What We Do When We Define Morality (and Why We Need to Do It)

Abstract

Psychological research on morality relies on definitions of morality. Yet the various definitions often go unstated. When unstated definitions diverge, theoretical disagreements become intractable, as theories that purport to explain “morality” actually talk about very different things. This article argues that we need to define morality and considers four common ways of doing so: The linguistic, the functionalist, the evaluating, and the normative. Each has encountered difficulties. To surmount those difficulties, I propose a technical, psychological, empirical, and distinctive definition of morality: obligatory concerns with others’ welfare, rights, fairness, and justice, as well plus the reasoning, judgment, emotions, and actions that spring from those concerns. By articulating workable definitions of morality, psychologists can communicate more clearly across paradigms, separate definitional from empirical disagreements, and jointly advance the field of moral psychology.

Keywords:

• Moral psychology
• moral philosophy
• definitions
• meta-theory

Acknowledgments

I thank Tal Waltzer, Margie Martinez, Charles Baxley, Mary Taylor Goeltz, and other members of the Developmental Moral Psychology Lab at UC Santa Cruz for their comments on earlier versions of this manuscript.

Notes

1 Though I build on the arguments of Social Domain Theory (Killen & Smetana, 2015; Royzman et al., 2009; Turiel, 1983, 2015), my arguments differ in at least two key respects. First, Turiel’s (1983) original arguments have been criticized for presuming too much agreement among philosophers about the definition of morality (Machery & Stich, 2022). My proposals about how to define morality presume no such philosophical consensus. Second, my proposed definition goes beyond Turiel’s original formulations by explicitly including emotions and actions (Turiel & Dahl, 2019).

2 As I explain later, I am not proposing that all matters of right and wrong should be called moral issues. For now, it is sufficient to note that people in different places differ in their judgments about right and wrong ways of acting. So that my conclusions in this section about the empirical literature is independent of my own definition of morality, I framed it more broadly, in terms of concerns and judgments about right and wrong.

3 I do agree with Greene (2007) that to examine everyday word use might be useful when you found a new science: “Biologists got their start, not by rigorously defining ‘life,’ but by studying the kinds of things that ordinary people regard as living” (p. 1). Without knowing exactly how the first biologists of Antiquity got going, I find Greene’s account plausible enough. When Sharp (1896) began his initial studies of moral psychology late in the 19th century, he might have benefitted from surveying what people call “morality.” Today, things are different. After over a century of data collection, the psychology of morality—by any definition—has outgrown its infancy, is no longer starting out, and is no longer a new science.

4 I thank a reviewer for suggesting that I discuss the scientific and non-scientific usage of the word “fish.”

5 To my surprise, the same descriptions of popular and scientific usage of "fish" remains in the contemporary edition of the Oxford English Dictionary (Simpson & Weiner, 1989). I have not met anyone who says whales are fishes.

6 The argument against defining morality by its consequences in psychological research is not an argument against consequentialism as a moral philosophy, nor is it an argument against other kinds of functionalist theories. The ethical tradition of consequentialism proposes that the morally right action is the action that leads to the best expected or actual consequences (Greene, 2014; Mill, 1879; Wiggins, 2006; Williams, 1973). Consequentialists readily grant the point that I have made in this section: It can be difficult to know the consequences of an act. But this is not a problem for consequentialism as a moral philosophy, because ethical consequentialism does not need to pick out a psychologically distinct set of phenomena that necessarily count as moral. The fact that we cannot always know what will generate the best consequences helps explain, for a consequentialist, why we have moral dilemmas. By contrast, a definition of morality for psychological research does need to pick out such a distinct set of psychological phenomena in order to study them. In psychological research, it would be unworkable if what counted as a moral phenomenon could only be determined long after it occurred.

7 I thank a reviewer for suggesting this example.

8 My argument does not imply moral relativism. As human beings, we all judge that some acts are morally right and other acts are morally wrong—whatever we mean by “moral.” But as a psychologists, engaged in empirical science, we have no special tools for deciding which human actions are right and which are wrong. Psychology cannot prove that we ought to donate 10% of our income to charity, any more than economics can prove that we ought to arrange our societies so as to maximize individual freedom (Friedman, 1962), or biology can prove that we ought to genetically modify human embryos.

9 In his influential 1983 book, Turiel defined the moral domain in terms of “prescriptive judgments of justice, rights, and welfare pertaining to how people ought to relate to each other” (p. 3, also pp. 35-44). Expanding Turiel’s definition, the present definition explicitly incorporates concerns, emotions, and actions to incorporate research on early moral development and the role of emotions from the past 20 years.

10 It would go beyond the scope of this article to further detail what makes concerns seem obligatory. For relevant discussions, the reader may consult deontological as well as consequentialist perspectives on moral philosophy (e.g., Kant, 1785; Korsgaard, 1996; Mill, 1879; Rawls, 1971; Scanlon, 1998; Wiggins, 2006). For the present purposes, it will suffice to say that individuals do deem some concerns to be obligatory. Saying that a concern is obligatory is different from saying that an action is obligatory. Even if people think we are always obligated to concern ourselves with the welfare of others, they do not judge that we are always obligated to help others or to otherwise prioritize others’ welfare in all our actions.

11 For my definition, there is nothing puzzling about the so-called “harmless moral wrongs” that have caused controversy among moral psychologists (Gray et al., 2014; Haidt et al., 2000; Royzman et al., 2015). The perception of actual harm to a specific individual is only one of the things that could lead someone to map an event onto a moral concerns, as I define them here. I thank a reviewer for requesting this clarification.

12 I thank a reviewer for suggesting that I include this discussion and another reviewer for promting me to discuss Gray’s Theory of Dyadic Morality.

13 Schein and Gray (2018b) do acknowledge Haidt’s definition of moral judgments as “evaluations (good vs. bad) of the actions or character of a person that are made with respect to a set of virtues held to be obligatory by a culture or subculture (Haidt, 2001, p. 817)” (Schein & Gray, 2018b, p. 35). Yet Haidt’s definition does not appear to circumscribe the morality that the Theory of Dyadic Morality aims to explain. Schein and Gray (2018b) refer to Haidt’s definition as “one popular working definition,” which is hardly a ringing endorsement in academese.

14 What if the Theory of Dyadic Morality took first approach I mentioned in this section: the normative approach to defining morality? It would then have defined moral judgments as any judgment of right and wrong. In principle, the theory could then make the empirically testable hypothesis that all judgments of right and wrong involve perception of harm. Like all normative approaches to defining morality, this would strain the overlap between scientific and everyday usage of “moral.” Gray and his colleagues might accept this definitional strain, though. In some cases, they treat as “moral” judgments that fall outside the realm of what most people would call morality. According to the Theory of Dyadic Morality, even perceptions that agents are harming themselves, by doing something dangerous or unhealthy, can generate moral judgments (Schein & Gray, 2018b, p. 59).

However, this normative definition of morality would expose the Theory of Dyadic Morality to a greater risk: unfalsiability. The theory defines “harm” as virtually any possible or actual negative consequence to any entity subjectable to negative consequences, from deceased souls to the natural environment. In effect, the theory would be saying that people only judge an act as wrong if they thought the act might bring about something negative. If launched, this claim would land somewhere between the trivial and the tautologous. For how could anyone judge an act as wrong if they thoroughly knew that nothing bad could follow from it, not even a loss of self-respect? As Chomsky (1959) wrote in his review of Skinner’s Verbal Behavior, there is “no possible grounds for argument here” (p. 27)—not even for Kant (1785). For these reasons, I ultimately doubt that the Theory of Dyadic Morality would adopt the normative approach to defining morality.

Additional information

Funding

The writing of this manuscript was supported, in part, by a grant from the National Institute of Child Health and Human Development [R03HD087590].