You Didn’t Have to Do That

Beliefs About Free Will and Determinism

The present research focused on gratitude as being influenced by the question of whether the benefactor did not have to provide the favor. In principle, either free will or determinism could affect that. Free will generally refers to the ability of the individual to act in different ways in the same situation. Determinism is a metaphysical doctrine that asserts the causal inevitability of every event. Although the two beliefs seem incompatible, many philosophers have found ways of reconciling the two in a position known as compatibilism (e.g., Holton, 2010). Perhaps ironically, few philosophers regard determinism as correct, although they continue to debate its compatibility with freedom of action. The compatibilist considers the free actions of the individual to be essential steps in the causal chain producing the outcome.

The present article focused primarily on belief in free will rather than determinism for a few reasons. First, the idea of free will is easily understood by most participants whereas the concept of determinism is less readily understood (Nahmias, Morris, Nadelhoffer, & Turner, 2005). Therefore, free will belief is likely to be more directly associated with the behaviors of the self and others and more amenable to experimental manipulation than belief in some type of determinism. Second, there is an existing literature (reviewed below) containing effective procedures for manipulating and measuring belief in free will. Beliefs in determinism have not been successfully manipulated to our knowledge.

Measures of belief in determinism do exist, and these have differentiated several different forms of that belief. (Moreover, beliefs in determinism are largely orthogonal to beliefs about free will, consistent with the compatibilist position.) Stroessner and Green (1990) showed that belief in two variants of determinism (psychosocial and religious–philosophical) and free will belief can exist independently. Similarly, Paulhus and Carey (2011) distinguished belief in scientific determinism from belief in fatalistic determinism and found that both exist independently from belief in free will. The two variants of determinism discovered by these researchers are conceptually similar: Psychosocial and scientific determinism refer to belief that biology/genes, environmental influences, and the laws of nature determine behavior. Religious–philosophical and fatalistic determinism refer to belief in fate, destiny, or some divine plan for life.

Meanwhile, beliefs about free will have been both measured and manipulated successfully. Vohs and Schooler (2008) found that those with a reduced belief in free will were more likely to cheat than those who did not have their free will beliefs reduced. Baumeister, Masicampo, and DeWall (2009) found that relative disbelief in free will increased aggression and reduced helpfulness. Alquist, Ainsworth, and Baumeister (2013) found evidence that a disbelief in free will increased conformity. All this work has been agnostic about the reality of free will and merely examined variations in belief. The social reality of such beliefs is well documented. For example, some degree of belief in free will was found across a range of cultures examined by Sarkissian et al. (2010). Even incarcerated adults report a strong belief in free will that is statistically equivalent to that of other adults (Laurene, Rakos, Tisak, Robichaud, & Horvath, 2011).

Given that belief in free will is such a widespread phenomenon, what do non-philosophers think having free will means? Monroe and Malle (2010) asked participants in the United States to explain their understanding of free will. The majority response, made by nearly two thirds of the participants, was “the ability to make choices” (65%). The next most common responses were “doing what you want” (33%) and “acting without internal or external constraints” (29%). Only 2% of the participants indicated that they did not believe free will existed. That proportion is obviously too small to furnish a meaningful basis for behavioral comparisons in smallish samples, but beliefs in free will can be conceptualized as existing on a continuum with varying degrees of certainty and conviction (e.g., Paulhus & Carey, 2011). Like much previous work, the present investigation examined relative differences in degree of belief in free will.

Gratitude

Gratitude has been defined as a positive emotion one experiences from the perception that one has benefitted from a costly, intentional, and voluntary action by another person (Bartlett & DeSteno, 2006; McCullough et al., 2008; Tesser, Gatewood, & Driver, 1968). Several gratitude researchers have used this person-to-person benefit triggered definition of gratitude in their experimental research (reviewed below). Although most research has focused on interpersonal gratitude, there are other possibilities, including transpersonal and theistic variants (Emmons & McCullough, 2004). That is, people may feel grateful to God for life and health, or they might feel grateful for pleasant weather or nature in a non-theistic way. The present research, in line with most experimental research on gratitude, is focused on gratitude as an emotion brought on by perceiving the receipt of a benefit from another person.

Models of gratitude developed by Tesser et al. (1968), and Wood, Maltby, Stewart, Linley, and Joseph (2008) suggest that the intensity of gratitude one experiences after receiving a benefit depends on three factors: (a) how costly it was for the benefactor (e.g., in time, effort, money); (b) how valuable the help was to the beneficiary; and, most relevant to free will beliefs, (c) how sincere the benefactor’s motivations were in providing the help.

Empirical findings have confirmed that gratitude is influenced by perceived motivational sincerity. Tsang (2006a) asked participants to imagine themselves in a scenario in which a friend did them a favor. In the benevolent condition, it was clear that the friend had benevolent intentions (the friend genuinely cared). In the ulterior motive condition, it was made clear that the friend was just doing the favor because he expected a favor in return. Participants in the benevolent favor condition expressed significantly more gratitude than those in the ulterior motive condition. Similar effects have been found when people consider favors from strangers. They feel more grateful when imagining that they had been helped by a stranger who had a benevolent motivation (i.e., the stranger really cared) compared with a self-serving ulterior motivation (i.e., the stranger wanted to feel like a good person for helping; Weinstein, DeHaan, & Ryan, 2010). Tsang (2006b) showed that participants who received a $9 benefit by random distribution reported significantly less gratitude than those who received the same $9 benefit ostensibly due to the voluntary action of a partner. Thus, people are most grateful when they perceive that a volitional agent intended to help them without having ulterior selfish motives. The more benevolent and volitional an action is perceived to be, the greater will be the perception of motivational sincerity. In this article, we focus on the aspect of motivational sincerity related to volition with the assumption that, without information suggesting otherwise, kind acts are perceived as having benevolent intentions.

The role of volition brings up the issue of free will—hence our hypothesis that gratitude will decrease insofar as people perceive that the benefactor’s actions were not the result of his or her free will. We reasoned that the intensity of gratitude would be related to perceiving that the benefactor had voluntarily acted to give the benefit—that is, had freely chosen to perform the favor based on wanting the recipient to benefit. Voluntarily doing what one wants is a factor that laypersons commonly associate with having free will (Monroe & Malle, 2010). If one’s belief in free will is reduced (vs. unchanged or increased), one would consequently view others (including benefactors) as acting less voluntarily and thusly be less sincerely motivated. On that basis, we predicted that people with a relatively weak belief in free will (be it dispositional or experimentally manipulated) would tend to experience less gratitude compared with those with a stronger belief in free will.

Present Research

We proposed that belief in free will influences how one perceives the behavior of others and that experiencing gratitude is based, in part, on a belief in free will. Specifically, we predicted that trait levels of free will belief would be positively associated with dispositional gratitude and that a manipulation that weakened belief in free will would lead to decreased feelings of gratitude triggered by receiving a benefit. We report four studies testing these hypotheses. The first study examined the prediction that trait free will belief would be positively correlated with trait levels of gratitude. The second study manipulated belief in free will and then measured self-reported gratitude from a recall task in which participants recalled a time in which they felt grateful for something. Study 3 manipulated belief in free will and measured feelings of gratitude after participants received a favor during the experiment. Study 4 manipulated belief in free will and measured self-reported gratitude using a scenario procedure. Studies 2 through 4 were experimental tests of the prediction that a reduced belief in free will would lead to a decrease in gratitude. We theorized that by reducing belief in free will, one would perceive the benefactor as having less free will and therefore view the benefactor’s actions as being less voluntary and thus less sincerely motivated. This in turn would lead to a reduction in feeling grateful.

Study 1a

Study 1a used trait measures of gratitude and belief in free will to test the hypothesis that a reduced belief in free will would be linked to a general pattern of not being grateful. Insofar as gratitude depends on perceiving that a benefactor acted voluntarily and could have done otherwise, we predicted that people with stronger belief in free will would have a more grateful disposition than people whose belief is weaker. We used a trait free will belief measure with four subscales, only one of which (Free Will subscale) directly measures belief in free will (pertaining to people generally). Therefore, we predicted that trait gratitude would be positively related to the Free Will subscale and not the other three.

Study 1b

The data for Study 1b were collected approximately 1.5 years after the collection of data for Study 1a. As in Study 1a, we predicted that there would be a significant positive correlation between trait gratitude and the Free Will subscale of FAD-Plus and no significant relationships between the GQ-6 and the other three subscales of the FAD-Plus.

Study 2

Study 2 used an experimental design to test the hypothesis that disbelief in free will would cause a decrease in gratitude. The use of experimental design will allow us to move beyond Studies 1a and 1b as we will be able to make causal claims. Study 2 manipulated participants’ belief in free will and then measured participants’ level of gratitude concerning specific events.

To measure gratitude, participants wrote about actual prior experiences from their lives for which they felt grateful. Then, participants rated how grateful they currently felt about that outcome. The prediction was that gratitude would be lower among participants who were induced to disbelieve in free will compared with those whose belief in free will was bolstered or unaltered. Mood was also measured (in the present and subsequent studies) to test the possibility that being induced to disbelieve in free will could be threatening to participants and thus induce a negative mood state or decrease positive mood. Because gratitude is a positive emotion, potential differences in negative or positive affect (compared with changes in free will belief) could account for any differences in gratitude. We predicted that mood would not be a relevant factor.

Results and Discussion

Gratitude

We tested the prediction that participants who rewrote anti-free will sentences would report feeling less grateful than those in the pro-free will and neutral conditions. There was significant variation among conditions in gratitude for the first recalled event, F(2, 47) = 3.27, p < .05, η² = .12. The pattern was similar but the difference dropped below significance for the second event measure, F(2, 47) = 1.97, p = .15. η² = .08, and it disappeared entirely on the third measure, F(2, 47) = 0.01, p = .99 (Figure 1). It is common to find that a manipulation’s effects are strongest immediately after its induction (perhaps especially a mind-set manipulation such as we used) and diminish on subsequent measures. Our results conform to that pattern, and so we emphasize the findings from the first measure.

OPEN IN VIEWER

Figure 1.

Mean gratitude scores by condition; Study 2.

Planned comparisons were made between the anti-free will condition and pro-free will condition as well as between the anti-free will and neutral conditions across the three events that participants furnished. For the first event, anti-free will participants (M = 7.12, SD = 2.76) were significantly less grateful than pro-free will participants, (M = 8.94, SD = 2.01), t(47) = 2.23, p = .03, d = .75, 95% CI = [0.04, 1.43]. Participants in the anti-free will condition also reported significantly less gratitude than those in the neutral control condition (M = 8.94, SD = 2.32), t(47) = 2.19, p = .03, d = .71, 95% CI = [0.01, 1.40].

On the two subsequent event reports, the differences grew weaker. On the second measure, the difference in gratitude between participants in the anti-free will (M = 7.53, SD = 2.83) and pro-free will conditions (M = 9.06, SD = 2.05) was marginal, t(47) = 1.82, p = .07, d = .62, 95% CI = [−0.08, 1.29]. Participants in the anti-free will condition were marginally less grateful than those in the neutral control (M = 8.89, SD = 2.40), t(47) = 1.58, p = .12, d = .52, 95% CI = [−0.19, 1.20]. No pairwise differences approached significance on the third and final event report.

Mood

We wanted to check whether the manipulation affected mood states, and so we tested PANAS scores as a function of condition. As expected (and in line with past research manipulating free will belief; e.g., Alquist et al., 2013; Baumeister et al., 2009; Vohs & Schooler, 2008), there were no significant differences among the anti-free will (M = 2.59, SD = 0.89), pro-free will (M = 2.78, SD = 0.86), or neutral conditions (M = 2.64, SD = 0.81) on the Positive Affect subscale, or the Negative Affect subscale (anti-free will: M = 1.51, SD = 0.51; pro-free will: M = 1.37, SD = 0.38; neutral: M = 1.43, SD = 0.46), Fs < 1. Direct comparisons between the anti-free will and pro-free will groups likewise did not reveal significant differences in positive or negative affect, ts < 1. Thus, the findings cannot be explained on the basis that the manipulations altered mood.

In line with our hypotheses, participants whose belief in free will had been reduced reported feeling less grateful than participants in either the pro-free will condition or the neutral condition. The difference was statistically significant for the first recalled event. For the second recalled event, the differences dropped just outside the significant range. By the third recalled event, there were no statistically significant differences among conditions.

Study 3

The aim of Study 3 was to examine the effects of free will belief on gratitude about the same beneficial act. In Study 2, participants selected events from their own lives to write about, so each event was necessarily unique. To rectify this potential drawback, Study 3 participants reported their gratitude toward the same event. Another goal was to measure the effects of belief in free will on gratitude for a real event that occurred during the experiment.

We assumed that most participants would prefer to be done with their research obligation and leave the laboratory sooner than later. Therefore, the present study incorporated a favor such that participants were relieved from a tedious task and thus released from part of their obligation. We anticipated that this favor would engender varying degrees of gratitude among participants.

Another aim of Study 3 was to examine variables that are related to the degree of gratitude one feels after receiving a benefit (based on Tesser et al., 1968; Wood et al., 2008). Participants answered questions about their perception of the value of the favor, the cost of the favor to the benefactor, and the motivation of the benefactor.

We predicted that participants in the anti-free will condition (vs. pro-free will) would report feeling less grateful and that this effect would be mediated by perceptions of motivational sincerity. Perceptions of value, perceptions of cost, and feelings about the tedious task they avoided were also measured. This was done in an exploratory fashion to help rule out potential alternative explanations and also to help attain a more complete picture of potential underlying processes.

Study 4

The aim of Study 4 was to extend the findings from Study 3 and address potential drawbacks from previous studies. Studies 2 and 3 relied on single items measures for the key dependent variables. Study 4 sought to address this issue by including multiple items for each construct measured. Another aim of Study 4 was to replicate and expand on the mediation found in Study 3.

In Study 4, participants read a vignette recounting a small favor and answered several questions about their perception of the value of the favor, the cost of the favor to the benefactor, and the motivation of the benefactor. This study added a new process variable measuring how much free will participants believed the benefactor had. Because belief in free will involves beliefs about whether others generally (the self included) have free will, we expected that experimentally reducing belief in free will would influence perceptions of the benefactor’s volition and therefore decrease reported gratitude. A double mediation model was predicted in which participants whose belief in free will had been weakened (vs. bolstered) would perceive the benefactor as having less free will and therefore perceive the benefactor as being less sincerely motivated (by virtue of behaving less voluntarily) in providing the favor, and therefore report feeling less grateful.

Results and Discussion

Two raters reviewed the essay summaries and determined with 100% agreement that all participants understood the main point of the essay (either that science supports the idea that free will does not exist, or that science supports the idea that free will does exist). Therefore, all participants received a small bonus payment. A manipulation check at the end of the study revealed that those who read the anti-free will essay (M = 6.85, SD = 1.99) reported a weaker belief in free will than those who read the pro-free will essay (M = 8.29, SD = 0.90), t(62) = 3.70, p < .001, d = .93, 95% CI = [0.40, 1.43]. There were no significant differences in the pleasant–unpleasant mood score on the BMIS between the anti-free will (M = 43.5, SD = 5.1) and the pro-free will (M = 44.9, SD = 5.4) conditions, t(62) = 1.06, p = .29, d = .27, 95% CI = [−0.23, 0.76].

We predicted that participants in the anti-free will condition would view the benefactor as having less free will, perceive the benefactor as being less sincerely motivated, and feel less gratitude toward the benefactor. Differences were not expected on perceptions of value or cost.

Participants in the anti-free will condition (M = 64.9, SD = 27.2) reported feeling less grateful than those in the pro-free will condition (M = 78.4, SD = 19.1), t(62) = 2.27, p = .027, d = .57, 95% CI = [0.06, 1.06].

An independent samples t test on the benefactor free will index revealed that participants in the anti-free will group (M = 77.4, SD = 19.7) perceived the benefactor as having less free will than those in the pro-free will group (M = 87.4, SD = 13.2), t(62) = 2.37, p = .021, d = .60, 95% CI = [0.09, 1.09].

An independent samples t test was conducted on the three-item index assessing benefactor motivation. Participants in the anti-free will condition (M = 53.2, SD = 18.4) perceived the benefactor as being less sincerely motivated than participants in the pro-free will condition (M = 63.6, SD = 10.4), t(62) = 2.77, p = .007, d = .70, 95% CI = [0.18, 1.19], which confirms our prediction.

T tests did not show significant differences between conditions on perceptions of value (anti-free will: M = 56.0, SD = 29.1; pro-free will: M = 65.8, SD = 23.7), t(62) = 1.47, p = .147, d = .37, 95% CI = [−0.13, 0.86] or perceptions of cost (anti-free will: M = 3.9, SD = 5.4; pro-free will: M = 9.4, SD = 15.9), t(62) = 1.89, p = .062, d = .46, 95% CI = [−0.03, 0.96]. Thus, consistent with our reasoning, the manipulation of free will beliefs altered perceptions of motivational sincerity but not of the other two determinants of gratitude (cost to benefactor and value to recipient). However, unlike in Study 3, perceptions of value and of cost between conditions were nearly significantly different and had a close to moderate effect size.

In summary, participants in the anti-free will condition (vs. the pro-free will condition) perceived the benefactor in a hypothetical scenario as having less free will. In addition, participants in the anti-free will condition perceived the benefactor as being less sincerely motivated and reported less gratitude compared with those in the pro-free will condition.

Mediation

Our proposed serial multiple mediator model was that participants in the anti-free will belief (vs. pro-free will belief) condition would perceive the benefactor as having less free will, which in turn would reduce perceptions of motivational sincerity, and therefore lead to less gratitude toward the benefactor. To test this model, we ran a mediation analysis using the PROCESS bootstrapping macro (Model 6, Hayes, 2013) for SPSS with 5,000 resamples and 95% bias-corrected standardized bootstrap CI. The anti-free will condition was coded as 0 and the pro-free will condition was coded as 1. These analyses are depicted in Figure 2.

OPEN IN VIEWER

Figure 2.

Serial mediation model; Study 4.

The effect of the free will belief condition on gratitude was significant without including any mediators, β = .28, t(62) = 2.27, p = .027, sr = .28; however, the direct effect was not significant when perceptions of benefactor free will and perceptions of motivational sincerity were respectively added as serial mediators, β = .08, t(62) = 0.72, p = .472, sr = .07. The specific indirect effect through both mediators (i.e., free will condition predicts free will of benefactor, which predicts perceptions of motivational sincerity, which predicts gratitude) was significant, as the 95% CI did not contain zero [0.019, 0.226]. The 95% CI of the total indirect effect also did not contain zero [0.061, 0.371] indicating a significant total indirect effect. Tests of alternate specific indirect effects whereby only one of the variables served as the mediator within the double mediator model were not significant, indicating additional support for the serial–multiple–mediator framework. The specific indirect effect of free will condition on gratitude through beliefs about benefactor free will was not significant, 95% CI = [−0.032, 0.133]. The specific indirect effect of free will condition on gratitude through beliefs about the benefactor motivation was not significant either, 95% CI = [−0.006, 0.217].

Thus, the free will belief condition indirectly influenced reported gratitude through its serial effect on beliefs about the benefactor’s free will and perceptions of the benefactor’s motivation. Specifically, those with a reduced belief in free will perceived the benefactor as having less free will and as a result perceived the benefactor as being less sincerely motivated, leading to weaker feelings of gratitude toward benefactor.

General Discussion

Beliefs about free will contribute to gratitude. Four studies with different methods found consistent, convergent support for that hypothesis. A personality trait measure of broad tendencies to feel grateful for things was positively correlated with a trait measure of belief in free will (Studies 1a and 1b). A reduced belief in free will induced by a laboratory manipulation caused people to report feeling less gratitude for an event from their lives in which they had reason to be grateful (Study 2). A different manipulation of free will belief likewise caused those with a reduced belief to report less gratitude for an actual favor performed to benefit the participant in the context of the laboratory session (Study 3). Another study showed that a reduced belief in free will led to less gratitude in response to a hypothetical favor in a vignette (Study 4).

How is free will involved in gratitude? We reasoned that gratitude is associated with perceiving that the benefactor did not have to give the benefit and thus freely chose to give the benefit. Hence, one is more grateful to the extent that one perceives that the other freely and voluntarily chose to provide the benefit. We predicted and found that perception of benefactor free will was positively related to a pre-existing construct in the gratitude literature called motivational sincerity. We found consistently that free will beliefs altered perceptions of motivational sincerity and that these changes mediated the effects on gratitude. That is, people who were induced to disbelieve in free will reported a weaker perception that the benefactor had genuinely wanted to benefit them, and those perceptions mediated the tepid feelings of gratitude they reported (Studies 3-4). The most thorough evidence came in Study 4, which found that the manipulated reduced belief in free will caused a reduction in the perception of the benefactor as acting freely, which in turn caused a reduction in perception of the benefactor’s motivational sincerity, which then caused a reduction in gratitude.

Thus, in this case, the relevance of free will applies to the other person’s free will rather than one’s own. Although free will belief is belief about people in general (the self and others), most previous work with free will belief has focused on how changes in free will beliefs produce effects by changing people’s sense of their own freedom to act: thus with cheating and stealing (Vohs & Schooler, 2008), helping and aggressing (Baumeister et al., 2009), conformity (Alquist et al., 2013), and volunteering (Stillman & Baumeister, 2010). Some work has shown that people are less prone to punish criminals and other wrongdoers insofar as they disbelieve in free will (Shariff et al., 2014). This package of studies may be the first to show that beliefs about free will have an impact on how one responds to positive behaviors by others.

Study 2 included a neutral control group. The pro-free will manipulation did not produce an increase in gratitude above the baseline indicated by the neutral control group. The anti-free will group did cause a drop in gratitude. The pattern that manipulations that reduce free will belief depart from the norm more than manipulations that seek to enhance or increase it has been found in multiple studies (e.g., Alquist et al., 2013; Baumeister et al., 2009; Vohs & Schooler, 2008) and may reflect a ceiling effect, insofar as baseline levels of belief in free will are generally high.

We noted that gratitude theories (e.g., Tesser et al., 1968; Wood et al., 2008) have emphasized three determinants of gratitude, one of which is most relevant to beliefs about free will. The current studies show that perceptions of the benefactor’s motivational sincerity is positively related to free will belief, whereas the other two (benefactor’s perceived cost and recipient’s value gained) do not to the same extent. Some research (e.g., Tsang, 2006b; Weinstein et al., 2010) has portrayed the benefactor’s helping motivation as having selfish ulterior motives (e.g., wanting a favor in return) or as sincere (e.g., being genuinely concerned). Free will belief is related to a different aspect of perceiving others’ motivations, namely the volition of the benefactor. For participants with a reduced (vs. increased) belief in free will, benefactor motivation was perceived as less sincere because it was seen as less voluntary, as opposed to being seen as having (intentional) selfish motives.

The current findings have relevance to the concept of counterfactual thinking. Counterfactual thinking involves “what might have been” thoughts, which means realizing that an outcome might have gone another way. They have been most frequently associated with negative outcomes (Epstude & Roese, 2008; Roese, 1997; Smallman & Roese, 2009), although they contribute to some positive feelings about positive outcomes too such as relief (Sweeny & Vohs, 2012). Free will is widely understood as the possibility to act differently, and some evidence indicates that belief in free will contributes to engaging in counterfactual thinking (Alquist, Ainsworth, Baumeister, Stillman, & Daly, 2012). It is plausible that belief in free will helps people appreciate that the benefactor might well not have done the favor, which would increase gratitude. Future work may benefit by elucidating the potential contribution of counterfactual analysis to gratitude and its relation to beliefs about free will.

The present work has some limitations that should be mentioned. First, some of the studies had a small to medium sample size, which may call the reliability of the findings into question. However, concern about reliability should be somewhat alleviated because multiple studies yielded similar findings. Second, gratitude in all the experimental studies was operationalized as self-reported reactions to a specific benefit from another person. As discussed previously, gratitude can be conceptualized in alternative ways as well (e.g., in a transpersonal or theistic sense). This article also focused specifically on experimentally altering free will belief and not beliefs about various types of determinism. Therefore, it would be beneficial for future research to explore how belief in free will is related to alternative conceptualizations of gratitude and also the role that beliefs about different types of determinism and their interaction with belief in free will may play. Finally, the process measures involved in how free will belief influences gratitude merit more attention. We found that perceptions of benefactor free will and benefactor motivation were important mediating variables in multiple studies, whereas perception of value and cost were not. However, perceptions of value and cost approached being significantly different (anti-free will condition perceiving less of both) in Study 4 (a monetary benefit) but not Study 3 (a favor saving the participant time and energy). Hence, it is possible that the nature of the benefit (e.g., a monetary vs. non-monetary benefit, or a small vs. large favor) could be a factor that implicates additional mediating variables (such as cost or value) and therefore could be a variable worth manipulating in future work.

Conclusion

Four studies provided evidence that belief in free will contributes to feelings of gratitude. Weaker dispositional and experimentally reduced free will belief was consistently associated with less gratitude. A reduced personal belief in free will led people to downgrade their perception of a benefactor as having free will and as having performed the favor as a voluntary choice based on a sincere desire to help. This significantly reduced gratitude. This work showcases the importance of belief in free will as facilitating the experience of the socially beneficial emotion of gratitude.

Grand questions about free will have prompted experts to engage in complex metaphysical, semantic, and even theological debates. Our work has nothing to say about these. In the social lives of ordinary people, however, freedom to choose and voluntary action are important categories of person perception that shape how people treat each other. Gratitude is one important positive lubricant of social interaction. Our findings suggest that the feeling and expression of gratitude are tied to the perception of whether the benefactor might indeed have done otherwise and instead freely choose to confer a benefit.