How Violent Youth Offenders and Typically Developing Adolescents Construct Moral Agency in Narratives About Doing Harm

Abstract

Experiences that involve having harmed another person tend to compel individuals to consider their own behavior in light of their understandings of right and wrong, thereby serving as an important context and source of moral development. Although this process begins in early childhood, adolescents become quite preoccupied with the type of person they want to become and are thus likely to be most fully engaged in constructing a sense of themselves as moral agents.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Experiences that involve having harmed another person tend to compel individuals to consider their own behavior in light of their understandings of right and wrong, thereby serving as an important context and source of moral development. Although this process begins in early childhood, adolescents become quite preoccupied with the type of person they want to become and are thus likely to be most fully engaged in constructing a sense of themselves as moral agents.

Research has demonstrated that most adolescents think it is wrong to hurt others (Turiel, 1998). Nevertheless, in the course of their normal interactions adolescents often act in ways that result in other people feeling hurt or mistreated and must negotiate the threat insinuated in their own harmful actions, namely, that they are the sort of person who sometimes causes harm to others. Hence, experiences in which they have hurt others, and the ways in which they make sense of these experiences, are laden with implications for adolescents’ views of themselves as moral beings (Wainryb, Brehl, & Matwin, 2005; Wainryb & Pasupathi, 2008). Whether this is also true for delinquent youth, many of whom chronically engage in extreme forms of violence against others, is less certain. Given their documented delays in moral development (e.g., Stams et al., 2006; Tisak, Tisak, & Goldstein, 2005) and deficits in empathy (e.g., Bush, Mullis, & Mullis, 2000; Robinson, Roberts, Strayer, & Koopman, 2007), these youth may differ from more typically developing adolescents in the extent to which they think of themselves as moral agents and in their motivation or capacity to consider their own harmful acts in moral terms.

We take adolescents’ narrative accounts of instances in which they have hurt others to be a window into this process. In telling about any sorts of events, adolescents (like most people) tend to talk not only about what actually happened, that is, about the concrete actions that took place in the physical world – what Bruner (1986) referred to as the “landscape of action” – but also about the varied thoughts and feelings that they experienced and that they believe others to have experienced – the “landscape of consciousness.” Accordingly, their narrative accounts typically include more than summations of the past; they also implicate their interpretations and evaluations of the past as well as their future prescriptions and commitments. This may be particularly so in the retelling of transgressive events, as such events require justification and tend to initiate a search for meaning (Bruner, 1990). Therefore we expect that in re-construing the full landscape of their own harmful actions, adolescents would consider not only the ways things were and the ways they behaved, but also their thoughts, feelings, regrets, and commitments about the ways things could have or should have been and the ways they could have or should have behaved. Narratives lacking in such materials might in turn be seen as reflecting a truncated process of meaning-making – a process suggestive of developmental deficits.

In this chapter we compare how a group of adolescents enrolled in a public high school and a group of incarcerated violent youth offenders talk about instances in which they have caused harm to another person. The typically developing adolescents spoke about instances in which they pushed and shoved their peers, lied to them, excluded them from activities, or betrayed them; violent youth offenders described stealing cars, beating people up to unconsciousness, shooting, and killing. We examine the organizing patterns reflected in their narrative construals and consider what these patterns reveal about how these two groups of adolescents make sense of these experiences and the extent to which they construe a sense of themselves as moral agents within the context of their perpetration.

Knowing Wrong and Doing Wrong

Research with typically developing samples of children and adolescents has reliably shown that, starting at a young age, children judge it to be wrong and unacceptable to hurt or mistreat others, not merely because of the potential for ensuing punishment but because of their concerns with fairness and the well-being of persons (Turiel, 1998). Nevertheless, most children (like most adults) engage, some of the time, in actions that hurt other people. While one might take this to mean that children (or, more generally, people) are morally flawed or hypocritical, living a moral life does not truly require “moral purity.” Morality is inextricably bound up with a range of nonmoral concerns that also make up people’s lives, and most people struggle to integrate their moral concerns with competing desires and needs such as friendship, autonomy, self-preservation, power, and retribution (Turiel, 1998; Wainryb et al., 2005). Therefore, developing an understanding that people can be hurt and that hurting people is wrong is just one part of becoming a moral person. In those instances when competing desires get the upper hand, the struggle to make sense of the experience of having hurt another person and to integrate that experience within a view of oneself as a moral agent is also part of living a moral life.

A recent study (Wainryb et al., 2005) comparing children’s (ages 5–16) narratives of instances when they hurt others (“perpetrator narratives”) and instances when others hurt them (“victim narratives”) suggests that perpetrator narratives may be particularly well suited for understanding how children integrate their own moral transgressions into a broader view of themselves as moral agents. When children spoke about instances in which they had been the targets of harm, their construal of the experience focused narrowly on what they themselves had thought or felt. When they spoke about times they had perpetrated harm, their narrative construals presented a more complex focus, shifting back and forth between a concern for the victim’s fate and welfare and a concern with their own goals, intentions, and beliefs.

Whereas some may interpret the “back and forth” shifts in children’s narration of their own perpetration as a dithering strategy designed to minimize responsibility and appear blameless (e.g., Baumeister & Catanese, 2001), explanations couched exclusively in terms of self-presentation are limited and tend to minimize the complexity of human experience. In their stead we propose that instances in which one person has hurt another present an opportunity for genuine moral learning and moral growth (Wainryb et al., 2005). We further speculate that the patterns characteristic of perpetrator narratives can be understood as reflecting children’s attempt at acknowledging and owning up to the negative consequences their actions had for others without entirely banishing themselves from the moral universe. Indeed we think that it is precisely by focusing not only on what they did and how they affected others, but also on their own subjective experience and mental life – that is, on their goals, intentions, beliefs, and regrets – that typically developing adolescents work to integrate the harm they had caused with a view of themselves as moral people.

It is possible, however, that this particular way of making sense of harmful acts does not apply to youth who chronically engage in more extreme forms of violence. In the United States, youth violence remains a foreboding challenge. According to the report of the US Department of Justice (Snyder & Sickmund, 2006), in 2003, the last year for which there are complete data, children under the age of 18 accounted for 15% of all violent crime arrests in this country, with a small proportion of these youth offenders being responsible for the lion’s share of violent offences. These statistics sharply underscore the importance of understanding how juvenile offenders think about right and wrong and about their own aggression.

Research has documented serious deficits in the social-cognition, moral thinking, and affective processing of delinquent adolescents. In general, these youth have been shown to perceive and interpret social behavior in ways that increase the likelihood of aggression (Larden, Melin, Holst, & Langstrom, 2006; Liau, Barriga, & Gibbs, 1998; Tisak, Lewis, & Jankowski, 1997) and retribution (Slaby & Guerra, 1988). In general, they exhibit consistent developmental delays in moral judgment, scoring largely at preconventional stages 1 and 2 – stages that are characterized by self-interest and the endorsement of retaliation (Nelson, Smith, & Dodd, 1990; Stams et al., 2006). While they judge moral transgressions as being wrong, they also tend to reason that such acts are wrong not because they hurt others but because they negatively affect their own well-being, as when they are punished or sent to a juvenile detention facility. Consequently, they are less likely to view moral transgressions as wrong in the absence of rules and sanctions (Tisak et al., 2005; Tisak & Jankowski, 1996).

Research on affective processes linked to morality complements the picture emerging from the social-cognition and moral development literatures, as it points to serious impairments in these youth’ abilities to appreciate the emotional significance of events. Their distortions in the perception of others’ feelings (Slaby & Guerra, 1988; Carr & Lutjemeier, 2005) and deficits in empathy (Bush et al., 2000; Robinson et al., 2007), in particular, are likely to have serious detrimental effects both for the way they interact with others and the way they interpret those interactions (Arsenio, Gold, & Adams, 2005; Wainryb & Brehl, 2006).

How youth offenders may apply their moral understandings to the real world and, in particular, to their own acts of perpetration, is not known, as most of the research about this group of adolescents has been conducted using hypothetical dilemmas and self-report measures. Evidence from two small qualitative studies with incarcerated adult offenders (Green, South, & Smith, 2006; Presser, 2004) suggests that some of these individuals try to claim “morally decent selves” in spite of their lives of crime by neutralizing the immorality of their actions (e.g., by framing their criminal behavior as fleeting or as atypical of their “true self”); others speak of their lives in incoherent ways, as though they were not capable of salvaging a sense of themselves as morally good or felt uncompelled to do so. These data coupled with data indicative of their concern with self-interest and with the endorsement of retaliation, and data concerning their deficits in empathic understanding of the victims’ plight, all suggest that delinquent adolescents construals of their own harmful actions may differ from that of their more typically developing peers.

In the next section we undertake a systematic analysis of narratives by both typically developing adolescents and violent youth offenders about instances in which they hurt other people, as a means for understanding how adolescents do, or do not, integrate their own moral transgressions into a broader view of themselves as moral agents, that is, as moral people who sometimes do “the wrong thing.”

Adolescents Speak About Having Harmed Others

The data we present below pertain to two separate samples collected in the same mid-size Western city. One is a group of male violent youth offenders (N = 40), between the ages 14 and 18 (mean age = 16.5 years), of varied ethnic background (54% Caucasian) who had been convicted of a violent offense and were serving time at a youth corrections’ facility. All had multiple previous arrests (mean number of arrests = 22, range 2–72) related to offenses such as truancy, drug possession, theft, and assault, with a mean age at first arrest of 12 years (range 8–17 years). As part of an interview about their family histories and social relationships (Cloward & Florsheim, 1995), these youth were asked to recount “a time when you became violent.”

The other is a group of male and female adolescents (N = 28), between the ages 15 and 17 (mean age = 16.2 years), largely Caucasians (71%), middle class, attending high school. (This group of adolescents was part of a larger sample of children and adolescents between the ages 5 and 17.) As part of an interview dealing with various aspects of moral development (Wainryb et al., 2005), they were asked to talk about “a time when you did or said something, and someone you know felt hurt by it.” Whereas the narratives of female participants were longer than those of males (mean number of words was 218 and 134, respectively), no other significant gender differences were found in the content or organization of the narratives. For our present purposes therefore we combine the narratives of male and female adolescents in the normative sample.

We note here that the data from these two samples were collected at different times, using protocols that were similar but not identical. The primary differences were that the youth in State’s custody were asked specifically about violent behavior (rather than harmful behavior) and were not required to pick an instance in which they had hurt someone they knew. It is also the case that the social milieu in which the interviews occurred was vastly different. While all interviews occurred in private rooms, the violent youth were interviewed within the confines of a juvenile lock-up facility and normative youth were interviewed in their schools. Moreover, the participants in the two groups were not matched on dimensions such as SES, ethnicity, or intellectual ability. Thus the comparisons between the two groups of adolescents should be interpreted with caution and used largely as a means for highlighting distinct patterns.

The Language of Mental Experience

We first consider the extent to which adolescents include in their narratives their subjective experience, by contrasting the proportion of “factual” and “interpretive” language. Factual language pertains to references about perceivable aspects of an event, that is, references about any information that would be available to the perceptual capabilities of a bystander (the label “factual” does not necessarily implicate veridicality or accuracy). Interpretive language pertains to the subjective aspects of an experience, that is, any utterances about people’s internal mental states and processes, including emotions, goals, beliefs, and inferences.

For the purpose of scoring, narratives were first divided into idea units roughly corresponding to verb phrases (Pasupathi & Hoyt, 2008), which were subsequently scored as either facts (“we were driving around”; “it happened in the summer”) or interpretations (“I was mad”; “he was an annoying kid”). Inter-rater reliability was 83%, κ = 0.68.

As shown in Fig. 10.1, the narratives of violent youth offenders featured more than twice as many facts as interpretations. The large majority of facts referred to actions (I ran; I said); indeed, most of these narratives read like action movies. By contrast, the narratives of nonviolent youth included equal proportion of facts and interpretations.

Fig. 10.1

Proportion of facts and interpretations in the narratives of violent and nonviolent youth

The relative dearth of internality in the narratives of violent youth and the preponderance of facts render their accounts fast-paced. Consider, as an example, the following account, which scored at 45 fact-units and 8 interpretation-units (names of people and places in all narratives have been changed to ensure confidentiality):

I just beat up on people just for the fun of it. I’d just be walking down the street. And I’d just be sitting in the yard and someone would walk by. I would just jump up and hit him. About a month before I came in here, I had a headache or something and I was just laying there looking at the sky. I heard someone walking. I heard the gravel and I looked over and this kid was walking by, looking at my car, and he stopped and looked at the stereo for a minute, looked around, and kept walking—I don’t think he saw me—kept walking, and I got up and started following him and I said “Hey!” and he looked back and kept walking a little faster and I said “Hey! Hey, puto, come here!” and the kid turned around and I just rushed him and started beating him up. I said, “Where you from?” He goes uh, I think he said Idaho or Nevada. I said, “Not that, what gang are you?” and then I started beating on him. And there was this house. I hit him against a fence, and behind this fence was a house, and he lived right there. So I started to walk away, and he turned and just went through the fence, I was like “Aaah”. He just ran into the house. I was like…I took off. [YO#1]

This narrative illustrates a construal that is rich in facts and short on interpretations. There is coherence in this account in the sense that the factual aspects of the event are well represented: it is not hard to understand what happened when or who did what. But there is also a sense of incoherence: the narrator’s actions – confronting, harassing, following, and beating up – are not organized around any thoughts, goals, or feelings. What did the narrator think “this kid” intended by looking at the car stereo? What did the narrator want to accomplish or think he would accomplish, by confronting this person and beating him up? How did the narrator feel prior to the event or after the beatings? What did he think the victim felt or thought?

Typically, the language of mental experience provides the psychological glue that allows people, narrators and listeners, to make sense of the who–what–when; it allows us to see actions as springing out of intentional agents. The dearth of internality in the above account renders the experience psychologically incoherent, not only to us as readers, but possibly, and perhaps more importantly, to the narrator himself as an agent, a moral agent. Contrast the youth offender’s account to the one below, given by a nonviolent adolescent, which scored at 19 fact-units and 19 interpretation-units:

Um…let’s see… probably...probably was, let’s see, a couple of months ago. Yeah it was probably a couple months ago. We were playing, let me think. No, we weren’t playing, we were going to a game and we stopped off to get something to eat and my friend left without paying. And so I was like, "Man..." So I … I like walked over to him and I’m like, "You had the most expensive thing, you don’t expect us to pay for your meal, right?" So I kind of said some bad words to him, like "get back over there," like "do that." So and I…I can see like in his face that he was hurt by it. But at the time, I thought it was okay because you don’t just walk out on something. So that was…that was probably the time that I said something to somebody that...that I feel that I hurt them. And later...later I found out that I kind of...I kind of didn’t get the whole story before I walked to him because later I found out that he didn’t have any money with him and one of his…like one of my other friends was going to pay for him, and he was going to pay him when he got back to his house, so I kind of didn’t get all of the situation before I took it...I walked over to him and talked to him so... [NV#1]

This narrative account clearly differs from the previous one in terms of the extent to which internality is represented. As was typical of most accounts given by nonviolent adolescents, the telling in this narrative is less fast-paced and more reflective. The narrator tells us not only what he and others did, or when they did it, but also what he thought and how he felt as well as what he thought the other people intended and how they felt; indeed the narrative hinges on what the narrator thought his friend intended and on his later realization that his belief about the friend’s intentions had been mistaken. As clearly distinct from the narrative of the youth offender, in this case the narrator’s actions are rendered coherent and sensible through a rich sense of the internal experiences of both the narrator and the person he hurt. Indeed, through this telling the narrator implies that beyond the actual actions lies a “moral lesson.”

The Contents of Their Experience

Whereas the scoring of interpretations and facts captures, quantitatively, how much of what adolescents said in any given account represents internality and how much of it refers to noninternal, observable, elements of the event, it doesn’t tell us much about the actual contents of their experience. What sorts of actions do narrators describe? What sort of mental states and emotions?

The scoring of the narratives’ content was two-pronged (Wainryb et al., 2005). First, we scored the presence/absence of references to nine specific narrative elements deemed relevant to understanding adolescents’ construals as moral events. References to the perpetrator’s harmful behaviors, the victim’s response, and the incident’s dénouement, as well as references to any precipitating events, made up the “landscape of action”; references to intentions, emotions, and other mental states (e.g., beliefs, desires) made up the “landscape of consciousness.” Next, the actual content of each narrative element (e.g., the specific types of harmful behaviors, the specific emotions) was also scored. For both scoring systems, inter-rater reliability ranged from 84 through 100%, with κ’s ranging from 0.81 through 0.97. (It bears noting that, while the distinction between facts and interpretations corresponds roughly to the distinction between landscapes of action and consciousness, the two scoring systems are only partially overlapping. Adolescents could, for example, speak about their own “harmful behavior” [scored within the landscape of action] in ways that suggest internality [scored as interpretation]. In spite of the differences in scoring, the proportions of action/consciousness were strikingly similar to those of facts/interpretations.)

The Landscape of Action

The components of the landscape of action represented in the narratives of violent and nonviolent youth are shown in Fig. 10.2. Given that participants in both groups were asked to recount incidents in which they harmed another person, it is unsurprising that virtually all narratives included at least one reference to their own harmful behaviors. References to the victim’s response were present in a majority of narratives as well. Nevertheless, there were three significant differences between the two groups’ narratives.

Fig. 10.2

Landscape of action in the narratives of violent and nonviolent youth

One major difference was in the nature of the interpersonal harm. Indeed, the universes of interpersonal harm depicted in the accounts of violent and nonviolent youth were almost entirely nonoverlapping. For the violent youth offenders, the bulk of incidents referred to assault with weapons (33%) or without weapons (36%), and robbery and property destruction (12%); their victims responded by pleading or asking for help (44%), engaging in verbal and physical confrontation (32%), or running away (12%). By contrast, youth in the normative sample spoke largely about incidents involving offensive behavior, such as name calling and making insensitive remarks (50%), and trust violation, such as breaking promises or divulging secrets (30%); incidents involving even minor forms of physical harm were extremely rare (5%); the most common responses by victims were verbal confrontation (38%) and withdrawal (24%).

It is surely unsurprising that adolescents who enter the juvenile justice system describe a world of interpersonal harm that is different from the world of interpersonal harm within which nonviolent youth exist. We will suggest that the severity and chronicity of the harm depicted by violent youth offenders may be related to the lack of internality and psychological coherence represented in their narratives – a relation likely to operate in multiple directions.

A second important difference between the landscape of action of violent and nonviolent youth was in their spontaneous discussion of the ways in which the conflict ended. As shown in Fig. 10.2, the majority of narratives by violent youth offenders (80%), but only half of those by nonviolent adolescents (50%), included references to the incident’s dénouement. It is possible that events experienced largely in terms of actions tend to have clearer endings than events construed in reference to internal states: internal events end less abruptly, as people continue turning things over in their minds.

In addition, the contents of the depicted endings also differed, with violent youth offenders emphasizing the consequences the incident had for themselves, such as arrest and incarceration (52%) or escape (28%), and nonviolent adolescents referring largely to the effects that their behaviors had on their relationships, such as positive resolutions (48%), damage wrought to their relationship with their victims (21%), or attempts at reparation (27%). It makes sense that the conflicts of nonviolent youth result in damage to relationships and those of violent offenders result in incarceration or escape. It is important to also note that in construing their harmful actions, nonviolent youth focused on the repercussions those actions had for others or for relationships, whereas violent youth focused on the repercussions for themselves. These findings are indeed consistent with findings of the moral development literature (Tisak et al., 2005; Tisak & Jankowski, 1996) and with research on empathy deficits among delinquents (Bush et al., 2000; Robinson et al., 2007), as well as with the generally self-referential focus that children in normative samples assume when they speak about themselves as victims (Wainryb et al., 2005).

A third difference in the landscape of action as depicted by violent and nonviolent youth was the extent to which they included, in their descriptions, references to events or interactions that, in their view, had precipitated their own harmful actions. Recall that participants were not asked directly to speak about what may have precipitated these events (or about any other aspect of the incident); they were merely asked to talk about an incident in detail. Whereas references to precipitating events were virtually nonexistent among adolescents in the normative sample, the large majority of violent youth offenders included at least one reference (and often multiple references) to events that, in their telling, precipitated their own violent behavior. Consider, as examples, the following two accounts by violent youth offenders:

Um, my dad got really, I was mad at my parents cause we had gotten in an argument and um…Yeah, and got mad… And I got mad, yeah, and left and then I went over…Um, I had gotten in an argument with my mom and went over and I got to one of my friends and I started getting in an argument with him and I got really mad and um, I had this butterfly knife that I carry around because there’s like a whole bunch of gangs um living in the area that I was living, and um I pulled out the knife and pulled out the blade and um, I remember I grabbed his arm and turned him around so his back was facing me, and I held the knife to his throat and I told him not to mess with me cause, that I was mad and that I didn’t want him to keep messing with me and uh, he left me alone for a while and um…I, um… left and then I guess he told his mom, or he told somebody. And then like an hour later the police came over to my house and talked to me, and I admitted to it, and they charged me with it. [YO#2]

There’s only one crime I really remember that I’ve done on my own. This one was up on Washington Street. You know where Pete’s Corner is, up on Washington street, that little gas station? There’s an alley across the street from there, and I was back there going to the bathroom because in Pete’s Corner the bathroom is broke. So I was back there, you know, going to the bathroom behind a trash can when I just noticed this car going down the alleyway a little further up. I didn’t know what was going on in the car, only that there was people in it. So I was going to the bathroom and people started getting crazy saying “What’s up?” you know, “What’s up?” And I looked down and there were all these Cuban members hanging out the car and stuff and they started yelling at me. So I pulled the gun out that I had and opened fire at them. And I hit the one kid six times right here in the shoulder, and he started crying, you know, and crawled back into the car talking about “Drive!” you know “Drive, drive, he hit me! He hit me! Go, go!” So I took off running back across Washington Street and fortunately the light was red so I could cross. And Pete’s Corner was just full of everybody who was on Washington Street at that time. And I just ran through everybody. And there’s a fence behind and I just ducked the gun over the fence. The cops came and started asking everybody, and after awhile everybody left. And I came back later that night and got the gun out of the bushes. Then I went home. [YO#3]

As suggested in the two preceding accounts, these youth tend to think of their own violent behavior in the context of, and as a direct or a displaced response to, previous frustrations and perceived insults, threats, or provocations. These cyclical construals wherein, in their eyes, their aggression blends into or is confused with their victimization, seem consistent with findings concerning these youth’ hostile attribution bias (Larden et al., 2006; Liau et al., 1998; Tisak et al., 1997) and endorsement of retribution (Slaby & Guerra, 1988). And yet, the construals are more fragmented than one might expect based on said findings. In YO#2, for example, the narrator tells of an argument with his mother, then with a friend; it is unclear what the arguments were over, whether they were over the same matter or how – if at all – they related to his pulling a knife and threatening his friend. YO#3 makes more sense, inasmuch as the yelling could be construed as a more direct insult or threat; nevertheless, it is hard to see the narrator’s shooting as being commensurate with (and thus explained or justified by) those verbal insults.

In addition to conveying a sense of fragmentation, their construal of their own perpetration in terms of embedded sequences of actions also betrays a sense of diminished psychological agency. This is not to say that these youth are not actors. Both narrators in the above accounts are clearly action oriented and, presumably, are also motivated by something. However, they do not, in construing their experiences, integrate their actions with their own goals and reasons; rather, they present their actions as being embedded in external circumstances or in other people’s actions – both beyond their control.

The Landscape of Consciousness

Consistent with the findings from the fact/interpretation scoring, the landscape of consciousness of violent youth offenders was thinner, less populated, than that of nonviolent youth. While the fact/interpretation scoring does not specify whose internality is represented, the content scoring revealed that the dearth of internality in the narratives of violent youth was characteristic of the way they talk both about themselves and, even more so, about their victims.

As can be seen in Fig. 10.3, nearly all adolescents in the nonviolent sample included in their accounts references to their own intentions and to other mental states, and about half included references to their own emotions. In fact, 100% of narratives in this group included at least one reference to their own intentions, emotions, or mental states, and the majority (83%) included four or more references. By contrast, less than two thirds of the narratives of violent youth offenders included references to their own intentions or their own mental states and about one third included references to their own emotions. Altogether, 10% of the narratives by violent youth included no references whatsoever to their own internality and another 33% included a single reference.

Fig. 10.3

Landscape of consciousness in the narratives of violent and nonviolent youth

The types of intentions that nonviolent youth and violent youth attributed to themselves were also different. For the most part, nonviolent youth spoke about their acts of aggression as being incidental to their pursuit of other, legitimate, goals; examples were breaking up a relationship or excluding one person to spend time with another. When violent youth offenders discussed their own intentions they talked overwhelmingly (43%) about the motivations behind their aggressive acts in terms of responding to a direct provocation, to a sense of threat, or as retribution for a past slight inflicted directly on them or on friends or gang members. To a lesser extent they explained their aggressive behavior in relation to their own emotional dysregulation (11%), such as when they were angry or frustrated, or spoke about their desire to hurt the victim (12%), as well as about instrumental goals such as to obtain money for drugs and to cover up crimes (10%), and social goals such as to impress a peer group (8%).

The dearth of references to internality becomes more marked when the focus is not the self but the other: 89% of nonviolent youth but only 10% of violent youth included in their accounts at least one reference to their victim’s emotions. Similarly, 54% of nonviolent youth but only 20% of violent youth speculated about their victim’s mental states. In general, in terms of their infrequent use of internal state descriptors, violent youth are similar to much younger children (ages 5–7) in normative samples. However, the scarce attention that violent youth pay to their victim’s emotions resembles not so much the way in which such young children attend to their victim’s emotions, but the extent to which younger children in a normative sample, speaking as victims, attend to the emotions of those who hurt them (Wainryb et al., 2005). Naturally, the meaning of such similarities between violent youth and much younger children drawn from a normative sample is not transparent; our data cannot speak about the functions or structures behind the dearth of internality among violent youth versus 5- or 7-year olds. Thus we do not, by these comparisons, imply that violent youth offenders are like 5- or 7-year olds; rather, we report these figures to underscore the extent to which violent youth appear to be developmentally off track.

Our findings concerning the general dearth of internality in these youth’ accounts are consistent with research documenting deficits in empathy and social-cognition among juvenile delinquents (Bush et al., 2000; Robinson et al., 2007; Tisak et al., 2005). To our knowledge, there has been no other research documenting how these youth describe and make sense of their harmful behaviors in terms of their own thoughts and feelings or the thoughts and feelings of their victims. The lack of attention to their victims’ internality, and in particular to their victims’ emotions, is especially troublesome, given the centrality that these attributions have for making moral decisions and, more generally, for the process of moral development (Arsenio et al., 2005; Wainryb & Brehl, 2006). Indeed, this seemed like such a serious concern that we decided to explore this question using a different and less conceptually demanding, definition of internality; we turn to this question next.

Implicit Psychological Concepts

The scoring of internality both in terms of the distinction between interpretations and facts and in terms of the presence and type of emotions and mental states attributed to self and other relied solely on the explicit utterances and statements made by the narrator in the course of accounting for the event. Internality, however, may be implicitly represented in narratives. For example, adolescents might speak about situations in which they hurt others in ways that implicitly convey an understanding of persons (self or other) as such that they can be disappointed, betrayed, or hurt, that is, in ways that convey a psychological presence.

To assess this more implicit representation of self/other, narratives were scored for the presence/absence of five aspects of personhood embedded in the narratives, including physical/material, psychological, relational, group, and identity (Pasupathi, Wainryb, & Bezemer, 2007). The scoring was done once for the types of concepts of personhood applied to the “self” (i.e., the narrator) and once for the concepts of personhood applied to the “other” (or others). The range of inter-rater reliability scoring was 86–100%, with κ’s ranging from 0.828 through 0.921.

The distributions of conceptions of personhood implicated in the descriptions of the narrator (“self”) and the victim (“other”) are represented in Figs. 10.4 and 10.5, respectively. As shown in Fig. 10.4, nonviolent youth spoke about themselves in ways that implicated their psychological and relational beings; Fig. 10.5 shows that they depicted the “other” in very similar terms. The following account serves to illustrate the aspects of self and other implicated in narratives given by nonviolent adolescents:

It was in ninth grade, and…I liked this girl and we ended up going out, and I made the mistake of telling her I wanted to go out with my ex-girlfriend again. So she got very upset and […] her name was Casey and, I don’t know, we’ve been kind of off and on for about two years. And, I left this private school because I had bad grades, so, I went out to public school and I met her again. And for about two weeks, it was kind of [mumbles], internet and stuff, and then I finally asked her out, and she said yeah and the next day, um, I think it was – it was a weekend. And we went to the movies and we came back and I said, "That reminds me of when I used to go out with my girlfriend.” And she said…she asked questions about her, and I told her…and just pretty much what was she like, and how was she. And I told her and then she said, "Do you still like her?" And I said, I said “yeah,” and she said, "Are you sure?" And I said, "I think." And she was like, "Do you like me?" and I was like “Yeah.” But she didn’t believe me, so she got really mad and that’s when she kind of started getting back at me, yelling stuff at me, kind of starting stories and…yeah. That was about it. [NV#2]

Fig. 10.4

Notions of personhood embedded in the narratives of violent and nonviolent youth: SELF

Fig. 10.5

Notions of personhood embeded in the narratives of violent and nonviolent youth: OTHER

Implicated in this account are two clearly psychological beings. The narrator is capable of experiencing feelings and attractions, reminiscing about past attractions and feelings, and of reflecting on his mistakes. The “other” is capable of feeling jealous and angry and of considering and suspecting the narrator’s promises. Inasmuch as the interpersonal conflict and the resulting hurt hinge on relationships, past and present, the relational dimension of both the narrator and the “other” is also implicated.

Like the nonviolent peers, nearly all violent youth offenders spoke about themselves in ways that implicated their psychological being. This finding is tremendously significant given the dearth of internality explicitly represented in their narratives. It appears that, while these youth did not use language such as “I thought that…” or “I felt…,” they nevertheless described their own actions and experiences in ways that implicated themselves as psychological sentient beings. Consider the following example:

Um, uh, me and a friend of mine were walking to my house…going through a school park and there were some kids playing basketball, starting shit with us, talking a bunch of shit, running their mouth. And me and my friend took their basketball, and my friend threw it at one of their heads, and I hit one. And the kid I hit, hit the ground, and my friend kicked him in the head. Then we took off running, chasing them. One of their friends, the one that my friend threw a basketball at his head, he took off running, and we started chasing him. [YO#4]

The narrator’s physical self is clearly involved in this account (as it is in the majority of narratives by violent youth offenders): the narrator (and his friend) pick up a basketball, throw it, hit, and kick; they also run and chase. This event, however, could not have happened without the narrator’s implicit, though obscured, motivations to act the way he did. Although this narrator does not use any psychological, or interpretive, language when speaking about himself, this account implicitly hinges on him becoming psychologically responsive to a perceived provocation. Unfortunately, we do not know whether the narrator was upset or angry, or perhaps excited and thrilled by an opportunity for violence. Either way, the dynamic described suggests that something psychologically relevant was happening.

The picture of the “other” emerging from the narratives of violent youth offenders (see Fig. 10.5) was profoundly different from both how these youth depicted themselves (Fig. 10.4) and how nonviolent youth depicted the other (Fig. 10.5). Only half of the narratives by violent youth featured an “other” depicted as a psychological being. Half their narratives included a “physical” other (see YO#4 above, where the “other” appears in the narrative as merely a collection of hurt body parts), and 25% included “others” depicted in terms of their belonging to a group (an example would be a narrative in which the only aspect of the “other” implicated in the account is his being a member of another gang). Most remarkable was the finding that, even as all of their narratives depicted situations in which another person had been the target of their harsh and ruthless aggression, nearly 20% of their narratives depicted the “other” in ways that included no discernable concept of personhood (coded as “none”). Consider the following example:

I was just beating up some kid. I don’t know, I don’t know what hit me. He said something to me, and my other friend just hit him. He hit him, and I don’t know, I just got into it, I started going at it, too…I was…My friend just hit him. That’s why I got into it. I see my friend doing something, I’ve gotta do it too. I’ve gotta be able to watch his back and stuff, you know? Then, it just got to the court offices. I guess some kids filed assault charges. [YO#5]

In this narrative, the physical and relational aspects of the narrator’s personhood are implicated; some of his utterances also implicate a psychological being – albeit one without insight into his own behavior. By contrast, the victim (“other”) in this narrative appears as “this kid,” “he,” or “him.” Although we hear that the narrator beats “this kid” up, we do not gain even a vague insight into who this person might be (is he the member of a specific group?) or what he did; not even this person’s body parts are represented in the narrative. In this respect, in this story, the “other” could be replaced by an inert object. Recall, again, that this was not an isolated occurrence. One fifth of all violent youth offenders depicted their victims – the targets of their severe aggression and attacks – in ways that failed to notice even the most rudimentary aspects of their personhood.

It is important to note here that adolescents in the normative sample were encouraged to discuss an instance when they hurt someone they knew, but the interview protocol used with violent youth offenders did not specify whether the victim should be known or unknown. As it turned out, at times it was difficult to know the nature of the relationship between participant and victim precisely because the relationship was not a matter of concern, as in the case of the previous narrative, where the victim is described as “some kid.” Of the narratives in which the victim was identifiable, about half referred to a person known to the participant and half against a person unknown to him. Notably, whether the victim was known or not known made a difference not so much in terms of the mental states explicitly attributed to the victim, but in terms of the implicit conceptions of person attributed to the victim. In general, known victims were depicted in ways that implied a psychological being more often than unknown victims, and most of the victims whose depiction lacked any attribute of personhood were unknown.

Conclusions

In this chapter we examined the narrative accounts of two groups of adolescents who were asked to describe a situation in which they had hurt another person. Adolescents drawn from a normative sample spoke about times when they made insensitive remarks, excluded their peers from activities, lied to them, or broke promises. Violent youth offenders told of times when they beat up people, threatened them with knives or weapons, or shot at them. Both groups of adolescents were clearly able to discuss their moral transgressions in some detail, but constructed their narratives quite differently. In each case, their narrative accounts provided invaluable information about how these youth construct, or fail to construct, a sense of themselves as moral agents within the complex landscape of their experiences.

Typically developing adolescents, drawn from a normative sample, situated their own acts of unfairness and aggression within a rich landscape of consciousness. That adolescents are attuned to victims’ internality, especially victims’ emotional responses, has long been recognized as being essential to their developing understandings of the intrinsically negative and hurtful consequences of moral transgressions (Arsenio et al., 2005; Shaw & Wainryb, 2007). It is therefore no surprise that nearly all adolescents in the normative sample included references to their victims’ emotional states; it is also unsurprising that their accounts were rich in conceptions of their victims as psychological and relational beings.

The study of moral development has been less attentive to the representation of the perpetrators’ internality, including their depiction of their own cognitions and emotions and their efforts to make sense of their actions. Our findings in this regard are straightforward: the overwhelming majority of nonviolent adolescents included in their accounts rich explicit and implicit representations of themselves as psychological beings. In so doing, these youth bare their assumption, or conviction, that their world is one in which agents, each with her or his subjective mental experience, interact, and come into conflict. This, we think, allows them to view their own wrong-doing as arising from conflicting, albeit at times opaque, goals, desires, beliefs, angers, and regrets. It allows them to not only acknowledge their own aggression as wrong, but also as explicably human. They have done wrong, but they are also more than only “bad”: their harmful acts have become integrated into a complex sense of themselves as moral agents who, like the narrator who hurt his friend’s feelings by jumping to conclusions (NV#1, p. 191), can acknowledge and regret the pain they caused, learn moral lessons, make future commitments and, possibly, also forgive themselves.

By contrast, the narratives of chronically violent youth were characterized by a pervasive emphasis on the observable and a relative dearth of the psychological. While their narratives conveyed a fairly clear sense of what they did, these youth did not explain their actions in relation to their own psychological processes, such as what they were thinking or wanting, feeling, or intending to do. Indeed, at times the psychological language was so impoverished as to create the impression that their behavior was incomprehensible even to them. The frequent references to the ways the events ended (e.g., arrest, juvenile detention) further contributed to the sense that these incidents were experienced as taking place in the external, not internal, landscape.

There is some research indicating that when delinquent boys are put into a room and left to their own devices, they often begin to speak about aggressive acts they have committed or would like to commit (Capaldi, Dishion, Stoolmiller, & Yoerger, 2001; Granic & Dishion, 2003). It is possible that the stories delinquent youth tell each other are deliberately stripped of their internality to give the impression of toughness and it is also plausible that this tendency to regale each other with bravado, known as “deviancy training,” may ultimately contribute to the scarce internality observed in the violent youth’ narratives. It is worth mentioning, however, that, as an audience, interviewers are distinctly different from delinquent boys and are trained to neither step away from nor boost lurid or dramatic details. In any case, the possibility that the stories told by violent youth are intended to convey toughness does not diminish the point that the relative absence of internality is problematic. Whatever a youth’s motive for telling a particular story in a particular way, we believe the narrative structure reflects something meaningful about how the event has been, and is being, encoded.

In our view, the dearth of internality explicitly represented in these youth’ accounts of their own experiences betrays a deeply undermined sense of agency. Other features of their narratives support this interpretation. For example, their recurrent reference to precipitating events presumed to have triggered their actions, and their portrayal of their own actions as having been motivated by the desire for retribution, by the need to strike preemptively, or by unavoidable affect or circumstances, all hint at a complex perpetration–victimization overlap, such that a given instance of aggression is experienced as a point in a cycle of violence rather than the endpoint of a linear pathway from perpetrator to victim. We recognize that to some extent these narrative features communicate an implicit appeal, on these youth’ part, to the larger time-flow of slights and injuries in their lives. We suggest that they also convey a compromised sense of agency. It is as though these youth cannot fully distinguish between what they do and what is done to them.

None of this implies that these youth are not actors. Clearly, these youth act in and on the world. Furthermore, they speak of themselves in ways that suggest, implicitly, that they are sentient actors. And arguably, were they to be asked directly, all would surely concede that they had done these deeds. Nevertheless, their construals of their own actions and experiences do not underscore their agency, but undermine it. In their telling, their actions are not fully integrated; they do not emanate from fully psychological agents.

To some extent, the argument could be made that these violent youth integrate their actions around moral concepts such as self-defense or retaliation. The fact that these youth develop such concepts has been documented in the moral literature (e.g., Tisak et al., 2005) and their concerns with self-interest and retaliation can indeed be gleaned from their narratives. We propose here that while self-preservation and retribution are eminently moral concepts around which narratives of perpetration can cohere, these notions tend to diffuse the location of moral responsibility and reduce the sense of alternatives or choices – all fundamental aspects of agency. This is not to say that it is impossible to construct a sense of moral agency – though perhaps a diminished one – around concepts of self-defense and retaliation. And yet, violent youth offenders in this sample do not actually do that. Consider the following passage:

… I was going to the bathroom and people started getting crazy saying “What’s up?” you know, “What’s up?” And I looked down and there were all these Cuban members hanging out the car and stuff and started yelling at me, so I pulled the gun out that I had and opened fire at them …. [YO#3]

When we, readers, read such a passage, we make sense of it by connecting the various actions – the people yelling, the narrator pulling out a gun and shooting – via some internal attributions: upon perceiving the yelling and such, the narrator felt afraid, angry, or excited, or perhaps thought he was being threatened. But these psychological statements – this psychological glue – are not present in the account. They are part of our understanding, not the narrator’s. Thus we argue that even as notions of self-preservation and retribution loosely organize these youth’ narrations, they do so without establishing a firm sense of psychological agents acting on the world.

Why such a diminished sense of psychological agency? These youth’ limited abilities to connect their actions to a clear sense of agency may be seen as a failure in meaning-making associated with the severity and chronicity of the violence implicated in their experiences. One possibility is that extreme forms of violence, and the consequent hyperarousal, interfere with these youth’ abilities to make sense of their experiences and construe them in a psychologically coherent fashion while the events unfold (Fonagy, Gergely, Jurist, & Target, 2004; van der Kolk, 1994). Even after the events are over, it may be hard for these youth to retrospectively reorganize such extreme forms of violence in ways that allow them to retain or construct some sense of moral agency. As we suggested above, construals based on self-defense or retribution might work to undercut their sense of agency.

Another possibility is that at least for some of the chronically violent youth, the lack of psychological coherence and internality in their understanding of violent experiences helps them maintain a view of an interpersonal world in which their own and other people’s violent behaviors are intrinsic to the circumstances themselves. Such a view, that violence is a fact of the world rather than a human response, makes the extreme violence in the midst of which they function more tolerable (Frick, Stickle, Dandreax, Farrell, & Kimonis, 2005; Loney, Frick, Clements, Ellis, & Kerlin, 2003). Further contributing to this view is the finding that these youth produced highly impoverished construals of their victims’ internality and agency.

Whereas adolescents in the normative sample represented self and other in fairly similar ways, the narratives of violent offenders betrayed a conspicuous discrepancy in the way the narrator and the victim were represented – a finding that was pervasive, especially in regards to unknown victims and even when internality was indirectly measured via implicit representations. It is hard to tell how generalized their impoverished representation of other people’s internality and agency may be. On the one hand, findings concerning these youth’ pervasive distortions in the perception of others’ feelings (Carr & Lutjemeier, 2005) and deficits in empathy (Bush et al., 2000; Robinson et al., 2007) suggest that this phenomenon may be quite generalized. On the other hand, the finding that known victims were depicted in more psychological terms than unknown victims suggests that their impoverished view of victims may not be indiscriminately applied to all others. These youth might construe other people in more agentic and fleshed out ways in the context of more positive experiences, such as when they help a family member or friend. However, even if their construal of friendly others were more adequate, the fact that they so often view their victims as nonintentional agents or unidentified representatives of a group, with no discernable human characteristics, is likely to both reflect and perpetuate their tendency to engage in continued violence.

In fact, this is true more generally. The representations contained in the narratives of these two groups of adolescents both reflect their organization and integration of past experiences and inform how they are likely to respond to future experiences (Noam, 1988; Toth, Cicchetti, Macfie, Rogosch, & Maughan, 2000; Wainryb & Pasupathi, 2008). Thus adolescents’ narratives about times when they hurt others help us understand these youth’s potential and limitations as moral agents. The distinctions found between the narratives of violent and normative youth suggest that violent youth have a deficit in their capacity to reflect upon internal states. It is premature to suggest that violent youth lack a fundamental, perhaps even biological, capacity for self-reflection or empathy (Raine, 2002). Indeed, some nonviolent adolescents can be highly externalizing and nonreflective when describing their moral transgressions and some violent youth are more reflective and empathic than others. It might be that the difference between groups is relative rather than fundamental or categorical. Nonetheless, the possibility that seriously violent adolescents have deficits (relative or fundamental) in their capacity to reflect upon the internal states of self and others seems clinically relevant and an important developmental consideration when devising treatment models and practices. Although we stop short of declaring that the violent youth offenders in this sample are sociopathic, as defined by Frick and others (Salekin & Frick, 2005), our analysis of their narratives provides a window into the development of sociopathic potential, just as the narratives of the nonviolent youth provide a window into the development of a more sophisticated understanding of the self as a moral agent, who must grapple with the occurrence of moral transgressions.

Our analyses also suggest that the stories typically developing and chronically violent adolescents tell about their own experiences may be an important conduit for moral development and clinical intervention. In the course of everyday interactions, typically developing youth tend to recount their harmful behaviors, telling stories about what happened in the school yard or the internet, often half-bragging about and half-confessing to a moral transgression. The telling of such stories and the spontaneous conversations that ensue from these stories, with parents and teachers, and sometimes with peers, serve as a context for youth to make sense of their transgressions and ultimately integrate their own harmful potential with a continued sense of themselves as people who make, or are capable of making, moral decisions.

But adolescents who commit acts of serious violence face a unique developmental conundrum: they are violent because they lack the capacity to manage their impulses, to comprehend the consequences of their behavior, and/or to empathize with their victims, but these same developmental deficits also interfere with their ability to think of their experience in ways that help them learn from it and develop moral agency. Moreover, these youth, who experience themselves as morally adrift, also rarely have the sorts of interpersonal relationships within which they might work to make sense of their transgressions and learn to regulate their aggression. While more research is needed to glean specific clinical implications from the narratives presented in this chapter, it seems possible that when adults are able to listen to and acknowledge these youth’ aggressive impulses and desires, as expressed in their stories about their very serious transgressions, an interpersonal structure for containing these youth’ destructive potential might emerge and develop and, over time, become internalized. Thus we believe that providing chronically violent adolescents with an interpersonal context that encourages them to narrate their moral transgressions while considering the thoughts and feelings of those involved and sorting through the possible antecedents and consequences might also help them learn to monitor and control their aggressive behaviors. Whether this interpersonal context is provided by parents, teachers, counselors, or probation officers probably matters less than whether it provides the mix of guidance, autonomy, and balanced compassion that typically facilitates the development of moral agency in more normative contexts.