Green Design Tools: Building Values and Politics into Material Choices

Introduction

An architect consults the manufacturer’s documentation about an energy-efficient fiberglass window that she is about to select for her latest green building project. Apparently the window sealant contains a solvent called toluene—a chemical name that sounds vaguely familiar, perhaps because it’s also an air pollutant emitted by petrochemical refineries in her local area. As a building product, the window has “green” credentials: it meets the environmental criteria of the highly regarded Living Building Challenge, a sustainable building certification program. But could it be toxic? Curious, she logs into a community website for architects concerned with safer materials and asks her colleagues for advice. Sure enough, toluene is linked to reproductive, developmental, and neurological toxicity from chronic low-level exposures—even though it does not appear on some of the many corporate “red lists” of chemicals to avoid in buildings. Should she be concerned? If she specifies this energy-saving window in her design, will it slowly release small quantities of toluene into the air that the building occupants breathe?

Architects and designers face dilemmas about how to attend to the harmful effects of the industrial materials that make up the built environment. As in our partly fictional example, they must navigate multiple incomplete and sometimes contradictory forms of environmental health knowledge. Even if well-informed, “green” design choices aiming to protect health and the environment frequently involve trade-offs between different impacts. For example, building materials and construction practices introduced for energy-efficiency gains have had the unintended consequences of exposing building occupants to worse indoor air quality (Steinemann, Wargocki, and Rismanchi 2017) and construction workers to substances that endanger their health (Guo et al. 2017). After decades of awareness that building materials can emit pollutants that affect human health, knowledge of these effects and how to effectively reduce them is still emerging (Murphy 2006; White and Pyke 2016; Dodson et al. 2017; Zimmer and Ha 2017). Moreover, designers have limited ways of influencing the larger network of technological choices in the industrial materials system that contribute to environmental health harm because these are often part of a sociotechnical structure that is impervious to deep changes.

Green design tools are emerging as a new response to these dilemmas and as an effort by a number of groups to reshape the material economy by intervening in design work. Environmental health advocates, scientists, and consulting firms are stepping in to provide building design practitioners with new tools to help them make informed decisions about the material consequences of their work (Goodwin Robbins et al. 2019). We use the term “design” broadly to encompass all aspects of the intentional development of material technologies, products, industrial processes, and built environments. Design involves work by scientists, engineers, product designers, company managers, architects, builders, and so on (Woodhouse and Patton 2004; Woodhouse and Breyman 2005). Green design tools include knowledge resources, science-based assessment methods, standards, databases, and computer programs that are all intended to help structure and inform decision-making in sustainable design. We focus on design tools that address the human and environmental health hazards of building materials because this has been a particularly active area of tool building.

We argue that green design tools play a key role in mediating whether and how designers’ values are translated into actual design choices; and at the same time, they embed the perspectives, values, and politics of their developers. Green design tools extend designers’ agency to intervene in the materials system by offering new capabilities to discriminate between potential material choices—for instance, based on evaluating and comparing their toxicity. What is less evident is how such tools mediate decision-making about materials in design processes by summing up toxicological and environmental knowledge in databases and algorithms, and by translating choices about what risks and impacts to prioritize into actual designs. Green design tools therefore function as “black boxes” that enfold complex scientific and interpretive work in ways that are not visible to their users, and that reflect the values, priorities, and assumptions of their makers.

When green design tools are contested, controversies that appear to be about their scientific validity may instead be rooted in tensions between implicit value judgments and interpretations of chemical knowledge. Our interviews with architects, interior designers, building sustainability experts, and tool developers active in the North American building industry revealed several such controversies. For example, some building product manufacturers and industry associations argue that green design tools overemphasize the health impacts of tiny quantities of harmful substances present in products, while environmental health advocates argue that those tiny quantities can represent real risks to workers and populations exposed through the supply chain or through waste streams. A common technical argument popular in the chemical industry is that health impacts should be understood and evaluated in proportion to the likelihood and magnitude of human exposure. But in a complex global industrial system, where the fate of long-lived synthetic materials is largely unforeseen (if not unconsidered) by industry and government actors, an underlying value-based question seems more relevant to ask: whose health are we trying to protect? And what does it mean to protect health: does it mean reducing harm to an acceptable level or aspiring to do no harm? We argue that all possible ways of evaluating toxic harm involve value judgments and that the prevalent discourse about hazard-based design tools misses this point.

In constructing green design tools, tool developers insert their own answers to the underlying questions that determine what a tool needs to do technically. As a result, they may inadvertently limit the ability of users to learn about chemical issues or critically appraise how the industrial chemicals/materials system might be different. This makes green design tools vulnerable in ways that echo many of the problems of regulatory science. Ultimately, selecting a given tool is itself a value judgment. To strengthen the legitimacy and credibility of green design tools, we suggest that designers and tool developers should be aware of and transparent about the values and politics built into the tools.

Values and Politics in Chemical Knowledge

Value-based decision-making permeates the production of scientific knowledge about chemicals. Jasanoff (1990) argues “regulatory science” is a hybrid of policy and scientific knowledge. The regulatory science that developed together with twentieth-century environmental policy has come to define and dominate understandings of how toxic substances cause harm, such as by positing the existence of acceptable risks and safe levels of exposure (Boudia and Jas 2013). The concept of risk is central to how regulatory regimes have sought to reduce and manage chemical pollutants without eliminating their industrial sources (Boyd 2012; Boudia 2014). Risk refers to measures of harm that express the probability and magnitude of health effects on human populations. Risk is commonly understood as a function of hazard (the inherent potential of a substance to cause harm), exposure, population vulnerability, and other factors—which can in principle be quantified using the tools and techniques of risk assessment.

How values influence the use of science in risk assessment and management is particularly relevant to green design tools. Assessing chemical risks for regulatory purposes calls for interpretation of scientific evidence typically characterized by high uncertainty, data gaps, and incomplete or evolving understanding of toxic effects. To act on such evidence, researchers and decision makers must resort to value judgments, interpretation of extant chemical knowledge, and ethical reasoning where the evidence is insufficient. For example, recent research examines the social, political, and economic influences on regulatory agencies’ risk assessments for determining acceptable levels of perfluorinated chemicals in drinking water (Cordner et al. 2019). Douglas (2000, 2009) identifies many legitimate and necessary, yet value-laden, methodological choices in toxicology and risk assessment. These include setting thresholds of statistical significance, classifying borderline experimental results (e.g., as normal or abnormal tissue samples), selecting appropriate models or assumptions for extrapolating quantitative dose–response relationships, and accepting or rejecting hypotheses about cause–effect relationships. These interpretive choices are often contested in contentious debates pitting the protection of public health against industry’s economic interests (e.g., concerning the safety of bisphenol A; Vogel 2013). Rather than dismissing regulatory science as “political,” this scholarship critically foregrounds the role of values and value conflicts in evidence-based democratic policy-making (Fernández Pinto and Hicks 2019). For example, policy makers can choose to adopt the precautionary principle as a guide for making decisions in the face of scientific uncertainty (European Environment Agency 2001, 2013).

Yet existing environmental policies have failed to prevent the permanent pollution of the biosphere (Boudia and Jas 2014), and the dilemmas policy makers face suggest even greater challenges when it comes to making and using tools to inform green design. The chemical knowledge systems of industrial and regulatory science—which privilege establishing direct, isolated relationships between individual molecular substances, industrial activities, polluted localities, and health effects—have been unable to grasp the complex reality of industrial materials in the environment, with their many social and biophysical relations (Murphy 2017; Boudia et al. 2018; Hepler-Smith 2019). Unexpected interactions, lengthy time scales, or confounding delays between cause and effect, ubiquitous presence, and the sheer scale of global pollution have again and again challenged established regulatory scientific models and definitions of harm (Liboiron 2016). Endocrine-disrupting chemicals, which invalidated key science/policy assumptions about health effects at low doses, are a case in point (Bergman et al. 2013). Policy makers’ reductionist compartmentalization of the world and our inattention to residual categories—“matter that is not supposed to matter” (Boudia et al. 2018)—has backfired, creating environmental health inequalities and institutionalizing ignorance. The technocratic politics of uncertainty and risk is just one facet of a deeper “toxic politics” (Liboiron, Tironi, and Calvillo 2018) where power relations are structurally embedded into how toxic harm is understood and acted upon (or ignored) through systems of scientific knowledge and governance.

Can green design tools exert a political agency of their own, and, if so, can it counter the prevailing toxic politics? In Winner’s (1980) analysis, technologies are political because specific technical arrangements can have real effects on the ordering and structuring of society. The processes of technological development are in turn shaped by a multitude of systemic biases, such as unequally distributed power to make value-laden technical choices, which can lead to arrangements that structurally favor some interests over others. Science and technology studies’ (STS) scholars have extended this analysis to investigate how social values and politics enter design, with or without conscious decisions by designers and participation by society (Nieusma 2004; Woodhouse and Patton 2004). Scientific tools are technical artifacts, and early STS research has explored how they are constructed and stabilized in scientific laboratory practices (Clarke and Fujimura 1992).

However, there has been little attention to technical tools used by designers. Liboiron has linked Winner’s insights to the politics of scientific tools, in her account of developing an environmental monitoring device that attempts to resist asymmetries in power and agency between local communities, polluting industries, and academic institutions (Liboiron 2017). Green design tools (as we will describe below) are more like scientific information systems; STS scholarship has shown how such systems can privilege particular sets of values and perspectives (Bowker and Star 1999; Bowker 2000). Tool developers in scientific fields increasingly make use of formalized rules and repeatable procedures (i.e., algorithms) to help automate information retrieval and classification and to produce decision-making aids that can be applied across many scenarios. Research by Benjamin (2019) and Noble (2018) raises serious concerns about the ability of developers to attend to the social agency of algorithms, particularly the ways that they can enact systemic oppression by amplifying racial biases associated with their inputs, outputs, and modes of application. Green design tools are unlikely to exhibit the same forms of social bias, but they may instead reproduce the uneven politics of chemical knowledge.

Green Design Tools in the Building Sector

Green design tools are knowledge resources that can inform decision-making in architecture and building engineering by enabling the prospective assessment of environmental impacts of designs and ultimately the selection of materials and products that are safer for human health and the environment. Many resources can function as design tools in green building, including assessment methods, rating systems, certification standards, databases, and computer programs (Haapio and Viitaniemi 2008; Zuo et al. 2017). These tools address many aspects of building design and construction ranging from the choice of products and materials to the operation and maintenance of buildings. Green building has traditionally focused much more on assessing and reducing energy and resource use than on the environmental health hazards of materials (Goodwin Robbins et al. 2019). Existing design tools reflect this focus, with architects and engineers until recently lacking actionable information to guide sustainable material selection (Franzoni 2011). Tools do exist for reducing the life cycle impacts of chemical substances in industrial processes (e.g., Bare 2011) and for reducing chemical toxicity at the molecular level (Faulkner et al. 2017), but these tools are too removed from building design to be realistically applicable. We therefore focus on green design tools that address chemical hazards and have been made accessible to building designers.

A variety of nonprofit organizations and private-sector firms develop green design tools in response to perceived gaps (or business opportunities) in industry decision-making. Some nongovernmental organizations (NGOs) active in environmental health and policy advocacy, like the Sweden-based ChemSec and the US-based Healthy Building Network (HBN), seek to influence industry practices, markets, and supply chains—to favor safer chemicals and materials—by furnishing the tools that industry and government have failed to develop by themselves. ¹ To make design tools, these organizations cultivate scientific and technical expertise among their staff: many of the individuals we interviewed who were involved in making green design tools had higher education in scientific and engineering fields, and most were also well-versed in public policy governing chemicals and materials. While the developers of green design tools do not necessarily have any training or industry experience in the design fields where their tools are intended to be used (e.g., architecture), their organizations do cultivate partnerships with industry actors and experts. Indeed, many NGOs that make tools also have close financial ties with businesses: some generate income from consulting or contract services; others, like the Health Product Declaration (HPD) Collaborative, are supported entirely by industry partners (HPD Collaborative 2020c).

A range of commercial firms also develop for-profit green design tools and market them as products or services that can help businesses outsource some of the work of green design. For example, the US-based company Scivera provides software platforms and services for toxicological assessment, business data management, and product stewardship. Designers in the apparel industry use Scivera’s tools to inform their chemical and material selections, such as selecting safer textile finishes and dyes (Rinkevich 2018). All of these tool makers—in both NGOs and firms—situate their work in a project to change industries, and they can be seen as participants in the “alternative industrial movements” of green building and green chemistry (Woodhouse and Breyman 2005; Hess 2007).

Green design tools share two broad functional characteristics. First, these tools organize and make sense of the information that is needed to understand the material health consequences of design choices. They identify information requirements (e.g., by collecting, or instructing users how to collect, detailed knowledge of a product’s chemical makeup), and they guide users to select the most relevant information. Second, they provide a means to measure success or failure in some aspect of green design. Tools typically include ways of interpreting data to enable assessments (e.g., technical criteria for safety), and ways of integrating and summarizing these assessments to produce decisions aids (e.g., a rating system for material health hazards). Tool developers may use knowledge from several scientific disciplines—such as environmental chemistry, toxicology, exposure science, and epidemiology—to construct green design tools, but this is no simple matter. Available scientific knowledge is typically incomplete, having significant gaps and many sources of uncertainty. Furthermore, the many dimensions of environmental health impacts mean that harm, or safety, cannot be reduced to simple metrics like efficiency ratios or numeric scores.

Three distinct but interrelated green design tools have become instrumental in green building: the GreenScreen for Safer Chemicals, HBN’s Pharos, and the HPD Open Standard. We analyzed these tools to understand how values and politics inform the technical choices made in their development. We interviewed professionals involved in the development and use of these tools to investigate the goals and underlying principles of the tools, the processes of their development, and how they are put into practice by designers. ²

GreenScreen is an open methodology for multi-endpoint chemical hazard assessment developed by the US-based organization Clean Production Action (2020b). As described by Heine and Franjevic (2013), GreenScreen is used to systematically compare different chemical substances based on their hazard properties. The methodology uses toxicological, biological, and physicochemical evidence to classify hazards from “very low” to “very high” across eighteen categories of human and environmental health effects. It then uses criteria and a decision logic to summarize those hazards in an overall rating on a scale from highest concern (Benchmark 1) to lowest concern (Benchmark 4). GreenScreen is not a piece of software or a service: human practitioners must carry out the work of conducting the assessment and generating the full GreenScreen report along with these hazard indicators. Once that is done, however, the GreenScreen Benchmark scores provide a simple, understandable indicator. Building design practitioners that we spoke with use Benchmark scores to identify hazardous chemicals in products and materials and to select safer substances, as do a range of decision makers in manufacturing industries, such as electronics (Holder et al. 2013).

Through the Benchmark system and its decision logic, GreenScreen enfolds complex assessment criteria and logic in simple pieces of information that can be communicated among scientists, designers, suppliers, and other actors. Performing a GreenScreen assessment requires knowledge of toxicology and chemistry, and it involves retrieving, organizing, and evaluating complex sets of chemical and toxicological information. In contrast, applying the Benchmark score in green design does not necessarily require any knowledge of how the scoring works. GreenScreen also includes a simpler methodology called GreenScreen List Translator (GSLT), which can be used to judge whether or not a chemical is already known to be of high concern. This is done by applying an algorithm based on the full GreenScreen criteria and using existing publicly available chemical assessments, which have been published by government and scientific sources that the GreenScreen developers consider to be authoritative (Clean Production Action 2020a). While easy to compute, GSLT scores are less informative than full GreenScreen assessments because they do not involve the open-ended and potentially deeper assessment process that practitioners carry out.

While GreenScreen offers a way of organizing evidence to generate knowledge that can help reach a conclusion about material hazards, Pharos aims to make these types of knowledge and evidence accessible to people who make material selection decisions. Pharos is a web-based tool that provides curated information about chemical substances from a large number of public sources, including information about toxicity, industrial use, and independently researched information about the chemical makeup of over a hundred categories of building products (HBN 2020). The US-based organization HBN originally developed Pharos as an information intervention (Kokai and Iles 2020) into the building products market: HBN believed that if designers could easily learn about the material hazards and environmental impacts of specific building products, they could more clearly express the green building sector’s market preferences for safer products to manufacturers.

Pharos presents information about building products in ways that are accessible to designers—for example, using the construction industry’s MasterFormat classification system to refer to generic types of products—but reframes products and materials through the lens of their health hazards and chemical relationships. Users can quickly discover whether a particular type of building product contains toxic substances by looking at a “roll-up” summarizing the most hazardous of its known chemical contents. Pharos also highlights chemicals that may be indirectly related to the product, as manufacturing inputs or trace impurities. In this sense, the developers of Pharos are encouraging designers to attend to the broader industrial lifecycle of building materials, not just what they are at the moment of installation. Broadening its scope beyond the building industry, Pharos serves as a general reference tool for understanding chemical hazards. Part of its general applicability is due to Pharos’ adoption of the GreenScreen as an information infrastructure. Pharos presents hazard information graphically, organized using the concepts and categories provided by the GreenScreen system; it also computes GSLT scores and provides a repository of publicly available full GreenScreen assessments of chemical substances.

Despite making it easier for building design professionals to find and understand hazard information relevant to material choices, neither GreenScreen nor Pharos addresses the issue that decision makers may not know the exact chemical ingredients of specific building products. Over the past decade, green building advocates and firms have increasingly promoted greater “transparency” in building products, contending that chemical ingredients should be tracked, evaluated, and disclosed by product manufacturers so that designers can choose less hazardous products (Geiser 2014). But achieving transparency is a struggle because manufacturers rarely know or even try to collect chemical and toxicological details about the materials they buy from their suppliers. In the United States, no regulations have required full disclosure of product compositions or full toxicological testing of chemicals. Instead, long-standing industry norms of trade secrecy have inhibited the flow of information through supply chains and prevented the accumulation of public domain data sets about how chemicals are used (Scruggs and Ortolano 2011; Scruggs et al. 2014; Krimsky 2017).

Stepping into this knowledge and policy gap, NGOs have developed voluntary chemical disclosure standards for the building industry, namely, the HPD standard developed by the US-based organization HPD Collaborative (2020a) and the Declare label developed by the International Living Future Institute (2019). We will examine the HPD standard because it includes explicit and stringent requirements for hazard disclosure. HPD is a standard for communicating the chemical makeup and associated environmental health hazards of building products. Manufacturers complete HPD documents describing their products, which are then distributed to architects, designers, and commercial clients along with traditional marketing information. Designers then use these manufacturer declarations to understand and compare the material hazard properties of each product, as well as comparing the level of disclosure that the manufacturers provided. The HPD standard requires manufacturers to disclose all known health hazards attributable to chemicals ingredients, as understood using the GreenScreen Benchmark system and using a set of “priority lists” designating chemicals of concern. However, manufacturers have some flexibility to choose how thoroughly they disclose chemical ingredients. ³ All information from HPDs is collected in an open-access public repository.

These three examples of green design tools are quite different in form and in purpose. But in practice, they are all used as a means to understand, evaluate, make, and rationalize material selection decisions. GreenScreen’s evaluative methodology is a tool to produce conclusions about substances and reliable comparisons between them. Pharos, a data system, is a tool for finding and comprehending chemical information through the lens of hazard assessment principles. HPD is a documentation standard, but in its actual use in the industry, it becomes a design tool in effect: a tool for collecting and summarizing information that can differentiate between competing building products and a conduit for manufacturers to present evidence of how their products meet the expectations of green building practitioners.

Creating Agency in Green Design

Industrial materials are developed and used in a sociotechnical system (Hughes 1987) that strongly shapes the possibilities for sustainable design. Designers seeking to reduce the life cycle environmental health impacts of buildings must trace those impacts back to building product design and manufacture, from there to the design and manufacture of industrial materials, and ultimately to the level of chemical design and production (Figure 1). Designers face barriers in exerting their influence through this multilayered network of technical and social relationships. The industrial materials system imposes what Dean Nieusma (2004) calls “agency-structure tensions” that function to preserve the status quo against the efforts of interventionist designers. These include dominant social assumptions about what designers should know and care about, as well as the dominant economic incentives of clients, product manufacturers, and chemical suppliers.

OPEN IN VIEWER

Figure 1.

Environmental impacts result from a multilayered network of relationships in the industrial materials system. ⁵

Green building designers need to push back against these structuring forces. Green design tools help them to exert new forms of influence through the materials system, while also mediating that influence in subtle ways. By using these tools, designers can make technical demands on products, materials, and their manufacturers in ways that were not previously possible.

First, tools give designers the capacity to access information about the chemical makeup and health hazard properties of building products. The HPD Open Standard has helped create the limited transparency that now exists, despite these deeply ingrained industrial norms and knowledge gaps. In 2014, a number of major design firms began demanding that manufacturers submit HPDs as a precondition for considering their products in new projects (Weeks 2013). A key driver for adoption of the HPD standard is the widely used Leadership in Energy and Environmental Design (LEED) green building rating system. ⁴ In 2014, LEED added a few “material health” credits that reward building projects if they use products with fully disclosed chemical ingredients or products that are documented to avoid certain high-hazard chemical ingredients (US Green Building Council 2019).

The HPD public repository now contains over 4,600 declarations (HPD Collaborative 2020b), but industry experts estimate that this represents “only a fraction of the tens of thousands of building product variants” on the market (Goodwin Robbins et al. 2019). One building product industry association, the Tile Council of North America, announced on Earth Day in 2020 that it would work with a consulting firm to generate full GreenScreen assessments of the thirty most common chemical ingredients of tiles and share this information with its member companies to help them prepare material ingredient disclosures using HPD and Declare (Tile Council of North America 2020). It may still be too early to judge the success of these transparency initiatives in institutionalizing chemical disclosure.

Second, green design tools enable designers to insert knowledge from chemical and environmental health sciences into their work practices. Whereas the chemistry of building materials has historically been opaque and off-limits to architects and interior designers, it now seems to be increasingly part of the territory in which they practice design. The architecture firm Perkins+Will developed one of the early and conceptually simple design tools in the industry: the P+W Precautionary List. Perkins+Will (2019) selected a set of substances of concern in collaboration with external scientific experts. The P+W Precautionary List still serves as a reference point for the firm’s designers about chemicals that should be avoided in buildings, and it is one of the several so-called red lists circulated throughout the industry. Many designers now deploy several green design tools—including GreenScreen, HPD, Pharos, and others—as part of a growing suite of knowledge resources. These resources enable them to interrogate the makeup of products and materials and to expose potential health hazards. ⁶

Green design tools that expand transparency are performing a political function. They seek to redistribute some knowledge and power from industry actors—who previously held exclusively authority over material design—to designers. Through these tools, designers now have the ability to make demands of those actors and to hold them accountable, at least in limited ways.

This can take the form of new information flows and feedbacks between designers and manufacturers, in which designers are able to exert new (albeit indirect) forms of influence. According to one informant, the process of completing an HPD can sometimes trigger manufacturers to reevaluate the product design: they may rather substitute or eliminate a GreenScreen Benchmark 1 ingredient than disclose the high hazard that it presents. To some extent, then, the market preferences of the green building sector are being transmitted through the use of disclosure mechanisms like HPD, with their intervening layers of interpretation. In a more common scenario, designers are unsatisfied with how little they still understand about product hazards based on the information manufacturers provide. ⁷ Designers often reach out to manufacturer representatives and engage them in a dialogue to fill gaps in the environmental health knowledge about the product. They use their own independent research—as well as knowledge gained from other green design tools—to investigate and push back on manufacturers’ lack of transparency or dubious claims, persuading representatives to track down more detailed information. Designers described essentially educating manufacturer representatives about why questions of chemical hazard are important. In one poignant account of such a dialogue, a sales representative became aware that the product they were marketing contained a toxic substance that had previously contaminated their own family’s drinking water.

By making scientific knowledge and analytical frameworks available to designers, the tools mediate whether and how designers’ values are translated into technical choices. Most architects interested in safer materials rely heavily on the outputs and interfaces of green design tools to inform decisions: chemical “red lists,” Benchmark scores, product ratings, and other decision aids. Some designers go beyond merely using the tools, cultivating their individual and organizational expertise in the larger scientific and political issues of environmental health. For example, some architects have set up “material health labs” within their firms to study the details of building product chemistry. People from design and engineering professions have engaged in lengthy discussions about chemical risks on the Pharos Project’s online forum (Kokai et al. 2020). They are motivated to attend to issues of occupational chemical exposure, material life cycles, and residual manufacturing impurities because they want to reduce ubiquitous chemical exposures. Still, most architects tend to accept the tools’ results as authoritative, without delving further into just how these results were derived.

Green design tools are thus intermediaries in the mechanism by which designers take on the responsibilities indicated by an ethics of material health and sustainability. Tools enable designers to operationalize values such as precaution, prevention of harm, and environmental justice, and translate these into making technical demands on products, materials, and their manufacturers.

How Tools Embed Values

At the same time, green design tools embed the perspectives, values, and politics of their developers. Our review of the chemical knowledge arena serves as a starting point for examining how this happens. Design tools, like regulatory science, assemble scientific evidence and policy framings to generate advice on highly uncertain questions, on which science itself may not necessarily offer the possibility of a clear conclusion (Jasanoff 1987, 1990). Likewise, both rely on constructing systems of definitions, conceptual models, and information infrastructures (Bowker and Star 1999) to ensure that the work of environmental protection can be coordinated. But this inevitably involves a narrowing of the many possible ways of knowing chemicals and their toxicities (Hepler-Smith 2019), choices about which aspects of a complex reality to attend to and which to leave out of the equation (Boudia et al. 2018), and power relations embedded in all of these aspects (Liboiron, Tironi, and Calvillo 2018).

Design tools are sites of analysis, calculation, and discernment about the possibilities for health and sustainability. Green design tools intentionally package chemical knowledge systems, with all their assumptions and contingencies, into simple interfaces and outputs. According to several tool developers, users demand simple and uniformly understandable indicators of chemical hazard. Developers prefer to make tools that are practically usable and immediately applicable: using a green design tool should be straightforward, not requiring users to fundamentally rethink the issues that the tool addresses—perhaps not even requiring a full understanding of how the tool works. Even if their methods are transparent, green design tools often effectively function as a “black box” that enfolds complex scientific and ethical reasoning out of sight of their users.

Constructing green design tools involves developers making a range of value-laden choices and introducing chemical knowledge politics into what may later appear to be a purely technical apparatus. First, all design tools are built on a set of core assumptions. These include value-based priorities and goals (such as reducing the toxic content of products, or occupational exposures) as well as possible design solutions such as product selection, chemical substitution, or risk management. These assumptions strongly influence what scientific and technical principles are then used to construct the tool. Pharos, for example, aims to help designers eliminate chemicals of concern throughout the full life cycle of building products—including products as well as their associated manufacturing processes and ultimate wastes. Pharos therefore uses hazard, or the inherent potential for a substance to cause harm, as the key organizing principle for measuring “healthy” or environmentally benign design. This is just one way to define what is “safe,” and it competes with the more dominant concept of risk—which, as a function of both inherent chemical hazard and human or environmental exposure, can theoretically be reduced by controlling exposure without reducing inherent hazard.

Second, green design tools integrate technical elements—data, metrics, categories, and criteria—to distinguish between design options based on the available scientific evidence. Values necessarily play a role in putting these elements into place. Tool developers must choose what kinds and sources of data (e.g., toxicological test methods, databases, computational models) are relevant and appropriate to evaluating environmental health impacts. They may set evidentiary thresholds to decide when sufficient scientific data exist to warrant a classification of harm (Douglas 2009). Classification systems are commonly used to “sort out” and weigh the many varied kinds of harm that might be controlled or prevented through design: cancer, reproductive toxicity, bioaccumulation, and so on. This can involve judgments about whether certain health effects are sufficiently important or well understood to evaluate as part of a policy program or a product design process. For example, GreenScreen largely follows the United Nations’ Globally Harmonized System (GHS) in specifying what health and environmental effects should be evaluated, and what kinds of evidence should be considered (United Nations 2019). But it adds several health endpoints that GHS excludes. In particular, GreenScreen addresses endocrine disruption, which growing—but highly incomplete and uneven—scientific evidence suggests can harm human development and reproduction. While increasingly recognized as a problem by European regulators, endocrine disruption has been heavily contested by industry and industry-funded scientists and somewhat neglected by US policy makers (Bergman et al. 2013).

Third, design tools contain assumptions, contingent decision rules, and other mechanisms that enable them to generate results under conditions of uncertainty and ignorance. In practice, most chemicals in commerce lack complete, detailed, and accurate information about their health hazards, their potential worker exposures, their end-of-life fate, and so on. Pervasive gaps in scientific knowledge about chemical hazards and exposures are well documented (Judson et al. 2009; Egeghy et al. 2012). Tool developers must decide how they identify, handle, and expose these uncertainties. For instance, the GreenScreen method dictates how to account for missing data in a chemical hazard assessment (Clean Production Action 2018, 79). The method permits larger data gaps in some kinds of health effects, reflecting both value judgments (e.g., chronic effects are elevated over acute effects) and pragmatic considerations (because some health effects, like endocrine disruption, are important but rarely studied). In the GreenScreen method, missing data are set to translate into a higher overall hazard assessment (a lower Benchmark score). In other words, unallowable data gaps are considered equivalent to evidence of potential harm. This embeds a precautionary bias into the tool that users may not realize exists.

Finally, for tools to be useful to designers, they must be configured to meaningfully distinguish between “good” and “bad” attributes of designs, materials, and substances. To do so, developers interpose multiple layers of interpretation in order to move from scientific evidence to actionable decision aids. This interpretation typically works by inserting into the tool evaluative criteria and decision logics that generate scores, ratings, or indicators. The GreenScreen method has two parallel systems of evaluative criteria: one system distinguishes between low, moderate, or high levels of concern for each of the eighteen hazard endpoints, and another appraises the level of confidence in the scientific evidence used (see Figure 2).

OPEN IN VIEWER

Figure 2.

A screenshot from Pharos showing the GreenScreen hazard summary table for the chemical DBDPE; superimposed: the GreenScreen criteria for persistence. Source: Clean Production Action (2018, 61).

The GreenScreen Benchmark system further distills the levels of concern for those eighteen hazard endpoints into a single Benchmark score that designers can use to evaluate the substance in question. This is done through a decision logic that explicitly prioritizes certain combinations of hazard properties to judge the overall level of concern. For example, substances are automatically scored as Benchmark 1 if they have a “high” level of concern in any of the Human Health Group I endpoints: cancer, mutagenicity, reproductive toxicity, developmental toxicity, and endocrine activity. These endpoints represent chronic human health effects, and GreenScreen emphasizes them because of the potential for severe irreversible effects on people from long-term low-level exposures—in contrast to other human endpoints, such as acute skin and eye irritation, or ecological effects such as aquatic toxicity. The Benchmark system’s logic and assumptions have been peer reviewed and are fully exposed to users in principle. In practice, users may not be aware that this information exists or be able to gauge its technical validity. The Benchmark system is meant to help designers by systematically identifying substances that fit certain patterns of hazard properties thought to be concerning. Yet this is not the only conceivable way of summing up hazard information. Scientists, regulators, and chemical industry actors can certainly raise objections over the value judgments that tool developers make. Is prioritizing a few chronic human health effects justified? What about ecosystem health or occupational exposure scenarios where acute hazards may be far more critical? In principle, design tools can include any kind of algorithm that weighs different types of evidence and environmental health impacts to produce an aggregate indicator or decision aid. The critical point is that trade-offs and value judgments are unavoidable.

Chemical Knowledge Controversies

Controversies about chemical knowledge can reveal more clearly how values are embedded in design tools. Even controversies that supposedly involve only technical questions are, under the surface, often rooted in tensions between different value judgments that tool developers make—which may or may not mirror those that users would favor. Value conflicts may arise in connection with environmental health and design problems such as:

What it means to protect health: are there acceptable levels of exposure and risk of toxic effects? Should there be no acceptable levels?

One particularly contentious issue relates to the outcomes that should be achieved by using design tools. Put simply, should hazard reduction or risk reduction prevail as the most desirable strategy to improve the environmental health performance of buildings? Take, for example, the Pharos hazard analysis of generic carpet tile product, shown in Figure 3. This product has a stain resistant treatment containing C-6, a chemical of high concern (GreenScreen Benchmark 1) at a level of 0.02 percent by mass. Based on this ingredient, what should designers do? Must they avoid the product outright because it is inherently hazardous, or can they justifiably calculate that the exposure—and hence risk—will be relatively low? If they can take the latter course, then the risk to whom—the building occupants, the construction workers, the manufacturing workers, or the people living around the factory?

OPEN IN VIEWER

Figure 3.

Pharos hazard analysis of carpet tile.

The use of GreenScreen, Pharos, and HPD in green building design reflects an underlying assumption by tool developers that hazard avoidance by designers is the most effective strategy to follow. The developers of Pharos deliberately chose to emphasize the hazard properties of a substance over its exposure potential, even if the chemical is present at trace levels or if it is only used during manufacturing and is not present in the final product. This is because they wanted designers to understand how their choices map onto a broader industrial system of chemical manufacturing that affects workers and frontline communities, not just building occupants. Making hazard the central focus shifts attention to why the product contains a hazard at all. Manufacturers are more likely to encounter demands to change or eliminate the product based on hazard avoidance.

In contrast, some scientists, companies, and chemical industry associations challenge the use of hazard-based design tools, arguing that environmental health impacts must be understood in terms of risk. They contend that GreenScreen and Pharos are flawed tools because these do not consider the magnitude of human exposure to chemicals, which might radically change the risk associated with their use. If a chemical is used at very low levels or is not widely used, then its risk could be minimal—even if it is hazardous. Finding ways to reduce exposure, then, might suffice. For example, designing a product to contain the toxic substance from being released, or training workers in safe use practices could diminish exposure. Two studies funded by the American Chemistry Council (a trade association) exemplify this argument. These studies compared several different chemical assessment tools and concluded that hazard-based tools are incomplete because they “lacked the capability to evaluate risk based on exposure” (Gauthier et al. 2015, 254) and that they are of questionable appropriateness because they involve “value judgments” (Panko et al. 2017). The makers of GreenScreen and Scivera Lens refuted these conclusions by pointing out technical and methodological flaws in the study but did not specifically respond to the claims of value bias (Palmer 2016).

Tool builders who focus on hazard reduction, such as HBN’s Tom Lent (2016), argue that evaluating risk is simply too complex and uncertain to produce effective and reliable green design tools. We gloss this argument but extend it to make the point that risk-based evaluations are just as value-laden as their hazard counterparts. Evaluating risk involves making many assumptions, without which it would be impossible to establish the quantitative relationships needed to make calculations (Douglas 2009). Estimating the magnitude and patterns of exposure encompasses a wide range of parameters (Greggs et al. 2019) and extends into highly uncertain territory such as predicting how materials will be used, managed, and captured as waste. Risk assessors typically calculate the incremental risk associated with individual substances in specific products or scenarios, but the reality is that we are all exposed to multitudes of toxicants that likely interact with each other (Callahan and Sexton 2007). Together, these assumptions reflect cultural and political biases about who or what needs to be protected from toxic threats. At the same time, with its compounded uncertainty and incrementalism (finding a “safe” level of exposure; Boyd 2012), risk assessment sends much weaker guiding signals for green design. Risk assessment can often exemplify what Jasanoff (2007) calls “technologies of hubris,” favoring prediction and control over a complex world. The incompleteness and partiality of scientific knowledge is a fundamental issue on which STS can provide at least some insight: the most appropriate response is humility.

Even if the hazard approach is favored, it can still be seen as limited in its dominant form, namely, appraising individual substances within a single product. Some practitioners and chemicals scholars argue this is a fundamentally reductionist perspective that overlooks the manifold complexities of how chemical substances move through a world of physical, ecological, and social relations (Liboiron 2016; Murphy 2017; Boudia et al. 2018). An alternative view could, then, frame hazard avoidance as beneficial only if it reduces the aggregate life cycle health impacts of the industrial systems involved. This critiques the system boundaries set by design tools, like HPD, which establish the product itself as the unit of analysis. For example, a business official from a major building product manufacturer told us that substituting a hazardous chemical with a safer alternative may not amount to a significant systems-level change, especially if the alternative material is “made in the factory right next door, owned by the same company, in Cancer Alley in Louisiana.” Moreover, this substitution may create even worse environmental health effects depending on the accompanying changes to the supply chain. One way that such unintended harm might happen is if a hazardous but recycled material is replaced with a less hazardous but primary petrochemical-based material whose use contributes to climate change and air pollution. ⁸

This more inclusive standpoint implies that single-chemical hazard properties may not be the most relevant organizing principle for constructing green design tools. Instead, population-level disease burdens, distributive justice, or greenhouse gas emissions may be more relevant. But this shift in system boundaries raises further value tensions between the many categories and scales of environmental health impacts that tool developers and building designers might consider. It also raises difficult questions regarding whether values and chemical knowledge choices should be embedded into a tool’s operations or should be kept open ended for designers to ponder and choose according to their own preferences. Should designers be expected to use tools to analyze the global impacts of every decision? Which parts of the industrial system should tools take for granted, and what should be left up to designers to imagine? Where does scientific analysis end and design intervention begin?

Many other value and knowledge choices permeate design tools. But designers may be oblivious to the decision-making that happens “behind the tool.” They may not fully realize that they are depending on results that reflect the “defaults” that developers have built into the tool. They may also not understand that they can consider a range of chemical concerns or even alternative framings.

Conclusions

Exposing the values and chemical politics of green design tools opens important questions concerning the agency of designers to intervene in an unsustainable and unjust materials system, and on what ethical basis. It is almost inevitable that the more designers and engineers try to comprehensively address environmental impacts, the more trade-offs they will need to make between incompatible approaches—trade-offs that require operationalizing some value preferences at the cost of others. How can designers weigh these trade-offs? What ethical basis do they have for setting priorities? Green design tools can quietly impose the ethical perspectives of their creators, rather than those of users, regulators, or even the public, to the extent that they implicitly encourage particular ways of prioritizing outcomes—such as elevating the apparent “greenness” of individual products over systemic sustainability and environmental justice. This process does not invite public debate or scrutiny, even though NGO staff may help produce the tools. Extensive social scientific evidence suggests technical experts can view environmental issues or risks quite differently from the public (e.g., Savadori et al. 2004). Values can thus be subtly and deeply embedded in the way the tool constructs assessments of problem-solving options.

In effect, green design tools have become a new “black box” species (Latour 1987). Their internal operations are understood as less important than their outputs for informing sustainable design. Nonetheless, this approach can generate environmental and social problems that are disguised by a sustainability imprimatur. That a design tool is intended to be “green” does not mean that it is, in fact, green in its effects. Yet, advocates—both NGOs and industry—are inserting these tools into voluntary standards, certifications, regulations, and government policies, to incentivize and reward designers for using them and to grow the market for green building materials and products. Without critical, thoughtful use, design tools can become de facto standards (Busch 2011) that are embedded into a variety of industrial and government practices. Like “molecular bureaucracy” (Hepler-Smith 2019), they can narrow possibilities without people noticing. They can become part of the invisible infrastructure that everyone takes for granted, or yet another part of the structural constraints that limit designers’ agency. In some ways, this is the goal: to integrate green design practices everywhere. In other ways, it can mean perpetuating policy failures—like the serial substitution of “bad actor” chemicals with new, unknown, and unconsidered harms (Geiser 2015)—and it can enable specific groups (industry, NGOs, governments) to push forward their particular approaches and interpretations without challenge.

We expose the politics of green design tools not to argue that they should have no politics or that they should be able to accommodate everybody’s value preferences. On the contrary, we argue that the most appropriate and powerful uses of green design tools can occur when designers intentionally invoke the tools with a clear understanding that they are choosing not just a tool but also a point of view. For this to happen, the values and politics that go into making green design tools cannot remain unexamined. These are part of the “questions before the question,” which Shapiro, Zakariya, and Roberts (2017, 586) argue both STS scholars, technoscientists, and humanistic practitioners alike should wholeheartedly engage with. Those authors make a call for “inviting apprehension”—for environmental inquiry and action that opens a deeper discourse around traditional scientific questions of quantifying harm or risk. Examining value conflicts and political debates, while certainly not an end goal, is one entry point into a conversation between the scientists who make green design tools, the practitioners who use them, and those whose lives and health will be affected by the resulting designs. As long as green design tools have their underlying assumptions, algorithms, and data properly interrogated and affirmed, they can play a valuable role in addressing the distributed harms of a toxic world.