In Study 1, we investigated the experience of agency in junior cadets and in a civilian control group. Similar to previous studies20,34,35, results showed a clear coercion effect for civilian participants. Interestingly, interval estimates of junior cadets were not reliably different in the coercion condition and in the free-choice condition. Importantly, for both groups, these results were not influenced by the identity of the experimenter, suggesting that coercion effects do not depend on the particular social status of the person giving orders, but rather reflect a more general difference between coercion and autonomy contexts34.
Further analyses revealed that the lack of coercion effect for junior cadets was due to long interval estimates in the free-choice condition, suggesting a reduced sense of agency9,10. This implies that junior cadets may show a global reduction of sense of agency, relative to non-military personnel, both when they can freely choose which action to perform and when they are coercively instructed. We also observed that the coercion effect did not appear to be influenced by any personality traits and that z-scores in both the free-choice and the coercion conditions were not influenced by the degree to which civilians thought that they would be suited to the military environment. It thus suggests a negative influence of the military environment on cadets’ sense of agency, rather than a predisposing trait.
We additionally observed a coercion effect on the amplitude of the auditory N1 suggesting that receiving coercive instructions reduces the neural processing of the outcomes of one’s own actions20. Importantly, this was the case for both junior cadets and civilian students, suggesting that outcome processing was not impacted by the military environment.
However, Study 1 does not represent a reliable sample of all the different categories of individuals working in the army, nor can it represent the effects of prolonged military training. In Study 2, we, therefore, compared three groups of military personnel, namely privates, junior cadets, and senior cadets. Comparisons between these groups could reveal how prolonged experience in the social environment of a military organization influences the sense of agency under coercion and how the different notions of responsibility enshrined in the training of officers and of ordinary soldiers might lead to modulations of the coercion effect. In comparison with junior cadets, seniors have been trained to be officers during 5 years in average and have reached the rank of lieutenant. They have thus worked for a longer time period within the military than junior cadets, and have been trained to be accountable for their own actions (including giving orders to others) during those 5 years. Privates correspond to troop soldiers. They have a lower rank within the military system and in comparison to cadets, accountability is less emphasized during their training and career, although they work for a similar number of years in this type of organization. One might predict that seniors regain a strong sense of agency when they are free to choose which action to perform given that they frequently command others and are accountable for those actions, e.g., refs. 36,37,38. Privates might, on the other hand, continue to exhibit this reduced agency, since they routinely obey orders. One might also predict that outcome processing could be influenced by military rank (i.e., officers vs. privates), as a result of 5 years of differentiated military training including differentiated responsibility. Since the function of privates is mostly to execute orders from the military hierarchy, downregulation of outcome processing could be observed. On the other hand, senior officers might show an upregulation of outcome processing, consistently with their function implying commanding subordinates and being accountable for their actions.
As in Study 1, in Study 2 junior cadets exhibited a low sense of agency in the free-choice condition. A similar result was also found for the group composed of privates, suggesting that the sense of agency is not positively influenced by the number of years spent in a military organization. Interestingly, we observed that seniors, despite working since a similar number of years in a military organization than privates, appear to have a higher sense of agency in the free-choice condition, similar to civilians. This result offers interesting insights on the possibility to train the sense of accountability, by restoring experience of agency. Indeed, while both junior cadets and privates exhibited a reduction in the sense of agency when they could freely decide which action to perform, hence underlying the negative effect of military hierarchy per se, being trained as an officer appears to block this effect and even reverse it.
We observed that junior cadets displayed a “coercion effect” on the amplitude of the auditory N1, similarly to Study 1. However, this was not the case for seniors and privates. Seniors displayed a high amplitude of the auditory N1 in both the free-choice and the coercion conditions, while privates displayed a low amplitude of the auditory N1 in the same conditions. It thus appears that 5 years in the military decreases the difference in cognitive outcome processing between receiving orders and being free to decide for yourself. Training as an officer appeared to protect against reduction of outcome processing in coercive contexts, while working as a private appeared to have a reduced outcome processing even for freely chosen decision.
Our experiment was designed to identify changes in N1 amplitude due to coercion, and the way that different groups respond to coercion. However, the amplitude of the auditory N1 can be modulated by other factors, such as age, education, and intelligence39. For instance, a higher amplitude of the auditory N1 was observed for a high number of years of education. Seniors and privates differ in educational levels: seniors have a university master degree while privates require only an elementary school certificate, thus potentially explaining differences in outcome processing between seniors and privates. However, the amplitude of the auditory N1 in the free-choice condition did not differ between junior cadets and seniors, despite 5 more years of military training and education. It thus seems unlikely that differences between other groups merely reflect differences in the duration of education prior to military service. Other studies showed that the amplitude of the auditory N1 could be modulated by perceptual, motor, and cognitive factors40. However, it is also unlikely that those factors influence our results. All participants were in the same age range41, without any physical disabilities, and heard a tone similar in both frequency and loudness42. Also, participants all used the same keyboard and the same fingers to press the buttons, ruling out the influence of motor performance. Attention has been previously discussed as modifying both late and early (i.e., the auditory N1) ERPs. However, it is unlikely that a difference in attention explains the difference between groups since the majority performed correctly the task of interval estimates and did not commit mistakes when pressing the buttons in the coercion condition. Therefore, it seems unlikely that the N1 results in our experimental design are confounded by these other factors, although it cannot be entirely excluded.
We did not investigate personality traits in Study 2. However, previous studies20,34,35, and supplementary analysis of the present Study 1 did not find strong evidence that personality traits could influence interval estimates and outcome processing.
In the present paper, we investigated the experience of agency in different groups of military personnel and in civilians in order to evaluate the respective influence of working in a military context and being trained to be accountable. We used an implicit measure, based on perceptual compression of action-outcome intervals, as a behavioral marker of sense of agency, and we used a reciprocal, transparent experimental design, in which participants administered electric shocks to another member of their dyad.
The fact that junior cadets and privates show a reduced sense of agency, even when they are free to choose when and what action to perform, sheds light on an important feature of human sense of agency. Being trained to follow orders appears to exert a negative impact on how one experiences agency for one’s own actions. It results in an adaptive behavior that reduces the distinction between “receiving orders” into “deciding for oneself”. Only if one experiences one’s action as voluntary, will one develop a sense of agency with respect to the action’s outcome43. Therefore, regularly receiving coercive instructions may create a new “normality”, in which the experience of acting voluntarily can come to resemble the experience of following orders.
Interestingly, although our data is cross-sectional, it raises the possibility that this effect can be reversed: we observed for the group of senior cadets that being trained to be accountable for one’s own action, and the actions of troops under one’s commands, was associated with an increased sense of agency, relative to privates and junior cadets who remained in the military system, at least when the individual is free to decide which action to execute. This result offers important insight on the potential for training the sense of agency, in order to avoid potential detrimental effects of a lack of agency. Not considering oneself as the author of an action could lead to moral disengagement, with negative effects on behavioral control44. Since the present study is cross-sectional rather than longitudinal, these suggestions can only be tentative, rather than conclusive. Future studies could confirm the role of training emphasizing responsibility on sense of agency and outcome processing by offering a longitudinal approach.
Although obedience is an essential aspect of the efficiency of armed forces, international law state clearly that military members must refuse orders if they do not fit in the interest of the service or if they imply committing a crime45. On the one hand, soldiers may have to disengage their moral control to follow orders and take distance from their personal responsibility46 because the organizational objectives may imply actions that society rejects in peacetime (notably, killing and wounding others). On the other hand, past and recent events show that refusing an illegitimate or illegal order is far from straightforward, and requires considerable personal courage1,47. In the present study, we could not investigate to what extent a strong sense of agency helps individuals to resist illegitimate orders, since disobedience rates were extremely low. This question would nonetheless be worth investigating to understand how to manage the risks associated with blind obedience.
Previous studies reported that the sense of agency decreased as the number of alternative possible actions decreased19,48. This could explain the reduced sense of agency in the coercion condition, where participants have only one action available, relative to our free-choice condition, where two actions are available. In the present experiment, the response set size was similar for all the groups tested, but only civilians and senior cadets displayed a coercion effect. This suggests that sense of agency under coercion is a matter of how context influences choice, rather than the strict availability of a number of choices. One might also suggest that in our study we did not control for causality between experimental conditions since participants received an auditory instruction in the coercion condition but not in the free-choice condition, which corresponds to real-life situations. However, previous studies involving a verbal instruction in both conditions yet observed the coercion effect on sense of agency, e.g., ref. 34. Further, a difference in causality or instruction could not explain differences between groups since all groups were instructed in the same way.
In this study, we did not evaluate the sense of agency for individual actions with an explicit measurement. Explicit trial-wise judgments of agency have generally been used when unexpected action outcomes49 or uncertainty about who caused the outcome50 is present. In our study, there is no doubt regarding the outcome of each action, nor is there any ambiguity regarding who causes the outcome of each action. For those reasons, we could not use explicit judgments of agency as our main outcome variable. To explore more reliably whether or not explicit sense of agency is also modulated by the military environment and by the military rank, other experimental designs should be developed.
We also observed that outcome processing, as measured by the amplitude of the auditory N1, was influenced by military rank. Supplementary analyses further revealed that those who persevered in the military system showed stronger reductions in outcome processing, possibility reflecting a capacity to adapt to one’s own environment. We suggest that working as a subordinate, as privates and junior cadets must do, may be associated with low levels of outcome processing. In contrast, our findings with senior cadets suggest that being trained to be accountable has a positive effect on outcome processing, since the amplitude of the auditory N1 was high in both conditions for seniors. Officer training may require upregulation of outcome processing, possibly reflecting accountability for one’s own decisions and actions, and for decisions and actions of those under one’s command.
Taken together, our results also highlight a dissociation between the implicit sense of agency and outcome processing. The psychological, conceptual and biological mechanisms linking both explicit and implicit measures of the sense of agency to neural outcome processing still need to be clarified and addressed in future studies.
Our results suggest systematic differences between groups in sense of agency and outcome processing, as a function of the daily working context but also specific training targeting responsibility. Junior cadets need to first downregulate outcome processing and sense of agency to persevere in the military system. They then start to upregulate those processes as they progress toward become officers. Those who failed in downregulating agency and outcome processing were less likely to remain within the military system after their first year. A capacity to adapt and modulate one’s own sense of agency appears to be a key factor for success in working within the hierarchical structures of military organizations. Our results also suggest that military training as an officer reverses this downregulation, perhaps through emphasizing accountability. As a result, officers may possess a sense of agency and processing of outcomes that distinguishes them from those they command. This possibility opens up socially relevant and optimistic perspectives for the development of a culture of responsibility within organizations. Cultures of responsibility within the workplace can enhance both civil society and the military organizations tasked to protect it.