Exploring Violent Cosmologies From a “Radical Interactionist” Approach

This article advances a theoretical perspective on violent crime, using interviews with male prisoners in Italy who had perpetrated violence. By drawing on Athens’ (1992, 1997, 2007, 2017) “radical interactionism,” we propose the concept of “violent cosmology” in order to counter linear explanations of cause and effect. In an effort to complement narrative criminologists’ contributions, we seek to recognize and understand the dimensions of meaning that are accessed by social actors when they prepare and carry out a violent act, exploring the psycho-social processes that animate violent social experiences from the perspectives of perpetrators. Specifically, we suggest that a “radical interactionist” approach, in dialogue with narrative criminology, can help (1) illuminate the sources of perpetrators’ narratives; (2) explore the interplay between individuals and social structures; and (3) investigate ambiguities in the narratives of violent actors. Finally, we examine how enhancing the reflexivity of violent actors and recognizing the specificity and integrity of their lives and social experiences is a necessary precondition for understanding violent crime.


Introduction: Imagining the Violent Actor as a Complex Cosmos
How do we imagine violent acts and their perpetrators? What emotions do violent actors feel before, during, and after the performance of a violent act, and do they engage in an inner monologue? As researchers and as individuals, our answers to these questions depend on how we view the world in which we live. The world, as a socially constructed reality, appears to us through our symbolic interactions with others. Over time, we create an entire cosmology in which our social negotiations take place, and by which they are legitimated (Inghilleri 2005).
In this article, "cosmology" becomes a "sensitizing concept" (see Blumer 1954: 7) directed at reassigning a meaning to (violent) human behavior beyond any rigid and formal distinction between "normality" and psychic suffering. 1 This concept suggests that each "violent" act, no matter how "crazy" and bloodthirsty it may seem, implies a cosmology. by this rich theoretical tradition" (Shaw 2017). In Athens' proposal, "radical interactionism" recognizes the principle of "domination" (roles of superordination and subordination) as a cornerstone of society and its basic institutions. 2 This notion replaces Mead's principle of "sociality" in order to better understand the symbolic and social interaction, in all social acts, and in violent crime, in particular (Athens 2007(Athens , 2013; see also Jacobsen and Picart 2019; Natali 2009, 2019). Moreover, in "radical interactionism," the Meadian notion of the generalized other (the "Me") is replaced by the one of the "phantom community"-a distillation of past and present experiences as interpreted by individual social actors throughout the course of their individual biographies and in dialogue with their internalized "significant others." This allows us to understand and explain violent behavior, taking into account the biographical uniqueness and the creativity of the individual. The inclusion of "domination" and of "phantom community" as a "sensitizing concept" in the analysis of violent offenders' narratives helps us to understand better how "individuals' narratives are both shaped by social structure as well as being creative and agentic" (Fleetwood 2016: 174), and it enables us to explore the interplay between individuals and social structures "not in opposition, but in a relationship" (Fleetwood 2016: 179).
One could say that narrative criminology (Presser and Sandberg 2015) looks to the ways those who commit criminal acts build their own sense of self, interweaving a series of narratives to make sense of the experiences lived and the socio-cultural world in which they are immersed. A critically oriented narrative criminology can focus on harm, rather than just "crime," and illuminate the dynamic aspects of harm-doing ; such an approach can consider the complexities of crime and harm in late modernity-such as the stories about violent crimes, which are structured and creative at the same time, and which compel researcher reflexivity.
If it is true that ambiguous and complex violent experiences have to be "inhabited" and considered as an important challenge aimed at re-observing (and not at surmounting) certain dualisms-such as mental illness and "normality" or agency and social pathology-a perspective, such as the one we propose, can represent a way for cultivating these sensitivities. Before describing our method, two additional points are in order.
First, we use the terms "violence" or "a violent act" to refer to an intentional physical attack that results in harm to others, such as in the case of assault, homicide or sexual assault. This choice is not, however, intended to dismiss or exclude other forms of violence (e.g., cultural, economic, psychological, structural, symbolic) (see also Auyero and Berti 2015;Henriksen and Bengtsson 2018;Jackman 2002). For analytical purposes, we decided to restrict our definition to the "behavior by persons against persons that intentionally threatens, attempts, or actually inflicts physical harm" (Reiss and Roth 1993: 35). 3 There are, in our view, clear analytical advantages in such a circumscribed definition of violence. Paraphrasing Tilly (2003), we could say that applying the term "violence" to all relations that we consider exploitative, oppressive or unjust undermines the effort to explain violence.
Second, we are well aware that "criminologists access the power of the state to identify violent persons" and that " [p]olitical and economic interests shape definitions of violence" (Presser 2005(Presser : 2069; see also Barak 2003). Even if research on violence is already conditioned by these dynamics of power-in which the researcher is still enmeshed-we do not see violence as an isolated physical encounter, but as a social experience and process, linked to logics of domination that may be unpacked thanks to the narratives that orbit around it (see Henriksen and Bengtsson 2018: 111). This aspect is compatible with a narrative and critical criminological approach that sees harm not as an isolated incident-in contrast with the tendency of mainstream criminology-and that counters the "individualism of the dominant approaches to the etiology of criminal behavior" (see Presser and Sandberg 2019: 136). As such, we start from the narratives of people who committed "violence" in the traditional sense of unlawful exercise of physical force, exploring how other equally damaging forms of violence and harm are at play because, from a "radical interactionist" perspective, violent crime is not simply an illegal act, but a by-product of the struggle for domination (Athens 2017).

Methodological Aspects of Undertaking Research in Prison From a "Radical Interactionist" Perspective
Our research involved a method that eschews linear explanations of cause and effect. Instead, we followed a so-called processual model according to which the violent action represents the result, which is never taken for granted, of a long and difficult interpretative and symbolic process that was developed by an actor and brought to fruition at a later point in time (Athens 1992(Athens , 1997; see also Lindesmith 1981). This processual model operates both from the point of view of the dynamics of violent criminal behavior-which concerns, above all, the interactions and the interpretations of violent situations by the social actors involved-and from the point of view of "criminogenesis," which refers to the formative process that leads an individual to embrace a violent self (Athens 1997(Athens , 2003. In this light, our research shows how the actor does not react to an external stimulus without any resistance but instead interposes-and in some cases opposes-reflexivity, which is understood as a symbolic filter of reality. In order to capture these tensions, we used the method we describe in this part. We undertook a study of twenty violent offenders between 2008 and 2020 at the Opera Prison in Milan and at the San Vittore Prison in Milan. All the participants had engaged in physical violence against others, resulting in prosecution, conviction and imprisonment for violent crimes (homicide or sexual assault). The participants were between twenty and fifty years of age at the time of the interview. All our research participants were men who, as Presser (2008) notes, perpetrate most violent crimes. Our group consisted of men who carried out physical attacks that could be considered "extraordinary" episodes of violencemany of whom were prisoners with violent criminal careers. We justified this decision by drawing on the "continuity thesis," according to which "people who commit heinous violent crimes always have some violence-related experiences in their background, although they may sometimes be deeply hidden from others and not apparent without a thorough and painstaking investigation of their biographies" (Athens 2017: 78). Different approaches and methodologies can be employed in constructing stories about people's (violent) lives, such as in-depth interviews, focus groups, life stories, and ethnographic research (see Brookman 2015: 228-229). Because our aim was to center our attention on the reflexivity and the interior dialogues that accompany violent acts before, during, and after their realization (see also Fernyhough 2016), we decided to use semi-structured interviews based on an interview guide that we developed. Each interview lasted from four to eight hours and included only the participant and ourselves. They were audio-recorded in their entirety and transcribed. The participants were interviewed on three to five separate occasions in order to afford the interviewee more time to tell his story and in order for us to elicit more detail. All the participants were guaranteed confidentiality. In order to grant primacy to the men's stories regarding their violent acts, we decided not to learn the details of their cases beforehand or to examine any documents regarding the legal proceedings surrounding their crimes. For us, it was of primary importance to extract the personal and relational truths that each actor was willing to allow to emerge in his narrative (see also Brookman et al. 2011).
We followed an interview guide that was intended to explore the following open questions: 1. Regarding the act: Can you describe what happened when you committed the violent crime for which you were sent to prison? Can you describe for us the context (actions, conversation/dialogue, gestures) that occurred prior to the act? 2. Regarding the self-narrative, self-perception and personal memories about the act: What were you thinking (about) while you were engaged in the violent act? What were you saying to yourself? Could you think of possible alternatives to the violent act? 3. About the participant's "phantom community" and emotions: How and in what emotional tone and with what words did your inner interlocutor speak to you? Did you feel any emotion(s)? If so, which? Were you able to control the emotion(s)? 4. And about the interpretation of the situation and image of the victim: What was the behavior and the attitude of the victim? Did you consider your victim superior or inferior to you? Did you watch your victim dispassionately or, on the contrary, did you feel strong emotions/feelings about/toward him/her? Did you feel threatened, thwarted, diminished, humiliated and/or dominated by the victim?
From there, we began an inquiry into the origins of the participant's violent cosmology based on: 1. His memories of past violence: What are your first memories linked to violence? Have you ever suffered violence? Have you ever witnessed violence suffered by your family, by your friends or by people you care or cared about? Can you describe for us these experiences-the situations in which you found yourself? 2. His "significant others": Was there a violent person in your family or among the people/ friends with whom you interacted frequently? Was there someone who told you that in life one needs to use violence-someone who was trying to "educate" you, so to speak? 3. His self-image: When you think of yourself, how do you see yourself and how would you describe yourself? Does/do this/those image(s) coincide with those of your family or friends? Have you always seen yourself in this way (and have you always been seen in this way) or can you recognize some changes in your way of seeing yourself? For how long have you perceived yourself and for how long were perceived by others in this way?

3
Do you remember which violent acts you committed during the period characterized by this specific self-image? Can you describe the situation, the thoughts, and the emotions that characterized them? 4. His social and cultural world: How would you describe the neighborhoods in which you lived and the groups/the people with whom you interacted frequently at this time? Was there a group or someone who dominated you, your family, your friends, and was some sort of model for others? Were there unwritten rules to be respected? Did you share them? Did you align with them or did you challenge them? Did you feel at ease or "out of place"? 5. His experiences or perceptions of self-change: Have you ever lived through periods of personal crisis and deep personal change? If so, were those changes caused by some particularly important and/or dramatic event-by some memorable experience that marked your life and your way of having recourse to violence?
Some of our interviews began with the story concerning the fact/event-the criminal action that put the participant in prison-that is with the first part of the interview guide. Other interviews started with the interviewee's earlier memories and past, and then moved forward in time to the event. Essentially, we allowed the interviewee to construct his story, without requiring him to trace a strict chronology.
Furthermore, the open-ended questions listed above helped us to focus on how the participants engaged in their violent actions, rather than why they decided to commit them (see Sandberg et al. 2019). In fact, we felt that asking participants why they committed a violent crime would seem to invite them to justify and excuse their actions and to respond with stories that would negate or reduce their culpability. 4 Finally, it is important to note that we proposed and developed with our interviewees topics that were highly significant to their lives, although they were not necessarily linked to the crimes for which they were incarcerated. These issues and topics emerged during the interviews in unpredictable ways, beyond the specific questions planned in our interview guide-such as those regarding drugs, friendship, gender identity, love, the perception of their own bodies, and their views on the future. 5 In the next part, we describe the notion of "violent cosmology" as the conceptual pivot in order to understand the reflexive and interpretative processes undertaken by violent offenders before, during, and after the performance of a violent act. 4 Asking participants why they committed a certain violent act seemed like a recipe for requesting them to engage in "techniques of neutralization" (Sykes and Matza 1957; see also Colvin and Pisoiu 2020;Maruna and Copes 2005;Sandberg et al. 2019). Hardened violent criminals no longer need to neutralize the immorality of their actions because they do not consider them serious anymore . On the contrary, they can use these techniques to make the interviewer believe that they are novices. 5 As Molotch andBoden (1985: 285, quoted in Presser andSandberg 2019: 137) observe: "Demands for 'just the facts', the simple answers, the forced-choice response, preclude the 'whole story' that contains another's truth."

A Personal Cosmology Oriented Toward Violence
The worlds that people experience, the events that befall them, and the significant individuals they meet all leave indelible traces on and in their bodies-lasting "evidence" (as it were) of their encounters (Minkowski 1936)-that tell their stories and demand to be heard and acknowledged. When we converse with ourselves, we try to impose some order, structure, or a "cosmology" upon that multiverse of images, representations and voices-some of them entrenched in our minds-that ask to be understood, listened to, and, at times, obeyed rigidly. As Presser (2009: 179-180) remarks, self-narratives and stories help us make meaning; they forge a sense of coherence that experience lacks and they are oriented toward a plot. Therefore, a "cosmology" is also the construction of a plot, directed at ourselves and others, that gives a personal narrative its meaning: "Everyone [in fact] is not a chance combination of elements but an organized system of meanings. Since meanings tend to be self-sustaining, each person is constantly striving to move themselves in decisive directions" (Shibutani 1961: 285). As Brookman and colleagues (2011: 417) elucidate: Essentially, people construct stories to account for what they did and why they did it. Narratives impose an order on our actions and explain our behavior with a sequence of events that connect up to explanatory goals, motivations, and feelings. They also act to shape and guide future behavior, as persons act in ways that agree with the stories or myths they have created about themselves….
Following Brookman and colleagues, we suggest the importance of understanding a violent actor's "personal cosmology." A "personal cosmology" can be defined as the multiple ways in which an individual constructs his/her universe and thinks and feels about himself/herself and the other human beings within it. Thus, it is an attempt to create order in their ethical, personal, and symbolic worlds in the midst of the chaos-and-order of the "external" socio-cultural world. This quest for order is not isolated from the rest of an individual's life in some philosophical "never-never land." Every personal cosmology is situated, and our knowledge of ourselves is the outcome of a process based on our reflexive negotiations with other social actors within situational contexts that are also structured in terms of domination (see also Presser 2005). Order is recreated continuously in the interpretation of a situation, in the preparation of an act, in experiencing the body, in the use of a personal vocabulary, and in the systematic performance of certain kinds of behavior. It is also expressed in the moral classification of human beings (see also Seeger 1981). This means that while social actors are narrating their own story, they are struggling to create their own universe.
That being said, how can an "organized system of meanings" orient itself in a violent direction? In order to answer this unsettling question, Athens (1994) elaborates on the sensitizing concept of the "phantom community," which could be described as the interlocutor that plays a leading role in our soliloquies. According to Mead (1934), it is generally accepted that the main interlocutor of the self is the "Me," which is the generalized voice that individuals assume during the course of social interactions. In Athens's view, however, it is the "phantom community" that plays this role. That is to say, the "phantom community" is composed of the inhabitants of a real or imaginary community, whose attitudes we normally take for granted and before whom each of us tries to control, protect, or improve our reputations (Athens 2007(Athens , 2013; see also Caviglia and Cecchini 2009). In short, while the "generalized other" may be depicted as a Greek chorus giving (a) voice to a unanimous and constant community, the "phantom community" resembles a "parliament" representing as many opinions as there are significant speakers that we have internalized in the course of our lifetime and to whom we anchor ourselves in every situation. The "phantom community" is therefore a distillation of past and present experiences as interpreted by individual social actors throughout the course of their individual biographies and as derived from their personal experiences of participation in social acts (Athens 2007).
Although this inner dialogue with our "phantom communities" precedes and accompanies our interpretation of a given situation, it does not refer simply to contextual immediacy. Indeed, the "phantom community" refers to something rooted far more deeply to that unbroken sense of self operating within neural circuits that are used repeatedly when adopting other people's attitudes over time and distilling moral maxims and guidelines to refer to when undertaking actions (see Athens 2007;Haidt and Bjorklund 2008). It is through these processes that "[w]e offer reasons for our actions that make sense to our real, imagined and potential interlocutors, including ourselves" (Presser 2009: 180). As "characters in search of an author," our "phantom companions" will suggest the answer to the compelling question of "What are we going to do?" To illustrate, we point to an excerpt from an interview with a participant who was found guilty of sexual assault and homicide. He describes how his soliloquy developed before the violent act took place: Interviewer: What did you think and feel when you were about to commit the act?
Did you in some way talk to yourself?
Zeno: I don't even know why I took out the knife, at once everything went wrong… the time then is short, fractions of a second, time is out of control… you have no time to think, reflect, or say: "Why? What am I doing?" It's as if I was not there… At one point, though, I sort of felt a voice telling me to stop, but there was also another voice telling me that the damage was already done and that I could not stop, could no longer draw back. It was my inner voice telling me: "Stop, stop! What are you doing!?," while another voice was saying: "You have done it now: you can't stop here." This last voice was still mine, but more sinister. The nastier voice was telling me that I could not stop and should kill her: "Finish her off," it ordered. This voice immediately overcame the one that was telling me: "Stop! What are you doing?" or even "Run!" Other things are "embedded" more deeply inside me and they are things that I cannot get out….
When we talk to ourselves in our heads, as Zeno describes, we do so by talking to other interlocutors. In this excerpt, Zeno starts by stating that he does not know why he committed the act; when given the opportunity to explain, however, he manages to describe clearly how his action was directed by the "voices" that animated his interior dialogue during those brief but most profound moments. Such voices are either specific individuals or embodiments of social rules from whom we imagine we hear exhortations, judgments, opinions, orders, praise, requests, suggestions and warnings. The words we attribute to them, however, are nothing more than a reformulation of our own words engaging in a constant semantic shift. It is in the interstices of these processes that deliberations, as well as the planning of future actions (including highly deviant ones), emerge.
Because society is less normatively homogeneous than implied by the concept of a "generalized other," the social actor has more degrees of freedom and therefore also takes on the burden of determining whose expectations to endorse personally and by what measures that person will satisfy them (Archer 2003). In this sense, violent actors are individuals oriented toward an organized system of meanings that can be highlighted by investigating the relational universe into which they are "thrown" from birth. What we call a "violent cosmology" thus becomes the invisible thread that ties together the multiple and complex narrative experiences that promote, accompany, and follow a violent act.
In the next part, we analyze how reflexivity, even if interwoven with several layers of "opaqueness" (which from our perspective does not equal the unconscious), makes the actors sensitive and responsive to their own emotional, physical and symbolic worlds and to their own actions. Athens' (1992Athens' ( , 1997 theoretical perspective contains innumerable elements which seem to uncover that the inner processes that lead to violent acts are made conscious by that symbolic process whereby the actor (1) assesses if and how certain elements (beliefs, desire, emotions or ideas) concern him/her; and (2) decides what to say and do in a given situation. It is this internal soliloquy-which has a psycho-social and relational nature (Archer 2003)-that confers meaning on one's actions.

Reflexivity and the Unconscious
If every human being is a "cosmos" that generates meaning, we suggest that reflexive deliberations (Archer 2003) may be seen as activities of which the actor is largely aware. It is in this context that the judgments, opinions, praise, and warnings of internalized "others," who advise or determine how to translate all this into (violent) acts, occur. The crucial point in this representation of the experience is the idea of a contact between the subject and himself. This contact is not truly a proper "reflection": it is a "listening" to one's presence (Jedlowski 2008: 11).
This epistemic-methodological lens, however, which is focused on the reflexive and interpretative processes of violent actors, does not diminish the role that the unconscious can play in the plotting of violent acts. As psychic individuals, we are, in fact, in contact with the unconscious, that manifests itself in this reflexive process between the lines of conscious speech. This relationship between the unconscious and the conscious mind implies a mutual presupposition, not a genetic connection. It is not the unconscious that produces a conscious mind, nor does the conscious mind replicate the unconscious; rather, it is the synergy between them that makes us what we are. With this in mind, the problem no longer consists in understanding how the emotional, instinctual, and unconscious spheres impact the "conscience" and "reason," but in understanding how they base themselves on uncertain and relatively unknown foundations. As Palermo (2011: 31) writes: Without the unconscious, the conscious mind would be a chaotic arena, inundated by numerous informative facts, requirements or rearrangements, and the self could not function. That is why the unconscious is also important. It is where the brain stores and uses information and prepares solutions for the signature of the self and where much thinking is done.
Although the conscious mind can theoretically expand indefinitely, in practice, it always finds its limits in the unknown. We cannot be reflexive about unconscious matters (see Archer 2003) because all the workings of the unconscious are beyond our control and cannot be reflected upon or evaluated directly.
All these observations lead to the conclusion that no one can generate a discursive sense of self merely by using the most suitable narratives. We are the creators of ourselves and of our stories in a series of events and circumstances that transcend us, not all of which we are able to control (Athens 1995a(Athens , 2013Harcourt 2006;Presser 2016). In the next part, we suggest how all these processes take place for violent actors.

Interpreting Other People's Attitudes
From our perspective, we discount the idea that there are social and individual variables that necessarily condition one to make deliberations that transcend the individual. Our research shows how it is through reflexivity that we become active agents-that is to say, people who have some degree of power to determine our own lives, to self-assess, and to take personal responsibility. But how does this process occur?
One aspect concerns the way we interpret other people's attitudes. Symbolic interactionism suggests that assuming other people's attitudes leads to knowing the definitions of the "others" and their states of mind-which means imagining the thoughts, feelings, and experiences of other people, adopting their perspectives, and blurring the boundaries between "us" and "them." This blurring may be deceptive, however. Mead's (1934) theory of "meaningful language," which is based on abstractions that people make when assuming the attitudes of others, of which they would be automatically aware in that moment, has been revised by Athens (2007). Indeed, Athens points out that our assumptions of others' attitudes ("role-taking") necessitate that we wait for other people's replies to our words because we can never take on their roles directly. The meaning attributed to a given gesture is ultimately always our own, and the expectations we have in terms of the behavior deriving from it is ours as well. In the following interview, the violent actor (Olmo, convicted of homicide) tries to assume the attitude of his victim-who was someone belonging to an organized crime group-by imagining what his own reactions would have been had he found himself in that position.
Interviewer: Can you tell us what you were telling yourself before you decided to attack your victim physically? How were you interpreting your behavior and theirs? What emotions accompanied these interpretations?
Olmo: I could see myself in the situation because I saw, thought… I had imagined how it could end. Get near… perhaps he, seeing the weapon, would be frightened and not react. The moment there was a physical contact, because he shoved me, I brandished the weapon to frighten him. I thought: "Perhaps he will be frightened and get into the car, go away, and never show himself again, or he might try to reason… or he might understand that what he has done is wrong, because he has not given me what he should have given me." On the contrary, everything went in a completely different direction from what I had imagined. My actions were meant to frighten him and make him understand this: "Pick on someone your own size." I only wanted to frighten him to make him believe that I was also the kind of person who could possibly react in a certain man-ner…. In this brief excerpt, Olmo tries to assume the victim's viewpoint in order to attribute meaning to his (Olmo's) gestures. Olmo, however, is not able to fully assume the other person's role because the meanings of all people's experiences are always personal and unique: it is an adhesion to life-a form of desire and network of resonances which "mean something" only to me (Jedlowski 2008). These are exactly the processes that lead interlocutors, in the most varied contexts, to ask themselves questions such as "What is on his/her mind?" To discover that the other person's answer is radically different from the imagined one, along with not being at ease in that role and feeling "out of character" (see also Sandberg et al. 2019), brings about a displacement that leaves one incredulous about the action committed, even if committed by oneself.

Experiencing the Body and the Role of Emotions
Humans are intelligent machines-creative and capable of adjusting by innovating within a limited register of possible responses. Myers (1986) refers to a peculiar aptitude of the "body-mind complex" to answer the questions we ask ourselves as a result of our sensual and emotional encounters with the external world (see also Henriksen and Bengtsson 2018). This kind of assumption recalls Mead's (1934) theory of the self (Blumer 1969) and the connected notion of "emotional work." More recently, Denzin (1984) has developed a symbolic interactionist theory of emotions which proposes that emotions are "social objects" in the flow of our actions: we recognize them while interpreting them. Besides experiencing emotions physically in a non-reflexive way (such as through an accelerated pulse and sweating), we essentially make them into reflexive experiences (see also Athens 1994). The personal interpretations and the social definitions of those inner physiological states are needed to confer on them an "emotional meaningfulness," or, conversely, to establish their irrelevance.
There are, however, different modes through which emotions can be expressed (see Mills and Kleinman 1985). In some situations, social actors will alter their reflexive activity to enter an emotional state that overpowers them.
Interviewer: How did you perceive your body in the moments before the physical attack?
Zeno: I am totally in the dark… I was not myself that night. No matter who was there, it would have ended in a knife fight anyway. She was there, but it could have been any other person…. The fact is this: that night, I was the one who was not there; I was not myself. I acted but was not in command of my body; I could see what was happening, but it was not me who commanded my body. I could not stop. My body was moved by anger… anger from years and years of putting up with it… anger, frustration… I had reached the limit. I worked, I was kind of the head of the family, I brought home the money… but no one considered me…. I did it because I had to… the others expected it, that I would…. I have never been thanked, praised, or anything…. I have never been given a gift in my life… rather, I have been kicked in the ass for oth-ers… but gifts? Never. The anger had always been there in the corner of my mind… every day, I had to swallow a wrong, day after day… for ten years…. How could I keep swallowing? I should have slapped her [his fiancée, not the victim of the crime], but I did not have the courage… with women I cannot be… that was the only time I have been violent with a woman….
Even in these moments where reflexivity is "partially darkened," socio-relational dimensions and the accompanying social expectations are still present and provide a powerful and convincing driving force for the completion of a violent act, such as the one described above. Frustration from years of perceived injustice or mistreatment causes anger that has been building over time in a corner of the actor's mind, looking for release. Thus, during the flow of intense emotional activities, in which we search for recognition and converse with our "phantom communities," we find the ability to channel our creative force (élan vital) within socially structured cosmologies. These processes do not occur in a vacuum, but rather within the "I-world" relationship.
With this complexity in mind, if, in the course of life, the self-that prism through which we construe ourselves and the external world reflexively (but never altogether transparently)-does not shift dramatically toward a violence-oriented cosmology (see also Athens 1995b), the actor will be able to continue to heed such warnings as: "Do not associate with people like these!" After embarking on what Athens (1992Athens ( , 2003 defines as a "process of violentization," however, and having internalized brutal "phantom others" and having become the custodian of principles that are definitely antithetical to the rules of "civil" society, an individual's inner dialogue might feature such thoughts as: "Cut him to pieces, without pity!" Such a dramatic shift may take place through an education in violence as well, capable of triggering flashes of disgust, intolerance, and resentment toward the "other," who is perceived as "diabolic," "evil," or "inferior." Here, the well-known "techniques of neutralization"-which we can find in some of the passages shown, such as "I did it because I had to… the others expected it, that I would" and "I was not myself. I acted, but I was not in command of my body"-are no longer necessary. Giorgio (convicted of homicide) did not feel the need to justify his actions for a very simple reason. He did not see his actions as wrong; rather, they were morally imperative (see also Colvin and Pisoiu 2020).
Interviewer: What were you thinking while the situation was developing? How did you perceive the victim? What were your feelings?
Giorgio: While he was talking, I was thinking: "See these shitty people who come here in Italy and want to mind their own business… here in Italy… it's not right." I was angry because he wanted to rape that girl. That was what triggered it… but I did not mean to kill him… the heat, the real evil came on later…. By now, everything had started. Something was taking place that I had not reckoned on but can happen, even though it does not mean it has to…. I was trying not to let it happen… because I am not the executioner who goes around killing people… even though many times I put myself in situations when I do judge people…. It's happened right in this case… a judgment was made because he is black. First, he is black… this is a deeply held belief I have about them. They are animals and I am a person. It may seem a contradiction, because you say things about blacks and then you listen to Bob Marley's music… but to be Jamaican is another thing… he [Bob Marley] is not a real black… blacks are animals. I see it also when it comes to Arabs and Moroccans. They make you racist. Even if the latter are white, they suck all the same because of how they behave, because they are not men who take a stand. Then, when you take them one at the time, they prove themselves utter fools. I, on the contrary, with others or on my own… I am always the same… because this is what I am, this is my attitude. These people, on the contrary, when they are a pack… and you start shooting at them, they turn into sheep and run away from even one person. For me, they are not men. They are animals. I am not talking about all blacks, I am talking about these animals here that are happy only in a pack. That's exactly what I hate! Because if I have to cause you mischief, I won't come at you in a pack, I come… as a person, and as such, I come and hurt you… be it evil or good.
It is in such moments that the social feeling of hatred-a relatively stable and deeprooted tendency to perceive others and their gestures as "diabolical"-festers. "Hatred" is often preceded and accompanied by a dynamic of demonizing the enemy and by the compelling desire to affirm one's reasons for domination. Those who cast a hateful glance-as Giorgio did-are unable to see the face of the other because something terrible has already happened to their ability to feel. In this case, "canceling" the enemy means attacking their bodies because one hates the "universal" they represent-because it belongs to a specific ethnic or religious group (De Monticelli, 2003) or a particular criminal organization. When one acts in this way, the aim is to affirm emphatically that set of values that one's own (malignant) corporeal community defines in opposition to an "enemy."

Organizing the Fragments in and of a Shifting Violent Cosmology
As we have suggested, during the interpretation of the other's attitude, the "emotional work" (mentioned above) and the preparation of the violent performance, the actor is engaged in an uninterrupted "inner conversation" that is unrivaled in terms of its reflexivity (see Archer 2003). During this process, the "inner conversation" flows toward a narrative reference pointwhich we call the "cosmology"-around which other internalized "phantom others" gather. 6 In this part, we draw on the concept of "shifting narratives," elaborated by Brookman (2015), to explore the last points we want to address in this contribution-the ambiguous and plural aspects of the narratives of violent actors and the complex interplay between individuals and social structures. It is precisely with respect to the narratives of violent actors that Brookman (2015) asks how violent actors "shift" from one kind of narrative to another over the course of the same interview while they recount their actions. For example, how do they move from a narrative that belongs to the socio-cultural world of violence and that vindicates their active role in the construction of the violent action to one that is linked more to being in the role of a victim-a victim of society, of alcohol or drugs, of the influence of a peer group, of provocation from other people during a social interaction, or of past humiliations suffered (Brookman 2015; see also Colvin and Pisoiu 2020;Winlow and Hall 2009)? And again, how can we make sense of violent acts and of their actors when the meaning of each story and perhaps the very stories themselves are constantly shifting "targets" as the cosmos transforms (Brookman 2015)? Brookman (2015) highlights how violent actors draw from different stories and from competing and potentially conflicting narrative scripts while they describe themselves and their acts, shifting from one narrative to another within the same story. The result is a sort of inter-discursiveness on more than one level that sees the co-existence of narratives that may appear incoherent only to others. These passages are very clear in the narratives of our interviewees. For example, Giorgio presents multiple narratives, which may appear jumbled, but these voices can express themselves in a sanitized, sugar-coated language, cleansed of all disturbing and violent elements; at the same time, they might contain the presence of "thrill discourses" (Katz 1988)-that is, narrative levels linked to excitement, pleasure, strong emotions, and to the sensual seduction of violent crime. These voices are, therefore, carriers of a co-incidence of apparently contradictory poles (see also Presser 2013; Sandberg 2013), but which can be included and accepted within a sufficiently complex cosmology. The recognition of these multitudes of voices that are embodied in the narrative of violent actors thus represents the natural counterpoint of the "radical interactionist" notion of the "phantom community" (Athens 1994)-the internal parliament that represents the principal interlocutor of our soliloquies and is comprised of the significant others we have internalized throughout our life.
With this in mind, we can also better explain the complex relationship between agency and structure (see also Fleetwood 2016;Sandberg 2013). In this regard, it is important to highlight that no social world or corporeal community is the model that informs directly and transparently the "phantom communities" of those who inhabit them. Thus, being "authentic" and finding the key to one's "authenticity" (see also Hochstetler et al. 2010;Kurtz and Upton 2017) does not mean adhering blindly to the modality created by the dominant models in a specific corporal community. One does not become oneself simply by embracing a "phantom community" that may prevail in some social worlds and that may seem absurd in others, but because one judges the "phantom community" to be reflexively consistent with the whole of one's evolving cosmology. From our perspective, the key to explaining the relationship between agency and structure lies in this cosmological relationship. The following exchange with Vincenzo (convicted of homicide) illustrates this theoretical consideration.
Interviewer: How would describe your relationships with the people you frequented in that period, and what about your relationship with drugs?
Vincenzo: A phrase that annoys me is: "They dragged me into it." Nobody drags you into anything. In my opinion in life… modern Italian life… nobody drags anybody into anything. In Sierra Leone, they drag in the children, as soldiers; they are dragged in, they give them cocaine. There, you are eight, they do drag you in… but when you are twenty-five, nobody drags you in.
I have done things in the drug scene that have caused people to stare and say: "That's impossible." A guy told me: "F**k, I don't have the money to get high, will you take this watch?" And I answered: "See, don't give me anything-have my hit"… It was ten thousand liras then; it cost very little in the Eighties. "You party, then, when you have the money, you can pay it back." And the others were like: "You're crazy." And I was like: "No, he wants to shoot up, let him shoot up!" And they were like: "But you have some?" And I was like: "No, I don't. I'll bang tonight or tomorrow!" The others were gobsmacked because heroin, it's like a magnet, it's something that pulls you… even if you wanted to quit it, you shiver, you feel ill, your nose runs, you laugh even at crap. When one wants a trip and needs a fix… he sweats, begs… I have always had a relationship with heroin that made me say: "I have to beat this thing." And in fact, the day it happened, I said: "Now I'll beat you, you have beaten me till now, now I will beat you and won't bang anymore." But this is a conversation between me and an object, not between me and a person, no one made me a heroin addict….
In this interview, in describing his drug use, Vincenzo shows the idiosyncrasy of "whatgoes-without-saying" tropes that might explain his actions in terms of passivity ("A phrase that annoys me is: 'They dragged me into it.' Nobody drags you into anything.") and asserts his own active role (see also Colvin and Pisoiu 2020). His words suggest that even during complex symbolic interactions, such as the ones described, we remain in control of our actions because of that very contact with ourselves-our fluctuating, multiple, and sometimes chaotic selves-that hold together our narrative identity 7 in the relationship between the self and the social worlds of reference.

Conclusion
In this article, we have introduced the concept of "violent cosmologies"-a notion directed at reassigning meaning to violent human behavior beyond any rigid and formal dualism (agency vs. structure; "normality" vs. psychic suffering). A "violent cosmology" may be described as the invisible thread that ties together the multiple complex narratives that promote, accompany and follow a violent act. The crucial point in this representation of violence is the idea of contact between the violent actor, his self, and the social worlds inhabited, which allows the violent actor to direct his actions. As we have suggested, this internal process has a deeply relational nature (Archer 2003: 94). It is the acknowledgment of the qualitative processes and of the ties intertwining the lives of violent actors and their worlds of reference-often forming relationships of "domination" (Athens 2002(Athens , 2015a)that allow us to overcome the belief that "violent people" are individuals whose violent actions indicate the presence of certain pathologies.
Admittedly, many other criminologists do not of conceive violent offenders as "mentally disabled" and are much more likely to focus on the pathology of communities/society in generating violence. Our approach also tries not to treat violent acts as the necessary consequences of a violent social world or of "subcultures" approving of violence. As Athens (2005:645) remarks, "people's social identities, including their genders, are by-products of their ongoing soliloquies with phantom others rather than vice versa." Athens' suggestion that (1) reflexivity actively mediates the socio-cultural and structural contexts within which meaningful experiences emerge (see also Auyero and Berti 2015) and that (2) narratives are rooted not only in socio-cultural worlds structured in terms of "domination" (social structure) but also in "phantom communities" seems to offer new answers to this conundrum.
Our research shows that the states of mind and the emotions that feature in our daily lives (those that we consider "normal") do not change in their essence when they lead to, and occur during, violent acts. Instead, the redefinition of the inner landscape that forms the backdrop of the unfolding of violent acts takes place through the acceleration of violent actors' inner conversations or "soliloquies," to borrow Athens's term, when they interpret the situation into which they are thrown. By "listening to" and "responding to" their inner experiences, they hurl themselves into the immediacy of the act. These inner conversations can be viewed as a "reawakening"-an intermittent return of the individuals to the materials stored in their memory, of which they are capable of deciphering some traits. As Jedlowski (2008: 145-146) puts it, "[t]his reawakening is not a final 'unveiling,' but part of a process of elaboration and re-orientation to which the individual can subject his or her life, thus revisiting its legacies".
Seeking out and describing those psycho-social paths that lead individuals to commit violent acts (such as assault, homicide, and sexual violence) means showing how violent acts can be placed within interpretative itineraries that are accessible and can be narrated, starting from the perspective of those who experienced them and recreating trails of meaning that are in some measure comprehensible and approachable. Very briefly, some of our qualitative research results include: the recognition of a full moral and symbolic complexity present within the narratives of violent actors; an ambiguity of symbols and values between violent and non-violent worlds, and therefore between a "we" and a "they" that are no longer hermetically separable (see also Presser 2013;Sandberg et al. 2019); and profound qualitative differences between diverse violent actors, on the one hand, and different kinds of violence, on the other (see also Athens 1992Athens , 1997Athens , 2017Collins 2008). It is, above all, on this level that the "radical interactionist" perspective we adopted works, complicating the moral (and symbolic) vision we have of violence and sensitizing us to new categories of social realities and experiences.
As described in this article, our view of narratives as a tool for personal meaningmaking and our reliance on interview material do not displace crucial questions about agency, domination, and structure. Moreover, it seems to us that narrative criminological approaches until now have not been able to say much about the sources of narratives and of their constraints other than to connect them to culture, broadly understood; consequently, the biographical uniqueness of these narratives-that is, how they were interpreted, negotiated or contested-is lost (Polletta et al. 2011). The crucial point here is that not all stories are heard the same way as people inhabit different "phantom communities." In this relational universe, stories might be persuasive when told by significant others, almost completely aside from their content. By recognizing agency, as well as constraints, we suggest that violent actors sometimes might be better off with multiple, even inconsistent, narratives that somehow, in their personal cosmology, seem to hang together while pointing in quite different normative directions (see also Polletta et al. 2011).
Our goal in this article has been to develop a dialogue between some of Athens' (1994Athens' ( , 2007 concepts and the ideas proposed by narrative criminology (e.g., Brookman 2015). We have also endeavored to orchestrate a new way to consider the multiplicity of voices that can find expression within a single narrative. 8 Voices are simultaneously personal/subjective, belonging to the social actor interviewed, and "external," meaning that they are borrowed/negotiated from the repertoire of pre-existing cultural stories and discourses in which we are all immersed, albeit in different ways (see also Sandberg et al. 2019).
In addition, we have attempted to contemplate the reflexivity of the criminologist when collecting participants' narratives. As Presser and Sandberg (2019: 139) have highlighted recently: the story that influences actors is almost certainly not the one that observers are in position to "collect." Stories do not stay the same from the time of action to the time of post-action reflection. Narrative criminologists insist that stories somehow precede actions, even though stories are told following action. Narrative criminologists must lay bare this sort of intellectual leap.
In fact, if the interviewees report having had certain experiences at the time of their violent acts, this does not imply that those experiences were the actual "causes" of their behavior. The issue is not about the "true" meanings ascribed to the subjects' behavior. It is clear that this would be a naive approach that does not consider the process of memory in selecting past experiences and the specific context of the interview where these meanings emerge. Because most narrative research is undertaken through interviews, proving that narrative must precede offending is a "logical impossibility" (Fleetwood 2016: 174; see also Brisman 2017Brisman , 2019aBrisman , 2019b. In this regard, our proposal suggests that the internal conversation-often elliptical-which precedes the act is not coincident with-and is not the same as-the story narrated after action. It is in this moment that the role of the criminologist in co-constructing the stories of the participants may be crucial to creating mutable and more open-ended narratives (see also Natali 2019; Natali and Budó 2019; Natali et al. 2020).
By focusing on self-narratives as a guide to (violent) behavior (see Presser 2009) and trying to reassign a meaningful dimension to a phenomenon (that of criminal violence) that has long been investigated by social scientists (but remains shrouded in a veil of incomprehensibility), we suggest that taking an approach that is capable of exploring the stories of violent offenders puts us in a position to discover the deeper roots of people's symbolic and moral worlds. Presser (2009: 178) writes: "Although the use of offenders' 'own' stories has a venerable tradition in criminology, criminologists have not exploited the potential of stories to theorize the etiology of crime" (see also Presser 2016). To adopt this narrative perspective also means to circumvent "the question of whether would-be offenders truly believe their stories or only tell them (to self or others) to enable harmful conduct. … Unlike other explanatory variables in criminology, we need not treat the stories that promote offending behavior as real or true in order to recognize their role" (Presser 2009: 190; see also Presser 2012). The concept of "violent cosmologies" helps us to describe and analyze how violent actors interact with themselves and with their worlds while preparing for and constructing a violent act. 9 The suggestion, here, is to search for, explore, and analyze the polyphonic dimension of the narratives of those who commit violent acts in order to situate them within the wider discursive frame that represents the bridge uniting the self to social worlds. Moreover, it is precisely because we occupy narrative worlds that we can be freed from one narrative by entering into another (Presser and Sandberg 2015). This also applies to those who live within symbolic and ethical worlds that promote the use of violence. Our research reveals that violent actors are always able to restructure their own narratives and can even manage to abandon (at some particularly significant times, which implies a real self-change) those voices that no longer resonate with or give meaning to their actual lives, thereby reorienting themselves to embrace new ones (see also Ishoy and Kruis 2019). From this perspective, if the world is socially constructed through stories, it can also be changed by stories . This is a crucial point for critical, narrative, and "radical interactionist" criminologies when we encounter violent actors, their narratives and their self-change.
Finally, we must remember that not only those who narrate their own violent acts but also those who listen to them (the researcher/criminologist, in our case) are inhabited by a multiplicity of voices that grant more or less bearing to certain narratives rather than others. Different researchers will identify different elements within the narratives of the same violent actors (Brookman 2015), and the outcome of the meeting between an interviewee and an interviewer is always a co-production and a co-construction of knowledge in the context of power relations (Presser 2005;Sandberg 2010). By listening intently to the interviewees' self-narratives, the researcher can hear different voices at the level of reflexivity but also on the "background" level, where the assumptions we take for granted about our everyday life ("tacit knowledge") reside (see Athens 2019). If trained interviewers are perceptive about what they hear interviewees say repeatedly and how they say it, they can document the principles of action and the multiple voices underlying the operation of their "phantom communities." 10 Moreover, there is always a real or an imagined audience-a "phantom community," in Athens's view-to whom the narrators tell their own stories (see also Presser and Sandberg 2019). This is another relevant area of dialogue between a critical and narrative approach-one that dispels the notion of interpretive neutrality-and "radical interactionism." Here, attentive self-reflexivity on the part of the criminologist as a social scientist is key.
In a similar vein, Brookman (2015: 230) writes: "reflexively interrogating one's own influence on the production and interpretation of narrative data, is therefore, an important ongoing consideration for narrative criminologists" (see also Presser 2005). This is why, as researchers who weave the stories of violent actors, we must allow each action and each experience space for ambiguity, accepting and, at the same time, instigating a silent dialogue among the events (Calvino 1988). As Sandberg (2010: 462) argues, "[i]nstead of always searching for consistency and rationality in the stories people tell, we should, sometimes, explore fragmentation and flux in language use." In order to do so, it is necessary to 10 Similar to the relationship between habitus and field (Fleetwood 2016: 181), the relationship between "phantom community" and "physical community"/"social world" may be dissonant. While the notion of habitus is pre-verbal and pre-reflexive, the process of internalization of the hierarchy of domination in the "phantom community" cannot operate without language and reflects the individual's position in social and cultural space. In this vein, we can say that "phantom communities" orient-and do not determine-action. develop methods and create concepts that are sensitive to the complex nature of language, including its weakest and most difficult-to-grasp signals. Recognizing the uniqueness, integrity, and ambiguity (see Sandberg et al. 2015) of the lives of those who have committed violent acts is among the preconditions for possibly transforming entire cosmologies.
Funding Open access funding provided by Università degli Studi di Milano -Bicocca within the CRUI-CARE Agreement.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.