Pragmatism: A practical philosophy for environmental scientists



Challenges to the credibility of the scientific community make it particularly important for environmental scientists to understand the bases for the authority of their science. We argue that pragmatism provides a defensible and effective scientific philosophy. It provides a transparent basis for justifying belief and a set of practices and concepts for inference. It makes the scientific community the author of scientific truth, which has implications for the opening of science in the age of social media and the communication of consensus positions on important issues. We describe how pragmatism acknowledges the social aspect of science without losing the scientific tradition of critical thinking. Integr Environ Assess Manag 2013; 9: 181–184. © 2012 SETAC


Do you feel unsure of the fundamental defensibility of your scientific work? Are you uncomfortably aware that philosophers of science complain that “most scientists are rather narrowly trained technicians who are largely innocent of the philosophical foundations of their chosen field” (Klee 1999)? We have good news for you. You are already practicing a perfectly legitimate philosophy known as pragmatism.

This is important to know because the philosophy of science is no longer just an academic concern, particularly in the environmental sciences. A false political ideology has arisen in the United States and is spreading elsewhere that claims that science is not a generator of neutral information, but rather is itself an ideology that is opposed to religion and capitalism and is dedicated to enriching itself at the public's expense (Oreskes and Conway 2010). Hence, environmental scientists must now justify the authority of their science. Defense of environmental science is particularly difficult because the complexity of the environment makes the confirmation of predictions ambiguous or even impossible except by waiting to see what happens after people modify the environment. Unfortunately, because most scientists have not been trained in scientific epistemology, they may have no appropriate response when science or its findings are attacked. A better understanding of the basis of our claims of credibility can give us an appropriate level of confidence and make our claims more defensible. It can also show that we are not just observers and testers but rather are natural philosophers who are providing the best available explanation of the natural world.


Pragmatism was developed in the late 19th century and early 20th century by a group of friends who met at Harvard University (Menand 1997, 2001). It originated with the governmental and academic scientist, mathematician, and logician, Charles Sanders Peirce, and was popularized by the academic scientist and philosopher, William James (1907). It has been applied to a variety of fields, most famously to jurisprudence by Oliver Wendell Holmes and to education and political philosophy by John Dewey. Although it is said to reflect the practical American character, pragmatism has persisted as an active school of philosophy in Europe as well as the United States (Bernstein 2010; Margolis 2010).

The central tenet of pragmatism can be condensed to “thinking is for doing.” That is, the truth content of an idea is determined by its correspondence with reality, which is determined by its real-world consequences. To scientists and engineers, this seems self-evident, but its application to religion, ethics, ontology, epistemology, and other fields has been controversial.


Science is concerned with generating justified beliefs about the natural world. For some, this is an attempt to know the absolute truth about nature. However, the pragmatists presented alternatives to idealistic ideas of truth (Posner 1990):

  • James argued that truth is what is good to believe in the sense that acting on the belief results in desired outcomes.

  • Holmes argued that truth is what survives in the competition of ideas.

  • Peirce believed that truth comes from the competition and synthesis of ideas. He believed in objective truth in the sense of real properties of the world that the scientific community's beliefs converge on. Therefore, the truth is what we tend to believe in the long term.

Note that these are not just definitions, but acceptance criteria for truth. This raises the question, who does the accepting? Putnam (1981) said that it is the right people—rational informed agents. Rorty (1979) said that it is all of us, when we are at our best. We suggest that the acceptors of a scientific truth (i.e., a justified belief) are experts who are fully informed and unbiased. That is, they have scientific training that makes them attuned to the evidence, demonstrated mastery of their field so that they have a deep understanding of the evidence, access to the relevant evidence, and an absence of personal financial or professional interest in the outcome.

In sum, pragmatists believe the truth is out there in the sense that there is a real natural world against which beliefs are tested, but the truth of beliefs is not absolute. Combining Peirce's and James's criteria, truth is what works best in the real world for the community of scientists. In that search for truths that work better, we can apply pragmatic scientific methods.


Rather than a single scientific method, science consists of a variety of methods including various mixtures of experimentation, observation, and modeling. It also includes techniques such as randomization and hypothesis rejection, attitudes such as objectivity, and ethics such as openness and honesty. Pragmatism's emphasis on doing what works makes it appropriate for this heterogeneous practice. Because pragmatism was intended to bring the scientific world view to philosophy and philosophical rigor to science, it is not surprising that it seems reasonable to scientists. The following pragmatic techniques are widely accepted by scientists, but their links to pragmatism are not generally recognized.

Pragmatism and inference

Peirce invented a new type of inference, abduction, which is reasoning to the best explanation. He proposed abduction as a way of framing hypotheses that would then be tested by deducing consequences and testing through induction from observations. However, his process is incomplete. Although induction, reasoning to general conclusions from the results of an experiment or set of observations, is necessary, it is unreliable, so it cannot be the last step. Peirce apparently did not recognize that ultimately science must return to abduction to synthesize and explain results. As the philosopher of science, Colin Howson (2000), has written, “Inference to the best explanation, also known as abduction, is the inferential method of science…” [emphasis in original].

Pragmatism and weighing evidence

Weighing multiple lines of evidence is a pragmatic method and an important tool of abduction. Peirce argued that any single line of reasoning, like a chain, is likely to fail due to a weak link, so science should be like a cable spun from multiple strands (Bernstein 2010). James introduced the term “pluralism” to the English language to describe the application of multiple methods to analysis of a case (Menand 2001).

The weighing of evidence is particularly important for evaluating whether or not a hypothesis really works. In many cases it is not clear what hypothesis works best or even whether a hypothesis worked because it is true or because of coincidence, technical errors, or some unknown factor. How applicable is that laboratory test to the field? Is the association in that field study representative of a general causal relationship? This is where the various conceptual tools of science must be deployed to go beyond an experiment or observation. Are the results consistent with a more general theory and with theory in related fields (i.e., is it consilient)? Is there a mechanism that explains the results? Is it confirmed by independent investigations, and is that confirmation insensitive to conditions that should be extraneous? In sum, pragmatic science must work in all relevant circumstances for any sincere and skillful investigator, and it must be consistent with accepted scientific truths. These considerations may be neglected by the individual investigator, but the scientific community must consider them before granting its concurrence.

Pragmatism and confirmation

The essence of scientific pragmatism is that a hypothesis must be confirmed or refuted by experiments or observations. Furthermore, those results must be confirmed by independent experiments or observations. Finally, when science is applied, its predictions concerning the outcome of actions must be confirmed by observing the results of the actions. For example, if we hypothesize that adding nutrients will increase the degradation rate of a chemical, we test that hypothesis in the laboratory, we confirm the results in other laboratories or in other test systems, and finally, if nutrients are applied to a chemical spill as a remedial action, we monitor the change in chemical concentration for real-world confirmation.

Pragmatism and scientific consensus

Peirce argued that science is a community enterprise that is converging on but never absolutely achieving the truth, but he did not explain how that consensus is built. Traditionally, consensus has arisen informally and organically, and its achievement has been reflected in awards to the originators of ideas, in the appearance of those ideas in text books and in the use of the ideas to take additional scientific steps.

In addition, mechanisms for establishing the scientific community's truth are built into the practice of science. Peer review is not only a quality assurance mechanism for publishers. It is more fundamentally a means for authors to determine whether their results are sufficiently compelling to obtain buy-in from their peers—that is, to change or expand the community's current version of truth. Consensus is more directly obtained from expert panels such as those convened by the US National Research Council to review contentious topics. The panelists, who are chosen for their expertise in the field of concern, reveal their potential conflicts of interest and are provided with literature, briefed by the stakeholders, receive public comments, talk among themselves, and exchange draft opinions. Their reports constitute a sort of facilitated scientific truth concerning a topic at a particular time. Workshops such as the Pellston series, currently organized by the Society for Environmental Toxicology and Chemistry (SETAC), are another mechanism to generate a scientific truth. Consensus is particularly strong when multiple independent and credible panels reach the same conclusion, as is the case with anthropogenic global warming. These formal mechanisms can speed the development of scientific consensus relative to the traditional organic process by facilitating discussion, providing a forum for identifying the current state of a scientific issue, and creating a vehicle for communicating that view more widely.

Pragmatism and scientific ethics

The concept that scientific truth is generated by the scientific community implies some important ethical standards. Scientists must be honest, open, and cooperative if they are to contribute to a scientific consensus. These ethical standards are taught to young scientists and most of us practice them. A pragmatic philosophy shows just how essential they are. Without consistent scientific ethics, there can be no trust and hence no grounds for consensus and no scientific truth.

Openness to the point of transparency is not only good for the advancement of science, it is important for gaining the acceptance of the wider society. When scientists withhold data and other information, they fuel distrust and the perception that science is a closed cabal. We scientists may have a right to go to court to keep our records private, but when we exercise that right, we do harm to the scientific community and to the broader society. Consider the deleterious effect from the exposure of emails between a few climate scientists who were withholding information and the subsequent court battle to withhold a climate scientist's records from a state official. These behaviors were legal and even ethical in a formal sense, but they did not help to gain the public's trust (Russel et al. 2010). Behaviors that earn trust are essential for scientific advancement and for society to use the knowledge generated by science.


The biggest problem with pragmatism is that it seems shallow and not very rigorous. After all, its method for determining truth can be condensed, only somewhat unfairly, to “whatever works.” Pragmatic science seems to surrender the high ground to its opponents: they seek Truth whereas we seek only a more workable approximation. Therefore, it is important to remember that there is no defensible basis for claims of truth other than determining whether an idea works in the real world (James 1907). Logically, there is no higher ground, but only a shared intellectual advance to more broadly and deeply useful scientific consensuses.


This article was inspired by a statement by a member of the editorial board of this journal that “science is not a matter of voting.” We suggest that it is a matter of voting, but in a special sense. Science is a search for knowledge about the natural world. When a scientific explanation consistently generates more accurate predictions than the alternatives, the explanation becomes generally accepted in the scientific community. Therefore, scientific knowledge is what the majority of the community of scientific experts believes at any time, tempered by the acknowledgment of uncertainties and unresolved issues. That is scientific pragmatism.

This pragmatic view of science should be taught to the next generation of scientists. Students are taught how to design and conduct experiments and observational studies and to apply inductive methods to the results. They are rarely taught how to apply abductive inference to determine the best explanation of the body of data that relates to their problem. In particular, they are not taught how to weigh multiple pieces and types of evidence in a defensible manner. Finally, they are not taught how to become members of a scientific community that is generating scientific truth and applying it to the world's problems. As a result, too many scientific articles conclude weakly: “Prior studies found w, x, and y and we found z. More research is needed.” We can do better in preparing future scientists.

We must learn to communicate the implications of science in a way that conveys its pragmatic properties without seeming to lack authority. Science communication has emphasized being clear to a lay audience, but has not addressed the interface between uncertain science, entrenched interests, political ideology, and public skepticism. How can we communicate the somewhat subtle concept that we do not have absolute knowledge, but we have some beliefs that are beyond reasonable doubt (e.g., evolution occurs by natural selection)? How can we convey the advice of the scientific community when the consensus is highly uncertain or when there is no consensus but there is a majority opinion and a compelling need to act (e.g., climate change or vaccination and autism)? How can we speak as a community when political and legal systems and the media reward contention?

Perhaps one way to enhance communication of the scientific community's truth is to develop a scale of confidence. For numerical results, confidence intervals from meta-analyses may be sufficient. However, for many scientific questions such as “Are greenhouse gas emissions warming the planet?” or “Is formaldehyde a carcinogen?” some sort of qualitative system is required. Such systems have been developed for specific applications such as evaluating medical treatments (e.g., the Cochrane handbook: or determining the causes of environmental impairments (CADDIS, An explicit scoring system can provide transparency and can ensure more rigorous abductive inference. A common scoring system for confidence in scientific findings could be an important communication tool both within the scientific community and with decision makers and the public.

Finally, we should move beyond peer review and workshops and find ways to behave more like a scientific community that is collaborating to generate the best approximation of truth and less like competitors or opponents. To an extent this is occurring. If you search the internet for the terms “open science” or “open research,” you will be impressed by what is going on: data sharing, code sharing, open access journals, collaborative research, research in the open (i.e., each day's activities and results posted on a blog), and Twitter reviews. This openness is not welcomed by all scientists and there are problems such as destructive “troll” reviewers and assigning credit for widely shared work. In the end, consciously or not, willingly or not, science is becoming more like Peirce's pragmatic community.

In addition to advancing science more rapidly, we believe that pragmatism will make it easier for scientists to present their case to the public. First, scientific truth is generated by a social process of consensus building. If you are challenged, an appeal to one or more of the indicators of consensus can provide your response: it was peer reviewed and other scientists concurred, it is consistent with the findings of a workshop, it has been replicated in other laboratories, etc. Second, the core of pragmatism is also the core of its scientific philosophy: scientific truth, like all truth, is what works in the real world. You might explain that the science is credible because it was confirmed in experiments or in a well defined survey or it is the basis for these real world applications. Hence, the consensus of the scientific community, based on evidence from testing and observation of nature, is the most reliable source of information for making real world decisions. It is that simple and that profound.


We thank Jason Lambert, Kate Schofield, and 2 anonymous reviewers for helpful comments. The article was written on our own time and received no financial or institutional support.