Consider a father who looks at his beloved daughter and thinks, ‘What I want most in life is just for you to be happy.’ In thinking this thought, the father makes use of a concept that is deeply important but also very difficult to adequately characterize – the ordinary concept of happiness. Our aim is to understand how this concept works.
One obvious view would be that the ordinary concept of happiness is just a matter of having certain psychological states. For example, it might be thought that the ordinary concept of happiness is a matter of feeling good, experiencing satisfaction with one’s life, and not experiencing negative affective states, such as pain, lonelinessor despair. On this view, when the father thinks that what he wants most in life is for his daughter to be happy, what he means is simply that what he wants is for her to have certain kinds of psychological states.
An alternative possibility is that the ordinary concept of happiness is something more normative. On this alternative, the concept is not just a matter of having certain psychological states; it also has something to do with having a genuinely good life (see, e.g., Annas, 1993; 2011; Foot, 2001; Kraut, 9179). The most extreme version of this second view would say that the ordinary concept of happiness simply is the notion of well-being.
So which of these two possibilities better captures what the father is thinking about his daughter in our example? Is the ordinary concept of happiness really just concerned with the psychological states one experiences, or is it normative in some way? At this point, a sizable number of studies have sought to shed light on this question, and the results are quite surprising.
In one recently published study, wepresented naive participants, academic researchers, and researchers who specifically work on happiness with different vignettes that described both an agent’s psychological states and the more general life that the agent was living. In all cases, these agents were always described as having extremely positive psychological states (feeling good, experiencing a lot of satisfaction, and rarely ever feeling bad). However, in some of the cases, the agents were described as living a life that we expected participants to see as normatively good (e.g., being kind to and caring for others), while in others, the agents were described as living a life that we expected participants to see as normatively bad (e.g., being cruel and harming others). We’ll refer to these as the ‘moral’ and ‘immoral’ agents.
After participants read about one of these two kinds of agents, we asked them a series of questions. First, just to make sure they accepted the key premises about the agent’s psychological states, we asked them whether the agent feels good, experiences general satisfaction, and does not feel bad. Fortunately, almost all participants agreed that the agent did have these states. Then came the real question: Is the agent happy?
On this latter question, we found a surprising result. Even restricting to the participants who agreed that the agent had positive psychological states, participants who got the story about the immoral agent were less inclined to say that the agent was happy. Moreover, we found this pattern to hold regardless of whether the participants were naive subjects, academic researchers, or even experts working specifically on happiness. In fact, we even found this pattern to hold in both a between-subjects design (where each participant only reads about one agent) and in a within-subjects design (where participants read about pairs of immoral and moral agents who were explicitly described as having identical psychological states). As you can see below, this tendency is extremely robust.
A second important piece of evidence is that this effect seems to be pretty specific to the concept of happiness and does not extend to other assessments of how the agent feels. Consider one study that provides some nice evidence for this. Just as in the study described above, participants read brief vignettes that described both an agent’s psychological states, and the life that the agent was living. In this case, the agents were (1) either described as having highly positive or highly negative psychological states and (2) either described as living a life we thought participants would find to be normatively bad or normatively good. After reading about an agent, participants were asked one of two questions about that agent.
One question asked them about the feeling that the agent experiences. Participants selected one of seven different faces which had been created using facial morphing to make a scale that ranged from highly negative affect to highly positive affect (like the one below):
The other question simply asked them to indicate the extent to which they agreed that the agent was happy (on a seven-point agreement scale that ranged from strongly disagree to strongly agree).
Participants’ judgments produced a pretty clear pattern: Assessments of the agent’s happiness were strongly affected by whether the agent was described as living a normatively good or normatively bad life, but in comparison, assessments of how this agent feels were not.
So what do these studies tell us about the ordinary concept of happiness? As far as the ordinary concept goes, is being happy simply a matter of having positive psychological states? The empirical evidence strongly suggests that the answer is no. Even when agents are described as having entirely positive psychological states, people show a reluctance to say that they are truly happy unless they also have morally acceptable lives. So then, should we go to the opposite extreme and say that the ordinary concept of happiness is just a matter of having a good life. Again, the answer appears to be no. Empirical studies on the ordinary concept have provided a wealth of evidence that assessments of happiness depend In a special way on the agent’s psychological states. In fact, changes in the agent’s psychological states tend to have a bigger impact on assessments of happiness than changes in the normative value of the agent’s life (you can see this in the above figure, for example).
In short, the current state of research on the ordinary concept of happiness doesn’t give us a straightforward answer to the question we started with. It’s clear that ordinary assessments of happiness depend both on psychological states and on normative value, but it’s not clear how these two factors are related or precisely what role they play in the ordinary concept. At this point, we suspect that what is needed is not just additional empirical studies but careful philosophical thought about what the existing data might be telling us.
Jonathan Phillips earned his Ph.D. in Philosophy and Psychology from Yale in 2015, and is currently at postdoctoral fellow in psychology at Harvard. Much of his work surrounds theory of mind, causal reasoning, moral judgment, formal semantics, and happiness.
Joshua Knobe is an experimental philosopher, appointed in both the Program in Cognitive Science and the Department of Philosophy at Yale University.1