102 Biological Bases of Personality
relationship between the P dimension and dopaminergic reac-
tivity although animal and clinical research would support
such a relationship.
High levels of testosterone and low levels of cortisol have
been associated with disinhibition and psychopathic traits.
But high levels of testosterone have also been associated with
sociability and low levels with neuroticism, as discussed in
previous sections. There is no necessary one-to-one relation-
ship between biological and personality traits. Neurotrans-
mitters like dopamine and hormones like testosterone may be
related to two or more of the basic dimensions of personality
or to a higher order dimensions like approach or inhibition.
Personality in the third dimension shows a high degree of
heritability compared to other major dimensions. A specific
gene, the dopamine receptor D4, has been associated with
the trait of novelty seeking, although replication has been
spotty. The association is supported by animal and clinical
studies. Disorders characterized by impulsivity like opiate
abuse, pathological gambling disorder, and attention-deficit-
hyperactivity disorder share the same form of the gene as
found in high novelty seekers. Mice with the gene removed
show decreases in exploration and responses to novel situa-
tions. The third dimension of personality has been a rich lode
of biological findings from the psychophysiological down to
the most basic genetic level.
AGGRESSION/HOSTILITY/ANGER/
AGREEABLENESS
Problems of definition confuse the fourth dimension of per-
sonality. Aggression refers to behavior, hostility to attitude,
and anger to emotion. One can be aggressive without being
hostile or angry, as in certain kinds of competition; or one can
be chronically hostile and angry without expressing the neg-
ative attitude and feelings in overt aggression. One may be
disagreeable without being aggressive or being aware of hos-
tile attitudes or anger. Hostility without aggression is more
closely associated with the N factor whereas aggression, with
or without hostility, is more closely associated with this
fourth factor.
Another source of differences is in the way aggression is
expressed. Aggression in other species is classified by the
source or context of the aggression: predatory, intermale, fear-
induced, maternal, sex-related, instrumental, territorial, or
merely irritable (Volavka, 1995). Human aggression is more
often characterized by the form of expression. For instance,
the widely used Buss-Durkee (1957) Hostility Scale (BDHS)
classifies aggression as assault, indirect hostility, verbal
hostility, irritability, negativism, resentment, and suspicion.
A new form of the scale has reduced the number of subscales
to four, using factor analyses: physical aggression, verbal ag-
gression, anger, and hostility (Buss & Perry, 1992). Although
the subscales are moderately intercorrelated, quite different
results have been found for the different subscales of the test
in biological studies. Another important distinction in the lit-
erature is whether aggression is impulsive. The impulsive
type of aggression seems more biologically rooted than in-
strumental types of aggression, but this confounds two differ-
ent dimensions of personality.
Although aggression and hostility are correlated in tests
and life, they are separated in two of the major trait classifi-
cation systems. Eysenck’s system includes negative feelings
like anger (moodiness) in neuroticism, but aggression and
hostility are at the core of the psychoticism dimension. Costa
and McCrae (1992a) have angry-hostility as a facet of neu-
roticism but regard aggression as the obverse of agreeable-
ness. My colleagues and I found that hostility and anger load
more highly on N and aggression on P in a three-factor
model, but all three correlate with a common factor in a five-
factor analysis (Zuckerman et al., 1991).
Aggression has been defined by several methods, includ-
ing self-report ratings or questionnaires, observer or ratings
by others, and life-history variables like membership in
groups characterized by violent acts or crimes. Aggression is
not a socially desirable trait and this may limit the usefulness
of self-report methods in some settings. Laboratory observa-
tions may be too specific to the experimental conditions.
Persons who committed a violent crime, like murder, may
differ depending on how characteristic their violent behavior
was before they committed the crime. All methods have
methodological problems, but in spite of this there are certain
consistencies in results across methods in the literature.Cortical Arousal and ArousabilityEarly studies of the EEG in abnormal populations, like vio-
lent criminals, used crude qualitative judgments of the EEG
records as “abnormal” or “normal” (Volavka, 1995). EEG ab-
normalities included diffused or focal slowing, spiking or
sharp waves in certain areas, and generalized paroxysmal
features resembling minor epileptic seizures. The incidence
of abnormal records found in samples of prisoners convicted
of homicides and habitually violent prisoners was quite high
(50–65%) compared to nonviolent prisoners or normal con-
trols (about 5–10%; Volavka, 1995). However, some other
studies found no differences between violent and nonviolent
offenders.
Studies using quantitative methods showed EEG slowing
in offenders, including slowing of alpha activity and an