Crowd Psychology in South African Murder Trials

South African courts have recently accepted social psychological phenomena as extenuating factors in murder trials. In one important case, eight railway workers were convicted of murdering four strike breakers during an industrial dispute. The court accepted coriformity, obedience, group polarization, deindividuation, bystander apathy, and other well-established psychological phenomena as extenuating factors for four of the eight defendants, but sentenced the others to death. In a second trial, death sentences o/five defendants for the "necklace" killing of a young woman were reduced to 20 months imprisonment in the light of similar social psychological evidence. Practical and ethical issues arising from expert psychological testimony are discussed.


Crowd Psychology in South African Murder Trials
Andrew M. Colman South African courts have recently accepted social psychological phenomena as extenuating factors in murder trials.In one important case, eight railway workers were convicted of murdering four strike breakers during an industrial dispute.The court accepted coriformity, obedience, group polarization, deindividuation, bystander apathy, and other well-established psychological phenomena as extenuating factors for four of the eight defendants, but sentenced the others to death.In a second trial, death sentences o/five defendants for the "necklace" killing of a young woman were reduced to 20 months imprisonment in the light of similar social psychological evidence.Practical and ethical issues arising from expert psychological testimony are discussed.
I recently testified as an expert witness in two South African murder trials that contributed to an important legal breakthrough and a significant development in the history of applied social psychology.In both trials the courts accepted social psychological phenomena such as conformity, obedience to authority, group polarization, deindividuation, frustration-aggression, relative deprivation, and bystander apathy as extenuating circumstances, enabling convicted murderers to escape the death penalty.Although these decisions are not binding on courts outside South Africa, they indicate a growing influence of psychology in the international legal arena and are of potential interest to psychologists and lawyers in all jurisdictions.

Capital Punishment in South Africa
Until very recently, the death sentence was mandatory for murder in South Africa, except in cases in which the accused succeeded in proving, on the balance of probabilities, the existence of extenuating circumstances.The concept of extenuation was introduced into South African law in 1935 to bolster public confidence in the legal system by shifting the discretion to exercise clemency from the executive to the jUdiciary.As General Smuts, who was then minister of justice, explained to Parliament: Only between 11 per cent and 12 per cent of our death sentences are actually carried out, and all the rest are reprieved.. . .Well, in a country like this that position is most unsatisfactory.The people sentenced for murder are mostly illiterate natives, and the present situation leads to a very unwholesome state of affairs [in which] the impression is gathered that the death sentence and all this frightening procedure of the white man's justice is so much bluff. . . .That sort of thing undermines all respect of justice. . . .The judge should not be compelled, in every murder case, to pronounce sentence of death; he should be The South African legislature did not define extenuating circumstances, nor did it place any limit on the factors that might be deemed to be extenuating.In principle, anything that tended to reduce the moral blameworthiness of a murderer's actions might count as an extenuating circumstance.In practice, the courts most often accepted as extenuating such factors as provocation, intoxication, youth, absence of premeditation, and duress (short of irresistible compulsion, which would exonerate the defendant entirely).The question of extenuation arises only after a defendant has been convicted of murder, which in South Africa, as in the United States, Britain, and elsewhere, requires the prosecution to have proved beyond reasonable doubt not only that the defendant committed the unlawful act (actus reus) but also that this conduct was accompanied by a prescribed mental state (mens rea) entailing criminal responsibility.
The mounting unrest in the Black townships of South Africa from the 1976 Soweto uprising onward led to numerous incidents of mob violence associated with industrial disputes, political demonstrations, consumer boycotts, and funerals of residents killed by the police.In several cases the violence escalated to the point at which murders were committed by rampaging mobs.The individuals eventually charged with these murders were typically only a small minority of those present at the time of the killing and were often not the ones who dealt the fatal blows.Under the legal doctrine of common purpose, individuals were often convicted of murder even if they took no part in the actual killing, merely by virtue of having associated themselves actively with the mob and its murderous goal: The acts of the mob were imputed to each individual member who was proved to have shared a common purpose with the mob.
The increase in township violence during the late 1970s and 1980s led to a huge increase in the number of executions in South Africa.From a base of about 40 executions per annum in the early 1970s, the numbers rose steadily, exceeding 100 per annum by the late 1970s and reaching a peak of 164 in 1987, a year in which 25 exe-cutions took place in the whole of the United States, a country whose population is nine times the size of South Africa's (Allen-Mills, 1988).There has never been a Black judge in South Africa, and trial by jury was abolished decades ago, but the vast majority of people executed for murder have been Black.The most recent figures regarding racial disparities in punishment are for the period from July 1982 to June 1983.During that fairly typical period, 47% (38 out of 81) of the Black people convicted of murdering White people, but none of the 21 White people convicted of murdering Black people, were subsequently hanged (Carlin, 1989).

Extenuating Circumstances
Most extenuation (penalty-phase) proceedings following convictions for murder are perfunctory affairs in which no fresh evidence of significance is introduced.In only 10% of capital cases heard in the Cape Supreme Court in 1988, for example, was expert evidence led in extenuation (Davis, 1989).In the late 1980s, however, in murder cases arising from mob violence, defense lawyers in South Africa began to lead expert psychological evidence in an attempt to introduce an entirely new class of extenuating circumstances, namely, social psychological processes operating at the time of the killings.The courts in South Africa adopted a liberal attitude toward the admission of such evidence, and expert witnesses were generally permitted to express opinions even about the intentions and motives of specific defendants at the time of the alleged offenses, although, as in the United States and elsewhere, the admissibility of expert evidence in the liability phase of a criminal trial is governed by strict common-law principles.
In 5. v. Motaung andOthers (1987/1990), members of a mob were convicted of murdering an alleged police informer at an emotionally and politically charged funeral.On the basis of the defendants' accounts of the incident and of their participation in it, Edward Diener of the University of Illinois testified that at the time of the killing the accused were experiencing severe deindividuation-Ioss of one's sense of individuality and personal accountability that can sometimes occur in large, noisy, emotional crowds (Dipboye, 1977;Festinger, Pepitone, & Newcomb, 1952;Zimbardo, 1969).According to Diener's testimony, it was quite plausible that the defendants were so deindividuated that they did not appreciate that death could ensue from their actions.The court rejected this evidence concerning the ultimate issue of mens rea on the ground that it rested on the veracity of the accused's own testimony, which the court disbelieved.The Court of Appeal upheld this rejection in 1990 but found Diener's evidence potentially "relevant and helpful" (5.v. Motaung and Others, p. 526) on the issue of penalty, in which less strict rules of evidence apply.In the event, however, the original sentences were set aside on an entirely unrelated point of law.
In the notorious case of the Sharpeville Six (5.v. SaJatsa and Others, 1988), the accused were convicted of the mob murder of a town councillor who was widely regarded as a collaborator with apartheid, and deindividuation was again put forward as an extenuating factor.Graham Tyson, a South African psychologist who was called as an expert witness, described the phenomenon of deindividuation and concluded as follows: I consider, on the basis of my assessment of the psychological literature, that it is highly probable that an individual in a mob situation will experience deindividuation and that this deindividuation will lead to diminished responsibility in much the same way as do the consumption of too much alcohol or great emotional stress.(cited in Skeen, 1989, p. 79) The court found that Tyson's evidence lacked probative value because he had testified in general terms, without addressing himself directly to the motivation or state of mind of each individual accused.Therefore, the case for extenuation was found to have not been proven, and the defendants were sentenced to death.This decision was subsequently upheld by the appeal court but, as in the earlier trial, neither the supreme court nor the appeal court ruled out the possibility that deindividuation, properly proved to have applied to the mind of an accused at the time of an offense, could constitute an extenuating circumstance.
The subsequent case of 5. v. Thabetha and Others (1988) arose from events that took place at the funeral of a popular civic leader who had recently been murdered.A crowd of more than 1,000 people, who had been singing, dancing, and listening to emotional speeches throughout the night, was attacked by a group of vigilantes.About 100 mourners, including the 6 accused, chased and killed one of the vigilantes, and the entire incident was recorded on videotape by a journalist, from whom the tape was later confiscated by the police and used in evidence.Graham Tyson, who on this occasion had studied the videotapes of the incident and interviewed each of the defendants, testified that in his opinion it was highly probable that all of them were deindividuated at the time of the offense.The court found that the defendants had indeed been deindividuated and that this amounted to an extenuating factor, and they were all given custodial sentences.

The SARHWU Trial
The extenuation proceedings in 5. v. Sibisi and Others (1989), a trial of eight members of the South African Railways and Harbours Workers' Union (SARHWU), which ran from the beginning of August until the end of November 1988, is believed by the lawyers who were involved in it to be the longest penalty-phase trial on record in any jurisdiction.

Background to the Killings
The events leading up to the killings were described by several witnesses during the trial, and the most important details were included in a set of "agreed facts" formally accepted by the prosecution and the defense after the liability phase of the trial.The murders were committed on April 28, 1987, during a bitter strike of railway workers October 1991 • American Psychologist and other employees of South African Transport Services (SATS), the largest public sector employer in the country.A Black van driver, Andrew Nedzamba, had been sacked for being dilatory in handing in a very small sum of money that he had collected during a delivery.Members of SARHWU came out in sympathy, demanding among other things the unconditional reinstatement of Nedzamba, improvements in the grievance procedures in SATS, and the elimination of numerous racially discriminatory employment practices involving rights and privileges reserved for White employees only.As the strike gained momentum, management steadfastly refused to negotiate with representatives ofSARHWU, which is affiliated with the left-wing Congress of South African Trade Unions (COSATU).They decided instead to lock the strikers out ofSATS premises.Many of the strikers who were migrant contract workers from the rural "homelands" living without their families in single-sex SATS hostels, in conditions reminiscent (as one press reporter pointed out) of concentration camps, found themselves without roofs over their heads.
Several thousand strikers decamped to Cosatu House, a trade union building in the center of Johannesburg.One of the men later convicted of murder with extenuating circumstances, Bongani Sibisi, acted as chairman in the main hall where throngs of strikers discussed the progress of the strike between spells of singing and dancing.Tension mounted when the police began to spray tear gas into Cosatu House and to attack strikers in the streets outside with sjamboks (rawhide whips).On April 22, 1987, police shot dead three strikers at COSATU's Germiston office.When news of these killings reached Cosatu House in Johannesburg, a group of strikers decided to go to Germiston to see for themselves what had happened.While they were walking toward the local Doornfontein railway station, police opened fire and killed three of them.The police then stormed Cosatu House, claiming (according to press reports) that there were "trained terrorists" in the building, and many more strikers were injured.
A few days later the strikers learned that they had all been fired by SATS.The group in Cosatu House, which was by now a frustrated and angry mob, decided unanimously to kill five workers who had refused to join the strike and had been kidnapped from their workplaces and brought there.Many people helped bundle the nonstrikers into a pickup truck and drive them to a nearby woodland area.On the way to the scene of the crime, a striker who was later found guilty of murder with extenuating circumstances had a change of heart and helped one of the intended victims to escape, although his assistance was disputed by the prosecution.The lynching party killed the other four by stabbing them with a long bread knife and dropping a massive concrete block on to their heads as they lay on the ground.One of the men who was later sentenced to death then doused the bodies with petrol and set them on fire.Ironically, it is unlikely that anyone would have been convicted of murder had the fifth intended victim not been allowed to escape.He went October 1991 • American Psychologist straight to the police and was the star prosecution witness at the subsequent murder trial.
Only four of the eight defendants in the murder trial participated directly in the killings.The prosecution conceded that Bongani Sibisi, the chairman in Cosatu House, was not even present at the scene of the crime.But all eight associated themselves with the unanimous decision of the vast mob in Cosatu House to kill the nonstrikers, and they were therefore found guilty of murder by virtue of common purpose.

Psychological Testimony
Two social psychologists testified for the defense: Scott Fraser, from the United States, and I, from Britain.Our evidence was supplemented by the testimony of two South African experts: Hoet Kotze, a social anthropologist, who testified about the "collective consciousness" that is characteristic of traditional African cultures and that may have made the defendants especially vulnerable to group pressures, and Lloyd Vogelman, a clinical psychologist, who outlined the social background and general personality characteristics of each defendant.After examining the facts of the case and the circumstances surrounding the killings, Fraser and I drew the court's attention to several phenomena, all firmly established by basic research in social psychology, that we believed influenced the eight defendants to varying degrees.The evidence given by the defendants themselves, together with the statement of "agreed facts" drawn up by the prosecution and the defense, provided the facts on which we based our inferences about these psychological phenomena and their probable effects on each of the defendants.The dense crowding in Cosatu House and the continuous singing and dancing that took place there appeared to have caused some of the defendants to become deindividuated and therefore less aware than they normally were of their individual identity and accountability (Dipboye, 1977;Zimbardo, 1969).By suppressing self-monitoring, deindividuation may make people especially vulnerable to external, situational pressures, which are, in any event, far more powerful than they appear to be.Experiments have shown that people generally underestimate the importance of such external pressures and overestimate the importance of internal motives and dispositions in interpreting the behavior of others.We explained this wellestablished bias, called the fundamental attribution error (Ross, 1977), and warned the court to be wary of it.
We presented evidence of conformity and obedience pressures (Asch, 1956;Colman, 1987;Milgram, 1974;Tanford & Penrod, 1984) operating in Cosatu House and at the scene of the crime.We discussed the relative deprivation (Crosby, 1976;Gurr, 1970;Masters & Smith, 1987) that the defendants seem to have experienced when they compared their wages, working conditions, and general quality of life with those of their White co-workers.Most of the defendants were living in SATS hostels, for example, and were not even allowed to have their wives on the premises as visitors.We cited evidence of extreme frustration among some of the defendants and pointed out that frustration, together with relative deprivation, has been shown to generate anger and aggression (Berkowitz, 1989;Masters & Smith, 1987;Olson, Herman, & Zanna, 1986).We explained the group polarization effect (Isenberg, 1986) that causes collective decisions, such as the mob decision to kill the nonstrikers, to tend toward greater extremity than do the individual opinions of the group members.We discussed the research evidence on bystander apathy (Latane & Darley, 1970;Latane & Naida, 1981) that helps to explain why some of the defendants stood idly by and allowed others to kill the nonstrikers.Finally, we suggested that Bongani Sibisi, the chairman in Cosatu House, showed the classic symptoms of learned helplessness-a passive, withdrawn condition resulting from exposure to repeated, inescapable aversive experiences (Kofta & Sedek, 1989;Seligman, 1975).
Asked by counsel for the defense lo sum up for the court all the psychological evidence, I pointed to A number of situational forces which we believe were operating in the lead-up to the killings ....There is a strong temptation, even when we know that external, situational forces were partly responsible for people's behaviour, to vastly underestimate their potency and to interpret the behaviour as being caused by internal, dispositional factors.Now this illusion is a very powerful one, as I showed in my earlier evidence, and it is very well founded in psychological research, and it is called the fundamental attribution error. . . .It requires a deliberate effort of will to avoid committing the fundamental attribution error when one is dealing with behaviour that is subject to external or situational influences.. . .I believe that the court, in judging the behaviour of the accused, ought also to be mindful of the fundamental attribution error, which might lead it to underestimate the potential power of the situational forces.(s.v. Sibisi and Others, 1989, court record, pp. 2158-2159) I went on to summarize the evidence already given about deindividuation, frustration and relative deprivation, group polarization, conformity, obedience to authority, bystander apathy, and learned helplessness.I told the court: I have no doubt in my own mind that these forces were very powerful in every case and that they go a long way towards explaining why the accused behaved in what, for all of them I think, was a wholly uncharacteristic manner. . . .Although none of these situational forces is irresistible, and that much is clear from the scientific evidence, their combined effect was in all cases so powerful, given the most unusual confluence of circumstances in Cosatu House, that it would have taken unusual personal qualities, I believe, to have resisted them altogether.(pp.2159-2160)I also drew attention to numerous examples in the evidence given by the accused of menacing threats and implied threats that must have made dissent not only very difficult but also probably dangerous.
My own evidence stretched over four days, two of them devoted to hostile cross examination.Most of the cross examination consisted of attempts by the prosecution to undermine the relevance of the psychological evidence to the specific facts of the case and to show that 1074 some ofthe defendants were leaders rather than followers.Part of the cross examination, however, focused on the psychological evidence itself.Dealing as I was with psychological phenomena and associated literature with which I was fairly familiar and findings that are firmly established by empirical research and in many cases well understood theoretically, I felt slightly more confident during this part of the cross examination.The following typical extract from the cross examination shows how the prosecutor sought to undermine the external and internal validity of the Milgram experiment by questioning its realism and the effectiveness of the deception involved (the prosecutor's questions are in italics): Concerning the Milgram experiment, here the subject was urged on by a confederate to administer shock, is that not so?-Yes,Is it an obvious possibility then that the subject thinks that the confederate knows what he is doing?-Well., . the confederate clearly does know what he is doing, I mean he is running the experiment, yes.Furthermore, the subject knows it is an experiment?-The subject knows it is an experiment, yes.And it is also true then that he knows he will not be prosecuted for his actions, [so] obviously he must think he will not be prosecuted, there will not be any criminalliability?-Well,you say it is obvious; it is not obvious to me. . . .I think it is very ambiguous.I do not know whether the subjects thought that they were criminally responsible or not.That could also be indicative of the fact that the subject does not believe that the person is being tortured?-Well, the subjects did, in fact, believe this.I mean, what happened after the publication of the original series of Milgram experiments was that an enormous debate took place in the psychological literature, and one of the questions that was raised was whether or not the subjects really believed, and in reply to this Milgram sent out questionnaires to all the subjects who had taken part in the experiment and the figure, the result that I happen to remember is 85 per cent of the people who took part. . .either completely believed or believed that they had been delivering painful electric shocks. . . .And then other experiments were carried out in order to put this beyond doubt, experiments in which the credibility issue does not exist.(s.v. Sibisi and Others, 1989, court record, pp. 2345-2348) I went on to describe an experiment performed by Kudirka (1966), in which the subjects were instructed to engage in the extremely disagreeable task of eating 36 quinine-soaked crackers.I pointed out that none of the arguments about the subjects failing to believe that the victim was suffering applied to Kudirka's experiment because the subjects served as their own victims and showed unmistakable signs of suffering, yet virtually all of the subjects ate all 36 quinine-soaked crackers when the experimenter was present, and 14 out of 19 (74%) were fully compliant in the absence of the experimenter.I expressed the view that most social psychologists now accepted that the high levels of obedience reported by researchers cannot be explained away as artifacts resulting from failed attempts to deceive the subjects.Nothing in my cross examination caused me to retract any of my earlier evidence about the psychological factors that helped to explain the killings.

Outcome
The court understood and accepted most of the psychological evidence: Broadly speaking, we find the phenomena [referred to] by the experts as having probably been present.We also accept that the accused were influenced to varying degrees by these factors. . . .In principle we find that none of the accused with whom we are now concerned was left completely unaffected by one or more of these influences.(s.v. Sibisi and Others, 1989, judgment, pp. 27-28) The court accepted that four of the eight accused had succeeded in proving extenuating circumstances, and they were sentenced to varying terms of imprisonment.In view of the "brutal," "fiendish," and "gruesome" manner in which the victims were killed, however, the other four accused, who had played a more active part in the incident, were found guilty of murder without extenuating circumstances and were sentenced to death.The antiapartheid movement immediately mounted a campaign to "save the SARHWU four" (Takalani David Mamphaga, Wilson Matshili, George Maungedzo, and Patrick Molefe), and appeals were lodged against their death sentences.
While waiting for judgment after the conclusion of the trial, one of the accused, Takalani Mamphaga, wrote to me from prison: In this case Doctor you rescued us from the mouth of a lion. . . .It was not our intention to commit such a serious crime before God, it was a pressure. . . .Dr. Colman I do not have a word to express my gratitude.(T.D. Mamphaga, personal communication, November 19, 1988) Despite his optimism, Takalani Mamphaga was soon to be sentenced to death; however in May 1991, the Court of Appeal set aside the death sentences ofMamphaga and the other three defendants on the grounds that the state had not succeeded in discharging its onus to disprove the psychological evidence.

The Retrial of the Queenstown Six
In November 1989, I testified along similar lines in the retrial (s.v. Gqeba and Others, 1990) of six men, five of whom had been found guilty of murder without extenuating circumstances in the original trial and had spent two years on death row.I was the only social psychologist called as a witness in the (penalty-phase) extenuation proceedings, although the social anthropologist Boet Kotze once again supported my evidence by testifying about "collective consciousness," and the Black South African clinical psychologist Pumla Gobodo Madikizela testified about the social backgrounds and personality characteristics of the individual defendants.

Background to the Murder
The trial arose from the "necklace" killing of a young woman in the township of Mlungisi in the Eastern Cape area of South Africa on December 8, 1985.The murder October 1991 • American Psychologist took place against a background of general unrest in most of the country's Black townships at that time.According to the uncontested evidence of several witnesses, Mlungisi was extremely overcrowded, poorly serviced, and generally deprived, even by South African standards.The quality of housing was appalling, and the level of unemployment in the township was extremely high.Apart from a central square that was floodlit, and was therefore nicknamed the "Golden," there was no electricity supply to the township.Toilet facilities consisted of about a dozen communal latrines, located on the outskirts of the township, serving a population of 35,000 residents.On Sundays, when most residents were at home, raw sewage often overflowed into the streets and caused the stench of human excrement to diffuse throughout the township.The township's water supply consisted of standpipes situated right next to the public latrines.
In August 1985, residents sympathetic to Nelson Mandela's African National Congress (ANC), which was still a banned organization at the time, decided to mount a consumer boycott of White-owned shops in the nearby urban center of Queenstown to persuade the authorities to make improvements to the township.The boycott was extremely effective, and 35 White-owned businesses quickly went bankrupt.The authorities responded by flooding the township with a specially trained unit of Black riot police loyal to Chief Buthelezi, the Zulu Inkatha leader and archenemy of the ANC, feared and hated by most of the Xhosa-speaking people in the area.Several witnesses described how these Inkatha police instituted a reign of terror, assaulting, intimidating, arresting, and not infrequently killing adults and children who appeared to be supporting the consumer boycott.

The Church Massacre
On November 17, 1985, a mass meeting was held in a Methodist church in Mlungisi to discuss the consumer boycott.The police arrived in force, parked an armored vehicle directly opposite the main door of the church, and issued an order over a loudspeaker for everyone to disperse within five minutes.Several witnesses testified that, without waiting for five minutes to pass and without issuing any further warning, the police began to fire tear gas canisters into the church.To escape the tear gas, people stampeded toward the doors and windows of the church.Those who managed to escape were met by a hail ofbullets from the police, and many of them tried to fight their way back into the church through the doors and windows.In the ensuing pandemonium, 11 people were killed and many more were injured.
The atmosphere in Mlungisi became extremely tense after the church massacre.A funeral for the victims, preceded by an all-night vigil, took place in Mlungisi on December 7, 1985.It was attended by 20,000 residents, many of them dressed defiantly in the distinctive black, green, and gold colors of the banned ANC.Press photographs taken at the funeral show coffins draped with ANC flags and several activists in the funeral cortege carrying replica AK-47 assault rifles.

The Murder of Nosipho
The necklace killing occurred the day after the funeral.The unfortunate victim, an 18-year-old woman called Nosipho Zamela, was suspected of being a collaborator and an informer after being accused of having sexual relations with a member of the Inkatha police unit, against which feelings of hatred and resentment had greatly intensified.The following account of what happened on the day of the murder is based on the "agreed facts" formally accepted by both the prosecution and the defense after the liability phase of the trial.A street committee of young activists decided to punish Nosipho by means ofa public whipping.She was immediately taken out into the open, beside a public lavatory, and flogged with a sjambok.While the beating was taking place, a crowd of spectators began to gather.Nosipho suddenly announced that she was not the only woman in Mlungisi having sexual relations with an Inkatha policeman, and she promised to point out where one of the other culprits lived.The crowd followed her across the township to several addresses, singing and dancing the traditional toyi-toyi, but it gradually became apparent that they were on a wild goose chase.The crowd had by this time grown in size to more than 200 and metamorphosed into an angry mob.In a horrific instance of group polarization, the opinions of members of the mob rapidly shifted in the direction of extremity, and eventually a unanimous decision emerged in favor of killing the impimpi (traitor) by necklacing.As darkness began to fall, the increasingly emotional and irrational mob processed toward the Golden, and Nosipho was forced to roll an ominous car tire in front of her as she walked.On arrival at the Golden, members of the mob placed the tire around Nosipho's body, doused her with gasoline, and set her on fire.The mob sang and performed a macabre toyi-toyi dance around her for some time while she burned to death.

Trial and Retrial
In June 1987, six members of the mob were convicted of murder through common purpose-they had not taken any part in the actual killing-and five of them were sentenced to death.Luckily for the Queenstown Six, as they became known, the trial was vitiated by a legal technicality: The Court of Appeal set aside the convictions and sentences and ordered a retrial because the trial judge had misdirected himself-although defense counsel had not objected at the time-in allowing one of his assessors (advisers), whose daughter was critically ill, to absent himself from court when sentencing took place.(In a murder trial in South Africa, the bench consists of a judge sitting with two legal assessors of his own choosing.) The social psychological phenomena to which I referred in the retrial were obedience to authority, conformity, group polarization, frustration-aggression, relative deprivation, and deindividuation.I also explained to the court the fundamental attribution error (Ross, 1977) and the law of social impact (Latane, 1981), according to which the impact of the social pressure on an individual is a multiplicative function of the strength, immediacy, 1076 and number of the social influence sources; the facts of the case suggested that all three parameters were very high in the circumstances prevailing at the time of the killing.
My evidence extended over three days, including cross examination.I explained all the relevant psychological phenomena to the court and then, in the light of the "agreed facts," the evidence the defendants had already given, and my consultations with them, I explained how the psychological phenomena could help to explain the conduct of each of the defendants.
In my consultations with the defendants before giving evidence, I had attempted to gauge their feelings of relative deprivation with the help of a crude version of Cantril's (1963Cantril's ( , 1965;;Kilpatrick & Cantril, 1960) selfanchoring Striving scale.The scale I used was simply a 10-rung ladder drawn on a piece of paper with the top rung representing the best possible life that the respondent could imagine and the bottom rung the worst possible.I reported to the court that, in terms of general quality of life, all but one of the defendants put their own group literally on the bottom rung of the ladder and the White group on the very top rung, with the "colored" (mixedrace) people, with whom they clearly compared themselves unfavorably, several rungs above them.I pointed out that this was suggestive of extreme relative deprivation and that there is evidence to show that high levels of relative deprivation are often associated with discontent that may be manifested in illegal and violent actions (Crosby, 1976;Gurr, 1970).
I summed up my evidence by saying, Each of these social forces on its own. . . is powerful and can lead people to behave in ways that are not characteristic of their normal behaviour.Taken together in these highly unusual circumstances, they all happened through some horrible combination of circumstances to come together.To have them come together like that, I would have thought as a social psychologist, is a recipe for trouble, and that the trouble was-although not this specific form of trouble-some form of trouble was almost predictable, given the background.If I knew all about those circumstances, but not what the consequence was, I think that as a social psychologist I would say something nasty is very likely to happen. . . .From that point of view, although I am not a lawyer, as a social psychologist I feel that this is relevant to the interpretation of the behaviour or the conduct of these accused. (s.v. Gqeba and Others, 1990, court record, p. 409)

Cross Examination
The cross examination started by establishing that my evidence had been based largely on my reading of published research.Although the common-law rule that excludes hearsay evidence does not apply to expert testimony about published scientific evidence, this line of questioning was presumably calculated to undermine my credibility by suggesting that I had no first-hand experience of research in some of the areas that I touched on in my testimony and no privileged standpoint from which to evaluate the literature.The prosecutor asked me whether his own evaluation of the literature would not be as valid as mine: "If I proceeded to read all these October 1991 • American Psychologist books, studies, literature, et cetera, and I understood everything that was said in those books, would I be in the same position as you are?"I replied, That is a difficult question.If you, Mr. Meiring, were to study psychology, to do a degree in psychology, to do postgraduate research in psychology, to reach the level of qualifications of an academic in psychology, and then to study all the literature on, let us say, obedience to authority, then you ought to be in a position to reach an evaluation, yes. (s.v. Gqeba and Others, 1990, court record, p. 411) The prosecutor pursued the argument: "Dr.The prosecutor persisted for some time along these lines, asking me eventually: "Have you done any experiments to test these various theories put forward by these various writers?"I was fortunately able to reply, Oh yes.I mean, some of these I referred to already in my evidence-in-chief, I think.I routinely do experiments as part of my teaching which illustrate some of these processes we have been talking about.Like the fundamental attribution error and group polarization, for example.I do experiments in class where you can actually see group polarization taking place and the fundamental attribution error.(p.413) A slightly different line of cross examination focused on the external validity of the experiments I had described: "Now as far as the experiments you have referred to are concerned, would I be correct in stating that there is a fairly large degree of generalization contained in those experiments?"I replied, Almost by definition an experiment on a selected sample of subjects involves generalization when you interpret it.There is always an assumption that has to be made that what you are observing has some significance beyond the people who are manifesting it.This will be equally true in medicine, where somebody is doing research into how the heart works, for example.You do research on a number of people and you make an assumption, a perfectly reasonable assumption in some cases, that the same thing ought to be true for other types of people.Now in social psychology it is true that sometimes that kind of generalization is more hazardous, because social attitudes and so on differ between people.In those cases it is necessary to replicate the research using different groups of people in order to specifically establish the range of applicability.And this is what is normally done in the best practice.(s.v. Gqeba and Others, 1990, court record, pp. 413-414) Most of the cross examination was devoted to establishing that the social pressures were not irresistible October 1991 • American Psychologist and that some of the accused showed autonomy or leadership qualities, and in trying to show that the social psychological phenomena did not apply to the specific facts of the case.I did not feel, in the end, that any of these lines of attack had seriously weakened my evidence.I

Judgment
The judge made it clear that he accepted the psychological evidence in respect of all the accused: Dr. Colman. . .explained the phenomenon of relative deprivation, which is to a certain degree related to the phenomenon of frustration-aggression.He further described the conceptions of conformity, polarization, and deindividuation in the context of social psychology .... Most important was the phenomenon of deindividuation. . . .I accept the evidence of Dr. Colman that all the accused were to a certain extent deindividuated at the time.(s.v. Gqeba and Others. 1990, judgment, pp. 480-481) Before passing sentence, the judge addressed the following remark directly to the accused: The sentence I am about to pass on the six of you may be regarded by some people as too lenient.Some politicians may even want to discuss it in Parliament. . . .I am passing this lenient sentence on you only because of the very exceptional circumstances in this case.(p. 485) Taking into account the fact that the accused had already spent about four years in custody and almost two ofthose years on death row, the judge sentenced all six accused to 60 months in prison, of which 40 months were suspended for five years.In other words, the same men who were sentenced to death in the original trial were sentenced to less than two years in prison for the same crime when extenuating circumstances based on social psychological phenomena were put forward on their behalf.

Discussion and Conclusions
Recent judicial decisions in South Africa, especially the judgment in the retrial of the Queenstown Six (s.v. Gqeba and Others, 1990), have legal implications that will be felt for many years to come.Since the introduction of the Criminal Law Amendment Act of 1990.the statutory constraint on courts in that country to impose the death penalty in cases of murder without extenuating circumstances has been removed.Courts are required to consider both aggravating and what are now called mitigating factors, in the light of which they have discretion to impose the death penalty or custodial sentences.The death penalty has not been (and is not likely to be) abolished entirely, and in murder cases the existence or nonexistence of mitigating circumstances will continue to be a key issue.Prior to the amendment, the onus rested on the defense to prove the existence of extenuating circumstances, and the required standard of proof was the "balance of probabilities."The effect of the amendment is to shift the onus to the state to disprove "beyond reasonable doubt" the existence of mitigating factors.There will therefore continue to be scope for expert psychological testimony in the penalty phases of murder trials in South Africa.
Anything that helps to explain the behavior of a defendant might reduce the perceived moral blameworthiness of that behavior, and there are many well-understood psychological phenomena that could be relevant in this regard, apart from those that have been cited in previous court cases.The central goal of psychological research is to explain behavior, and a court oflaw has essentially the same goal when it is passing sentence on a convicted criminal.In order to judge the moral blameworthiness of an individual's actions, it is helpful to understand the causes of those actions.Almost by definition, psychology is devoted to deepening our understanding of human behavior, and to that extent its concerns coincide with those of the judge (or in some jurisdictions the jury) whose function it is to pass sentence.The difference, of course, is that psychology relies on carefully controlled research, whereas judges and juries rely chiefly on intuition and common sense.
Psychology could, it seems to me, playa much larger role than it has hitherto done at the point of sentencing, not only in death penalty cases in South Africa, but also in cases involving lesser offenses in other countries.Wherever a court has discretion to decrease the severity of a judicial punishment in the light of extenuating or mitigating circumstances, expert psychological testimony is potentially illuminating.The law tends to be extremely conservative, however, and it is another matter altogether whether courts will, in practice, consider such testimony to be relevant or even admissible.I have found, for example, that courts in Northern Ireland are rather reluctant to admit expert psychological evidence.The South African judiciary has, at least, helped to break the ice.
In conclusion, it is worth commenting briefly on some of the ethical problems that arise for psychologists who are asked to testify as expert witnesses (see also McLoskey, Egeth, & McKenna, 1986).The first moral dilemma often centers on the problem of whether to testify.Should an expert who is convinced that a particular defendant is guilty as charged agree to testify on that defendant's behalf?Is it right to testify on behalf of a defendant charged with an utterly repellent and morally indefensible crime?One answer to these questions is that according to the rule of law, people are innocent until they are proven guilty and, inasmuch as it is not the function of an expert witness to decide whether an accused person is guilty or innocent, the testimony should not be withheld in such circumstances.A second argument in favor of testifying is that people are often wrongly accused, and surely even those charged with the worst crimeswho therefore face the heaviest sentences if they are convicted-are entitled to at least the same standard of defense as others.In spite of the apparent force of these principled arguments, I should be personally reluctant to testify on behalf of certain defendants, for example, extremist White vigilantes accused of lynching Black 1078 people in South Africa, or anywhere else for that matter; I should also be reluctant to testify on behalf of former Nazis accused of crimes against humanity.The principled arguments are, in any event, irrelevant to testimony in penalty-phase proceedings, because by then there is no question of the defendants' innocence.Would it be ethical to offer mitigating evidence about crowd psychology along the lines described in this article in penalty-phase proceedings on behalf of defendants who had been convicted of gang rapes or sadistic mob killings inspired by entirely base and vicious motives?This question cannot be answered straightforwardly, but it is worth expressing the opinion that in such circumstances there can surely be no moral imperative to testify.
A different set of moral dilemmas relate to the content and presentation of expert testimony.Having agreed to testify, perhaps only after considerable soul searching, is it a psychologist's duty to draw attention to the inherently probabilistic, and therefore uncertain, implications of the relevant research findings and to present both sides of controversial psychological issues?Is it right for expert witnesses to tailor their testimony to benefit particular clients?Problems such as these arise from a clash between two quite different ways of seeking the truth (Loftus & Monahan, 1980).Courtroom proceedings are adversarial processes in which each of the parties advances all of the arguments and evidence that support his or her case and vigorously attacks all those put forward by the opposition.The judge or jury-more generally, the trier of fact-acts as a kind of referee and decides in the end on which side of the argument the truth lies.Psychologists, at least in theory, adopt a neutral and detached viewpoint and seek the truth through the dispassionate application of established research methods.An expert witness has a duty to alert the court to the uncertain implications of research findings, in cases in which this point is relevant, and to give a fair and balanced account of controversial issues.Biased testimony does a disservice not only to the expert's discipline and to the truth, but also in the long run to the credibility of all future expert witnesses.The dilemma about content and presentation arises from the common misconception that it is an expert's function to try to help the client.The client's lawyers hope and expect that the expert's testimony will help their case-they would not otherwise have sought the expert's testimony-but while giving evidence, the expert's proper function is simply and solely to help the court to discover the truth.It can be difficult at times for an expert witness to maintain this attitude of detached neutrality, as I discovered in the South African trials described in this article, but it is an ideal worth striving for.
It is rather like a court case: At the end of it there are arguments on both sides and somebody who is in a position with the background and experience of research can arrive at an evaluation of what a fair conclusion is, and usually it is a conclusion which all experts in the area generally agree about.(p.412)