Reader Comments Agentive Power in COVID-19 Digital News Articles: Challenging Parascientific Information?

Abstract

The recent COVID-19 pandemic has triggered an enormous stream of information. Parascientific digital communication has pursued different avenues, from mainstream media news to social networking, at times combined. Likewise, citizens have developed new discourse practices, with readers as active participants who claim authority. Based on a corpus of 500 reader comments from The Guardian, we analyse how readers build their authorial voice on COVID-19 news as well as their agentive power and its implications. Methodologically, we draw upon stance markers, depersonalisation strategies, and heteroglossic markers, from the perspective of discursive interpersonality. Our findings unearth that stance markers are central for readers to build authority and produce content. Depersonalised and heteroglossic markers are also resorted, reinforcing readers’ authority with external information that mirrors expert scientific communication. Conclusions suggest a strong citizen agentive power that can either support news articles, spreading parascientific information, or challenge them, therefore, contributing to produce pseudoscientific messages.

1. Introduction

1.1. Parascientific vs Pseudoscientific Information

In order to contextualise our research, it seems necessary to depart from some assumptions about two concepts that need disambiguation: parascience and pseudoscience. Scientists nowadays play a variety of roles related to the skills associated with scientific work. They often act as experts [1] on issues of social relevance in the media or at events with public impact, a practice we could name as parascientific. The issuing of judgements based on their expertise in a scientific field can help to shape public opinion or to guide the behaviour of citizens [2]. In this vein, scientific communication in general has undergone important changes over the last decades. Many scientists have begun to practice what is called open science, where “research materials are provided through an “open” (online) lab notebook, data collections are made available, and some scientists even blog about the research progress” [3].

Simultaneously, generalistic media, complying with their task of publishing news related to current scientific topics such as COVID-19, also issue parascientific information with a documented professional focus, relying on scientific data. This practice of journalists publishing scientific information can be said to have triggered parascientific genres that operate somewhere between expert discourses and popularizations. Some of these parascientific genres are popular news articles and their comment sections. We consider them parascientific and not just popularizations, and they are related to the ability for citizens to comment, adopting certain markers of scientific discourse. Comments are written in a space provided by the digital newspapers, in an attempt to improve audience loyalty [4], making readers feel protagonists as commenters. However, this phenomenon is also related to social networks, as happens with Facebook, where mainstream newspapers offer to follow their profiles, as is the case with The Guardian.

From this perspective, parascience is closely linked to digital news, as a result of several variables, the first one being the outburst of social networking that has allowed worldwide digital communication. The arrival of Web 2.0 technologies and social media networking have resulted in a paradigm shift, giving rise to the novel practice of produsage (producing plus using), where citizens not only use technically-mediated genres, but also produce information, or user-generated content (UGC) [5]. Indeed, the research community has claimed a shift in power to the benefit of users through the rise of a new “participatory culture”, where boundaries between production and consumption become blurry [6].

Another variable worth considering is the increasingly agentive role that citizens have undertaken in practically all areas of society, as a result of the current customer service society [7], which has empowered individuals by giving them a voice. This financial mercantilism of today’s postmodern society [8] has also permeated the digital news and media system, where readers and their comments have reached a new dimension. Hence, citizens now have the possibility to take part in an array of e-genres that deal with knowledge dissemination through comments they make on fields such as climate shift, medical discoveries, natural disasters, politics or global health, to name but a few.

Nevertheless, this practice of laypersons commenting on scientific news in digitally mediated newspapers can also lead towards a dangerous consequence: that of creating and spreading pseudoscientific information. The media are, at the same time, the main allies of science in their desire to disseminate news about relevant issues, but also their main enemies by giving shelter to pseudoscientific messages. Due to the great power of the media as disseminators of reality and guarantors of that which is included/excluded in the system, their responsibility is not a minor issue, as it can multiply the threats posed by pseudoscience [9].

Moreover, the media intoxication resulting from the dissemination of pseudoscience can lead to a normalisation of the phenomenon on a social scale, while at the same time establishing a false understanding of the problem. Thanks in part to the media complicity, pseudoscience, in its various manifestations, invades the social body and forms a dangerous scenario, which exploits the individual uncertainty of the liquid modernity [10] as well as the gap between expert knowledge and popular knowledge. In fact, many readers may be incapable of decoding information, since they lack references that allow them to make a free choice based on real knowledge.

As for the description of pseudoscience, three features are attributed to it [9]: it usually belongs to a discipline that is related to science; (2) it is not epistemologically warranted; and (3) its proponents try to create the impression that it is scientific. In addition, it manipulates the emotions of the public at large, such as hope or fear of the future [9]. The fact that pseudoscience has gained access to the media sphere reflects one of the main characteristics of our society today: its paradoxical component. This paradox can be seen in the parallel paths of science and pseudoscience. While science has reached goals as well as fulfilled objectives and expectations, pseudoscience opposes its own growth and development to the contemporary scientific boom, with an omnipresent sense of crisis [10,11].

1.2. New Discourse Practices of Citizens/Readers

Reader comments of a great range of media may support and spread positive news, but they can also challenge journalists and produce pseudoscientific messages, understood as a downgrading of scientific and parascientific knowledge because readers are laypersons and thus, supposed non-experts. Research has been conducted in many UGC genres such as online consumer forums and online reviews [12,13,14]. In consumer forums, readers reply to writers, turning into writers themselves as wreaders [13], since they hold both roles and they produce legitimate content. In addition to that, they do not challenge authors, since all the interactants have an equal, generically defined peer status, their common aim being to share opinions about services or products. Another, similar case, are technologically-mediated platforms among professionals, such as medical blogs, where interactants are experts; thus, readers and authors also hold a similar status, afforded by the genre and its discourse community. Yet, it is not the case with digital news articles. In this genre, readers reply to news articles, showing a powerful agency [15], responding and at times challenging journalists who are professional writers holding documented information. The difference lies in that readers cannot qualify as journalists with expert knowledge, since the genre and the related discourse community [16,17] does not endow them with that status. This is the case with The Guardian, a mainstream newspaper addressed to lay people. In other words, it is the genre that provides its interactants with the status of experts or non-experts and, therefore, gives them authority or not.

At this point it is important to clarify how we regard the concepts of genre and discourse community. Following Swales [16], a genre comprises a set of communicative purposes that must be recognized by the expert members of the discourse community. This rationale shapes the structure of its discourse and influences and constrains choice of content and style. If we consider reader comments as part of the digital news article, described as a genre that produces true information, we are accepting one single discourse community behind, where content and style are accepted by all its members. Reader comments can respond to journalistic information through opinions, as long as readers accept this expert knowledge as true. However, this is not always the case. Commenters of news articles on health issues such as COVID-19 may reject writers as truth guarantors and at times challenge them, therefore, issuing pseudoscientific information. Besides, journalists, as members of the same discourse community, cannot accept commenters’ opinions if these trespass the discursive boundaries set by the genre. This is why the limits of writers and readers’ interaction in parascientific communication through the Internet may look eroded [18].

In Trench’s [18] words, the Internet “is turning science communication inside-out” and, as a result, the boundaries between expert and non-expert science communication are “eroding”. These boundaries have long been blurred by parascientific genres such as trade magazines [19] but also by scientific journalism in generalistic media. The COVID-19 pandemic has been an extraordinary, unprecedented situation, with citizens taking active roles on different aspects of this new illness, such as vaccines, prevention norms or government measures. In digital media news related to health issues such as COVID-19, readers accomplish the principle of shared knowledge only to some extent, since they can react to news in unexpected ways.

In view of this scenario, we asked ourselves which mechanisms could help COVID-19 news article commenters to write convincingly, challenging expert journalists and creating messages that can influence citizens. Since this kind of communication entails strong interpersonal discourse characteristics, two aspects seemed worth analysing in the construction of commenters’ authority: authorial voice and content. The authorial voice usually takes stance markers that favour assertiveness and closeness to readers, such as self-mentions, but also other markers such as hedges, that mitigate a too strong stance [20]. Likewise, commenters can resort to boosters as grammatical enhancers, or attitudinals, which provide a subjective, persuasive shade, principally in the form of qualifying adjectives. They can also use external references in the form of impersonalisations, depersonalisations and heteroglossia [21] to produce comments with scientific, external support, imitating the scientific discourse and thus creating content. This is what we attempt to unveil through the current study.

2. Materials and Methods

2.1. What We Hypothesise

We hypothesise that readers may support but also challenge writers on COVID-19 related matters, construing an authorial voice that defies that of journalists who write about scientific issues. To do so, readers produce judgments and opinions, taking an active role as citizens in societal matters such as health, and having the possibility to influence other readers positively, but also in negative ways, contributing to rumours or false knowledge.

To this end, we would like to test the following two hypotheses:

HP1. Readers build their discursive authority through an authorial self-construction that responds and at times contests news article writers through assertiveness and also mirroring expert scientific discourse strategies.

HP2. Readers’ authorial self-construction of authority may include closeness and distance pragmatic strategies [20], using stance markers, but also resorting to impersonalisation, depersonalization or heteroglossic strategies, in an attempt to reduce an excessive personal assertiveness, thus projecting a more convincing authorial voice.

2.2. Methodology

2.2.1. Corpus Description

Our corpus collection was carried out from November 2020 to February 2021, and a sample of 100 news articles along with 500 reader comments were retrieved from the British digital newspaper The Guardian by accessing the social networking site Facebook. Our corpus collection has, at all times, been carried out following the University of Valencia Code of Good Practices in Research: (can be found at https://www.uv.es/hrs4r/Code_Good_Practice_Research.pdf (accessed on 10 September 2021)). We extracted our corpus through this networking site, which is massively used by all kinds of people, instead of collecting it from the digital The Guardian platform that is read by fewer citizens, at times only via subscription. With this decision, we tried to assure a more diverse number of commenters in the corpus. So, news articles dealing with the COVID-19 pandemic as well as their reader comments were selected. These comments contest the COVID-19 information of each article.

As regards the size of the corpora, the news articles data set contains 40,400 words and that of the reader comments contains 42,330 words. It is worth underlining that we have only focused on analysing the comments made by readers and, when needed for contrasting purposes, the information given by the article writer. As far as comments held among readers are concerned, these were excluded, since they are out of the scope of this research.

2.2.2. Framework of Analysis

The perspective followed in our analysis is eclectic and draws on the concept of discursive interpersonality [22], which addresses interpersonal communication—here of readers towards writers—from the viewpoint of discourse and not from that of genre. This viewpoint of interpersonal communication allows to transcend strict metadiscursive approaches [23] and acknowledges the use of lexico-semantic elements or discursive strategies dictated by each specific text. Therefore, it is the discourse of reader comments with concrete social conditions (the COVID-19 pandemic in general) that has enabled us to identify interpersonal discursive strategies, features and resources that shape interaction in particular ways. To analyse how readers construe their selves, we have resorted to Hyland’s [23] stance markers, as well as to depersonalisation strategies: personifications, nominalisations, passive sentences, impersonal expressions + infinitive, and there + be clauses [21]. Finally, we have relied on heteroglossic categories, namely endorsement, attribution and distance [24], to complete the analysis.

An observational analysis has been the main methodological approach followed in this research, in turn supported by quantitative data. As for the procedure of analysis, it has been as follows: as shown in Figure 1 below, we have departed from the assumption that reader comments entail strong interpersonal characteristics, making readers construct their authority not only through their authorial voice, but also through content. The authorial voice is materialised through closeness and distance pragmatic functions, and these, in turn, through discourse strategies such as stance [23], depersonalisation [21] and heteroglossia [24]; all of them are realised through different lexico-grammatical uses:

In order to have a full picture of the readers’ discursive self-construction, we also found it necessary to analyse that of the journalists as writers, so both discourses could be compared and visualized, as they are shown in our Section 3.

The Ant.Conc 3.5.8. concordancing tool has been used to work out the absolute frequency of the different closeness and distance pragmatic strategies found in the two corpora analysed as well as their lexico-grammatical uses. To do so, we have followed the categories and linguistic realizations previously dealt with in the literature [20,21,25,26,27]. Due to the constraints imposed by the AntConc Software to identify all the strategies considered in the analysis, some of them have been manually counted. It is also important to point out that, as both corpora are inevitably of a different lexical density, the absolute corpora are also of a different lexical density. The absolute frequency of each interactional marker and their lexico-grammatical realizations have, therefore, been divided by the total amount of words contained in each data set and then computed per 1000 words, which has been regarded as a conventional way of standardising results of corpora of unequal size.

3. Results

This section is devoted to describe the most relevant findings, illustrated through different examples drawn from the corpora selected for the research.

Concerning closeness and distance pragmatic strategies, Hyland [28] defines closeness as “a writer’s control of those rhetorical features that display both authority as an expert and a personal position towards issues in an unfolding text”. In other words, this pragmatic strategy is intrinsically associated with promoting subjectivity, visibility and friendliness. Nevertheless, given that reader comments not only trigger strong interpersonal characteristics but also specific content, we have considered that closeness does not only reinforce readers’ authority through stance markers such as first-person singular pronouns, but also through attitudinals such as adjectives, verbs, etc. Following Hyland’s model of stance [23], the markers and lexico-grammatical realisations that have been selected for the analysis are the following: a) boosters (e.g., emphatic adverbs and phrases); b) attitudinals (e.g., attitudinal adjectives, verbs, and adverbs, deontic verbs, exclamations); and c) self-mention markers (e.g., first-person singular pronouns).

In regards to distance, [20] (p. 78) describes this pragmatic strategy as “a specific way through which authors project authority by means of diverse linguistic choices including objective and/or depersonalised realisations with the deliberate intention to show credibility”. Therefore, in sharp contrast with closeness, distance can be linked to objectivity, invisibility and professionalism. In a similar vein, distance pragmatic strategies can characterize authorial voice as well as content, as manifested in the use of hedges or inclusive first-person plural pronouns, depersonalizations and heteroglossic markers, which mirror the conventions of scientific discourse. Drawing on stance markers [23], depersonalisation strategies [21] and heteroglossic markers [24], the following resources and lexico-grammatical realizations of closeness and distance pragmatic strategies have been considered for the analysis: a) hedges (e.g., approximators, modal verbs, semi-auxiliary verbs, conditionals, probability adverbs, verbs, adjectives, nouns, concessive linking words); b) first-person plural pronouns (e.g., inclusive ‘we’ forms); c) depersonalization markers (e.g., personifications, agentless passive sentences, nominalisations, impersonal passive constructions, there + be clauses); and d) heteroglossic markers (e.g., endorsement, attribution and distance).

Table 1. Total distribution of closeness and distance pragmatic strategies in the two corpora.

Total Distribution	Writer News Content		Reader Comments
	AF n/40,040	RF * 1000	AF n/42,330	RF * 1000
Closeness	131	32.42%	220	61.50%
Distance	274	67.58%	163	38.50%

AF stands for ‘Absolute Frequency’; RF stands for ‘Relative Frequency’. * Due to the different lexical density in the two corpora analysed, the absolute frequency was calculated per 1000 words in order to standardize the results obtained.

below shows a quantified evidence of how total closeness and distance pragmatic strategies are distributed in the two corpora analysed.

As seen in Table 1 above, the total occurrence of closeness pragmatic strategies in the reader comments data set outnumbers those related to distance. Nevertheless, the frequency of the latter is clearly higher in the news articles corpus. At first sight, these general quantitative findings show that news articles written by journalists who provide information on COVID-19 and comments made by readers towards the content seem to follow the genre conventions. On the one hand, journalistic discourse and the news article genre have been traditionally characterized by the canon of objectivity and impartiality, encouraging the use of distance pragmatic strategies. On the other hand, given the colloquial linguistic nature of reader comments, a higher occurrence of closeness pragmatic strategies is expected, as the table shows. Despite these general findings, we have found it necessary to carry out a more thorough analysis of the different interpersonal markers, through which these two pragmatic strategies are realized along with their different lexico-grammatical uses to study whether there is some sort of linguistic variation between both data sets that may confirm or refute our hypotheses. Results in the use of closeness and distance pragmatic strategies, their categories and their lexico-grammatical realizations are shown in

Table 2. Frequencies in the use of closeness pragmatic strategies in the two corpora.

Closeness Pragmatic Strategies	Writer News Content		Reader Comments
	AF n/40,040	RF * 1000	AF n/42,330	RF * 1000
Boosters	-	-	11	2.59%
Emphatic adverbs (e.g., actually, really, certainly, etc.)	-	-	7	1.65%
Emphatic phrases (e.g., it is clear that, in fact, indeed, etc.)	-	-	4	0.94%
Attitudinals	131	32.42%	189	44.64%
Attitude adjectives (e.g., interesting, terrible) best)	81	20.04%	88	20.7%
Attitude verbs (e.g., know, guess, feel, think, etc.)	33	8.16%	20	4.72%
Attitude adverbs (e.g., fortunately, personally, etc.)	14	3.46%	48	11.3%
Deontic verbs (e.g., must, need, should)	3	0.74%	20	8.22%
Exclamations	-	-	13	3.07%
Self-mention Markers	-	-	20	4.72%
First-person singular pronouns (e.g., I, me, my, etc.)	-	-	20	4.72%
Total	131	32.42%	220	51.97%

AF stands for ‘Absolute Frequency’; RF stands for ‘Relative Frequency’. * calculated per 1000 words.

and

Table 3. Frequencies in the use of distance pragmatic strategies in the two corpora.

Distance Pragmatic Strategies	Writer News Content		Reader Comments
	AF n/40,040	RF * 1000	AF n/42,330	RF * 1000
Hedges	91	22.5%	82	19.3%
Approximators	10	2.47%	1	0.23%
Modal verbs	24	5.94%	26	6.14%
Semi-auxiliary verbs	9	2.22%	7	1.65%
Conditionals	6	1.48%	16	3.77%
Probability adjectives, nouns, adverbs and verbs	13	3.21%	9	2.12%
Concessive linking words	29	7.17%	23	5.43%
Self-mention Markers	13	3.21%	47	11.1%
First-person plural pronouns (inclusive ‘we’ forms)	13	3.21%	47	11.1%
Depersonalisations	90	22.27%	13	3.07%
Personifications	20	4.95%	-	-
Agentless passive sentences	53	13.1%	8	1.88%
Nominalisations	10	2.47%	2	0.47%
There+be+clauses	7	1.73%	3	0.70%
Heteroglossic Markers	80	19.8%	21	4.96%
Dialogic contraction: proclaim-endorsement	34	8.14%	6	1.41%
Dialogic expansion: attribution-acknowledge	7	1.73%	8	1.88%
Dialogic expansion: attribution-distance	39	9.65%	7	1.65%
Total	274	67.82%	163	38.50%

AF stands for ‘Absolute Frequency’; RF stands for ‘Relative Frequency’. * calculated per 1000 words.

below:

This preliminary quantitative analysis reveals that both corpora show a variation in the use of stance, depersonalised and heteroglossic markers. In other words, the corpus of news articles on COVID-19 pandemic written by journalists include implicit self-mention markers such as depersonalisation strategies and heteroglossic markers to a greater extent than readers do in their comments. On the contrary, the occurrence of attitude markers and explicit self-mention markers is higher in reader comments. These findings support the idea that journalists and commenters follow different communicative constructions of their authorial voices, with different effects on the readership. This is clearly manifested in the objectiveness and impartiality through which journalists write on health issues such as COVID-19, that distinctly promote the use of depersonalised and heteroglossic markers. Contrarily, stance markers predominate in the reader comments, as a possible result of the colloquial modality of this communicative context. Notwithstanding this, readers do not seem to overlook the use of depersonalised and heteroglossic markers, when making their comments, albeit with much lower frequencies than those of the writer news articles. This is one of the chief reasons why we have found it convenient to carry out an analysis of a more qualitative nature with the two corpora, that may aid to develop these preliminary findings.

Even though the news articles corpus is not deprived of stance markers such as attitudinals, through which journalists provide a subjective view of the informational content, these are frequently combined with stance markers such as hedges as well as with a high frequency of heteroglossic markers, leading to a clear depersonalization of the journalistic discourse. By the same token, stance markers such as attitudinals, boosters and first-person singular pronouns used by readers also appear in combination with hedges, implicit self-mention markers and heteroglossic markers, seemingly producing a more distant, objective discourse. In the following paragraphs, we offer some examples drawn from the two corpora, to better illustrate our observations:

Writer News Content

• Despite this, the arrival of the Oxford/AstraZeneca vaccine would undoubtedly be a game-changer, said Helen Buckingham, director of strategy at the Nuffield Trust. “Then staffing will become the primary constraint,” she added […]
• If a new virus sounds scary, a new mutating virus sounds scarier still. In Kent in September, scientists now believe, somebody with Covid was the unlucky first person to pass on a variant form of the coronavirus that is maybe as much as 70% more transmissible than the version we have been used to […]

Reader Comments

3.

I’m reassured that the current vaccines would still fight B117 variant […]

4.

[…] The media seems wedded to narrow oversimplification of vaccine trial data. They never emphasise the critical point that all of these vaccines currently seem to offer protection against serious illness […]

In example (1), the writer of the news article uses an attitudinal adverb (undoubtedly) in the form of a stance marker to provide their own subjective view towards information on the Oxford/AstraZeneca vaccine. In spite of this fact, several distance strategies realised through hedges (Despite this, would) or even the use of an attributional marker (said Helen Buckingham, …) through which the writer literally reproduces the words of a third person to support this information, helping them to disassociate the proposition uttered from the text’s internal authorial voice, as seen in the aforementioned attitudinal adverb, leading to depersonalisation. The same interpretation can be applied to example (2) in which attitudinals (scary, scarier) are used by the writer to offer their view on a variant form of the coronavirus. Even so, this stance is depersonalised by means of grammatical realisations acting as hedges, as seen in the semi-auxiliary verb (sounds) to attenuate the opinion conveyed. This depersonalisation is also reinforced through an endorsement strategy (In Kent in September, scientists now believe…), in which the writer’s proposition is supported by making reference to external sources to prove the information transmitted.

In example (3), the explicit self-mention marker in the form of a first-person singular pronoun allows the reader to provide their thoughts on the vaccines to fight the B117 variant. The modal verb (would) acting as a hedge attenuates the implication of the reader toward the opinionated information. Example (4) deserves particular attention, in which we can perceive how the reader makes use of stance markers in the form of attitudinal nouns and adjectives (oversimplification, critical) to put forward their view about trial data on vaccines. Nevertheless, the recurrence to heteroglossia by means of attributional makers referring to a third person (media, they) together with the use of a semi-auxiliary verb (seems) perform a mitigating/attenuating effect, allowing readers to distance themselves from the opinion stated and consequently making their discourse more depersonalised.

No examples of boosters were found in the writer news content. Some relevant examples on boosters drawn from the reader comments corpus are shown below:

Reader Comments

5.

[…] I certainly won’t be taking the vaccine. It’s been rushed and there’s no guarantee it will be effective if the covid19 virus mutates […]

6.

[…] the current phase III trials are not actually set up to prove either […]

In example (5), the reader includes in their comment a booster realized through an emphasising adverb (certainly) to give prominence to their personal objections to get vaccinated. Nonetheless, this type of booster seems to be depersonalised in the adjacent sentence by adding both an agentless passive construction (It’s been rushed) and a there + be clause (there’s no guarantee). Something similar occurs in example (6) where the reader includes another emphasising adverb (actually) to stress their opinion on the information given in the article on vaccine trials. Nevertheless, this is inserted within an agentless passive sentence (are not actually set up).

As for the use of attitudinals, the following examples extracted from both data sets illustrate our findings:

Writer News Content

7.

Against the background of rapidly rising infection rates, Boris Johnson’s abysmal public rating appears to have bottomed out. According to the latest Opinium poll for the Observer […]

8.

[…] The results, reported in Nature Medicine, suggest the variant’s key mutation, known as E484K, does not dramatically undermine the vaccine’s protection […]

Reader Comments

9.

Maybe we should add sore throat to the list of symptoms requiring isolation and a test then; just a thought […]

10.

[…] Most members of the public believe that a vaccine against Covid.19 will prevent infection and save lives. Unfortunately, that is a very naïve belief as Dr. Peter Doshi explains in this publication by the peer-reviewed medical trade journal BMJ […]

11.

[…] This is brilliant news, yet we must be cautious as still early but any negativity is bonkers […]

Examples (7) and (8) reveal how the journalists include attitudinals to provide their stance towards the informational content, such as intensifying adjectives and adverbs (abysmal, dramatically). However, by means of both an attribution marker (According to the latest) and an endorsement strategy (The results, reported in Nature Medicine), these writers appear to distance themselves from the proposition conveyed.

In example (9), the reader includes a deontic verb (should) to give advice on the symptoms caused by COVID-19 that are in need of isolation. Yet, by adding the probability adverb functioning as a hedge (maybe) followed by the inclusive pronoun (we), the reader’s opinion appears to be highly minimised. In addition, when introducing (just a thought) at the end of the comment, the dialogic space seems to be opened to further comments from the expert writer or from other readers. Example (10) is also worth commenting. In this case, the reader makes use of both an attitudinal adverb (unfortunately) and an attitudinal adjective (naïve) to provide their stance towards the role played by vaccines to save lives. Despite this, it can be observed that he/she adds an attributional marker (as Dr Peter Doshi explains…), through which the writer seemingly distances themself from the proposition uttered, making their intervention more depersonalised. Finally, in example (11) attitudinals are also used (brilliant, cautious) to refer to the news that Moderna will show positive results by the end of 2021. Nevertheless, these stance markers change in the second part of the comment, in which the reader adds the collective pronoun (we), in an attempt to protect their own space and avoid responsibility.

The ensuing examples give evidence of the results yielded on the use of self-mentions:

Writer News Content

12.

[…] The UK is not the only country with B117 cases. It may be because we do more genome sequencing than others that we have picked up so many […]

13.

[…] Moderna has agreed to provide the US with 100 million doses, with an option to buy 400 million more […]

14.

[…] Devolved nations were advised to bring forward their own national lockdowns […]

Reader Comments

15.

[…] Not my opinion this is what the science is telling us and is backed up by 12-month experience now in places like Australia and NZ […]

16.

[…] Our body’s internal defence mechanism works well when we follow the natural path […]

17.

[…] Yes the science now is different and more advanced from previous pandemic in the past but this was rushed in my opinion to risk it […]

As observed above, the news articles include a higher occurrence of implicit self-mentions markers through depersonalisation strategies, such as the use of a collective pronoun (we) in example (12), the use of a personification (Moderna) in example (13) and a passive agentless sentence in example (14). These strategies reflect the specific genre conventions of the news article, characterized by distance, objectivity and professionalism.

Even though the reader comments include explicit self-mention markers such as the use of first-person singular pronouns, this personal opinion is mitigated by means of an attributional marker (what the science is telling us) as reflected in example (15). We can also notice the use of inclusive pronouns (our, we) in example (16), as well as passive constructions (example 17). By using all these types of depersonalised markers, the reader’s view is highly attenuated, giving room for alternative opinions that other readers may have in mind.

Finally, in relation to heteroglossic markers, findings drawn from our quantitative analysis revealed that the news articles include a higher occurrence of these markers due to the genre conventions. Despite this, the observational analysis indicates that some heteroglossic markers are used to back up reader comments towards the information provided by journalists. Examples drawn from both data sets may shed some light:

Writer News Content

18.

[…] In the UK, the NHS on Saturday revealed that a further 161 people who had tested positive for covid.19 have died […]

19.

[…] Earlier the UK health minister Edward Argar acknowledged there would be a need for booster shots […]

20.

[…] More than 1 billion people could be immunised against coronavirus by the end of next year with shots from the first two companies to reveal positive results, after the latest vaccine was shown in trials to have an efficacy of nearly 95% […]

Reader Comments

21.

[…] The fact that they have embarked on postcode testing in some areas shows the level of concern about the SA variant […]

22.

[…] Does it protect against severe cases of Covid? Hospitalisation and so on? Yes, according to data. Then it’s as good as it gets. The rest is media noise […]

23.

Bill Gates is on video record stating it takes FIVE YEARS to develop and test a vaccine for maximum effectiveness and SAFETY. These guys have NO IDEA what the long term side effects on the human body might be. I’m just…wow […]

In example (18) the writer makes reference to the findings revealed by the NHS, which is an instance of endorsement. An attributional marker can be seen in example (19), in which the writer makes an explicit reference to the UK health minister. Lastly, in example (20), the writer also resorts to the use of an endorsement strategy, as it is the case with the results of the vaccines shown in trials.

In example (21), an endorsement strategy is used to support the opinion towards the SA variant. The rhetorical questions made by the reader in example (22) are supported by an attribution marker (according to data). In the last example (23) another attributional marker (Bill Gates…) is included to support the opinion on the side effects Moderna may have throughout time (These guys have no idea what…). Once more, heteroglossic markers are of great help for readers to avoid responsibility towards the opinions introduced through their comments.

In order to conclude this Section 3, we can say that after comparing both corpora data sets, journalists in their news articles resort to a depersonalized and heteroglossic style to a greater extent than readers in their comments, complying with the genre conventions concerning impartiality and objectivity. Oppositely, readers show a stronger stance in their comments than journalists, not only with an important use of closeness strategies made of abundant self-mentions and attitudinals but also with some distance strategies, using hedges. Besides, depersonalisations and heteroglossia also appear in reader comments, even though to a lesser degree than in news articles. However, when observed as a whole, reader comments show a greater amount of external support strategies, in the form of scientific quotations or references, than that of stance strategies.

4. Discussion

As our findings reveal, readers’ authority construction is realised through closeness pragmatic strategies in the form of stance markers, mostly self-mentions and attitudinals, surpassing the use that writers of news articles make of these strategies. A subjective dimension is, thus, favoured. Nevertheless, readers also resort to frequent distance strategies in the form of depersonalisation and heteroglossic markers, which may appear intertwined with stance markers. Very importantly for our study, reader comments seem to contain more external support references than article news do, therefore, strategically reinforcing readers’ authority. This combination of stance markers, depersonalization and heteroglossic strategies aids to build a powerful authorial voice that projects assertiveness and closeness, together with the authority that depersonalization provides [20], deploying a strong persuasive power. As far as content is concerned, the reader comments in our study contain information on COVID-19 that can either be in line with the news article, thus reinforcing its positive effect on other readers, or contest what documented journalists say, therefore, entailing a possible negative effect that may contribute to creating pseudoscientific messages. What follows is a detailed discussion of how reader comments authority is built, aided by the different discursive strategies and including a reference to the content, labelled for each comment as supporting or challenging:

Concerning hedges as a stance category, reader comments reflect the use of some attenuating realisations, not only through modal verbs and conditional sentences, but also by addressing propositional content, sharing the discourse conventions of expert journalists. Readers appear to use hedges to strategically attenuate responsibility over their opinions, therefore, not openly invading the news writer’s personal space [29,30]. Hedges may also appear in combination with a strong, subjective authorial self-mention or with depersonalisations:

3. I’m reassured that the current vaccines would still fight B117 variant […] SUPPORTING COMMENT.

4. […] The media seems wedded to narrow oversimplification of vaccine trial data. They never emphasise the critical point that all of these vaccines currently seem to offer protection against serious illness […] CHALLENGING COMMENT.

As for attitudinal markers, they are frequent in the reader comments, especially in the form of qualifying adjectives or adverbs as well as adjectival and noun phrases. Through these markers, readers provide closeness and subjectivity [20], something aided by this type of discourse, characterized by concise messages and a colloquial modality [29,30,31,32], where attitudinals are bound to prevail:

10. […] Most members of the public believe that a vaccine against Covid.19 will prevent infection and save lives. Unfortunately, that is a very naïve belief as Dr. Peter Doshi explains in this publication by the peer-reviewed medical trade journal BMJ […] CHALLENGING COMMENT.

11. […] This is brilliant news, yes we must be cautious as still early but any negativity is bonkers […] SUPPORTING COMMENT.

22. […] Does it protect against severe cases of Covid? Hospitalisation and so on? Yes, according to data. Then it’s as good as it gets. The rest is media noise […] SUPPORTING COMMENT.

Taking into account the aforementioned colloquial nature of reader comments, the same interpretation is valid for self-mention markers, especially first-person singular pronouns. As the findings have yielded, this stance marker is more frequent in reader comments and non-existent in the news article corpus:

3. I’m reassured that the current vaccines would still fight B117 variant […] SUPPORTING COMMENT.

5. […] I certainly won’t be taking the vaccine. It’s been rushed and there’s no guarantee it will be effective if the covid19 virus mutates […] CHALLENGING COMMENT.

15. […] Not my opinion this is what the science is telling us and is backed up by 12-month experience now in places like Australia and NZ […] SUPPORTING COMMENT.

23. Bill Gates is on video record stating it takes FIVE YEARS to develop and test a vaccine for maximum effectiveness and SAFETY. These guys have NO IDEA what the long-term side effects on the human body might be. I’m just…wow […] CHALLENGING COMMENT.

The use of first-person plural pronouns in reader comments deserves special attention. Indeed, readers seem to resort to an inclusive “we/us”, behind which they seem to hide, so as to collectivize the utterance authorship and its related responsibility:

9. Maybe we should add sore throat to the list of symptoms requiring isolation and a test then; just a thought […] SUPPORTING COMMENT.

11. […] This is brilliant news, yes we must be cautious as still early but any negativity is bonkers […] SUPPORTING COMMENT.

15. […] Not my opinion this is what the science is telling us and is backed up by 12-month experience now in places like Australia and NZ […] SUPPORTING COMMENT.

As for impersonalisations and depersonalisations, their use is central in reader comments as a whole, and importantly enough, they could also be regarded as a kind of attenuation strategy, in tune with what [33,34,35] refer to as impersonalisation, shields and depersonalisations, respectively, thus giving a positive effect to the authorial image. Attenuation has been traditionally understood as a minimization of both the illocutionary force and the roles of participants in order to be effective in social communication and manage social image [36,37]. However, according to these linguists, depersonalised markers such as inclusive pronouns or agentless passive sentences do not seem to diminish the illocutionary force of the speech act—in our study, comments in the form of opinions. Something similar occurs with the use of agentless passive sentences. By means of both grammatical uses, readers resort to external support for their comments, helping the personal enunciator to make them invisible, thus producing a more distant, expert-like discourse, which is a powerful tool to convince other readers:

4. […] The media seems wedded to narrow oversimplification of vaccine trial data. They never emphasise the critical point that all of these vaccines currently seem to offer protection against serious illness […] CHALLENGING COMMENT.

6. […] the current phase III trials are not actually set up to prove either […] CHALLENGING COMMENT.

10. […] Most members of the public believe that a vaccine against Covid.19 will prevent infection and save lives. Unfortunately, that is a very naïve belief as Dr. Peter Doshi explains in this publication by the peer-reviewed medical trade journal BMJ […] CHALLENGING COMMENT.

21. […] The fact that they have embarked on postcode testing in some areas shows the level of concern about the SA variant […] CHALLENGING COMMENT.

22. […] Does it protect against severe cases of Covid? Hospitalisation and so on? Yes, according to data. Then it’s as good as it gets. The rest is media noise […] SUPPORTING COMMENT.

23. Bill Gates is on video record stating it takes FIVE YEARS to develop and test a vaccine for maximum effectiveness and SAFETY. These guys have NO IDEA what the long term side effects on the human body might be. I’m just…wow […] CHALLENGING COMMENT.

Summing up, as our results show, reader comments reveal a twofold nature in building authority, referring to discourse and content. On the one hand, from a strictly discursive viewpoint, a strong authorial self-construction is shown, enacted by closeness strategies that include self-mentions and attitudinals, creating an assertive persona. Likewise, distance strategies made up of depersonalisations, impersonalisations and heteroglossia, acting as attenuators that mirror scientific discourse conventions are importantly used. These results confirm what discursive interpersonality [22] claims, transcending conventional metadiscursive approaches [23] and showing lexico-semantic realisations and strategies dictated by each specific discourse. On the other hand, from a content viewpoint, reader comments contain information that is not guaranteed to be expert knowledge, since readers are not endowed with this power by the news article genre. As a result, reader comments are either in line with news article information on COVID-19, disseminating a positive effect, or contain some challenging information, frequently supported by external, not always scientifically accredited sources, which can influence other readers, spreading possible pseudoscientific information.

5. Conclusions

The findings drawn from our study lead us to interpret that, as in other Web 2.0 communication platforms, digital news article readers are now empowered with an agentive role as news consumers [36,37]. This has paved the way for numerous changes in the production and circulation of information, since the reader can now be considered a producer of texts and information [38]. In these participatory journalism cultures, engaging the public is assumed to assist journalists in “telling stories more honestly” [39]. This means that news is no longer a finished project that claims exclusive access to the truth [15]. In this vein, audience participation is seen as a new agency that signals a shift in power in favour of social agents [40], as well as a challenge to journalists, who so far have been gatekeeping power and media monopoly in “defining social realities” [15] (p. 23).

Following our analysis and results, it seems that our two hypotheses can be confirmed. Regarding the first one, our results prove that readers build their discursive authority through a strong authorial self-construction, creating an interaction with journalists that contests them at times and that challenges the conventional role of news article readers as audience. These results contravene the principle of recipient design [38], since the digital news article genre was fundamentally meant to allow readers to give opinions but not necessarily to let them create content that could influence other readers in ways that could harm their wellbeing.

As for the second hypothesis, and also according to our results, readers’ authorial self-construction includes closeness and distance pragmatic strategies [20] using stance markers and yielding a subjective and friendly discourse that can be very persuasive because it shares a colloquial language that other readers may recognize and accept as reliable. However, they also importantly resort to depersonalized and heteroglossic resources, meant to project a more convincing authority, based on external, scientific evidence, mirroring the discourse conventions of scientists.

It is needless to say that further research that analyses from a similar perspective a more extensive body of digital news articles and reader comments on matters of paramount importance for citizens would be required to extract more solid conclusions. However, our results already suggest a partial change of the readers’ role in the digital news article, partly because this technologically mediated genre has allowed readers to make comments, in an attempt to keep their loyalty. As a result, citizens feel entitled to produce knowledge information and to contest documented news article writers, something that endows them with a new power status in this participatory journalism. Our results seem to be in line with what [15] claims to be telling stories more honestly, with news articles not being a finished project until readers complete them with opinions and content. Parascientific communication might thus have turned inside-out [18], showing a positive dissemination of scientific knowledge where citizens can openly intervene and reinforce it with their opinions but also entailing dangerous implications when readers resort to content that may be pseudoscientific.

Although most media make an effort, with varying results, to disseminate parascientific information that provides citizens with true knowledge, some kind of defence against the spread of pseudoscientic messages seems necessary. Although it is a complex topic that entails ethical aspects dealing with citizens’ freedom of expression, and also with public wellbeing and health, parascientific practitioners might consider acting as gatekeepers and deploying mechanisms that can stop pseudoscience [2]. As mentioned above, this phenomenon of pseudoscience may be related to the social turmoil our society is going through [10], in which many people take seriously what lay people write about health issues, persuaded through a combination of discursive assertiveness and closeness, as well as an imitation of what scientific or parascientific writers would express in terms of content, impersonal style and external references. Concerning the question posed in the title, about a challenge of reader comments to parascientific information in news articles, it seems that some possibility exists, in light of our findings. Undoubtedly, ignorance is a breeding ground for pseudoscience so to counteract this ignorance and defend parascience, practitioners should learn from the discursive strategies that commenters use. Besides, an important tool is educational action, which can take place in the various science subjects at school, where the difference between science, parascience and pseudoscience should be addressed.