Models of justice evoked in published scientific studies of plastic pollution

“Developmental justice” is a term we devised specifically for trends we saw in this project, as it is not a term used in the wider EJ literature but is the most prevalent model emerging from our analysis. The developmental model of justice understands production and growth as inherently good, natural, and (or) the best place for intervention. This model of justice was evoked in several ways, outlined in turn.

Within studies, models of developmental justice often appear in the introductory or concluding paragraphs when the article discusses but does not evaluate the rise of plastic production over time. For example, “Plastics have become increasingly dominant in the consumer marketplace since their commercial development in the 1930s and 1940s” or “Our waste will continue to grow with increased population and increased per capita consumption associated with economic growth” (Jambeck et al. 2015, p. 768, 770). In some cases, the naturalization of industrial activities is explicit: “Plastic waste generation is an inevitable product of human activities” (Kumar et al. 2018, p. 781). When such statements are left without evaluation, we cannot tell if the authors are for or against developmental justice – the right of industry to produce and grow. However, when industrial phenomena are framed as “a given” or inevitable activity in the natural order of the world, they become naturalized, meaning that they are made to seem normal and unremarkable, and thus much less available to debate or change (Bijker 1997).

Developmental models turn plastic pollution into a downstream issue: the production of plastics is left untouched in favour of locating justice issues and interventions in the pollution itself, after production has occurred, which allows production to continue (Crang et al. 2013). This tends to be most strongly articulated in concluding sections that include recommendations such as recycling (15 papers) or circular economy (12 papers), terms that have low betweenness centrality so do not organize the conversation, but are sprinkled throughout the corpus. Other recommendations for solving plastic pollution through development include plastic credits (e.g., Hardman 2021), “life cycle consideration” (Soares et al. 2020, p. 7), education about consumption (Dalu et al. 2020, p. 2), changes in design including the circular economy (Zota et al. 2014, p. 239) and ingestion by microbes or mushrooms (Guo et al. 2020, p. 8; Caruso 2015, p. 1).

There is mixed evidence that actions based in developmental justice create change, particularly at scale (e.g., MacBride 2013; MacBride 2019; Gregson et al. 2015). Yet since plastics only come from industry sources, interventions into industry have the potential (though never the guarantee) to mitigate plastic pollution. Though we agree with other researchers that harm reduction rather than significant changes to the system is the most likely mode of change in a developmental approach to justice (Mah 2021, p. 111). For example, among the general U.S. population’s exposure to different phthalates, Zota et al. (2014) found that “reductions in DEP [diethyl phthalate] exposures have been the most pronounced, possibly because of changes in the formulation or use of personal care products, which are an important source of exposure to DEP” (p. 239). In the case of DEP, a phthalate, NGO advocacy, industry change, and regulation often moved hand in hand (e.g., Canada Minister of Health 2016). As such, development provides an important but limited model of justice, particularly when it is aligned with procedural justice linked to regulations and law.

A few texts, mainly on aquatic plastic pollution, used a development justice model to frame plastic pollution as a source of harm to economic stability or growth. For example, the United Nations Environmental Programme reports that, “in Thailand, where plastic waste import increased by 1,000 percent since the Chinese ban on plastic waste imports in 2018, shrimp farms have especially taken an economic hit from plastic pollution” (2021: 42, see also Jang et al. 2014; Laglbauer et al. 2014; Shim et al. 2018). Further examples are summarized in Table 3.

In another example, Beaumont et al. (2019, p. 194) argue that articulating harm to industry is a crucial mechanism for change: “Drawing on previous experiences of global pollutants (Van den Bergh and Botzen, 2015), we propose that the calculation of the economic costs per tonne of marine plastic is fundamental in future global negotiations to change the way plastics are designed, produced, used, reused and reprocessed. […] While explicitly recognising the limitations of the economic cost estimate presented here, we propose this as a foundation on which a Social Cost of Marine Plastic could be calculated”. Here, as is common in the corpus, authors describe a developmental theory of change as crucial yet limited.

Notably, some texts – almost always grey literature summarizing primary research or NGO-authored primary research reports – offer an explicit critique of the developmental justice model and its use by plastics industries in particular. For example, a report by the United Nations Environmental Programme noted that,

“the pandemic has been viewed as an opportunity to drive consumption of single-use plastics. For instance, during the early weeks of the pandemic in March 2020, the U.S. Department of Health and Human Services received a letter from the Plastics Industry Association requesting a public announcement from the Department praising promoting the health and safety benefits of single-use plastics and speaking out against bans of single-use plastic materials. The letter was sent a week after a peer-reviewed study was published demonstrating that the novel COVID-19 virus could survive on plastic surfaces for up to 72 hours, compared to up to 24 hours on cardboard surfaces” (UNEP 2021, pp. 7–8. See also, Schlegel 2020)

A smaller number of texts critique developmental justice methodologically. For example, Plastics Exposed (2019) by the Global Alliance for Incinerator Alternatives (GAIA) uses brand audits to indicate which companies produce shoreline plastic waste. They write that this research method is designed to “unmask how the industry has passed on the blame for the waste they produce to the consumers of their products, and the responsibility for clean-up of their packaging to governments” and to “reinforce the need for corporations to accept liability for the full life-cycle impacts of their products and the packaging in which their products are sold” (GAIA 2019, p. 6). This method is increasingly used in peer-reviewed shoreline studies as well (e.g., Okuku et al. 2021; Aragaw 2021). Other implicit methodological critiques of production include biomonitoring and epidemiological studies of chemical additives to plastics (e.g., Autian 1973, pp. 5–6; Zota et al. 2014; Rillig et al. 2021) and occupational health studies broadly.

We saw the model of developmental justice applied to science itself. These were most often calls for more data, more studies, more refined scientific standards or methods paired with the often explicit argument that more science and more data will lead to change. For example, it is frequently recommended “to undertake further research […] to enable the efficient development of future policy and regulation” (Beaumont et al. 2019, 194) or that further “Investigations […] will also enable prioritizing resources and to focus and steer conservation measures” (Thiel et al. 2018, p. 13). The theory of change implicit in this use of the developmental model is that more data or information is able to lead to change, whether by influencing publics, policy makers, or industry. Although this theory of change is common in knowledge-producing professions such as science and journalism (Callison and Young 2019; Singh et al. 2021), Singh et al. (2021) argue that more science does “not inherently lead to sustainable outcomes” but may in fact contribute to “unsustainable and inequitable development”, since powerful actors have long-used science to exploit the ocean and other environments (2021, p. 2). They emphasize that solution-oriented research based on evaluation and testing of existing efforts to make concrete change, rather than more basic science aimed at understanding ecosystems are “more effective in contributing to policy goals” (2021, p. 2). While this paper focuses mainly on models of justice, it is important to recognize that different models of justice (what is most important and right) are aligned with different theories of change (how systems change or shift).

There are different approaches to understanding where the differential distribution of plastic pollution harms comes from and thus where interventions might best occur. We frame these approaches as “upstream” or “downstream” because they locate responsibilities for harm variably along the value chain, from oil and gas extraction (upstream) to consumer choice (downstream), and variably between systems that produce unevenness such as racism (upstream) to symptoms of those systems such as racialization (downstream). Upstream approaches are closer to how distributive justice is understood in the EJ literature outside of the sciences, where harm is a result of complex and interlocking systematic oppressions and their material effects. Alternatively, downstream articulations of distributive justice locate the problem in waste disposal practices and consumer behaviour, often naturalizing both the production of plastics (pro-developmental justice) and the unevenness of its effects (distributive justice).

Within EJ literature, there are two main types of distributive justice: one that focuses on distributing benefits and burdens more equally so certain groups are not consistently overburdened with harms while others consistently gain the benefits, and another that seeks to guarantee protection from environmental degradation of any kind for all people (Clayton 2000; Blue et al. 2021). EJ professional Mike Ewall states that the two strains “represent the fundamental difference between the concepts of ‘poison people equally’ and ’stop poisoning people, period’” (Ewall 2012, p. 4). We find these two strains in our corpus. Upstream distributive justice mainly aims to decrease overall plastic or chemical production and even occasionally challenges the structural oppression that leads to social stratification to begin with. Downstream distributive justice focuses on evening out or mitigating the harm experienced by certain populations, such as through appeals to public education.

For decades, fence-line distributive justice has been the classic framing used to address the inequitable distribution of harms associated with chemical production and pollution (Bullard 2018). It emerged as a way to show how industrial production facilities, mines, pipelines, and landfills – and the pollution they produced – tend to be sited near low income, racialized, and Indigenous communities. A classic example is the 1987 report, Toxic wastes and race in the United States: A national report on the racial and socio-economic characteristics of communities with hazard waste sites, which found that three of every five (60%) Black and Hispanic Americans and approximately half (50%) of all Asian, Pacific Islander, and Native Americans lived in communities with uncontrolled toxic waste sites (Chavis and Lee 1987). The report showed that race was the most significant variable – greater than socio-economic status – in the siting of hazardous waste facilities in the United States (see also Tishman Environment and Design Center 2019; UNEP 2021). Because of this focus on production, fence-line distributive justice is often aligned with critiques of developmental justice.

A key characteristic of fence-line distributive justice is its use of a structural analysis that attributes harm to discriminatory systems rather than irresponsible individuals or events (Bullard 2018). In other words, fence-line distributive justice distinguishes between race and racism, class and capitalism, and gender and patriarchy in its diagnoses of the problem. Instead of framing the condition of being Black or poor as risk factors for certain exposures or forms of harm – which essentializes difference and locates responsibility in the individual – traditional fence-line distributive justice emphasizes the systems of exploitation (e.g., racism and capitalism) that create and distribute risks in inequitable and predictable ways (Pulido 2017). For instance, the report U.S. Municipal Solid Waste Incinerators discusses how the siting of plastic incinerators in poor and racialized communities “is not a coincidence but rather it is a product of historic residential, racial segregation and expulsive zoning laws that allowed whiter, wealthier communities to exclude industrial uses and people of colour from their boundaries” (2019, p. 13). Table 4 outlines examples of this type of harm in the corpus.

Related to fence-line distributive justice, we identified a specific upstream articulation of global geographies of injustice in articles that discussed relationships between the Global North to the Global South (a useful but imperfect categorization for looking at distribution [Dirlik 2007, 15]). According to a UNEP plastics report, “the increasing disconnection of economic benefits and ecological costs of the global economy has created opportunities for exploitation by more powerful actors from the Global North, while also complicating the attribution of liability” (2021, 14). GAIA’s report Discarded: Communities on the Frontlines of the Global Plastic Crisis (2019) emphasizes how waste exports from wealthy countries overwhelm the infrastructures of, and offload an array of environmental and health problems to, lower-income countries in Southeast Asia. In their article on global plastic and climate governance, Stoett and Vince underscore that it would be constructive to consider the historic plastic footprints of European and North American countries where industrial processes and large-scale plastics production were first innovated, rather than just contemporary snapshots (2019, 350). This framing of distributive justice illuminates the globalization of the fence-line by highlighting not only the concentration of global plastic disposal in the communities of developing, coastal, and island nations but also the upstream culpabilities of overdeveloped Global North governments and industries. Although there are many articles in our corpus that investigated global distributions and sources of plastic pollution and might have aligned with this model of justice, we found arguments of global fence-line distributive justice mainly in grey literature, social science, and journalistic sources (GAIA 2019; Petrlik et al. 2021), despite many articles that investigated global distributions and sources of plastic pollutions that had findings that would have aligned with this model.

The global or fence-line distributive models of justice become complicit in mis-attributing disproportionate responsibility to more disempowered Global South agencies. An example of this misattribution can be found in Jambeck et al. (2015) article, which argues that 83% of marine plastic debris is generated by 20 countries, most of which are developing countries. Jambeck et al. go on to frame the global plastic pollution problem as an issue of developing countries’ waste “mismanagement” (2015, 770). GAIA (2015) and Liboiron (2021) specifically critique this article’s misdiagnosis of injustice for overlooking the ways in which historical and contemporary processes of colonialism and imperialism craft such data, narratives, and waste flows to begin with. Thus, global economies of both plastic production and waste export complicate where upstream and downstream occur, as well as provide challenges for the use of global data.

Occupational health is the most central keyword in our corpus (Fig. 2). This field has an implicit framing of distributive justice in that it focuses on exposure to harm for mainly blue-collar workers and looks upstream in terms of investigating whether/when industry is the source. These studies emphasize the disproportionate disease burdens borne by workers through exposure to polymer-associated chemicals and particulates in specific industries and informal sectors, including synthetic textile manufacturing, plastics manufacturing, construction, waste picking, and beauty care work (Huang et al. 2011; Wright and Kelly 2017; Fucic et al. 2018; UNEP 2021; Varshavsky et al. 2020).

In its most upstream articulation, occupational distributive justice in the literature considered inequitable occupational distributions of harm in the workplace in relation to the racialized, gendered, and classed nature of labour divisions, particularly in the production of plastics and plastic additives (e.g., DeMatteo et al. 2012). These studies often found that women, Indigenous peoples, racialized, and working-class people tend to face higher occupational exposures due to the kinds of jobs available or assigned to them not only in production and manufacturing but also in blue collar service and waste disposable industries (N’dri et al 2015, p. 16; Varshavsky et al. 2020; UNEP 2021). In their pilot study on Vietnamese American nail salon workers, for example, Varshavsky et al. found that Vietnamese nail salon workers in California – most of whom are low-income women – are disproportionately exposed to phthalates in comparison to the general population (2020, p. 10).

However, some studies straddle upstream and downstream approaches. Although these studies concede that exposure is linked to occupation, they forward a developmental model of justice by advocating for downstream solutions: more rigorous occupational health surveillance via personal biomonitoring tracking systems, educating workers and doctors, conducting more research, providing better protective gear to workers, and improving workplace safety practices (Fucic et al. 2018, p. 5; Varshavsky et al. 2020, p. 12). That is, these studies do not align the problem (distributive) with the solution (developmental), which allows distributive injustice to continue, even if mitigate).

There is high agreement in the literature that there are strong race, class, sex, gender, and age disparities in terms of which populations carry the highest risks and burdens of plastic pollution (for a summary, see Bergman et al. 2013, as well as Adeyi and Babalola 2019; Belova et al. 2013; Nelson et al. 2012). In all cases, these studies use cultural categories of difference to identify uneven harms. At the same time, rather than pointing upstream to issues of racism, classism, sexism, ageism, colonialism, capitalism, and other systems that organize populations through differential privilege and oppression, they almost always direct their attention downstream, either to consumption behaviour and choices or innate biology. For example, in nearly all cases, the analysis of racialized disparities stops at ethnic differences and expresses uncertainty as to why such differences occur in product consumption and thus chemical exposure (e.g., Kobrosly et al. 2012, 15). This lack of explanation and conflation of race and racism leaves room for potential blame of exposure on racialized individuals for using certain consumer products. When researchers use racialized categories for comparison but do not take into account that plastic packaging is often related to affordability, cultural practices, and the types of stores that surround communities of colour, they collapse racism into race or ethnicity. By decontextualizing distributive injustice into categories of difference rather than systems of differentiation, studies can mis-align a distributive model of justice with developmental justice.

This trend is compounded when disproportionate burdens of plastic pollution along lines of race, sex, gender, and age are attributed to biology. For instance, multiple sources attributed the larger plastic toxicity burdens borne by women to “particular biological vulnerabilities” (DeMatteo et al. 2012, p. 428), “women’s specific biology” (UNEP 2021, p. 19), or to their use of certain consumer products (Arbuckle et al. 2015, 283), rather than to patriarchal social arrangements that constrict women to occupations of increased exposure like the household labour or informal waste picking (Lynn et al. 2017). That is, the problem is located in women rather than sexism, which risks legitimizing uneven harms (Moore et al. 2003). Like the trend in occupational health, many of these studies turn to the scientific development model and call for more research or enhanced biomonitoring of “vulnerable” populations (Arbuckle et al. 2015, p. 281; Kobrosly et al. 2012, p. 16), accruing resources to those least impacted (white collar researchers). This is another example of misalignment of models of justice within a single text (see Table 5).

The final type of distributive justice turns its attention (as we did above) to uneven power relations in scientific research, joining a wider multidisciplinary trend (e.g., Carter et al. 2021; Liboiron 2021; Mervins 2022; Robertson and Hairston 2022). A number of articles tacitly pointed upstream towards university institutions, funding agencies, and dominant research cultures as the sources of inequitable distributions of scientific attention, studies, personnel, and resources, and the negative impacts of this maldistribution. In their review paper on women workers’ chemical exposures in the plastics industry for example, DeMatteo et al. assert gendered biases in occupational health research. They argue that when occupation is considered without also considering gender, they find that the most impacted workers – women – are missed (2012, p. 440). In our literature review, there were only a few studies, DeMatteo et al (2012) among them, that considered the unique intersection of two or more markers of difference (Crenshaw 2017). While many studies considered multiple markers, they used a one-at-a-time approach.

Because this finding has direct ramifications for research design and methodology, we highlight two articles that take an intersectional approach using geography and either socio-economic status or indigeneity to analyse research biases and gaps. Blettler et al. (2018) argue that not only are freshwater environments massively understudied in plastic pollution research compared to marine environments, but that existing research on freshwater pollution does not even address the most polluted rivers. They argue that these research biases are produced through socio-economic differences between developing and developed countries (2018, 422. Also see Melvin et al. 2021 for a similar argument around plastics research in icy regions). Likewise, Liboiron et al. (2021) found that 100% of research leads who studied surface water plastics in Inuit homelands were not Inuit and were not from Inuit homelands “despite the existence and excellence of Inuit researchers” (2021: 21–22). Liboiron et al. posit that this results “in an overall regional skew of knowledge” (2021: 23) divergent from local research needs (also see Inuit Tapiriit Kanatami 2018).

Rather than advocating for better decision-making representation of marginalized or often-harmed groups, it was exceedingly common for researchers to make recommendations on behalf of these groups. Many advocated for scientists to be better represented in plastics governance so they could better represent harmed groups. Yet grey literature originating in social movements critiqued this model, arguing that scientists often speak over and obfuscate marginalized populations’ own priorities and knowledge in direct contravention of traditional ideas of procedural justice (GAIA 2015).

Another model of procedural justice in the corpus makes more explicit critiques of governance structures, infrastructure, and power. Direct criticism of weak governance systems, such as how insufficient regulations and enforcements create asymmetrical risks and harms, appeared in several articles (e.g., Njeru 2006). For instance, Korfali et al.’s (2013) study on phthalates in children’s toys on the Lebanese market criticizes the Lebanese Standards Institution for failing to set appropriate national criteria for levels of toxic chemicals (2013, p. 380). In the North American context, scientists have critiqued current dose-based regulatory conventions for failing to capture the low dose toxicities of endocrine disrupting chemicals (EDCs) (Charboneau and Koger 2008; DeMatteo et al. 2012). For example, Charboneau and Koger’s paper on the connection between EDCs and developmental disabilities condemns the affinity between government and industry, and the US EPA’s reliance on chemical manufacturers’ data (2007, p. 125). Social science research such as Njeru’s critiques of how economic instruments and anti-littering campaigns in Nairobi, Kenya individualize systemic issues of colonialism and capitalism is also part of this model of procedural (2006, p.1047). While critiques such as these do not always posit methods for achieving procedural justice, they do articulate areas of procedural injustice to be addressed.

Due to the frequency of environment-first statements in science, researchers have investigated the pervasiveness of implicit critiques of animal or ecosystem harm in studies that do not investigate such harms. Rochman et al. (2016) found that 27% of claims of harm in scientific articles are what they call “perceived” (stated without scientific evidence) rather than “demonstrated” (through the findings of the paper) and that many perceived impacts are articulated at the ecological level (rather than organelle, individual, or species levels). That is, scientists are making claims that plastics have negative impacts on ecosystems when those claims are not scientifically proven within the study. Our interpretation of Rochman et al.’s findings is that nearly a third of scientists in the study (n = 366) are expressing an environment-first orientation regardless of their research questions or methods (2016, p. 306).

In an interview, conservation scientist and plastic pollution researcher Alex Bond has said that, “As a scientist, you have to take a step back and almost detach yourself from the situation. But as a human being, when I cut open a dead bird and see bottle caps, tetra-pack lids and balloon clips gushing out of the stomach, it just breaks my heart. You think, ‘God, there’s no way that this has not severely affected this bird.’ But that’s a different thing than the scientific weight of evidence” (Bond and Liboiron 2018). Rochman et al. and Bond illuminate the same tension, that scientists who forward an environment-first model of justice often do so without toxicological or other scientific measures of harm in contradiction to scientific models of evidence that do require such measures.

There is a small but overt conversation in the scientific literature on plastic pollution that science is a potential cause of harm to animals, nature, and (or) ecosystems. Some scientists actively develop non-invasive methods to research plastic ingestion without the need to kill or even disturb animals, such as using ematics (pharmaceutical agents that induce regurgitation) and lavage (pumping the stomach with water) (Bond and Lavers 2013), visual identification of plastics in bird’s nests (Grant et al. 2021), boluses (Bond et al. 2021), and scat (Donohue et al. 2019). Some, though not all, of the articles cite animal welfare as the reason behind researcher’s choice to develop or use non-lethal methods.

Most sources articulated plastic pollution as a colonial injustice, arguing that plastic and “EDCs are materially a form of colonial environmental violence” (Shadaan and Murphy 2020, p 1) and that plastic waste management is part of an ongoing “struggle to overcome the inequalities of colonialism” (Njeru 2006, p. 1056). In a special issue on plastic pollution in Science, historian Rebecca Altman makes the case that bio-based plastics were not only “extracted under colonial regimes” but also were core drivers of Indigenous dispossession and sites of resistance (2021, p. 48). GAIA’s #StopWasteColonialism Campaign is based on “the effects of waste colonialism in the African continent. … Where our resources are returned to us, in the form of waste and cheap products made from toxic recycled materials. Where plastic waste has infiltrated its way into our land, oceans and physical bodies, severing our cultural connections with the earth and violating our rights to a clean and healthy environment” (GAIA 2021). Other sources argued that plastic pollution is a form of colonial injustice because it interrupts Indigenous food sovereignty and Indigenous relationships to fish and other aquatic life (Ngata 2018; Ngata and Liboiron 2020).

One scientific study on marine plastics called for “reconciliation science”, where “existing and ongoing Indigenous research relations should never be divided from scientific study and reporting” and should be characterized by “respecting Indigenous sovereignty” (Liboiron et al. 2021: pp. 2, 11). This text outlines how access to Indigenous land by non-Indigenous scientists, even when that access is for benevolent or well-intentioned environmental science, perpetuates non-Indigenous entitled access to Indigenous land (Liboiron 2021). The authors provide examples from existing plastic pollution research where researchers do not or may not seek consent, permissions, or permits for conducting plastic pollution research on Indigenous lands. They note that “of the 18 peer reviewed articles collected in our systematic literature review where Indigenous permits were required… five (28%) mentioned whether a permit from an Inuit Nunangat [Inuit homelands] research center was obtained” while more mentioned permits from settler Canadian government agencies even though both are required (2021, p. 8).

The same study found that all research on plastic pollution on surface water and in ice in Inuit Nunangat was led by non-Inuit researchers, resulting in all research on plastics stemming from the research interests and skills of white, settler, Western-trained scientists (Liboiron et al. 2021, pp. 8-9). These trends are in direct contravention of both the United Nations Declaration of the Rights of Indigenous Peoples (UNDRIP) that outlines Indigenous peoples’ right to self-determination in any aspect of knowledge or policy concerning their lands (United Nations 2008) and the Inuit Tapiriit Kanatami’s (ITK’s) National Inuit Strategy on Research (NISR), which seeks to change the norm wherein “the primary beneficiaries of Inuit Nunangat research continue to be [non-Inuit] researchers themselves, in the form of access to funding, data and information, research outcomes, and career advancement” and seeks to change this norm (ITK 2018, p. 5). Outside of Indigenous lands, the same finding is reflected between freshwater plastic pollution research conducted in the global north versus global south with the majority of research “performed in Europe and North America (67%)” and 69% in developed countries compared to developing ones (Blettler et al. 2018, pp. 418, 420). While this paper critiques power relations of dominant science institutions and colonial geographies, it did not examine whether the few studies in developing nations were conducted by “parachute” researchers from developed nations or by local researchers.

We also saw a consistent trend in published literature that implicitly worked against the Indigenous Sovereignty models of justice: many peer-reviewed articles claimed to be the first to know about plastic pollution in Indigenous lands, even when they also acknowledged Indigenous people were part of the research who lived in the area and thus knew about plastics (e.g., Laglbauer et al. 2014; Lusher et al. 2015; Mallory et al. 2021; Huntington et al. 2020; Pinzone et al. 2021). In almost all cases, these Indigenous co-researchers were not named or listed as authors. Attribution experts Anderson and Christen write that, “Firsting [stating that a researcher is the first to do or know something] is a linguistic act that supports and makes possible the physical act of taking: it is, fundamentally, an act of settler-colonial attribution. Firsting names something in order to erase what was before it—eliding both a previous existence and continued presence. Firsting, then, is a mechanism that supports a colonial property paradigm of possession through taking, naming, and attributing” (Anderson and Christen 2019, p. 121; see also O’Brien 2010). As such, taking up Indigenous Sovereignty models of justice in science would require changing cultural norms in the field, including but not limited to claims of novel knowledge in Indigenous homelands and attributing Indigenous knowledge from hunters, fishers, guides, and associations to non-Indigenous authors (see ITK 2018 for more).

In the classic EJ literature, distributive justice problems (uneven harms and benefits) are often addressed with procedural justice solutions (structural changes such as policy or participation by those harmed in governance generally) (e.g., Chavis and Lee 1987, p. xv; UNEP 2021, 13; Whyte 2017, p. 116). We found the same trend in our corpus: 20% of our closely read documents used these two models in combination exclusively, and another 18% included a third model (Fig. 4). These combinations most often occurred when identification of structural discrimination in the findings was accompanied by calls for stricter policies, laws, regulation, and oversight in the conclusion of the study (e.g., Ruiz et al. 2018; DeMatteo et al. 2012, p. 440; Varshavsky et al. 2020, p. 2).

However, most of these papers do not invoke procedural justice in its traditional definition, which emphasizes fairness in political representation and decision-making powers for different groups. That is, although upstream models of distributive justice critique weak governance structures and make expert appeals to policy makers for new policies or treaties (e.g., Charboneau and Koger 2008, N’Dri et al. 2015; Korfali et al. 2013), rarely do they directly address the need for more working-class, Global South, racial minority, women, gender minority, and Indigenous representation in politics, research, or grassroots advocacy work (exceptions include GAIA 2019, p. 40; Liboiron et al. 2021)

There are some instances where multiple models – three or more – are aligned in the same work. This trend is confined almost exclusively to justice-oriented research groups and grey literature. For instance, GAIA’s plastic pollution research is fairly unique in that their work encompasses nearly the full range of models of justice outlined here, including distributive justice by showing how plastics uniquely harm certain groups of people (e.g., Southeast Asian countries in the global waste trade, GAIA 2019), procedural justice in the demand to have waste pickers and other key rightsholders participate in decision making (GAIA 2019, p. 40), critiques of developmental justice in recommendations that foreground the need to reduce the production of plastics and hold companies accountable (e.g., Liamzon et al. 2020), and implicit invocations of environment-first justice (Liamzon et al. 2020, p. 6). GAIA clearly aligns these various models of justice together against developmental justice by showing how industry growth, profit, and waste practices are at the root of these injustices.

A common form of contradiction occurred in papers that paired developmental justice with nearly any other form of justice. These papers would often use a non-development model of justice in the introduction or design of the article and later recommend solutions that aligned with developmental justice. For example, Huysman et al. (2017, p. 46) gesture to oil extraction as the root of plastic waste when they write that, “The production of plastics consumes yearly 4 to 8 % of the global crude oil extraction”. This is an argument against developmental justice. But they then advocate for a circular economy where plastics are brought into waste-to-energy structures, which leaves the issue of extraction and production untouched. Similarly, Fucic et al. (2018) frame the production of endocrine disruptors as the core problem in how construction industries lead to increased occupational and regional exposure to endocrine disruptors: “In Europe, 20% of plastics production is used in the construction sector … However, there is no knowledge of the possible interaction of plastic materials with building materials, which may cause biological effects (synergistic and/or additive effects), and consequently no occupational safety protocols for work with such complex mixtures containing ED” (Fucic et al. 2018, p. 3). However, rather than proposing decreases in the production of endocrine disruptors, they recommend enhanced occupational surveillance by biomonitoring workers as the way to reduce exposure (see also DeMatteo et al. 2012, p. 440 for a similar argument that recommends substitution or re-engineering rather than biomonitoring). In all cases, both the root of the problem and the solution lie in the development models. The dominance of developmental justice in the framing of the problem of plastic pollution leads to solutionism, or the idea that all solutionism, or the idea that complex social, cultural, environmental, and economic problems have a technical fix (Rittel and Webber 1973).

From the perspective of justice movements and theories of change, placing both the problem and the solution within the same paradigm “means that only the most narrow of parameters of change are possible and allowable… For the master’s tools will never dismantle the master’s house. They may allow us to temporarily beat him at his own game, but they will never enable us to bring about genuine change” (Lorde 2012 [1979], pp. 111–112, emphasis in original). This is one reason that understanding models of justice can be useful to scientists and users of science: the intervention into the problem must address the injustices specific to the problem, rather than reach for incompatible solutions such as recycling, avoidance, or behaviour change that do not address the problem specifically (Rittel and Weber 1973; Liboiron 2014).

Though 8% of papers used distributive and developmental justice models together, in its most upstream forms, distributive justice is antithetical to developmental justice. Distributional justice models distinguish between the unequal responsibilities for and vulnerabilities to plastic pollution. That is, distributive justice shows us that consumers are not as responsible for plastic pollution, environmental harm, and individual harm compared to production, extraction, and corporations (e.g., GAIA 2019; Ruiz et al. 2018, p. 201), and that low-income groups in the Global South are not as responsible as peoples in the Global North (e.g., Petrlik et al. 2021, Stoett and Vince 2019). Developmental justice, on the other hand, supports the role of industry and economies in producing more of these products and effluents, which in turn accrue harms and benefits unevenly, and thus perpetuate distributive injustice. In other words, they erase the systems that cause unevenness by focusing on and naturalizing the symptoms of those systems instead.

We found a striking contradiction in the corpus in how particular developmental interpretations of “downstream” distributive justice were used to naturalize uneven harm. Most of these were studies of the impacts of plastic additives on different ethnic, racial, socio-economic, gender, and other groups, and they worked to locate the unevenness within those being harmed rather than the production of harmful chemicals. For example, Trasande et al.’s (2013) project on uneven phthalate levels in children from different ethnicities and races explains this correlation by speculating that that this correlation emerges from different consumer purchasing habits and genetic and epigenetic differences among the children’s families rather than the excessive production of polymer associated chemicals, siting decisions, systemic poverty and (or) racism, or corporate negligence (2013, p. 505). In this way, studies that use a distributive justice framework can fail to observe any of the tenants of distribute justice itself.

Some forms of justice are so antithetical to one another that the use of one can implicitly argue against other forms of justice even if they are not overtly discussed. One case is the relationship between developmental justice and Indigenous sovereignty. Indigenous sovereignty explicitly challenges developmental justice by highlighting the negative impacts of industrial extraction, waste, and growth on Indigenous or local peoples. For example, GAIA’s #StopWasteColonialism Campaign challenges “the effects of waste colonialism in the African continent…. Where our resources [extracted from colonial states/industries] are returned to us, in the form of waste and cheap products made from toxic recycled materials. Where plastic waste has infiltrated its way into our land, oceans and physical bodies, severing our cultural connections with the earth and violating our rights to a clean and healthy environment… We demand that the health and wellbeing of our communities be prioritized over profit” (GAIA 2021).

The Indigenous sovereignty model also challenges narratives about the relationships between plastics and “humankind” as a category, which requires a complete absence of distributive justice that turns all people into one equally accountable category (Liboiron and Lepawsky 2022b). Many scientific studies make claims such as, “Human behaviour is the sole source of marine litter, and changing perceptions and behaviour is key to tackling litter escaping into the natural environment” (Hartley et al. 2018, p. 945) or that “Humans are involved in plastic pollution and micro(nano)plastic presence in the environment in different dimensions of the problem: they caused it, can help address it, and may suffer from its impact” (Soares et al. 2020, p. 9). In these arguments, a particular type of human comes to stand in for all of humankind: a polluting, ignorant, usually affluent human in need of education. Yet none of these narratives account for the people represented by GAIA’s work, nor Indigenous peoples and their conservation efforts. Industrial and economic processes premised on constant growth and disposability are relatively new and come from specific cultures (Meikle 1995; Strasser 1999; D’Ailsa et al. 2014). Arguments about “humans” or “the” economy are “universalizing project[s], [that] serve to re-invisibilize the power of Eurocentric narratives, again re-placing them as the neutral and global perspective” (Davis and Todd 2017, p. 762). “Humankind” or “people” and their relationship to plastic, then, are “not precise enough terms for many Indigenous peoples, because they sound like all humans are implicated in and affected by colonialism, capitalism and industrialization in the same ways” (Whyte 2017, p. 159). Thus, when scientists use arguments that foreground humans, people in general, or consumers as a general category, they are also arguing against both Indigenous sovereignty and distributive justice models.