classical Chinese to vernacular Chinese. The revolution caught on quickly, leading to a switch to the vernacular in newspapers and in primary school education. President Xi’s 2015 proposal of Toilet Revolution (廁所革命), which aims to improve the sanitary conditions of toilets in the country, may be considered another example [58].

In contemporary Hong Kong, a prominent use of the term 革命 (revolution) in a non-political context is digital revolution (數碼革命) or information revolution (信息革命), which refers to the change in how we store and process information [23]. There are also plenty of news articles that discuss financial revolution (金融革命) [37], scientific revolution (科技革命) [8], and fashion revolution (時裝革命) [53]. Vegetable Revolution (蔬菜革命) is the name of a vegetable retailor in Hong Kong and also the name of a book published in Shanghai in 2005 (which discusses “revolution at the dining table” [5]). Another recent use of the term is climate revolution (氣候革命), which is a political movement against climate change that has gathered strong momentum around the world [11]. These are all common, everyday usage of the term 革命 (revolution). We tend to remember unusual uses of words associated with memorable events (such as a full-blown political revolution) better than common use of words associated with everyday events (such as these everyday depictions of revolutionary changes). We tend to overestimate the frequency of the unusual uses.

Returning to the political context, there are two salient recent events in Hong Kong that carry the label of 革命 (revolution). The first is the Umbrella Revolution (雨傘革命) (alternatively known as ‘Occupy Central’ or ‘Umbrella Movement’) in 2014, a civil disobedience movement that called for the implementation of universal suffrage [52]. The protesters tried to pressure the Hong Kong government to reform the electoral system. Although the occupation was illegal, it emphasized non-violence. It was part of an international occupy movement that opposes inequality and seeks to advance democracy. The 79-day occupation ended when court injunctions were issued.

The next notable event is Fishball Revolution (魚蛋革命) (alternatively known as Mong Kok Unrest) in 2016, which was a clash about street food hawkers during Chinese New Year. Although the clash turned violent and led to arrests, the social issue that triggered the clash was hawker control by the Food and Environmental Hygiene Department. Although culturally significant, hawking as an industry is dying out in Hong Kong – no new hawker license had been issued in recent years [6]. The unwritten practice to relax hawker control during Chinese New Year appeared to be ending around 2014, triggering activists to show up and ‘protect’ street food hawkers.

Secessionism is not an apt characterisation of either incidents. It is important to note that in both incidents, the term 革命 (revolution) was first used by western media rather than by the protesters (for example, after a BBC reporter tweeted a picture entitled ‘Umbrella Man’, western media started calling the movement the ‘Umbrella Revolution’ [18]). This is unlike the liberation movements discussed earlier. There may be an element of exaggeration in the use of the term 革命 (revolution) to dramatize these incidents, especially in the digital age, where dramatic headlines or phrases are used to catch eyeballs. The term was then adopted by some protesters for self-glorification.

There is also evidence that 革命 (revolution) may be used in a metaphorical sense. The term ‘Umbrella Soldiers’ (傘兵) has been used to refer to parties established by the 2014 activists that focus on community and grassroots work and running in local elections [55]. Of course, the term cannot be taken literally, as they are anything but soldiers. But the metaphor of revolution/war emphasizes the resilience or strength of the activists towards their cause.

In sum, 革命 (revolution) denotes a major change; it can also be used metaphorically to refer to the spirit of change. It enjoys wide usage in the political and non-political context. As mentioned above, the meaning of 革命 (revolution) hinges on its collocation. In the slogan, 革命 (revolution) is modified by 時代 (era). 時代革命 (literally, revolution of an era) does not articulate clearly the target object of change—any major social and political change necessarily cuts across time and can be a marker of an era. The phrase is therefore capable of divergent interpretations. Just as revolutions named after a colour can be violent or non-violent, contrary to the prosecution expert’s conclusion, there is nothing inherently violent in the meaning of the phrase 時代革命 (revolution of an era).

2.7 Interpreting the Meaning of the Slogan

The usage analysis suggests that the slogan as a whole refers to a need to rectify a problem and to return to the original, a more desirable state of affairs for Hong Kong, without specifying what problem there is and what the desirable state of affairs looks like. This interpretation is consistent with denotations listed in dictionaries and is broader than that of the prosecution expert. I warn against interpreting the political slogan too literally, and I argue that conveying a desire for change is not only a permissible but an ordinary meaning that the slogan carries in Hong Kong today.

In both the election and the social movement contexts, people have interpreted and used the slogan to express a diverge range of views. The slogan, being open to divergent applications and interpretations, may be used to challenge the existing political order, but it may also be used to address social issues without challenging the existing political order. To pin the slogan down to a singular meaning is to disregard the range of meanings it carries and the subjectivity entailed in its use.

On 2 July 2020, two days after the NSL came into place, the Hong Kong government issued the following statement: “The slogan ‘Liberate Hong Kong, the revolution of our times’ nowadays connotes ‘Hong Kong independence’, or separating the Hong Kong Special Administrative Region (HKSAR) from the People's Republic of China, altering the legal status of the HKSAR, or subverting the State power” [51]. The government is clearly entitled to apply its interpretation to the slogan and impose laws that restricts its usage. It is however worth noting that the government statement emphasized “nowadays”, leaving room for divergent applications of the slogan prior to its official characterisation.

3 Some Critical Issues

In this section I will discuss the challenges that I encountered during cross-examination and broader questions about the ethics of expert witnessing. The goal here is pick out issues that are worthy of academic discussion for an audience interested in intersections between law and language.

3.1 Misuses of Dictionaries

Dictionaries “can be useful for defining unknown terms, showing the range of potential meanings, or attesting contested meanings” [25: fn 165]. But they are not reliable evidence of how ordinary a meaning is. Although the prosecution expert and I have both consulted dictionaries, he arrived at a much more narrow interpretation of the slogan than I have. How did this happen?

3.1.1 Etymological Fallacy

The Etymological Fallacy is “one of the most pernicious of popular idées fixes” about language [39: 87]. The idea that historical roots inform the interpretation of words today ignores the fact that meaning changes over time. “It cannot be overemphasized that a word’s etymology is not its meaning” [45: 131–132]. The common argument that an English word comes from Greek, Latin, or Arabic and therefore the correct meaning must be what it was in the language of origin is fallacious [32]. Mouritsen illustrates the danger of the fallacy this way: “(t)he notion that we may accept a given meaning as valid simply because its etymology is consistent with our proffered meaning is unsustainable because it would lead to absurd results: December would quite literally mean October, anthology would mean a bouquet of flowers”[34: 1915].

The prosecution expert’s account of the phrase 光復 (liberate) is based on dictionary definitions, as well as historical usage from the period of the Three Kingdoms (220 to 280 AD) up to the late 1940s. Similarly, his analysis of 革命 (revolution) is based on dictionary definitions as well as historical data from the Qin dynasty (221 to 206 BC) up to around 1911. The etymology and historical antecedents he provides ignore the capacity of language to change over time and fail to consider the abundance of contemporary examples of their usage in Hong Kong which are more directly related to the current case. In fact, chances are that people who chanted the slogan in 2019 have limited knowledge about the meaning and usage of these keywords in Chinese antiquity. As shown in Section II above, the keywords光復 (liberate) and 革命 (revolution) have enjoyed popularity in usage in contemporary Hong Kong since 2012. Contemporary usage of the keywords in Hong Kong show that they have a broad meaning, even within the political context. Although prosecution expert’s report also referred to Edward Leung’s 2016 election materials, he did not attend to contemporary uses of the term 光復 (liberate) in Hong Kong as reflected in the same materials, which provide important context to its sense in the slogan.

3.1.2 Sense-Ranking Fallacy

But the broader meanings are listed in the dictionaries as senses of these keywords. Why were they ignored? Sense-Ranking Fallacy describes the common mistake of inferring frequency information from the order of senses listed in dictionaries, contrary to the fact that many dictionaries base the order on historical appearance rather than frequency. Even where they arrange senses by frequency, it is done through human intuition rather than language data. In addition, “human intuition about the frequency of lexical items is often unreliable” [34: 1936], at least in part because human notices unusual occurrences more than typical occurrences. Except for a few recent attempts, lexicographers have not relied on statistical information about frequency of word senses in organising dictionary entries.

Chinese dictionaries are no exception. The Chinese lexicography literature suggests that etymology is still the preferred way of organising senses. According to Zhao, who edits the Modern Chinese Standard Dictionary (Xiandai Hanyu Guifan Cidian),

Therefore an ideal dictionary should follow the evolution of senses in its arrangement of entries, and state the derivative relationships among the entries, in order to objectively show the structure of the senses and help users learn them [60: 87; my translation]

In his report, the prosecution expert has mistakenly assumed that the first sense listed in the dictionary is the most frequently used sense, and that the second and third senses are necessarily rarely used. The dictionary (Xiandai Hanyu Cidian) that he used does not provide any information about the principles of its sense order ranking, nor does it compile the statistical frequency of a given word or word sense. Dictionary makers are not in the business of establishing ordinary usage; instead “they seek to instantiate the realm of permissible or possible usage” [34: 1945].

Moreover, a less commonly used meaning may well also be an ordinary meaning. The fact that a second meaning occurs less frequently in a corpus may be “merely a matter of which of two kinds of events, equally describable in the same language, happens to occur more often” [47: 33].

3.1.3 Divergent Readings

The content of the dictionaries consulted is rather consistent. While I used them as a tool to establish semantic range, which I then used to triangulate usage data, the prosecution expert extracted one salient and specific meaning of the keywords and argued that it is the only possible sense they carry in the slogan.

It is also worth noting that the slogan was uttered in Cantonese and in Hong Kong. While referring to Chinese dictionaries published outside of Hong Kong to sample the semantic range of keywords is not necessarily an inadequate starting point of analysis, it would not be prudent to base one’s analysis solely on such dictionaries or on historical usage that has little relevance to contemporary Hong Kong.

3.2 On Quantification

Ordinary meaning could be understood as the most frequently occurring meaning, or as a meaning that is comfortable to the speech community (and therefore multiple ordinary meanings are possible for the same phrase). In trying to establish the ordinary meaning of the slogan, my report is largely based on qualitative analysis of authentic language use. Could quantitative methods have been helpful in establishing the ordinary meaning of the slogan? Corpus analysis, largely based on the frequencies of occurrences and co-occurrences, is generally seen as a reliable way of establishing ordinary meaning. One benefit of this method is its replicability and falsifiability. It has been used to establish ordinary meaning in statutory interpretation in the US [49].

Let us begin by exploring what corpora would form a suitable basis for such a quantitative analysis. Considering that the slogan was used in elections and was chanted during a social movement in Hong Kong, a suitable corpus should have data on language use in Hong Kong; corpora consisting solely of the use of Chinese in mainland China or overseas territories will not do. This ideal corpus should contain language use in the political context, such as the genres of political speeches, relevant news reports, government publications, etc. Temporally, it would be most useful if the corpus contains data for the ten years preceding the case (2010–2020). Unfortunately, such a corpus does not exist. It may also be possible to gauge relevant keywords are used in Hong Kong from a general corpus. Existing general corpora from Hong KongFootnote 5 are unfortunately rather out of date. A Hong Kong Chinese equivalent of the NOW corpus, a database of web-based newspapers and magazines which grows every day (i.e., a monitor corpus), would have been immensely helpful. The Internet, as accessed through search engines, provided the next best alternative, though for reasons of transparency,Footnote 6 finiteness, and replicability, it is not a real corpus and does not provide a good basis for quantitative analysis. I focused therefore on using it to identify possible and attested usage of the keywords in contemporary Hong Kong.

The prosecution has relied on quantified data provided by the police, in a separate part of the trial. An 8-member investigation team at the Hong Kong Police Force downloaded 2177 video clips from the Internet, and reported the co-occurrence of the slogan with any ‘violent’ or ‘unlawful’ behaviour, and with any ‘separatist or subversive elements’ identified in the videos. The police report was never made available to me, and I was not aware of its existence until the first day of the trial. Commenting on the police report was beyond the scope of what I had been commissioned to do. Without knowing all the details of how the investigation was conducted, I will take this opportunity to say a few general words of caution. First, there is the obvious question of representativeness in data sampling, not just on the part of the police in what they have or have not downloaded, but more broadly about what videos were made and shared in the first place. Countless videos capturing the protests have been uploaded onto YouTube by news outlets and netizens, in which various slogans chanted are audible. Even though the police have not edited the videos, how a footage was framed and shot, and what gets uploaded onto the Internet involve an editorial choice, which is likely driven by newsworthiness or attention-worthiness. When a protest was peaceful, a common technique was for the camera to zoom out to show the scale of the movement. When violence broke out, close-up shots became desirable. It is questionable that these videos on the Internet constitute a representative sample of what went on during the protests, given that it is likely to be dramatic moments that draw the attention of the videographer and get shared in the attention economy. Measuring correlation based on these videos is a textbook example of the base rate fallacy, as it ignores the fact that the slogan was chanted tens of millions of times, peacefully, during the prolonged protests (in other words, with limited prevalence of illegal behaviour). The actual correlation between the slogan and illegal behaviour is much weaker than claimed. Second, there is the further and deeper question of what conclusions one can reasonably draw from these correlational findings. While corpus analysis also relies on correlations, the search for correlations in linguistics is much more constrained and structured; for example, collocation refers to two or more words occurring within a short space of each other in a text. During the protests, tens of thousands of people marched at the same time while being miles and hours apart; violence (such as the burning of a trash can) may break out at one street corner that the rest of the crowd have nothing to do with. The fact that a small fraction of people shouted the slogan while damaging government property does not mean that every protester who chanted the same slogan on the same day would have agreed with their action or would have interpreted the slogan in the same way. Violent or illegal behaviour is not necessarily indicative of seditious intention. Furthermore, correlations require interpretation. The words ‘police’ and ‘criminals’ are likely to frequently co-occur, but this correlation does not suggest that police are criminals, or that they are similar to criminals. There is a reasonable explanation of why these words co-occur: it is the police’s job to catch criminals. Correlation does not imply equivalence or causation: they may be a coincidence [see 10], or there may be a common source of causation (i.e., a confounding factor). In the current case, people with seditious intentions and those who do not may both be drawn to the protests because of a common source of anger (such as the proposed extradition bill and police brutality during the protests).

There was a similar claim by the prosecution expert based on co-occurrence. He asserts that because footages collected on 21 July 2019 showed that when some protesters defaced the exterior of the Chinese Liaison Office, the slogan ‘Liberate Hong Kong, Revolution of Our Times’ could be heard in the background, and therefore the slogan must be seditious. Among tens of thousands of appearances of the slogan, this incident was deemed particularly symbolic to the prosecution because it involved the defacing of the national emblem. In my review of the same footages, other slogans could also be heard during the incident. From the Now TV YouTube video [35] the prosecution expert referenced and the ViuTV’s YouTube videoFootnote 7 of the 21 July 2019 rally, one can hear protesters shout other slogans such as ‘Be Water’ and ‘Hong Kongers, Add Oil’. The fact that the slogan ‘Be Water’ was heard during the Chinese Liaison Office incident does not mean that ‘Be Water’ is always subversive wherever and however it was used in the 2019 protests. When asked by the defence lawyer during cross-examination, whether he thought ‘Hong Kongers, Add Oil’ can also be seditious, the prosecution expert answered in the positive.

His answer invites commentary. ‘Hong Kongers, Add Oil’ was in fact frequently heard in Hong Kong during the 2022 Beijing Winter Olympics and is clearly not seditious all the time. However, there is an element of truth in the claim that any utterance can be seditious. Just as the apparent greeting ‘How’s David?’ can be a threat when analysed in the context,Footnote 8 the intention of an utterance cannot be determined by its semantic content alone. This certainly does not mean that the police should go around arresting anyone who utters ‘Hong Kongers, Add Oil’ or ‘How’s David?’. Instead, contextual interpretation is necessary.

No correlational analysis can tell us about what the protestors had in mind when they chanted the slogan. I believe that qualitative analysis of the meaning of relevant keywords in contemporary Hong Kong and ethnographic data of protesters discussing what they thought the slogan meant as the protests were happening are better clues to intentionality and ordinariness of meaning.

My analysis does not show which interpretation of the slogan is the most frequent. I do not disagree that the keywords ‘liberate’ and ‘revolution’ conjure up prototypical historical moments of regime change, but I have shown that these words have broader and more basic meanings in the political context of contemporary Hong Kong that are attested in usage. When constructed as a whole, taking into account collocational meanings within the slogan and the contexts in which it was used, I have come to the conclusion that the slogan refers to a need to rectify a problem and to return to the original, a more desirable state of affairs for Hong Kong, without specifying what problem there is and what the desirable state of affairs looks like.

3.3 The Ethics of Expert Witnessing

Much of forensic linguistics looks inward to focus on the linguistic analysis itself. Correspondingly, discussions about the ethics of forensic linguists often begin and end within the field of our expertise. The primary ethical concern is tied to the objectivity of the analysis, which may be compromised by confirmation bias arising from prior knowledge of the result that the party engaging the expert would like them to reach [48: 353]. Moral considerations beyond the analysis itself have been mentioned in passing, such as by Ronald Butters, who expressed concerns about how the rich may use their wealth to manipulate the legal system “to censor or otherwise bully the poor” in civil cases such as trademark disputes [2: 378], and by Roger Shuy, who opines that while forensic linguists are free to avoid cases for moral reasons, he sees no ethical issues as long as the expert focuses on the data analysis and does not join in the advocacy of either side [43: 123–124].

As the field of forensic linguistics spreads from the UK and the US to other parts of the world, how geopolitical context affects the value and integrity of our work requires thoughtful reflections. Most people in the world do not work and live in liberal democracies with a strong rule of law foundation and an independent judiciary. As China increased its imperatives for regulating national security in Hong Kong, many have expressed the concern that the rule of law in Hong Kong is eroding [3]. With democratic backsliding happening in parts of the world, an ethical dilemma for experts is—regardless of which side they serve and the expert’s ability to stay neutral—whether their participation lends credence to the legitimacy of a judicial system. Scholarship in language and law rarely confronts legal systems that do not fully respect the rule of law or exercise judicial independence, or even those that stage political trials.Footnote 9 While such an ethical concern applies to any expert witnesses, forensic linguists need to exercise particular caution, given how susceptible language crimes (such as incitement and sedition) are to political abuse.

There may be value in expert witnessing work beyond influencing the outcome of the case. One hope is that expert voices will trigger public debate, and different perspectives will be recorded in the public domain. What I had not anticipated was the speed at which erasure of public records happened in Hong Kong. Between my writing of the expert report and the conclusion of the trial, major news outlets were shut down and their online archives deleted. These include some sources that I have used in my analysis, and news reports on the trial I participated in itself.

4 Postscript

The judge found the defendant guilty in 11 of the 14 charges [20] and sentenced him to 40 months of imprisonment. Paragraph 5 of the English press summary reads:

The court ruled that the offence of uttering seditious words under sections 9 and 10 of the Crimes Ordinance Cap 200 is constitutional and prescribed by law. It is also ruled that the political slogan of “Liberate Hong Kong, Revolution of our times” bears a close semantic connection and cannot be construed separately. The slogan, when uttered and/or displaced [sic], was capable of inciting others to commit secession. An attack on the HKSAR government can be taken as an attack on the Central Authorities. [26]

By focusing on what the slogan “was capable of” meaning rather than what it ordinarily means, the judge has adopted the logic of the NSL case that adjudicated on the slogan. In HKSAR v. Tong Ying Kit, the court relied on the undisputed fact that the slogan is “capable of” carrying a secessionist meaning and sided with the prosecution. In that case, the judges found the defendant guilty on both charges and sentenced him to 9 years of imprisonment.

In the end, even though the judge clearly sided with the prosecution, the expert reports did not matter, because both sides agree that the slogan is “capable of” carrying a seditious meaning. To base a criminal conviction on whether a text is “capable of” carrying criminality rather than actually carrying it would be akin to holding all utterances of ‘How’s David?’ as a threat.