1How might an experience of violence affect evaluations of situations involving aggression?1 More particularly, how are those situations evaluated by prisoners and by non-prisoners who have been victims of violence?
2These questions on attitudes towards punishment are addressed in the two studies presented below. These attitudes derive from the larger domain of morality, which consists of separating good from evil, truth from falsehood, legitimate from illegitimate, legal from the illegal (Vidal, 1998).
3The first psychological research on moral judgement was in the area of psychological development and moral ontogenesis. The most comprehensive theories-those that provided the foundation for later research-were those of Piaget (1932) and Kohlberg (1976).
4Piaget (1932) presented scenarios to young children that varied according to intentionality and consequences of the act. This act was either intentional without consequences or accidental with consequences. The subjects’ task was to determine, after reading the two scenarios, which of the two perpetrators was the more evil.
5Relying on the verbal explanations of the subjects-children under 10 years of age-Piaget distinguished between two mutually-exclusive moral attitudes : one is called « morally objective, » where the children make their judgements based only on the consequences of the act (which Piaget refers to as “moral realism”), while the other is called “morally subjective,” since judgements are based only on the intention. Thus, the children produced uni-dimensional judgements, based on a single variable. For Piaget, these two attitudes, being mutually exclusive, forced the children to use only one element (Intention or Consequence) in judging the actor, thus preventing multidimensional judgement (Intention and Consequence). In addition, Piaget was able to show that the importance given to intention increases with age and symmetrically, that the weight given to consequence decreases.
6Continuing in the Piagettian line of research, Kohlberg (1976) defined moral judgement as a journey through cognitive structures of increasing complexity. His preferred method for determining levels of development was the «moral dilemma ». It was the structure of the arguments used to resolve moral problems that interested Kohlberg. He defined six stages, organized into three levels (pre-conventional, conventional, and post conventional) considered as structures that were global, invariant, hierarchical, and universal. The claim to universality can nevertheless be challenged to the extent that the subjects of these studies were mostly male. Thus, Gilligan (1982) criticizes the models’ neglect of the moral experience specific to women. It is believed that the female moral outlook is characterized by an “ethic of solicitude” based on connections to other people.
7Following the developmentalist models, other approaches, such as the theories of socialization, developed and critiqued the constructivist orientation as having neglected the social processes governing the transmission of morality.
8According to the theory of social learning (Bandura 1977), the development of a moral sense depends less on the age of the subject than on exposure to adequate models. However, the theory of internalization (Hoffman, 1983) emphasizes the limits of this « observational learning », the acquisition of moral values being essentially related to the kinds of disciplinary interactions between parents and children. Hence, several studies have shown that a child’s level of moral development is related to the style of parental authority (Boyes & Allen, 1993).
9For Hogan (1973), the different psychological theories did not take into account the multidimensionality of moral judgement, which brought him to propose a model linking five distinct dimensions of personality and moral conduct: moral knowledge, socialization, empathy, autonomy and a component of moral judgement that is equivalent to a bi-polar dimension (personal ethical conscience versus an ethic of social responsibility).
10Finally, according to Forsyth (1980), it is possible to categorize individuals according to their « ethical ideology ». Thus, it is the differences among implicit moral philosophies that distinguish individuals.
11All of these studies show not only that many fields of psychology have studied moral judgement, but also that different methods have been applied. The first studies in the area of developmental psychology used qualitative methods based on interviews. Other experimental methods have subsequently examined moral judgement through quantitative analysis, based on scales of judgement. A significant number of studies have been conducted in this context, notably those using the Theory of Information Integration (Anderson, 1982).
12This theory, developed by Norman Anderson at the end of the 1950s, has origins in the psychophysical and aims to highlight the laws of psycho-cognitive processing and the integration of information. The premise of this theory is based on the ability of individuals to integrate manifold information to produce a unitary response. According to Anderson, judgement consists of three steps : (1) « evaluation » : attributing a psychological value to physical stimuli, (2) « integration » : combining these values. (3) “judgement”: producing an observable response on a linear response scale.
13Research conducted in the context of the Theory of Information Integration (TII) was at first concerned with the classic variables of Intention and Consequence. Léon (1980, 1984) and Surber (1977, 1982) demonstrated that children were capable, from a very young age, of integrating these concepts (Intention and Consequence) in their judgements of punishment.
14Subsequently, other variables were studied in order to understand the factors that influenced moral judgement. Later studies demonstrated the significant influence of variables such as the justifications for the act (Leon, 1982), the reward (Anderson & Hommers, 1985) as well as the age of the aggressor and the age of the victim (Przygodzki & Mullet, 1997).
15Leon (1982) then used not just the two usual variables-the intention and the consequences-but also information relating to the « reasoning » of the actor: he could truly regret his act, or simply admit his guilt, or be completely indifferent. The judgements of the subjects show that this third variable has a dampening effect on punishment. We find the same moderating effect on judgements with the variable « reparation », considered by Anderson and Hommers in 1985, following Berschied and Walster in 1967. This observable is important to the extent that even if not explicit, it is at least implicit in any moral judgement: we expect that someone responsible for an injury will make “reparation”, whether freely or by coercion.
16Przygodzki and Mullet (1997), wanted to analyze the integration of the ages of the aggressors and victims with the traditional elements of the « Intention and Consequence » framework. The results of this experiment show that the judgements regarding punishment vary in relation to these two factors of age. More precisely, the adult aggressor is punished more severely than an aggressor that is younger or older. The severity also increases with the age of the victim: the older the victim, the more the aggressor is punished. In addition, the age of the victim interacts with the age of the person judging: with adults and older adults, it is when the victim is an adult or an older adult, respectively, that punishment is maximized. Children escape this phenomenon of identifying with the victim. In the context of this study, the authors have observed another phenomenon-that the subjects’ judgements became increasingly severe. Roberts and Edwards (1989) had already highlighted this phenomenon: the subjects, after having read a scenario describing a serious crime, saw the subsequent offenses in a more serious light, judged the guilty more negatively, and punished them more severely. Does this mean that the more we judge, the more harshly we judge? In order to answer this question, Przygodzki and Mullet (1999) created a small experiment, the results of which corroborate the following assertion: as the process of judging the punishment due an aggressor is repeated within a short time frame, the more severe those judgements tend to become. Other studies have highlighted the possible impact of ethnic origins or gender on judgements about punishment. Thus, Cortese (1989) conducted a study among American men and women with African, European, and Mexican backgrounds. The results obtained in this research, and in many subsequent studies, support the hypothesis of a universal structure in moral judgement, with no demonstrated differences based on gender or ethnic origin.
17The impact of the experience of violence (as the perpetrator or as the victim) on the judgements regarding punishment has also been the subject of experimental studies. A study by Przygodzki and Mullet (1993) compared a group of people sentenced to imprisonment with a control group. The subjects were exposed to two pieces of information-the intention of the perpetrator and the consequences of the perpetrator’s action-in order to determine the degree of punishment to impose upon the perpetrator. This research demonstrates that the prisoners, giving less importance to the intention, are more gentle in their judgements than the control group.
18Conversely, the subjects who had been victims of aggression generally judged the committed offenses more severely. This severity manifested by the victims echoed the blame attributed to victims by external observers. Thus, some studies, such as those of Lerner and Simmons (1966), show that we have the tendency to denigrate victims, especially if we are unable to help them as they are suffering. Suffering refers to the harm suffered by the victim, with physical harm clearly being considered more serious than material harm. Howe (1994) proposed a classification of various offenses according to their degree of dangerousness, the most dangerous being « pre-meditated murder » and the least dangerous being « disturbing the peace ». In his second experiment, he combined the gravity of the act with its frequency in order to study their effects on the judgements of dangerousness: a recidivist being judged more dangerous than a first-time offender, and particularly since the acts committed were of increasing severity.
19In line with the earlier work, we decided to conduct a comparative study of the judgements of punishment made by prisoners and by non-prisoners who had been victims of verbal and/or physical attacks.
20To the extent that our studies rest on Anderson’s « methodology of the functional measure », they respect the two experimental conditions that are inherent to it, namely: (1) the need to submit to the judgements of the subjects the combinations of stimuli corresponding to all of the cells of an orthogonal factorial plane and (2) the use of numeric scales.
21We compared the results of our two studies, realized in two different professional contexts, in order to understand the different ways the experiences of violence impacted upon the judgements.
22It is anticipated that the judgements of punishment of the prisoners and the rail service employees would be different: the employees of the SNCF that had suffered a verbal and/or a physical attack would want to punish more severely than the control-subjects (Cusson, 1998), and the control subjects themselves would be more severe than the prisoners. (Przygodzki & Mullet, 1993).
23It is also anticipated that judgements of punishment would be influenced by : the intention of the aggressor (Leon, 1980), the aggressor’s age (Przygodzki & Mullet, 1997), the type of violence (Howe, 1988), and by the consequences of the act (Hermand et al., 1992) More specifically, these judgements should be higher (more severe punishment) when related to a physical aggression, not justified, committed by an adult, and causing significant consequences.
24Finally, an interaction between the type of respondent and the information regarding the intentions of the aggressor is expected (Przygodzki & Mullet, 1993): the prisoners should give less importance to this information than the employees of the SNCF or the subjects from the control group.
25This study focuses on 61 people who are divided into three groups equivalent in terms of age and level of education.
26The first group consists of 10 women and 18 men, service personnel (collectively known as « controllers ») on commercial rail lines in the north of France-who had been assaulted verbally and/or physically.
27The second group consists of 6 women and 11 men incarcerated in a detention centre for a variety of offences.
28Finally, the third group, which can be considered a control group, consists of 7 women and 9 men. These people had never been in trouble with the law and had never experienced a physical assault. In addition, their jobs do not regularly expose them to verbal violence.
29The material is made up of 32 different pages, each of which consists of a scenario and a response scale.
30Each scenario integrates 4 pieces of information that are our independent variables.
31 They are : the intention of the aggressor (justified and unjustified)2, his or her age (10, 16, 30 and 55 years)3, the type of violence (verbal and physical) and the consequence of the act (consequence and no consequence).
32In combining the degrees of each variable, 32 scenarios are presented identically to all the subjects in what is apparently a random order.
33The participants evaluate the same scenarios twice, applying the ITT methodology, with the result being 64 judgements. One sample scenario: « Following a pleasant family gathering, Guy, a 30 year old, leaves his home in a relaxed frame of mind. Meeting Hugue along the way, he insults him without reason. Hugue is not affected ». This scenario displays the following: unjustified intention-30 year old adult-verbal aggression-no consequence.
34Below each scenario is a response scale of 20 levels that allows the participant to express the degree of punishment to be applied to the aggressor. The left limit of the scale corresponds to « no punishment » and the right limit to « very strong punishment ».
35The experimental design is: 3 x 2 x 2 x 4 x 2 x 2 x 2 with the type of subjects (detainees or controllers verbally attacked and/or physically attacked or subjects-control), the gender of the participants (masculine or feminine) as inter-subject factors and the intention (justified or non-justified), the age of the aggressor (10, 16, 30 and 55 years), the type of violence (verbal and physical), the consequence (no consequence and consequence) and the repetition of the questionnaires (1st and 2nd questionnaire) as factors of repeated measurement.
36The experiment’s procedure is individualized, anonymous and composed of three phases.
37The first phase is the explanation and presentation of the material to the participants. This phase allows the participants to ask questions and familiarize themselves with the material. The participants then make some judgements.
38The other two phases are the experimental ones, during which the participants work alone and are not able to return to earlier answers to compare and/or correct them. The participants have unlimited time to complete their tasks. These two experimental phases are identical, except for the order of the scenarios, which differ from one questionnaire to the next.
39A numerical value was given to each participant based on the number of steps between the value at the left edge and the participant’s response value.
40The analysis of variance shows the non-significance of the participants’ gender (F(1, 59) < 1; ns).
41However, it does show the significance of the group factor (F(2, 58) = 7.3; p<0.05). The prisoners are the most lenient towards immoral acts to the extent that their average judgement regarding punishment is found at 7.1 (on a scale of 20) ; they are followed by the control group (M=9.3) and then the employees of the rail company, SNCF, (M=11.3). An analysis of the differences shows that the prisoners tend to differentiate themselves from the control group in their judgements (F(1, 58) =3; p=0.09). The average judgement of the control group also tends to differ from that of the SNCF rail employees (F(1, 58) = 3.3, p=0.07). The difference between the prisoners and the rail employees is statistically significant (F(1, 58) = 14.5; p<0.001).
42The difference between the control group and the prisoner group conforms to the results obtained by Przygodzki and Mullet (1993). We can also state that people who are regularly faced with verbal and/or physical violence tend to be less tolerant.
43The analysis of variance also shows the significance of the four factors used in the scenarios : F(1, 58) = 6.6 ; p < 0.05 for the factor of intention, F(1, 58) = 32.6 ; p < 0.001 for the age of the aggressor, F(1, 58) = 51.5 ; p < 0.001 for the type of aggression and F(1, 58) = 51.9 ; p < 0.001 for the consequence.
44Thus, the participants punish the aggressor harder when the intention is unjustified and when the act results in a consequence, which confirms the earlier results relative to the influence of these two variables.
45However, the intention is modulated by the type of violence (F(3, 57) = 8,7 ; p < 0.05). Indeed, the distinction between justified intention and unjustified intention is only made when the aggression is physical. Thus, discriminating between the two degrees of intention becomes crucial when the act is serious. The importance of this concept of intentionality refers to the need to understand the genesis of aggressive behavior, which is also manifest in any criminal trial.
46In addition, the severity (of judgement) increases with the age of the aggressor. Indeed, the degree of punishment on the scale is 7.6 for a 10 year old child, 9.2 for an adolescent of 16 years, 9.9 for an adult of 30 years and 10.1 for a person of 55 years. The participants are more severe towards an aggressor of 16 years than towards one of 10 years (F(1, 58) = 51.3 ; p < 0.001) and more severe towards an aggressor of 30 than towards one of 16 years (F(1, 58) = 9.7 ; p < 0.01). However, the degree of punishment does not differ relative to the aggressors of 30 and 55 years (F(1, 58) = 1.5 ; ns). This hierarchy of levels of judgement regarding punishment is not surprising if we consider the attitude of the French justice system towards young offenders. Immoral acts are considered, in part, as being the result of a « bad » upbringing, which explains the stress generally placed, at least in European countries, on the prevention rather than the repression of juvenile delinquency (Blatier & Robin 2000). The severity manifested in relation to adults can be explained by the intellectual and moral maturity that has normally been acquired by that stage, thus making the individual conscious and responsible for his or her acts.
47Logically, the subjects are more lenient in their judgements when the aggression is verbal (M = 7.8) rather than physical (M = 10.6). Also, the difference between physical and verbal aggression is more pronounced when the act is followed by a consequence (F(1, 58) = 25.4 ; p < 0.001). We can assume that the subjects attribute consequences from a verbal aggression to the fragility of the victim. Conversely, since physical harm is « quantifiable », an understanding of the consequences of physical aggression is necessary in order to evaluate the punishment that should be inflicted on an aggressor.
48To the extent that the participants judged the same scenarios twice, we have studied the factor of questionnaire repetition. This factor is significant (F(1, 58) = 10.6 ; p < 0.01). The judgement averages are 9 for the first questionnaire and 9.5 for the second. Nevertheless, the increase in the observed severity at the level of the second questionnaire was only observed for the aggressors at 30 and 55 years of age (F(3, 174) = 10.4 ; p < 0.001). The subjects probably interpreted the repetition of the questionnaires as the recidivism of adults conscious of their actions.
49Contrary to our hypothesis, the interaction between the group and the intention is not significant (F(2, 58) < 1 ; ns). This absence of results may be ascribed to the formalization of intention in our study. The two types of intention used in our experiment are characterized by degrees of justification for the act that are quite subtle and possibly do not emphasize their differences enough.
50This first study has allowed us to verify that the experience of violence as a victim or as an aggressor creates differences in the judgements regarding punishment. We have therefore decided to continue our study in a second experiment, both to introduce new variables able to modulate judgements regarding punishment, and to conduct this research in relation to victims of aggression in another professional context.
51A person’s profession influences his or her interpretation of an aggressive act. For the employees of the SNCF (French Rail Service), violence is often incomprehensible and can cause trauma. This incomprehension can explain the severity of their judgements. When the aggressive acts intensify, do the victims become more indulgent?
52In order to answer this question, we have studied the punishment judgements of nursing staff from a psychiatric hospital. Since the acts of aggression are generally the result of changes in mental faculties, these caregivers show an excellent comprehension of the aggressive behaviors that victimize them.
53However, is this indulgence, linked as it is to an understanding of the transgressive act, able to resist the repeated victimizations? To provide an answer to this question, we compared the judgement of punishment of caregivers who had been attacked once with those who had been attacked several times. To follow our reasoning to its conclusion and treat the totality of the issue, it seemed pertinent to study two groups of prisoners, first-time offenders and recidivists. The objective here is to see whether the indulgence typical of the punishment judgements of prisoners is increased by the number of incarcerations.
54In order to promote the process of identification of the subjects to people presented in the scenarios, we have introduced, in addition to the classic variables of intention, consequence and the nature of the act, information concerning the eventual recidivism of the aggressor and information on the number of victimizations.
55There is an anticipated graduation in the severity according to the type of participant : the caregivers who had suffered multiple aggressions would be the most severe, next, the caregivers victimized only once, next, the one-time prisoners, and last, the recidivist prisoners.
56We can also anticipate that intention will have an effect (Leon, 1980) as well as: the recidivism of the aggressor (Howe, 1988), the type of aggression (Howe, 1988), the number of victimizations, the consequences of the act (Hermand et al., 1992), and the repetition of the questionnaire (Roberts & Edwards, 1989). More precisely, the judgements of punishment would be most severe in the second questionnaire, for a physical aggression, intentional, committed by a recidivist aggressor, and where there are consequences.
57For the variable « number of victimizations » (the number of aggressions suffered by the victim), two hypotheses can be proposed. The subjects can experience compassion for a victim that has suffered multiple aggressions and judge the aggressor more harshly. Conversely, thinking that a victim of multiple aggressions is partly responsible for them, the subjects will show less severity towards the aggressor (Lerner & Simmons, 1966).
58This second research focuses on 72 people who are divided into four groups of subjects.
59The first group is composed of 14 women and 10 men, acting as caregivers (caregivers and assistants) in a Psychiatric and Mental Health Facility and who have each suffered one physical aggression in the course of the job.
60The second group is composed of 13 women and 10 men acting as caregivers in the same establishment, but they have suffered multiple physical aggressions.
61The third group combines nine women and six men incarcerated for the first time in a detention centre in the north of France.
62Finally, the fourth group is made up of six female and four male recidivist4 prisoners incarcerated in the same establishment as the third group.
63The mechanism consists of 32 pages, each containing a scenario and a response scale.
64Each scenario has five pieces of information that act as our independent variables. They are : the intention of the aggressor (an accidental or intentional act), the degree of recidivism of the aggressor (first time or repeat offender), the type of aggression (rudeness or verbal aggression), the number of victimizations (whether or not the victim had been attacked before), and the consequence of the act (whether or not there was a consequence). By combining the different degrees of each variable, the result is 32 variables which are presented to the subjects in a random order.
65As in the Experiment 1, the subjects were asked to make 64 judgements –since the researcher presented them with the same scenario twice.
66One sample scenario : « In a deliberate manner, and not for the first time, a person strikes someone. The victim has never experienced this type of act before and is upset by it ». This scenario is expressed in the following formula : deliberate intention-recidivist-physical aggression-victim never attacked in that manner-consequence.
67The same type of scale used in Experiment 1 is used by the subjects to express their judgements.
68The experimental plan is therefore : 4 x 2 x 2 x 2 x 2 x 2 x 2 with the group (caregivers attacked once or caregivers attacked multiple times or first-time prisoners or recidivist prisoners) as inter-subject factors and intention (non-intentional and intentional), recidivist (non-recidivist and recidivist), the type of violence (rudeness and physical violence), the number of victimizations (never suffered this type of act and already suffered), the consequence (no consequence and consequence) and the repetition of questionnaires (first questionnaire and second questionnaire) as factors in repeated measurements.
69The procedure is the same as in the Experiment 1.
70The numeric conversion is the same as in the Experiment 1.
71The subjects do not reach different judgements of punishment (F(3, 68) < 1 ; ns). Indeed, the average for the judgements of the caregivers attacked only once is 8.7, for those attacked multiple times it is 9.5, for the first-time prisoners it is 7.6 and for the recidivist prisoners it is 8.4. Even if one finds that the averages are in line with expectations, that on average the caregivers’ judgements are more severe than those of the prisoners, these results reveal nothing of any statistical significance. Thus, the caregivers do not make more severe judgements for punishment. The role of caregiver as well as the position of judge that the subjects are put into for the experiment can explain this result. Indeed, the role of the caregiver is not to judge but to understand these acts of violence that are sometimes the only way that a sick person can express his or her pain. Because the caregivers understood the aggressors in the scenarios as being mentally ill, they showed a pronounced leniency towards them. The lack of a significant difference between caregivers assaulted just once and caregivers assaulted multiple times tends to re-inforce the idea that a comprehension of the aggressive act is characteristic of caregivers regardless of how many attacks they have endured.
72Furthermore, the recidivists are not more lenient than the first-time prisoners. The personal characteristics of the people interviewed may be the reason for the results that contradict our hypothesis. Indeed, the questionnaires being strictly anonymous, we did not know the length of the sentence and the offense that lead to incarceration. In any case, these two pieces of information do not consist of perfect indicators of « criminal roots » of the people interviewed (Blatier, 1999).
73The analysis of variance shows the significance of our six independent variables : intention, (F(1, 68) = 59.7 ; p < 0.001), the eventual recidivism of the aggressor (F(1, 68) = 115.4 ; p < 0.001), the type of aggression (F(1, 68) = 21.3 ; p < 0.001), the number of victimizations (F(1, 68 )= 5.8 ; p = 0.02), the consequence of the act (F(1, 68) = 69.8 ; p < 0.001) and the repetition of the questionnaires (F(1, 68) = 4.7 ; p = 0.03).
74Indeed, as in Experiment 1, the subjects are more severe when the act is deliberate (M = 9.8) than when it is accidental (M = 7.4), more when the aggression is physical (M = 9.5) than when it consists of rudeness (M=7.6), and when the act results in a consequence (M = 9.3 versus M = 7.8 when there is no consequence). The subjects also reached judgements that were a bit more severe in the second questionnaire (M = 8.7) than in the first one (M = 8.4).
75In addition, the punishment is more significant in relation to the recidivist aggressor (M = 9.9) than for the first-time aggressor (M = 7.2). This result agrees with those of Slovic and Monahan (1995) which clearly indicated that a person is considered more dangerous when he or she has already committed aggressions in the past. This information about earlier aggressions committed by a given person is of prime importance, as much for judgements about the probability of recidivism as for judgements about dangerousness.
76Nevertheless, the effect of recidivism is affected by the type of participants (F(3, 68) = 4.9 ; p = 0.004). The analysis of contrasts indicates that the information relative to recidivism has the same impact upon, on the one hand, the caregivers, whatever the number of victimizations suffered (F(1, 68) < 1 ; ns) and on the other hand, the prisoners, whatever their penal status (F(1, 68) < 1 ; ns). However, the caregivers and the prisoners are not influenced in the same manner by this information (F(1, 68) = 13.7 ; p < 0.001) : in their judgements, the prisoners make a more marked distinction between the conditions of « non-recidivist » and « recidivist ». The precise knowledge of the penal code acquired by the prisoners as well as the questioning of the expertise of the caregivers when a patient repeats an act of violence can explain the result.
77When the target has had multiple experiences of this type of aggression, the severity towards the aggressor is greater (M = 8.7) than after a single experience. While this information regarding the number of previous victimizations was not already known as he committed the act, the subjects nevertheless wanted tougher punishment. The work of Alicke and Davis (1989), recently repeated by Hermand, Mullet, Tomera and Touzard (2000), has also shown this.
78Information unavailable to an individual at the time an action is occurring nevertheless very strongly influences the later judgements regarding punishment when the consequences are serious. These studies show especially that when a person, in legitimate self-defense, acts against a person that is depicted as dangerous, he is sanctioned less than when he acts against someone who is not depicted as dangerous. However, the person defending himself does not always know how dangerous his attacker is.
79Finally, the variable of consequence is modulated by the recidivism of the aggressor (F(1, 68) = 6.4 ; p = 0.01). The subjects are more severe for the condition « recidivist and the act having consequences » than for the other three conditions. The presence of just one of these two factors can create some leniency. However, if the negative consequences and the recidivism occur together, the recommended punishments are very severe.
80This study therefore shows that caregivers and prisoners do not prescribe different punishments. One of the criticisms that can be directed at this work concerns the limited number of subjects and in particular, the number of prisoners. Since participation in the experiments was voluntary, we had to do the work with this limited number of candidates. It is important, however, to emphasize that the statistical analysis performed on these population samples allows for the control of this limitation. Furthermore, the heterogeneity of our subjects as well as the multiplicity of the variables manipulated in these studies, although possibly appearing to create bias, were the results of our research objectives and hypotheses and in the end allowed for interesting results. It is true that we could have ignored the gender of the subjects, since this variable has generally proven to be insignificant in the various studies relating to moral judgement. Because this variable proved to be statistically insignificant in our first research, we chose to ignore it in our second.
81While the results of our first research allowed us to confirm both the leniency of prisoners and the severity of victims regarding punishment judgements, our second study modulated these first results by showing how important is the meaning attributed to victimization. Indeed, even if the professionals interviewed in the context of their work had suffered the same type of aggression (verbal and/or physical), some were very severe while the others were so tolerant that their judgements resembled those of prisoners.
82It seems essential to consider the source of the violence. The attacks aimed at the employees of the SNCF generally can come from any customer, which makes them unpredictable, gratuitous, and incomprehensible. The attacks on caregivers come from patients generally considered not responsible. Since violence is sometimes the only means of expression available to some patients, health professionals do their best to better understand these acts. It is revealing that for the same type of victimization, there is a different phrasing used : contrary to the SNCF, the psychiatric caregivers do not speak of aggression but rather of a « violent interaction » (Rioton, Peltier & Guibert, 1998). Also, because this outburst of violence is not generally aimed at the caregivers per se, they therefore do not consider it to be « aggressive behavior ». This interpretation completely agrees with the position of most psychologists who, whatever their specialty, make a distinction between « violence » and « aggressiveness ». Leyens (1979), for example, defines aggression as « a behavior made with the intention to hurt another person ». With regard to this motivational dimension, Bergeret (1992) describes aggressiveness as the actual « perversification of violence ». It seems clear, in terms of our results, that it is not the behavior in itself that is important but rather the representation that is made of it. Faced with an event such as an attack, it is obvious that a quest for meaning is required: depending on the available information, there are causal explanations made allowing, or not, the attribution of meaning to such behavior. The intention of the attacker, inferred by the victim, seems determinative to the extent that it has a direct influence on the understanding of the situation and the judgement made against the aggressor.
83Parallel to these interpretations relating to the personal characteristics of violent and aggressive individuals, it seems obvious to us that the social status and role of professionals interviewed in these studies influenced their understanding of these situations of aggression. The function of « caregiver » does not give rise to the same representations and social expectations as the function of « controller » and one can think that the helping orientation of the health-care professions creates an incompatibility with any judgement in favor of punishment. We may ask ourselves whether our results do not simply reflect the social norms that usually see care and punishment as being opposed to each other.
84We may question the strength and applicability of this position. Is the caregiver’s understanding of and leniency towards acts of violence, even after multiple victimizations, still intact where there has been a medical leave of absence resulting from the violence of a patient? Since the results of our second study have not shown a difference in judgement according to the number of victimizations suffered by the caregivers, it is logical to ask to where such indulgence might lead? Would this not be desirable, even necessary for the caregivers? Is the understanding shown by the caregiver, even when regularly confronted by violence, also not a way of preserving a positive image of the job and so of oneself? Finally, does not this behavior constitute a sort of strategy, more or less conscious, in order to adapt to attacks, allowing « to better cope with them ».
85Furthermore, does this permanent caregiver attitude of benevolence towards people displaying aggressiveness even help them? This again poses the complex question regarding the « treatment » of people who show both psychological problems and transgressive behaviors. It seems that today the problem is translated onto the social and political levels by this question: must irresponsibility inevitably lead to impunity?