During the Covid-19 pandemic, whether to digitally track the spread of the virus sparked a controversial debate in the United Kingdom (Kretzschmar et al. 2020; Mbwogge 2021; Williams et al. 2021) and beyond (Kleinman and Merkel 2020; Lo and Sim 2021; Shahroz et al. 2021). In the UK, after relaxing Coronavirus restrictions, the number of users notified to self-isolate by the NHS Covid-19 App reached an unprecedented peak, when 689,313 individuals were ‘pinged’ between July 14th and July 21st 2021 (BBC 2021). These alerts were meant to be sent to those who came in contact with someone who had tested positive for the virus, hence exposed to a contagion risk (Abbasi 2021). The term ‘pingdemic’ was coined to describe this situation, drawing attention to the classic notification sound and the high suspected infection rate (Rimmer 2021).

While existing research has investigated the state of digital contact-tracing (Williams et al. 2021; Samuel et al. 2022; Dowthwaite et al. 2021), and how language changed during the pandemic (Tsapro and Sivaieva 2021; Sulalah 2020; Oderf¨alt 2021; Oktaviana 2020), there has been limited focus on the specific term pingdemic and its utilisation by the UK press, especially when compared to the term pandemic. Therefore, investigating the adoption of pingdemic would provide insight into how the media portrayed this period in the Covid-19 timeline in the UK. While previous explorations have primarily approached pingdemic from a public health perspective (Marsh et al. 2021; Smith et al. 2022), computational linguistics offers an opportunity to delve into its broader societal impact and public attitudes towards Covid-19 (Verspoor et al. 2020; Glandt et al. 2021; Li et al. 2022; Ning et al. 2022). Thus, this investigation aims to address this research gap and looking at the social impact of the pingdemic in the wider context of the digital contact tracing system and Covid-19 in the UK more generally.

To examine the connotations acquired by pingdemic, this study focused on semantic prosody (Sinclair 1987, 1991; Louw 1993; Liu 2020), which refers to examining words that frequently co-occur with the term of interest to understand attitudinal meaning (Hunston 2007; Stewart 2010). Building on research into the negative semantic prosody of pandemic (Tsapro and Sivaieva 2021), this study uses Corpus Linguistic (CL) tools to investigate how pingdemic has been used and whether it exhibits a different semantic prosody to pandemic.

Methodologically, CL can efficiently reveal frequency patterns in large amounts of data (Jaworska 2017; McEnery and Hardie 2011; Leech 2014). To compare the semantic prosody of pingdemic to the benchmark of pandemic, we examined keywords in a corpus of printed press newspaper articles against a corpus of English broadsheet newspapers 1993–2021 (SiBol) (Partington 2022) and integrated it with the collocation analysis and a comparison of the visualisations provided by the Word Sketch Difference tool (Kilgarriff 2004). Ultimately, we investigated the following research question: is the semantic prosody of the word pingdemic different from that of pandemic, as portrayed in UK printed news outlets during a peak of Covid-19 cases (i.e., from July 2021)?

Although the term pingdemic had only been commonly used for a short period of time by July 2021, the impact it had on printed press discourses was pronounced, since this word featured in many articles (Rimmer 2021; Lea 2021; Leno¨el et al. 2021). Consequently, it is possible that its semantic prosody was different from that of pandemic, as pingdemic had a stronger focus on political rationale and decision- making. Arguably, this difference may have been influenced by the media focus on then-UK Prime Minister Boris Johnson and other members of the government in light of the relaxation of Coronavirus restrictions (Young 2021; Calnan and Douglass 2022; Wilman 2022).

Summarising, our work adds to the existing literature on the social impact of the Covid-19 app (Dowthwaite et al. 2021; Pepper et al. 2022; Kent 2020; Samuel et al. 2022). With the aid of CL tools (Tsapro and Sivaieva 2021; Sulalah 2020; Oderf¨alt 2021; Oktaviana 2020), we intend to contribute a fuller insight into the evolution of linguistic terminology during the Covid-19 pandemic and how the connotations of the words may reflect attitudes and perceptions of the UK public.


Covid-19 in the UK and the pingdemic

During the Covid-19 pandemic, in an attempt to track active cases and curb the spread of the virus, the UK government launched the NHS Covid-19 App, a Serco-created digital contact tracking application (NHS England 2021). This impacted the lives of several Brits due to backwards incompatibility, incorrect alerts and false positive tests (Kretzschmar et al. 2020), which meant that users had to self-isolate even when test results were incorrect, impacting income and well-being (Kent 2020). The uptake of the application was less than expected at 20.9 million downloads, with no specific figures published on active users (Pandit et al. 2022; Wymant et al. 2021).

A growing body of research focuses on attitudes towards digital contact-tracing in the context of Covid-19 and has been preceded by other research into attitudes to contact-tracing (Williams et al. 2021; Samuel et al. 2022; Seto et al. 2021). For instance, Dowthwaite et al. (2021) surveyed 1001 UK adults and found that half of the participants installed the app and with 60% of these claiming to comply with it on a regular basis. They also noted issues surrounding trust and understanding that hindered the effective adoption of the app. In a follow-up study, Pepper et al. (2022) identified five main themes, characterising perceptions of the app: flaws, usefulness and functionality affecting trust, low trust in the UK government, varying degrees of trust in other stakeholders, and public disinterest. These factors may have reduced compliance. Samuel et al. (2022) echoed these views, specifically around privacy.

In July 2021, the media focus was placed on the app once more as the number of alerts to self-isolate sent to users increased sharply. The growth of positive Covid-19 cases in the UK and the relaxation of government restrictions meant more frequent contacts via the app due to the higher exposure logging (Abbasi 2021). As a result, pingdemic was coined (Rimmer 2021)—a blend between pandemic and the word ‘ping’, i.e., the onomatopoeic term to denote the noise an exposure notification is likely to sound on a mobile phone.

Although research interest in digital contact-tracing has grown (Dowthwaite et al. 2021; Pepper et al. 2022; Kleinman and Merkel 2020; Lo and Sim 2021; Shahroz et al. 2021), a focus on the pingdemic is currently lacking and may be beneficial for understanding how the media reacted to the surge in notifications. Research into the rise in Covid-19 cases in the UK in July 2021 has focused on infections from a public health perspective (Marsh et al. 2021; Smith et al. 2022), more than a sociological standpoint. Simultaneously, a gap is also apparent with regard to linguistic research into the blended term pingdemic itself and its use in context.

Language, the Covid-19 pandemic and corpus linguistics

Computational linguistics techniques have been used to investigate the impact of, and attitudes towards, Covid-19 (Verspoor et al. 2020; Glandt et al. 2021; Li et al. 2022; Ning et al. 2022). Moreover, CL approaches have been employed to examine the use of language during the pandemic and semantic prosody. Our study builds on the findings by Tsapro and Sivaieva (2021), who looked at the semantic prosody of pandemic in broadsheet and tabloid newspapers. Their findings reflected the negative semantic prosody of pandemic that the dictionary definition of the term presupposes. Specifically, the pandemic depicted a metaphorical image of disaster, having collocates as start, hit, lead, cause, force, mean and devastate, that must be battled against, as suggested by other collocates like exacerbate and stop. Similarly, Sulalah (2020) explored the semantic prosody of the word increase. In light of the Covid-19 pandemic, this originally neutral word can be found associated with lexical items such as risk, rate, levels, number and significantly, implying a negative collocation.

Examining the semantic prosody of Covid-19-related terms, Oderf¨alt (2021) found cases of words having negative semantic prosody in a corpus of related media, such as lockdown and distancing and very few with positive semantic prosody, like easing. However, many of the words examined had neutral semantic prosody, like zoom and phase, due to the variations in language patterns discerned through the analysis. Therefore, the semantic prosody of Coronavirus-related terms was mixed, perhaps more negative.

More specifically, Oktaviana (2020) looked at the semantic prosody of distancing in journals, when referring to social distancing and physical distancing. Their investigation showed that the semantic prosody of social distancing was mixed. This was different to physical distancing, which showed collocates that were mostly positive, such as ‘effective’ and ‘working’. Despite this, social distancing was much more promnently used. Arguably, collocation analysis can help determine the semantic prosodies of pandemic-related terms, offering further insight into the changes in language use during the Covid-19 waves.


Data collection

Our corpus was gathered from multiple UK newspapers, representative of the wider UK printed newspaper media, including broadsheets and tabloids. The corpus spans from 1st July to 31st July 2021 to include the lead-up to, and aftermath of, the greatest surge in self-isolation notifications between 14th and 21st July 2021. Data was collected via Lexis Nexis, using the search term pingdemic to include articles directly relevant to the topic investigated. This was chosen as the datatbase as it was freely available. The articles were filtered to include English-language UK-printed media only.

Results were further filtered to include seven newspapers (The Daily Mail, The Guardian, The Independent, The Mirror, The Sun, The Telegraph and The Times). These were chosen as the sample as they are all national newspapers that print daily, thus providing a national representation. For consistency, all articles were accessed and downloaded on 4th May 2022.

In total, 783 articles were selected and downloaded. After removing duplicate files and live-reporting documents, that had not been automatically classified as online texts, the final total of texts was 628. Table 1 provides further information about the corpus and the sub-corpora of each newspaper. Because of availability and resource constraints, the corpus is not balanced. For instance, the corpus is weighted towards the Independent and away from The Sun and The Mirror. Nevertheless, the relative representations from different newspapers, alongside how the data might reflect particular audiences, were carefully considered when interpreting results.

Table 1 Distribution of the corpus and subcorpora

Analytical approach

The analysis focused on three key elements in this study: keyword, collocation and Word Sketch Difference to inform semantic prosody. This was undertaken using Sketch Engine, a CL software tool (Kilgarriff 2004).

For the keyword analysis, the corpus was compared to a reference corpus of English broadsheet newspapers 1993–2021 (SiBol) to determine ‘keywords’, distinguishing the dataset collected from such a comparable sample, belonging to the news genre, too. The focus was set to ‘word’ to process word forms of the same lemma individually and on 1000 (common). Therefore, words which were very frequent in general language or in the target corpus were considered. The keyness score (KS, hereafter) was calculated using simple maths for keywords (Kilgarriff 2009). Although stopwords could be filtered out, and usually are, of this analysis, it was decided that these would be kept in the keyword analysis in order to show prepositional relationships between concepts through manual analysis of concordance grids afterwards.

LogDice (LD, hereafter) was used as the statistical measure of collocational strength (Mautner 2007) and it scales well on different corpus sizes (Rychly´ 2008). We examined the co-occurrence of words within a span of 5 words to the left and right of pingdemic or pandemic to gather information about the broader context where words were used, rather than relying on (possibly incorrect) assumptions.

Word Sketch Difference, a tool within Sketch Engine software, is utilised to contrast existing collocations found and facilitate meaningful comparisons. Sketch Engine uses a colour-coding system to distinguish search words and generate two word sketches (Almaz´an-Ruiz and Orrequia-Barea 2020). The tool then compares the use of each collocate with both search words in each grammatical relation separately, with the shade of the colour indicating the strength of the collocation, ultimately forming a distinction of collocates between the two terms.

Regarding interpreting semantic prosody, Liu (2020) stated that semantic prosody refers not only to the meaning of a single word but the collocational meaning arising from the interplay between a given node (the searched term) and its most frequent collocates. This supports Louw (1993), according to whom it is a ‘consistent aura of meaning with which a form is imbued by its collocates’ (p.157), building upon the notions of semantic prosody set out by Sinclair (1987, 1991). To interpret the semantic prosodies of pingdemic and pandemic, we used the findings from:

  1. 1.

    The keyword analysis (Kilgarriff 2009);

  2. 2.

    The collocation analysis (Hunston 2007; Stewart 2010);

  3. 3.

    Word Sketch Difference, a visualisation tool that distinctly compares collo- cates from two different terms (Almaz´an-Ruiz and Orrequia-Barea 2020).


Keyword analysis

Table 2 showed the top ten words with the highest KS compared to SiBol. Since the dates of publication were included in many articles, july (KS: 3.65) was likely to exhibit a high frequency. Other prominent words were pingdemic (KS: 2.33) and pinged (KS: 2.29) and other pandemic-related terms, such as covid (KS: 3.59) and app (KS: 3.10). Therefore, the discourse focused on the pandemic, exploring the following aspects:

  1. 1.

    the source of the notification (i.e., the app);

  2. 2.

    the action itself (i.e., pinged);

  3. 3.

    those responsible for implementing the automated tracking system (i.e., government);

  4. 4.

    those affected by the surge in people self-isolating (i.e., staff, nhs, workers). These themes provided a foundation to examine the term pingdemic in context.

Table 2 The top 10 words with the highest KS

Collocation analysis


In total, 745 occurrences of the word pingdemic featured in the corpus. A collocational strength analysis was undertaken, with the top ten scores, measured by LD, shown in Table 3.

Table 3 Top 10 collocates of pingdemic

The strongest collocate, so-called (LD: 11.28), appears to challenge the legitimacy of the term pingdemic (‘business leaders warned that the economy could “grind to a halt” due to the so-called pingdemic’) and question whether the phenomenon behind the word pingdemic, the increasing number of notifications sent to app users to self- isolate, was genuine (‘large swathes of the UK economy are now being affected by the so-called pingdemic, with warnings about services such as rubbish collection being disrupted’). However, an alternative interpretation could be that the authors of the articles analysed were commenting on the novel term and its unclear etymology.

This also co-occurred with words that indicated resistance, such as revolt (e.g., Conservatives ‘revolting’ against the pingdemic), complaints (e.g., complaining about Covid being brought by international travellers to the UK) and warnings (e.g., about disruption to services or a synonym of notifications). In over half of the cases of pingdemic, the use of quotation marks around so-called seemed to support this interpretation.

Caused (LD: 10.17) was used in active and passive constructions surrounding pingdemic. This also applied to causing (LD: 10.09) (although it occurred more frequently with pingdemic being the active agent). The pingdemic was reported to have caused chaos (LD to caused: 8.72), shortages (LD to caused: 9.51), problems (LD to caused: 7.55) and disruption (LD to caused: 8.75). Examples of this included ‘criticism grows over the chaos caused by the “pingdemic” ‘, ‘travel disruption and food supply woes as a result of worker shortages caused by the so-called “pingdemic” ‘, ‘Pret a Manger, Nando’s, Greggs and McDonald’s are all suffering extra staffing problems caused by the NHS self-isolation policies’ and ‘Plans to make NHS England ‘s contact tracing app less sensitive to reduce disruption caused by mass isolations have been pushed back’. Shortages (LD: 9.84) co-occurred regularly with caused and causing. The shortages in question were of food and staff (in organisations such as the NHS, Santander and ‘labour’), in examples like ‘Santander has closed around 25 branches in the UK, out of a total of 465, for a week due to staff shortages caused by the “pingdemic” ‘ and ‘a lot of the issues with labour shortages were being caused by the NHS Covid app “pinging” workers’. Many of the uses of shortages, in conjunction with pingdemic, talked about how this worsened, even though the pingdemic shortages could be combated and would likely fade over time.

Instances of chaos (LD: 9.92), showed the pingdemic as a source of disruption. Some of these saw pingdemic chaos together, with the word pingdemic modifying chaos. This referred to disruptions to ambulance services and supermarket shortages (‘Sir John’s intervention came as the pingdemic continued to cause chaos—and result in empty shelves in supermarkets’). Interestingly, the word crisis (LD: 9.60), semantically related to chaos, was also discussed in a similar way, with pingdemic used as a modifier. With crisis, though, a more significant focus was on the hospitality sector (e.g., staffing shortages in examples such as ‘The SNP leader confirmed “limited” changes to self- isolation rules for Scotland to help key industries cope with significant staff shortages, as the country struggles with the so-called “pingdemic” crisis’).

By (LD: 9.79) regularly occurred with caused, especially in passive constructions. This implied the foregrounding of the object in the sentences was the focus of these articles, rather than the pingdemic itself, which may indicate that the context of the pingdemic was apparent and established by July 2021. This was seen in examples such as ‘Johnson also confirmed that crippling self-isolation rules are nailed on’ to end on August 16 for the double jabbed, bringing hope of an end to the chaos caused by the Pingdemic’ and ‘Currently, we are urgently looking for 2,000 extra store assistants and van drivers to fill the gaps created by the pingdemic’.

Is (LD: 9.68) was used where pingdemic acted as a noun. The most common verb after is was causing (LD: 11.22). Additional insights included its association with the cause of staff shortages (‘Recruiters are more confident than ever in the future of their business and hiring decisions—but the pingdemic is causing shortages’), disruptions to family lives (‘Everyone is only too aware of the disruption that self-isolation is causing to business and people’s lives’, and additional hardships (‘I ‘ve come to the view that the current pingdemic ‘ is causing more hardship, unhappiness and harm to the health and welfare of the nation than the app was devised to prevent’). Is also collocated strongly with ongoing (LD to is: 10.96); however, ongoing had been used as an adjective to qualify pingdemic rather that an additional verb. On July 21st, many newspapers quoted Andrew Opie, of the British Retail Consortium, referring to the pingdemic as putting increasing pressure on the retailers’ ‘ability to maintain their original opening hours and shelves stocked’. Hence, this recurring reference is likely to have skewed results.

End (LD: 9.73) saw a call for the end of the pingdemic. This included journalists calling on Boris Johnson to end the pingdemic (‘For weeks, the Mail has urged Mr Johnson to end self-isolation for the double-jabbed who are pinged by the NHS contact tracing app’) and Keir Starmer calling for it to end (‘WHEN even Sir Keir Starmer calls for an end to the pingdemic, it really must be time for a change’). The Daily Mail even mentioned launching a campaign to end the pingdemic, encouraging readers to join it.

To summarise, pingdemic collocated with terms that either appeared to question the legitimacy of the term or event or showcased the supposed damage that it had done to families and professional sectors alike.


The collocational findings related to pandemic differ from those reported for pingdemic. pandemic occurred 577 times in the corpus. Its top ten collocates are displayed in Table 4.

Table 4 Top 10 collocates of pandemic

Many of the collocates of pandemic related to time (e.g., during, since, throughout, start, before, began, behind), which will be explored in further detail. During (LD: 11.18) collocated with pandemic when reporting events occurred since the Covid-19 outbreak,for example, infections and hospitalisations rising. Since (LD: 10.25) referred to the same idea, but also to this being the busiest period of travel since the pandemic began.

For throughout (LD: 10.17), Andrew Opie also said, ‘Retail workers and suppliers, who have played a vital role throughout this pandemic, should be allowed to work provided they are double vaccinated or can show a negative Covid test, to ensure there is no disruption to the public’s ability to get food and other goods.’ The reporting of this quote by multiple newspapers had, once again, affected the collocational strength. Yet, the fact that this quotation appeared across different articles and newspapers may suggest the prominence and wide audience that this comment reached.

Behind (LD: 9.31), although a proximal preposition, was used in the corpus to refer to time. Frequently, articles questioned whether the worst of the pandemic was ‘behind us’, in structures such as ‘Professor Neil Ferguson predicted Britain would have “the bulk of the pandemic behind us” by late September’. Other instances talked about the UK ‘emerging from the worst effects of the health pandemic’ but ‘falling behind its EU rivals in international trade’ by slowly removing restrictions, referring to the UK competitiveness suffering, perhaps due to external factors (e.g., Brexit) (Analytica 2022). Thus, behind was used with pingdemic in both positive (the worst ‘is behind’ us) and negative (the UK ‘is behind its rivals’) contexts.

Wave (LD 9.33) referred to the pandemic phase, characterised by fluctuations in infection rates, mentioning the first (‘Industry leaders and the Government insisted there was still “ plenty of food “ in the supply chain, adding that there was no need for a repeat of the panic buying seen during the first wave of the pandemic’), third (‘he increase is fairly obviously correlated with the third wave of the pandemic’) and latest (‘The current wave of the pandemic sweeping across Britain was also likely to get worse before it gets better’) waves. These were usually contextualised through the numbers of infections and deaths in each wave. Therefore, although factual reporting, the reference to death and infection had negative connotations.

To summarise, the temporal collocates with pandemic indicated that these articles may have been contextualising the pingdemic in relation to the long-standing pandemic. However, some, such as behind were more negative. Other collocates, such as wave were also contextualising the pandemic with reference to previous infection increases.

Word sketch difference

Before using The Sketch Engine’s ‘Word Sketch Difference’ tool, results from the collocation examination indicated that pingdemic appeared to co-occur more frequently with negatively connotated terms (e.g., chaos and shortage) than pandemic, which mostly collocated with time-related vocabulary, which seemed neutral.

Using Word Sketch Difference, three grammatical variations of words surrounding pingdemic and pandemic became apparent:

  1. 1.

    Words that modified pingdemic/pandemic;

  2. 2.

    Pingdemic/pandemic as a subject;

  3. 3.

    Pingdemic/pandemic as an object.

These were chosen to help establish how pingdemic and pandemic were presented in terms of agency (e.g., through active and passive constructions that indicated cause) and to further investigate their semantic prosody (e.g., through adjectives).

The Word Sketch Difference for words modifying pingdemic and pandemic, shown here in Fig. 1, included some that also featured in the previous collocation examination. For instance, so-called pingdemic appeared to question the legitimacy of the name or the event itself, once more. However, this tool revealed further collocates that were not statistically significant to meet the minimum threshold of the LD measurements used previously. An example was the word good, which collocated with pandemic.

Fig. 1
figure 1

Word Sketch Difference visualisation to show modifiers of pingdemic and pandemic

Although this might be seen as a collocational clash, further analysis exposed this as predominantly used when discussing Keir Starmer, leader of the Labour opposition party in the UK, stating that he had a ‘good pandemic’ until the vaccine rollout. Concordance lines showed that there were repeated sentences in multiple articles, espe- cially quotes, as previously mentioned. Unsurprisingly, occurrences like coronavirus, covid-19, covid and global all appeared with pandemic. Severe also collocated with pan- demic, but only appeared twice in the corpus. This may suggest a similar presentation of the pandemic, across the different media outlets.

The Word Sketch Difference for pandemic/pingdemic as subjects in the clause, shown in Fig. 2, also included shared collocates. Although verbs such as to be and to have did not reveal much about the semantic prosodies of these words, others, like hit and strike, appeared to be used negatively, justifying the pejorative effects that both the pingdemic and the pandemic had. As seen in Fig. 2, collocates of pingdemic, like threaten, disadvantage, exacerbate and deprive could be deemed as negative. Other collocates of pingdemic appeared to have a negative connotation, such as devastate and cripple.

Fig. 2
figure 2

Word Sketch Difference visualisation to show pingdemic and pandemic as a subject

In Fig. 3, pandemic and pingdemic were explored as grammatical objects. End, again, appeared as a frequent collocate of pingdemic, particularly when attempting to end the pingdemic. Both stress and blame, collocates of pingdemic, emphasised the impact of the pingdemic, yet removed such an impact from tertiary sectors, through a variety of positively and negatively connotated collocates. For example, Lift was perhaps more positive, speaking of the lifting of pandemic-related restrictions, whilst fear had a negative connotation and referred to the public worrying about the duration of the pandemic. Overall, The Word Sketch Difference tool gave us a more holistic view of the semantic prosodies of pandemic and pingdemic in newspaper articles published in July 2021, as will be detailed.

Fig. 3
figure 3

Word Sketch Difference visualisation to show pingdemic and pandemic as an object


In the following sub-sections, we focus the results of the analysis undertaken, separately. We also evaluate these and recommend future research developments.

Semantic prosodies of pingdemic/pandemic

First, the keyword analysis established the importance of pingdemic in the corpus through the frequent use of the word itself, and also pandemic through lexicon relating to Covid-19. This offered a strong rationale for continuing this investigation with pingdemic and pandemic as the key lexical focal points. The collocation analysis of pingdemic showed a strong association with negatively connoted words, such as chaos and shortage. These could be interpreted to broadly refer to the governance of the country, rather than as a personal problem or nuisance. Possibly, this revealed report- ing priorities of the news organisations considered. Additionally, collocates appeared to question the legitimacy of either the term itself or, perhaps less likely, the sociological phenomenon occurring, such as so-called and the quotation marks. In this sense, the collocation analysis of pandemic differed from that of pingdemic, whose collocations seemed mostly neutral in meaning, with some exceptions like behind.

Further investigation using the Word Sketch Difference tool corroborated existing ideas about both terms added further detail. For example, many more collocations of both terms appeared neutral, perhaps suggesting that further investigation of the words following those would be useful. For instance, pandemic showed a stronger association with virus-related terms, while pingdemic had more associations with words semantically tied to putting an end to the growing numbers of notifications (beat, tackle, end). Results for pandemic indicated it was something to be ‘beaten’ (Tsapro and Sivaieva 2021). Specifically, both terms showed collocations with words more negatively connoted, particularly verbs such as cause and strike.

This analysis was meant to find out whether pingdemic and pandemic had any shared semantic prosodies, as per our research question. In this context, some similarities were shared but findings ultimately differed. Pingdemic offered a more overtly negative semantic prosody. Firstly, its collocation with so-called and quotation marks showed a question of its legitimacy, whilst implying it was out of control and inflicting damage, strongly associating the impact it had and discussion about how to end it. This was different from the semantic prosody that pandemic exhibited. The latter was more concerned with the contextualisation of the evolution of Covid-19, neither overtly positive nor negative.

Perhaps the articles collected tried to contextualise the pingdemic in relation to the long-standing pandemic. However, given that the corpus explored was from July 2021, rather than the beginning of the Covid-19 outbreak, the word pandemic may have changed in semantic prosody, as it grew to be a part of everyday discussion. The hypothesis that the semantic prosody of pingdemic would be more politically motivated turned out to be only partially true, appearing more socially motivated and focused on the impact rather than intention.

These findings build upon existing work within social research regarding digital contact-tracing apps and language employed during the Covid-19 pandemic waves. For the former, this study demonstrated that the perception of the Covid-19 app held by several members of the public (Williams et al. 2021; Samuel et al. 2022; Dowthwaite et al. 2021) was reflected in printed press discourses. However, no lexicon relating to privacy that had high collocational strength with pingdemic, differing from other findings (Samuel et al. 2022). For the latter, the terms related to the Covid- 19 pandemic appeared mainly negative in their semantic prosodies, such as increase (Sulalah 2020), lockdown (Oderf¨alt 2021) and distancing (Oktaviana 2020). This could have been due to the subject matter.

Ultimately, with terminology related to the Covid-19 pandemic evolving quickly, the semantic prosodies of new terms may be expected to change over time. Thus, temporal rather than overtly negative words collocated more strongly with pandemic, showcasing ‘seasonal’ more than consistent collocates.


Even though CL contributed considerably in narrowing the search for informative lexical and semantic associations characterising the large dataset analysed in this study (Jaworska 2017; Tognini-Bonelli 2001), more data could have substantiated these findings more, especially those of the Word Sketch Difference analysis. As the sample size of 628 articles is relatively small, to expand the types of newspapers to include in the corpus, local printed press or a longer time-span could have been considered.

Additionally, CL can usefully pinpoint a descriptive analysis of the function of concordances, although it can also draw away from the interpretation or critique (Baker and Levon 2015). Like Rose (2017) claimed, using CL provides strong evidence, but struggles to justify it. Using a complementary approach, such as Critical Discourse Analysis (CDA), could overcome this limitation.

While CL analysis tools struggle to pinpoint different perspectives and meaning shades, CDA examines texts for nuance (Van Dijk 1997). As CDA can be underpinned by theoretical frameworks, this study would ensure that the corpus findings highlight the insights into how pingdemic is presented. In this sense, the combination of approaches could provide fuller insights into the presentations by media outlets, as well as implications on public trust, such as giving a better account of public reporting of governmental (digital) responses to the pandemic itself.

Recommendations for future work

Further research in this area should expand the corpus to online news articles. Considering the increasing consumption and adoption of digital media (Fletcher and Nielsen 2018; Twenge et al. 2019; Mihelj et al. 2019) and the growing audiences of digital news (Cornia et al. 2016; Newman et al. 2021), focusing on both print and digital news sources would provide a fuller account of news with a higher readership reach.

Another suggestion is to focus on the different uses of pingdemic in different newspapers individually, using sub-corpora. This could also be done diachronically, supported by CL tools (Baker 2010). Examining the collocates of words, which appear semantically neutral in isolation, in conjunction with pingdemic or pandemic could expand our findings and substantiate the study. Further research could examine the instances of pingdemic when used as a subject or object. All these recommendations would shed additional light on the grammatical and physical agency of the digital contact-tracing app, ultimately proving valuable insights into the public perception of AI-driven systems.


This study used a corpus of 628 articles from seven national UK newspapers, published in July 2021, to compare the semantic prosodies of pingdemic and pandemic. This was done using a CL-based approach, including an analysis of keywords, collocations and Word Sketch Difference visualisations, with the aid of The Sketch Engine (Kilgarriff 2004).

According to our results, pingdemic held negative semantic prosody overall, supported by trends. First, the corpus exhibited strong collocates such as so-called, which appeared to question the legitimacy of the term pingdemic or the event itself. Second, causing and chaos reiterated the idea of Covid-19 inflicting damage, especially given the context in which these terms were used, such as the impact on commerce, health- care and families. Such semantic prosody was found to differ from pandemic, which, in July 2021, seemed to aim at the contextualisation more than the construction of an overt positive or negative semantic prosody, contrasting with earlier research (Tsapro and Sivaieva 2021).

These conclusions enabled us to answer our research question as pingdemic and pandemic had different semantic prosodies. Therefore, the initial hypothesis that the semantic prosody of pingdemic would be more politically motivated was only partly true, as it instead appeared to be more socially motivated and concerned with the impact of its disruption.