This is going to be long...
First, let me address the question of definition of "evil" and why psychiatrists and psychologist don't like to use this word. The word "evil" can, as has already been pointed out in this thread, be interpreted in different ways. It is a word with a lot of religious, moral and cultural connotations. Thus, like most laymen terms, it is unsuitable for science and even more for diagnostics that determine the fate of people's lives, because it lacks an exact operational definition that can is measurable and quantifiable.
However, there are many human behaviours that are destructive and harmful for others. Instrumental narcissism, the idea that you have superior rights which means you can use and abuse other people as much as it please you and this is just your right, is one such behaviour pattern. Another is sadism, that you take pleasure out of other people's suffering and even induce suffering in them for your own well-being.
Genocide politics of the kind we saw in nazi-Germany certainly belongs to the narcissism-spectrum. You hold one race or one group of people superior, and think it's right to get rid of others. On a smaller scale, you can see it in many ordinary people (for instance on this board) who believe they have superior rights because they come from a certain country or belong to a certain culture. Narcissism is inself only a elaboration of the proximity-principle: people care most about what is closest and most immediate to them. Narcissism, like many other behaviour patterns, can not be classified into absent = healthy, present = disorder. The same with sadism, sadism can be seen in everyday behaviour, people take pleasure and gloat at other's misery, especially when they feel they want to revenge something. Few however would take pleasure in kidnapping another person and torture the person for days before eating the persons body parts, to take an extreme example.
In any case, extreme acts based on instrumental narcissism and sadism, belong to the group of behaviours that most people would agree on labelling as "evil". On the other hand, bisarre delusions and visual hallucinations belong to a group of behaviours that most people would agree on labelling as "mad", but that doesn't mean "mad" is a useful concept in diagnostics or in psychiatric research. When Michael Stone claims that it's a good idea to re-introduce the word "evil" in psychiatry, it is to me an analogy to re-introducing the word "mad" or "idiot". My question is: what for? Why strive for using loaded laymen terms with moral, religious and stigmatising connotations that makes them open for all sorts of subjective opinion? Why not develop the diagnosic system into deeper detail? It's increased detail knowledge about causes and mechanisms we need for development of better treatment (or any treatment) and for risk-assessment of criminal offenders, not more words that are vague and sweeping generalisations.
Neurological, neuropsychiatric and psychiatric disorders have an observation-based classification where we look at symptomatology and functional behaviour. It's not like infectious diseases; there is no schizophrenia virus or ADHD-bacillus. We are working with the body's most complex organ and the complexity of human behaviour. Diagnosis are made base on a pattern of set symptom criteria, and subjective labels like "evil" or "good" are irrelevant here. There are however many other problems here, but I need to explain a bit about the diagnosis system in order to demonstrate what the problems are.
Psychatric disorder is diagnosed based on a system which is described in the manual DSM-IV. This system was developed in order to release psychology and psychiatry from the various opinions and ideologies that previously influenced assessment of patients (ie psychoanalysis and anti-psychiatry back in the 1960's). The system is divided into 5 axis, including symptomatology and level of functioning, but I will only go into the first 2 axis. The full DSM-IV is only available for professionals, but those who wish to read more, can find some information here:
http://allpsych.com/disorders/dsm.html
Axis I covers the clinical syndroms, which includes the well-known disorders
schizophrenia, bipolar, depression, phobias. There are no major problems with these diagnosis. Then, we come to the problem child of psychiatry, axis II, the personality disorders. To make a long story short, the personality disorder system was constructed as a consensus compromise between several different schools of psychiatry. The basis for this bad solution was simply that much less in known about personality disorders than of psychosis, affective disorder and anxiety disorder. We are not even sure of how to conceptualise what is currently called personality disorder. Current personality disorder diagnosis overlap each other a lot, and also overlap with normal personality traits a lot.
Now, with this background we can look at the diagnosis
anti-social personality disorder (ASPD), an axis II diagnosis that has replaced the older psychopath and sociopath labels, and is associated with criminal behaviour. The criteria for ASPD are a horrible mix of behaviours and symptoms that overlap partly with the different psychopath concepts (yes, there are more than one), developmental disorders ADHD and autism, conduct disorder, narcissistic personality disorder, borderline personality disorder, and axis I dissociative syndroms. It's actually quite worthless, and it's not connected to "evil" either.
Although brief and simplified, I think this little article illustrates the lack of connection between "evil" and psychopathy as well:
http://www.sfgate.com/cgi-bin/article.c ... B9A8D1.DTL
If anyone is specially interested in reading more about the concepts of anti-social PD, psychopathy and criminal behaviour, I have a good essay written by a collaborator of mine who is a forensic psychiatrist. It's part of a research program, but it contains some useful information for the interested layman, so if you wish to know more, PM me.
A behaviour that can be labelled as "evil" have a multifactorial background that differs between individuals, but also, the labelling will differ depending on context and observer. I understand the frustration from forensics people and laymen alike, when they are looking for a development that would help in risk-assessment and treatment and thus reduce "evil" acts and protect people from it. However, starting to use the term "evil" will not assist us in this task even though Stone claims so. There is no support for the existence of a concept of "evil" as an individual trait or behaviour pattern. People committ "evil" acts for all sorts of reason, in some areas and not others, at some points in time and not others, during some conditions and not others. It is simply a useless concept for any sort of serious take at psychiatric disorders, or human behaviour in general.
Second, it is important to note that the concept of psychiatric disorder is completely different from the legal question of whether a person can be held responsible for their acts or not. Most people with a neuropsychiatric disorder never committ any crimes, and lots of people with no neuropsychiatric disorder committ crimes. These are two disparate questions. Then, from a health perspective (which is not unimportant), it can be discussed whether a person perhaps must suffer from a psychiatric disorder in order to develop such a deviating and destructive behaviour like for instance, sadistic murder like I described above. The health issue is however not relevant from a legal perspective, but it's highly relevant for the question of
treatment.