Commentary: Parricides—Unanswered Questions, Methodological Obstacles, and Legal Considerations

Most parricides fall into one of two categories. Adolescent parricides tend to be cataclysmic reactions to enduring, severe physical abuse, perpetrated by an individual who is typically neither conduct disordered nor psychotic. Adult parricides tend to be tragic conclusions of highly conflictual relationships between untreated psychotic individuals and their parents. The many questions about the killing of one's parent go unanswered in the modest body of scientific literature on parricides. The existing studies are limited by several methodological problems. Herein, we describe the problems, propose remedies, and review the legal ramifications of parricide.

Parricides are fortunately rare offenses, estimated to make up 1 to 4 percent of all homicides^1,2 and 20 to 30 percent of homicides committed by psychotically ill individuals.² This low prevalence makes it difficult to investigate a sample with a size and lack of bias that allows for generalization of the findings.³ Bourget et al.⁴ deserve our gratitude for describing one of the largest samples of parricides to date. Their report stresses the role played by psychosis, nonadherence to treatment, and the lack of prodromal signs. It also shows the challenges that parricide studies pose. Most studies describe a small number of cases, constituting samples of convenience where case selection is not systematic. One strength of their work is that they studied consecutive cases.

In light of the low prevalence of parricides, how can investigators access a sample of adequate size? One strategy is to combine several data sets using the same data collection method. Weisman and colleagues⁵ used this technique to study the particularly rare phenomenon of double parricide (the killing of both parents). Combining cases from multiple sites by using a common method of data extraction, they generated the largest extant data set on double parricides. Another strategy, increasingly popular in many fields,⁶ is to design qualitative studies that emphasize phenomenology. The depth of analysis seen in such studies compensates for their limited generalizability. Ideal participants in such studies are individuals who have killed a parent, have been adjudicated, have been successfully treated, and are sufficiently improved to describe cogently their experience of parricide.

Another research strategy consists of widening the scope of inquiry from parricides to the entire continuum that lies between completed parricides and nonlethal acts of child‐on‐parent violence.^7,8 One argument in favor of this strategy is the following: Whether a parent survives an attack by his or her child depends on many factors, including the quality of emergency services in that community (e.g., promptness of responses to 9‐1‐1 calls). Distinguishing completed from attempted parricides may thus be arbitrary. Our research group described individuals found not guilty by reason of insanity (NGRI) of attempted murder of a parent as presenting a very similar clinical picture and the same challenges in psychotherapeutic treatment as individuals found NGRI after murdering a parent.⁹ Marleau et al.² also found attempted and completed parricides to be quite similar, whereas Weisman and Sharma¹⁰ found significant differences. Investigating the entire continuum of child‐on‐parent violence is likely to reveal qualitative differences among different types of violent events: those in which potentially lethal means were used with the intent to kill (e.g., stabbing), events involving potentially lethal means without the intent to kill (e.g., slashing), and events involving less dangerous means and no intent to kill (e.g., beating). Future empirical investigations will shed light on the question of which segments of this continuum are sufficiently similar to warrant the same treatment, both from the judicial system and from the mental health system.

The subject of victim gender and offender gender illustrates how generalization from any one study of parricide (or any other rare phenomenon) is limited. Parricide is predominantly a male‐on‐male (son‐on‐father) crime, though this predominance has faded in recent decades.¹¹ In the extant literature, male patricides outnumber male matricides about 2:1 (a ratio similar to Bourget's 3:2 ratio), and male parricides outnumber female parricides by about 5:1¹ (a ratio considerably lower than the 15:1 ratio in the data of Bourget et al.⁴). Though it is always true that data aggregated across studies generate better estimates of population parameters than individual studies permit, it is especially true in fields of inquiry in which typical sample size is small.

What are the main questions about parricide that remain unanswered? We know that most children do not kill their abusive parents and that most individuals with a psychotic illness do not kill their parents. It is the particular nexus of biopsychosocial factors connecting abuse and parricide or psychosis and parricide that remains to be established. In other words, does everyone who has a psychosis and has conflicts with parents pose at least a slight risk of parricide? Does every abused child pose at least a slight risk of parricide? Bourget and colleagues⁴ point to an important feature of this nexus: treatment nonadherence. Our research group has also found nonadherence to play a crucial role in parricides.^12,13

In contrast to the study by Bourget et al.,⁴ we found “warning signs,” such as excessive risk‐taking by the parent. For example, the parent of a psychotically ill individual who refused treatment invited his child to live with him despite their extremely contentious relationship, which involved threat of harm to the parent and the child's insistence on possessing a gun. Beyond the fact that identifying “warning signs” involves a powerful retrospective bias, the difference between the findings of Bourget et al.⁴ and ours illustrate once again the limitations of small samples.

Another unanswered question concerns the relationship between adult parricide and prior abuse of the parricidal offender at the hands of the victim. There is a strong consensus in the literature that child and adolescent parricides typically follow lengthy, severe abuse by the parents.^1,14–17 The prevalence of prior abuse among adult parricides is unknown. The lack of data is all the more amazing in that several investigators have described the offender‐victim relationship and have used terms such as “disturbed rearing patterns,”³ “hostile and dependent‐aggressive” relationship,¹⁷ or “cruelty.”¹⁸ It is as if they stopped short of perceiving the behavior as abuse, maybe for fear of blaming the victim. It would be important to know how commonly abuse precedes adult parricides. If abuse were found to be a common precursor of adult parricides, it would be possible to educate patients, their families, mental heath providers, emergency department staff, and others about this risk factor, with the goal of prevention.

Parricides raise legal questions with regard to criminal responsibility. In parricide cases, the facts tend to be more emotionally salient, and it is conceivable that a judge or jury might be persuaded to arrive at a more drastic outcome than in another homicide. In the event of an insanity defense for parricide, a judge or jury may be more likely to be persuaded of a defendant's mental incapacity because of the relative inconceivability of the crime.

Courts seek to determine an offender's level of intention to kill or harm the victim or victims. For example, the courts try to determine whether the offender willingly and knowingly killed the victim or victims, whether the offender should have known the risks inherent in his or her actions, or whether the offender failed to take reasonable care in his or her actions. Each of these levels of intent demands a different level of punishment.

Parricide offenders may try to raise legal defenses relating to mental state at the time of the crime in an effort to show that they lacked mental capacity. Two possibilities suggested by Bourget et al.⁴ are mental incapacity in the form of psychosis and self‐defense. In many jurisdictions, insanity may serve as a legal defense if it can be established that the offender, at the time of the crime, could not appreciate the wrongfulness of his or her actions or was unable to control his or her actions due to a mental illness or defect.^19,20 Self‐defense is also a legal defense to murder. To plead self‐defense, a perpetrator in many jurisdictions must establish that he or she had a reasonable perception of imminent harm to self or others, that the use of force was necessary to avoid the danger, and that the force used in self‐defense was justified by the degree of threatened harm. Although not widely recognized in the legal community, academic arguments have been made for expanding self‐defense theory, using what has been labeled battered‐child syndrome as a defense to parricide.²¹ The theory of battered‐child syndrome is modeled after battered‐woman syndrome and suggests that in the case of the severely abused child, parricide is an act of desperation, as the child sees the death of the abusive parent as the only way out of an intolerable situation (even if the abuse is not occurring at the time of the offense). It has been suggested that current definitions of self‐defense are too narrow and should be expanded to include battered‐child syndrome as a legal defense to parricide.²²

An additional consideration with regard to mental state legal defenses is posttraumatic stress disorder (PTSD). A defendant may assert that the crime occurred in the midst of a PTSD‐related dissociative state, or “flashback,” and thus claim insanity (as reflected in an impaired ability to appreciate the nature of his or her actions with regard to the law at the time of the offense). This defense is more likely to be persuasive in the event of a documented prior history of flashbacks, particularly under circumstances mirroring those preceding the crime.²² Specifically, three causal connections must be supported: between the traumatic event and development of PTSD symptoms, between PTSD symptoms and the offense, and between the traumatic event and the offense.²³ While PTSD‐based insanity defenses are not particularly successful, in the case of abuse‐motivated parricides, a PTSD defense appears to be especially conceivable when a perpetrator with a diagnosis of PTSD murders a parent in the home where the abuse occurred. Overall, PTSD is most often raised during sentencing as a mitigating factor.²⁴

Bourget et al.⁴ suggest that three of the parricides in their study may have arisen out of compassion for the victims. Euthanasia, the intentional killing of a person for his or her alleged benefit, is not a legal defense to murder. The issue of physician‐assisted suicide continues to be hotly debated. In 2006, the United States Supreme Court upheld Oregon's physician‐assisted suicide law for terminally ill patients.²⁵ Oregon is the only state that allows physicians to provide information, guidance, and the means to take one's own life, with the intention that the suicide will be carried out. This ruling applies in very narrow circumstances that do not include mercy killing by family members. One reason for the illegality of “compassionate” killings is the social concern that ulterior motives (e.g., financial gain) may play a role in such homicides. In contrast, depending on the facts of the case, if it can be proven that the victim was elderly, infirm, and wanted to die, these circumstances may serve as mitigating factors at sentencing.

In their study, Bourget et al.⁴ found intoxication to be the third leading cause of parricide. As with euthanasia, voluntary intoxication is not a legal defense to murder, nor can it be used as a mitigating factor when considering the appropriate level of punishment. Public policy does not support the use of voluntary intoxication as a legal defense to one's actions or as a mitigating factor at sentencing, as it would suggest that willfully losing one's self‐control is socially acceptable. Nonetheless, some jurisdictions may recognize substance abuse, secondary to a PTSD diagnosis, as a mitigating factor when the offender is perceived as self‐medicating PTSD‐based emotional distress.

Methodological improvements in child‐on‐parent violence research is likely to enrich our understanding of this phenomenon. Such improvements will enhance our ability to address more effectively the prevention, treatment, and judicial disposition of cases involving parricide or attempted parricide. Of particular importance is the matter of how to treat parricidal offenders. What contributes to their recovery? What level of supervision do they typically require? What are common postcrime developmental trajectories? The answers to these questions will allow us to improve the treatment of those who commit this unusual form of killing.

• American Academy of Psychiatry and the Law