Corpus-assisted ecolinguistics and media framing: mapping climate narratives in Bangladesh

4.3.1 Frequency analysis

The first stage involved generating a word frequency list to identify the 50 most frequent content words (nouns, verbs, adjectives, and adverbs). Function words were filtered out to focus on semantically loaded lexical items. The minimum frequency threshold was set at 100 occurrences to ensure that the words analyzed were representative rather than incidental. Such thresholding is consistent with corpus-based convention that frequency cut-offs should balance representativeness with interpretability, as seen in prior studies that adopt thresholds around comparable scales depending on corpus size and research goals (Biber and Barbieri 2007; Cui and Kim 2023).To capture related lexical forms that did not appear in the initial top 50 list, the lemmas of these frequent words were also used to retrieve morphologically related variants (e.g. using economic* to include economically). This ensured a more comprehensive representation of semantically connected terms within the corpus. Frequency counts allow an overall idea of thematic salience of what topics dominate climate coverage and serve as entry points for framing interpretation.

4.3.2 Collocation analysis

The second stage examined collocational patterns surrounding key lexical items central to climate discourse, including climate, environment, carbon, disaster, and policy. Collocation refers to the statistically meaningful co-occurrence of lexical items within a defined span, which can reveal ideologically salient semantic associations and discursive framing tendencies. A collocational window of five words to the left and right (5L–5R) was applied, following established practice in corpus-linguistic research (Gablasova et al. 2017). To ensure reliability and interpretability of results, only collocates occurring at least 20 times in the corpus were retained, consistent with the approach recommended by Baker (2016) to focus on reasonably high-frequency patterns while avoiding spurious low-frequency co-occurrences.

Collocational strength was evaluated using the Log-Likelihood (LL) statistic, which estimates whether a co-occurrence is significantly more frequent than would be expected by chance (Dunning 1993; Gries 2013). LL was selected as the primary association measure given its robustness with medium-to-large corpora and topical convergence (Gablasova et al. 2017). Values above 6.63 (p < 0.01) were treated as statistically significant.

In line with Corpus-Assisted Discourse Studies (CADS) methodology, which combines quantitative corpus outputs and qualitative interpretation (Ancarno 2020), the quantitative collocate list was subsequently subjected to qualitative interpretive filtering to assess thematic relevance to climate-related frames including finance, governance, adaptation, and vulnerability. This combined quantitative–qualitative procedure supports the identification of lexicogrammatical patterns through which agency, responsibility, and environmental relations are constructed.

4.3.3 N-gram analysis

Four-word clusters (4-grams) were extracted from the corpus with a minimum frequency threshold of 20 occurrences to identify recurring multiword patterns in Bangladeshi newspaper coverage of climate change. These clusters capture extended lexical sequences beyond simple word co-occurrence, allowing for systematic examination of phraseological patterns. Following standard practice in corpus-based bundle studies, a threshold of 20 was adopted to balance the inclusion of meaningful recurring patterns with the exclusion of low-frequency, potentially idiosyncratic sequences (Liu et al. 2023). Frame identification was conducted inductively by examining the co-occurrence and thematic proximity of these clusters. Concordance lines and surrounding co-texts were iteratively reviewed to ensure that recurring patterns were captured consistently. These patterns were then prepared for further analysis to examine how linguistic choices may reflect relationships between humans and the environment, particularly regarding anthropocentric and ecocentric perspectives.

5.1.1 Key framing patterns in climate change coverage

To extend the lexical findings, this section analyzes how the most frequent content words function within their immediate textual environments. Using concordance lines extracted from the top-ranked items in Table 2, dominant thematic frames such as security, food and agriculture, gender, migration, and economic loss were identified. This concordance-based reading uncovers how the lexical items participate in constructing distinct discursive frames. Collectively, these patterns represent the key interpretive frames through which climate change is articulated in Bangladeshi media.

These narratives in Examples (1) and (2) present climate change as a direct threat to national security and stability, intensifying its discursive weight. Stibbe (2021) notes that security framing often elevates climate discourse’s prominence.

In Example (3), climate is clearly tied to food production and security, water resources – not just environment but subsistence. This framing makes climate intimately relevant to rural livelihoods and state stability.

The lens in Example (4) shows how women have to suffer more due to climate change. This aligns with global feminist climate discourse, which consistently demonstrates that women experience heightened exposure and reduced adaptive capacity due to socio-economic inequalities and gendered labor divisions (Arora-Jonsson 2011; Dankelman 2010; MacGregor 2010).

Examples (5) and (6) reveal that climate change is linguistically framed through an economic-loss narrative, where impacts are quantified in terms of financial depletion, productivity decline, and investment needs rather than ecological degradation or ethical accountability. Consequently, the “solution” to climate change is portrayed through a financial lens, positioning mitigation and adaptation as budgetary challenges instead of ecological imperatives.

This reflects what Olausson (2014) terms the economic rationalization of climate risk, where media prioritize cost, funding, and compensation as the dominant grammar of climate action. Similarly, Boykoff and Boykoff (2007) show that news discourse often translates environmental processes into policy-finance problems to engage institutional actors. Carvalho (2008) further notes that such economic framing reinforces administrative rationalism. Collectively, these tendencies reveal that Bangladeshi media mirror a global pattern where economic security overrides ecological security, sustaining what Stibbe (2021) calls an anthropocentric “story we live by”.

In Examples (7) and (8), “erratic rainfall” and “rising waters washed away homes” linguistically position natural forces as grammatical agents where rainfall and water perform destructive actions while humans are depicted as passive victims. This grammatical construction effectively naturalizes the crisis, attributing agency and blame to nature rather than to socio-political or industrial causes.

Such agentive representations of nature are common in media discourse. Carvalho (2008) argues that this framing depoliticizes climate responsibility by diverting attention from systemic human drivers to natural phenomena. Stibbe (2021) also observes that these “destructive stories” grammatically erase human accountability, turning ecosystems into threatening agents rather than living systems in distress. In the Bangladeshi context, such language reproduces a “victimhood narrative” (Alexander and Stibbe 2014), where people appear as helpless targets of an uncontrollable natural enemy. This narrative not only obscures anthropogenic responsibility but also hinders ecocentric understanding by linguistically constructing nature as the source rather than the sufferer of crisis.

5.2.1 From ecology to economy: the financialization of climate discourse

The collocational pattern of Climat* and Financ* clearly restates that Bangladeshi climate discourse is linguistically dominated by an economic frame. The Climat* node most strongly collocates with impacts (LL = 443.41), finance (LL = 416.00), action (LL = 233.03), justice (LL = 180.73), fund (LL = 139.75), budget (LL = 67.67), and adaptation (LL = 135.72). Similarly, Financ* frequently co-occurs with adaptation (LL = 106.74), support (LL = 82.42), green (LL = 60.17), and sustainable (LL = 20.07). These statistically strong collocations prove that climate discourse is portrayed through monetary and administrative frames, where the problem and its solution are both represented through financial discourse rather than ecological renewal.

Example (9) linguistically constructs finance as the missing agent of progress, reinforcing what Stibbe (2021) terms the “more-is-better” story we live by – the belief that growth and funding inherently represent improvement. Such discourse naturalizes expansionary economics as the route to climate recovery, even though this same logic contributes to ecological degradation (see Example 10).

This again strengthens the tie of solution with economy.

In Examples (11) and (12), justice and adaptation have been commodified by reframing moral and ecological objectives in economic terms. This aligns with Boykoff and Boykoff’s (2007) finding that media narratives frequently equate environmental responsibility with financial compensation, and with Carvalho’s (2008) argument that policy discourse tends to administratively rationalize ecological crises. Here, justice is achievable through money, not through structural or ecological transformation.

A related subcluster emerges around Financ* and green (LL = 60.17); and Governmen* and development (LL = 12.53).

Examples (13) and (14) reflect a national-interest orientation, where climate solutions are incorporated within the framework of economic modernization. The appeal to “green bonds” and “integration within development” foregrounds financial innovation over ecosystem repair. Nerlich et al. (2010) describe this as the “green growth” narrative, in which sustainability is domesticated within neoliberal logic.

5.2.2 Constructing human vulnerability through climate change narratives

Beyond the financial and disaster framings, the collocational results reveal that Bangladeshi climate discourse frequently ties climate change to essential life domains such as education, health, and social vulnerability. Strong associations such as education with children (LL = 87.49), health with public (LL = 174.58), health with mental (LL = 206.71), and women with girls (LL = 135.80) highlight that climate change is linguistically framed through its effects on human welfare rather than on ecological systems. Similar human-interest framings have been noted across global media, where climate coverage increasingly emphasizes human impacts to enhance emotional engagement and moral salience (O’Neill et al. 2015; Schäfer and O’Neill 2017). In doing so, it can mobilize empathy, public concern, or calls for action in ways that pure economic framings might not.

This pattern is exemplified in the Dhaka Tribune’s report:

In Example (15), the education and children collocation transforms a climatic event into a social crisis, showing how a fundamental right like education is jeopardized by environmental change. Such framing aligns with Boykoff (2011), who observes that journalists often domesticate climate narratives by linking them to everyday life domains.

A similar strategy appears in gendered framings. For example:

Example (16) positions women and children as emblematic victims, echoing findings from Arora-Jonsson (2011) and Dankelman (2010), who argue that media and policy discourses often construct gendered vulnerability to evoke empathy and justify intervention. The strong co-occurrence of women with children thus reproduces what Pepermans and Maeseele (2016) call “humanization framing,” where social categories stand in metonymically for climate impacts on society at large.

Overall, these collocational trends demonstrate that Bangladeshi newspapers not only frame climate change as an economic and environmental issue but also as a threat to everyday life such as education, health, and gendered well-being. These positive emotional framings support O’Neill et al.’s (2013) view that human-centered narratives can enhance perceived relevance and motivate engagement with climate action.

5.2.3 Disaster narratives and moral displacement in climate reporting

The collocational profile of Flood with cyclones (LL = 331.90) and droughts (LL = 109.19) reveals a dominant frame of disaster victimhood, where natural phenomena are foregrounded as causal agents of human suffering.

In Examples (17) and (18), nature is seen as the agent of destruction and human loss. Through such material process constructions, climatic elements (droughts, floods, erosion) perform destructive actions, while human subjects are grammatically positioned as patients or victims.

Similarly, human displacement is frequently attributed to climatic forces, which are linguistically foregrounded as the causal agents of social disruption. For instance:

In Example (19), hurricanes, as natural phenomena, are explicitly described as “deadly” and “destructive,” which positions them as the main drivers of migration. By attributing agency to hurricanes and other natural events, the press effectively portrays these occurrences as causal forces. This approach aligns with Hulme’s (2008) analysis, which identifies the prevalent framing of climate change as a catastrophic event. Such a discourse emphasizes the immediate and severe impacts on human societies, thereby reinforcing an anthropocentric perspective where ecological events are considered significant primarily for their consequences on human populations, rather than their intrinsic ecological value.

However, a subtle change in agency occurs in Example (20):

Here, human agency emerges, but in a distal and abstract form of “developed-country emissions” which deflects culpability from local to global actors. Such distributed accountability, while ethically charged, remains grammatically passive, framing humans as sufferers rather than participants in ecological processes. This linguistic pattern reflects a moral displacement within Bangladeshi climate discourse: nature is the proximate villain, and humanity, though implicated, remains rhetorically insulated from direct blame (Stibbe 2021).

5.2.4 From crisis to care: constructing ecological responsibility in climate narratives

Beyond crisis narratives, solution-oriented framings are also evident in the corpus, supported by prominent collocational patterns. The Energy node shows high LL values with renewable (783.61), clean (174.57), efficiency (135.92), transition (82.84), and solar (89.98), reflecting a discourse of technological optimism and sustainable transition. Likewise, Emission collocates with greenhouse (745.43), reduce (278.74), zero (119.92), and cut (99.14), signifying an emphasis on mitigation and responsibility.

As the Example (21) illustrates, agentive nominalization (e.g. factories, emissions) linguistically attributes accountability to human activity rather than abstract natural processes. This pattern is consistent with ecolinguistic analyses of environmental responsibility (Halliday 2001 [1990]).

Similarly, Example (22) foregrounds human agency in mitigation, aligning with the reconstructive discourse Stibbe (2021) advocates, where language promotes ecological care rather than destruction.

These positive framings echo findings by Fløttum et al. (2016) and Nerlich and Koteyko (2009), who note that climate journalism increasingly blends risk and responsibility frames, presenting energy transition as both a necessity and an opportunity for sustainable reform.

5.3.1 Framing climate change: from financial solutions to moral struggle

The 4-gram cluster profile reinforces the economic preoccupation already evident in the lexical and collocational analyses. Financially charged clusters such as loss and damage fund (41) and climate change trust fund (36) highlight that the imagined “solution” to climate change is constructed through fiscal mechanisms rather than ecological repair.

Example (23) situates environmental recovery within the lens of budgetary circulation and national gain. This pattern corresponds with what Boykoff and Boykoff (2007) call the economization of climate discourse, where environmental action is linguistically tied to finance, policy, and compensation. It also resonates with Stibbe’s (2021) critique of the “more-is-better” narrative, which conflates financial expansion with ecological progress.

Simultaneously, the high frequencies of impact(s) of climate change (366) and effect(s) of climate change (228) index a crisis frame grounded in vulnerability.

Example (24) reveals how existential threats link climate to food security and human survival.

Parallelly, Example (25) brings climate change under gendered precarity discourse, aligning with global feminist climate narratives that foreground disproportionate female suffering (Dankelman 2010). Collectively, these clusters depict climate change as an ongoing humanitarian emergency rather than a planetary imbalance.

Another significant frame evident in the corpus is the war or battle metaphor, where climate action is conceptualized as a collective struggle demanding courage, leadership, and moral commitment. This metaphor is linguistically realized through recurrent lexical clusters such as to tackle climate change (41), to combat climate change (33), and the fight against climate change (27). Examples (26)–(28) illustrate this frame clearly:

These examples collectively frame climate action as a battlefront, where human agency is valorized through verbs such as tackle, combat, and fight. The metaphor unites diverse actors like governments, communities, and individuals under a shared sense of moral and strategic commitment. It constructs climate response not merely as policy compliance but as a collective duty, thereby promoting solidarity and national resolve. The Dhaka Tribune excerpt, in particular, extends this metaphor by positioning Bangladesh as a “leader” in this global struggle, reinforcing a sense of national responsibility and pride.

As Nerlich and Koteyko (2009) observe, such militarized rhetoric can be persuasive in mobilizing collective action by evoking urgency and moral purpose. However, this framing also risks simplifying complex ecological dynamics into anthropocentric struggles, emphasizing human control and heroism over ecological interdependence.

A smaller but ideologically significant cluster is human-induced climate change, which highlights anthropogenic accountability.

Examples (29)–(31) collectively contribute to what Stibbe (2021) terms “beneficial discourse” – language that foregrounds human responsibility while fostering awareness of interspecies interdependence. They linguistically position humans as causal agents rather than passive victims, signaling a discursive movement toward self-implication and moral accountability. This tendency, though limited in frequency, resonates with critical-ecological framings (Stibbe 2021), which advocate recognition of human agency as a prerequisite for genuine environmental transformation.

In sum, Bangladeshi newspaper coverage of climate change is dominated by anthropocentric lexical and collocational patterns, with economic, security, disaster, and gender frames foregrounding human concerns. While the discourse largely positions humans as beneficiaries, victims, or moral agents within financial and administrative frameworks, there lies a glimmer of hope in the form of scantly lit beacons of the existing frames, where humans are still at the center but at the center of blame and responsibility that they have to take and fulfill.