Cutting emissions, halting biodiversity loss: climate change solutions in the language of UK politicians and UK climate activists, 2018–2020

2.1 Climate change activists and politicians

There is both space and a need for more research examining the language of climate change activists and politicians, especially in terms of corpus-assisted discourse studies. To the best of my knowledge, there is no research to date which directly compares the language of UK activists and politicians, except for Cunningham et al.’s (2022) small corpus-based study. Both activists and politicians frequently discuss acting against climate change (e.g., Cunningham et al. 2022; Currie and Clarke 2022; Van Eck and Feindt 2022), but there are differences in their linguistic strategies and tendencies. Activists are more direct compared to politicians in how they discursively construct climate change and solutions, emphasising the urgency of the issue, and rejecting hegemonic, neoliberal framings, and economic concerns (Cunningham et al. 2022; Mangat and Dalby 2018; Molek-Kozakowska 2021; Penz 2022; Van Eck and Feindt 2022). This directness can be displayed through war metaphors and the framing of fossil fuels, elites, and industrialisation as “enemies” (Mangat and Dalby 2018; Molek-Kozakowska 2021).

However, this kind of language may have negative consequences, inducing fear and panic, and has been criticised by certain activist groups as demotivating and demobilizing (Cassegård and Thörn 2022: 3; White et al. 2019: 29). Alarmist and pessimistic “stories” are found to be prevalent in climate activist blogposts between 2009 and 2015, although hopeful narratives are also found towards the end of this period, discussing opportunities of addressing climate change (Van Eck and Feindt 2022: 205). Moreover, research specifically examining youth climate activist websites reveals a tendency to take a potentially hopeful, “forward-looking approach” to climate change policy (Falcone 2024: 60).

In contrast to climate activists, UK politicians have traditionally struggled to articulate the urgency of climate change. Interviews with UK politicians (Willis 2019) and corpus analyses of parliamentary debates (Cunningham et al. 2022; Kirk-Browne 2021; Willis 2017) show that climate change is often discussed as a technical, scientific issue, and in terms of economic costs – of climate change effects and solutions – thus perpetuating the kind of hegemonic discourses climate activists seem to reject. Moreover, impacts on the natural environment and on humans are often downplayed, or ignored (Cunningham et al. 2022: 15; Kirk-Browne 2021: 9; Willis 2017: 222–224) along with human agency as a cause of climate change (Currie and Clarke 2022: 601–602).

While Kirk-Browne (2021: 5) finds that the issue of climate change decreased in use in the UK Houses of Parliament between 2006 and 2018, other studies show UK politicians may be engaging more with the issue in recent years. Currie and Clarke (2022) show how climate change was discussed with increasing urgency in UK parliamentary discourse between 2015 and 2019 via the use of conflict metaphors. Moreover, Ebrey et al. (2020: 7), comparing the Tweets of UK MPs before and after the 2018 IPCC special report and subsequent large-scale climate protests in the UK, find that MPs became more inclined to tweet about climate change both generally and in terms of urgency and solutions, suggesting that the actions of climate activists might have influenced politicians’ language.

2.2 Semantic transitivity

Having reviewed previous research on how activists and politicians discuss climate change and solutions, I now outline semantic transitivity in the tradition of Systemic Functional Linguistics. This is usefully adopted as a tool in this paper to analyse how responses to causes and effects of climate change are articulated. Transitivity analysis involves categorising language into process types and the various participant roles which occupy these processes, thereby helping to reveal how people “encode in language their mental picture of reality and how they account for their experience of the world around them” (Simpson 1993: 88). The six process types in English identified by Halliday and Matthiessen (2004) are material, mental, relational, verbal, behavioural, and existential. The transitivity analysis in this paper is concerned with material process types (§3.2.3), in which someone or something does something in the physical world either literally or metaphorically.

Previous research employing transitivity analysis to examine discursive constructions of climate change solutions includes Wei (2023), who investigates agency and responsibility regarding causes of and solutions to air pollution in the English language newspaper, the China Daily, 2008–2018. Solutions to air pollution are mainly construed through material processes, with “social agents” such as places, governments, and humans employing verbs including “tackle” and “disperse” (pp. 369–370). Schleppegrell (1997) and Roman and Busch (2016) also examine agency and responsibility, but on smaller scales and in the context of US school textbooks about climate change and biodiversity. Agency can be obscured grammatically (for example, using nominalisations) or socially (for example, using “generic nouns” such as “people”) (Schleppegrell 1997: 57; Roman and Busch 2016: 1171). When solutions are discursively constructed, removing individual responsibility could lead to students believing that climate change is only a problem for scientists (Roman and Busch 2016: 1174–1175). On the other hand, implying that the readers of textbooks have agency in terms of solutions to biodiversity loss can lead to students accepting responsibility, but looking for solutions at the wrong level as they falsely believe it to be an individual rather than systemic problem (Schleppegrell 1997: 63–64). While these studies have focussed on agency and responsibility (the doers), there seems to be a lack of relevant literature examining the done-tos, regarding climate change.

3.1 Data collection

Two corpora of UK parliamentary and UK climate activist data have been compiled. The UK parliamentary data comprises debates and speeches from Hansard – an online record of what is said in the UK Houses of Parliament, from the nineteenth century to the present day. This was collected using a Python script which scraped all transcripts from the House of Commons between 2018 and 2020 via the website They Work For You (theyworkforyou.com). There is a considerable body of work on Hansard and what its data characterises; a dataset representing a range of voices, i.e., parliamentarians, reporters, and editors, it is nonetheless a faithful and accurate account of topics discussed in the UK parliament (Alexander 2023; Alexander and Struan 2022). While it is unsuitable for certain linguistic analyses such as hedging, it is considered wholly suitable regarding “primary sense-carrying parts of speech such as nouns, verbs, and adjectives” (Alexander and Struan 2022: 484).

The Activist corpus comprises data from the UK versions of the websites for Extinction Rebellion, Friends of the Earth, and Greenpeace. These organisations were chosen for several reasons, including the fact that all three are prominent within the UK and during the period between 2018 and 2020. As Extinction Rebellion began in 2018 – over forty years after Greenpeace and Friends of the Earth were established in the UK – the selection also provides a mixture of new and more traditional climate activism (De Moor et al. 2021). Moreover, all three activist groups’ websites provide comparable forms of data to analyse: blogs, updates, news, and press releases. This corpus thus also contains a range of voices as with the Hansard corpus, including different writers, editors, and occasionally interviewees. The lack of comparable data meant that the climate youth movement Fridays for Future – perhaps an obvious choice for the Activist corpus, given the time period under study – was not included. Again, a Python script was written to scrape all data from these sections of the websites published between 2018 and 2020.

Although the corpora represent different communicative situations – institutionalised, formal language of MPs and the personalised language of activists, aimed at mobilisation – care was taken to ensure that the analysis encapsulated the different language of both corpora in, for example, the creation of input strings (§3.2.1) and the selection of verb collocates (§3.2.2). The word counts of both corpora are displayed in Table 1.

3.2 Data analytic procedure

4.1 Goals – climate change mitigation

Goals referring to causes of climate change are categorised as follows: ‘emissions / greenhouse gas’ (e.g., “emissions”, “CO2”), ‘pollution / waste’ (e.g., “plastic waste”, “air pollution”), ‘fossil fuel’ (e.g., “support for fossil fuels”, “coal”), ‘human effect on land’ (e.g., “cutting down trees”), ‘food production’ (e.g., “intensive farming”), ‘other’ (e.g., “energy demand”). The frequency of Goals categorised is shown in Table 3.

There appears to be considerable overlap regarding the category ‘emissions / greenhouse gas’, especially in 2019 and 2020. This category occurs most frequently for both corpora and across all time periods – peaking in 2019 – except for 2018 in the Activist corpus, when it is second to ‘pollution / waste’. The most common verb occurring in both corpora and across all three time periods to discuss acting against ‘emissions / greenhouse gas’ is “reduce”, with “cut” as the second-most frequently used. While this suggests climate mitigation is similarly discussed in both corpora, i.e., that ‘emissions / greenhouse gas’ should be “reduced” more so than, say, “stopped”, or “eliminated”, the phrase “to (net) zero” occurs as part of the Goals here more frequently in the Activist corpus (367 of 666 instances) than in the Hansard corpus (9 of 681 instances). For example:

The category ‘fossil fuel’ occurs consistently more frequently in the Activist corpus than in the Hansard corpus. In 2019, however, the category peaks in use in the Hansard corpus while decreasing as a percentage share in the Activist corpus. 2019 is the only year when specific types of fossil fuel occur as Goals in the Hansard corpus (“coal”, “fracking”), whereas “coal”, “fracking”, “oil” and “gas” appear in the Activist corpus across all three years. Moreover, 2019 is the only year the verbs “ban” and “stop” occur in the Hansard corpus alongside ‘fossil fuel’; in 2018 and 2020, the verbs used are “divest from” and the perhaps semantically weaker “reduce” and “phase out”. In the Activist corpus, “reduce” occurs only three times with ‘fossil fuel’ and “phase out” is absent. The most frequently used verbs include “stop”, “end”, “ban”, and “divest from”.

In terms of both frequency of use and the way it is discussed, then, this category is most similarly represented in the two corpora in 2019.

‘Human effect on land’ occurs significantly more frequently in the Activist corpus in 2018 than for any of the other three time periods across either corpus; thereafter, the percentage share of this category between the two corpora is similar (though the raw frequency in the Activist corpus is higher), following the same trend of increasing in use between 2019 and 2020, so that in terms of frequency and percentage share, there is similarity between the two corpora in 2020. Almost all Goals occurring in this category are grouped into two subcategories referring to ‘deforestation’ and ‘pesticides’:

A surprising omission is the lack of reference to ‘(over)consumption’ being addressed as a form of climate change mitigation, despite these terms being included in the CQL input string. It may be the case that these phenomena are discussed in different ways. The category ‘food production’ touches on a similar area to overconsumption, concerned with lifestyle changes, but there are only five instances of these, all within the Hansard corpus.

4.2 Goals – climate change adaptation

Goals referring to effects of climate change are categorised as follows: ‘extreme weather’ (e.g., “floods”, “drought”), ‘effect on biodiversity’ (e.g., “biodiversity loss”), ‘temperature rise’ (e.g., “warming”), ‘effect on humans’ (e.g., “fuel poverty”), ‘other’ (e.g., “impending climate disaster”). The frequency of Goals categorised across the two corpora and all three time periods is shown in Table 4.

‘Extreme weather’ occurs with a consistently low frequency in the Activist corpus, although it is one of only two categories to appear in all three time periods. The category occurs frequently and consistently in the Hansard corpus, seemingly increasing in prominence over time, and occurring in over 60 % of all Goals concerning climate change effects in 2020. Aside from five instances in the Hansard corpus, and one in the Activist corpus, every occurrence of this category relates to flooding. The category is, moreover, almost entirely concerned with extreme weather events occurring in the UK:

The category ‘effect on biodiversity’ is most noticeable in the Activist corpus, where it changes from not being mentioned at all in 2018 to accounting for more than 80 % of Goals concerning adaptation in 2019 and 2020. In the Hansard data, this category is the third most frequently occurring in each time period. For both corpora, there is a peak in frequency in 2019. The high frequency in the Activist corpus seems to be largely due to a demand of Extinction Rebellion, repeated on multiple pages in 2019 and 2020:

In the Activist corpus, the specific Goal in this category is “biodiversity loss” in 293 of the 323 instances, whereas it accounts for 3 of 31 Goals in the Hansard corpus. In terms of how ‘effect on biodiversity’ is addressed, the verbs primarily employed in both corpora seem to metaphorize the category (Lakoff and Johnson 1980) as something advancing, and the role of humanity is to stop or reverse the journey:

“Halt” occurs 311 times in the Activist corpus, with “reverse” occurring 3 times. In the Hansard corpus, “halt” does not feature, but “reverse” is the most prevalent verb in each year and occurs 18 times.

The category ‘effect on humans’ is not represented in 2018 in the Activist corpus and remains infrequent in 2019 and 2020. In the Hansard corpus, it is the most prominent category in 2018 but decreases in 2019 and again in 2020. Every instance of this category in both corpora is triggered by the Goal “fuel poverty” or a variation such as “the scourge of fuel poverty” (except for two instances in the Activist corpus in 2019: “crop failure”, “human extinction”). “Fuel poverty” is itself an economic consequence of climate change. In addition to this, a common trend in both corpora is that the act of addressing “fuel poverty” is described as an economic opportunity or side-benefit of climate action:

4.3 Actors

Categories for Actors are applied to returns from both the causes and effects input strings. These are ‘human’, ‘human activity’, and ‘nature’. Subcategories are not presented in the tables due to space constraints. The category ‘human’ has the following subcategories: ‘nations / the world’ (e.g., “countries”), ‘the UK’ (e.g., “this country”), ‘activists’ (e.g., “Greenpeace”), ‘government’ (e.g., “the government”), ‘politicians / leaders’ (e.g., “ministers”), ‘industry’ (e.g., “retailers”), “we”, and ‘other’ (e.g., “people”). The subcategory “we” is used whenever “we” alone is the Actor, as in many cases it is impossible to discern from the context who “we” is referring to, even when expanding the co-text (e.g., if used in the Hansard corpus, whether it refers to all humans, everyone in England, in the UK, a political party, the UK government, etc.).

The category ‘human activity’ has the following subcategories: ‘technology / alternative fuel’ (e.g., “new engine technology”), ‘infrastructure / public transport’ (e.g., “a bypass”), ‘natural solutions’ (e.g., “increasing tree cover”), ‘laws / treaties’ (e.g., “legislation”), ‘measures’ (e.g., “policy measures”), ‘plans / policy’ (e.g., “our plans”) and ‘other’ (e.g., “our ability”). The category ‘nature’ has no subcategories, but includes “trees”, “peatlands”, “forests”, etc.

The frequency of Actors categorised across the two corpora and all three time periods is shown in Tables 5 and 6.

There seems to be greater similarity regarding climate change mitigation than adaptation. Actors observed from the causes input string are primarily categorised as ‘human’, followed by ‘human activity’, with ‘nature’ rarely featuring. In terms of percentage shares, ‘human’ increases over time in the Activist corpus, whereas ‘human activity’ is decreasing. In the Hansard corpus there is a slightly different pattern, with 2019 being the year when ‘human’ peaks and ‘human activity’ is at its lowest, so that the percentage share of these two categories is almost identical across the two corpora in 2019. The effects input string shows a marked difference. Here, ‘human activity’ is consistently the most frequently occurring in the Hansard corpus, followed by ‘human’. This is also seen in the Activist corpus in 2018, before there is a shift in 2019, continued to 2020, whereby the category ‘human’ becomes the most frequent, accounting for more than 80 % of Actors. In 2019, ‘human’ peaks in frequency and percentage share in the Hansard corpus, though this trend does not continue into 2020.

I now highlight the major trends in the distribution of Actors. There are similarities in both corpora regarding where most responsibility is attributed in the category ‘human’ (Table 7).

The subcategory 'government' is the most frequently occurring in the Activist corpus and the second most frequently occurring in the Hansard corpus. The subcategory “we” occurs most frequently in the Hansard corpus. In contrast, the subcategories ‘activists’ and “we” occur infrequently in the Activist corpus, while ‘activist’ does not occur at all in the Hansard corpus.

Most responsibility, then, appears to be placed on governments to enact climate change solutions, although the “we” subcategory primarily used in the Hansard corpus may be an example of agency being socially obscured (Roman and Busch 2016; Schleppegrell 1997), as it is not always possible to identify who the “we” refers to.

A noticeable trend in the category ‘human activity’ concerns the subcategory ‘technology / alternative fuel’, which primarily occurs in the Hansard corpus as an Actor for climate change mitigation, as the second most frequent subcategory of ‘human activity’. Table 8 shows that this subcategory appears to be decreasing over time in both corpora.

This trend occurs only once in the Activist corpus (19). In other instances, it is common that the benefits emphasised are not related to money or employment but to contributions towards reducing individual emissions (20):

Although this subcategory occurs predominantly in the Hansard corpus, the fact that when it does occur in either corpus it is used almost exclusively alongside causes of climate change, suggests a level of agreement. Climate change mitigation can be perceived as disruptive, so technological solutions may be more palatable than lifestyle changes (Lamb et al. 2020). Replacing a diesel car for an electric car, for example, is probably less disruptive than abstaining from using a car at all.