Full Report of Research Activities and Results
Background
Self-immolation can be contrasted with other types of altruistic suicide. A suicide attack is
intended to kill others or inflict material damage. The rationale is clear: by inflicting harm on
your opponents, you hope to force them to yield. Self-immolation is far more theoretically
puzzling. Wherein lies its efficacy? The same problem is raised by hunger strikes, which use the
prospect of death as a ‘threat’ to win concessions. But self-immolation poses a more extreme
puzzle, because it pre-empts any concession. “No hungerstriker aims at death,” observed a Jesuit
theologian in 1920, after an Irish patriot starved himself to death in a British prison. “He aims at
escaping from unjust detention, and, to do this is willing to run the risk of death, of which he has
no desire, not even as a means” (quoted in Sweeney 1993, p. 428). Self-immolation should be
distinguished from cultic suicides. When such deaths are chosen voluntarily, they aim at
‘transfiguration’: death leads to “a state considered by the subject to be infinitely more
delightful” (Baechler 1975, p. 63). The Hindu tradition in which a wife (sati) joins her husband
on his funeral pyre is the subject of controversy, but the intent is not to advance a collective
cause (Hawley (ed.) 1994).
As a concept, self-immolation demarcates a conceptually coherent species of social action. It is
also an historical lineage, with Quang Duc as its progenitor. Indeed, his choice of method—
fire—led to a variant meaning of the word ‘immolation’; it strictly means sacrifice, but it is now
often used as a synonym for fiery death. In the four decades since 1963, there have been at least
a thousand acts of self-immolation (including attempts which did not prove fatal), and perhaps as
many three thousand. The vast majority of them—and almost certainly every case in which fire
was used—were directly or indirectly modeled on his act.
Needless to say, we can find earlier instances which resemble self-immolation. Most familiar to
Western readers is Emily Wilding Davison, the English suffragist who died in 1913. During the
Derby, she ran onto the race track (with suffrage flags in hand) and attempted to stop the King’s
race horse; she received fatal injuries. Actually, she did not intend her death; she had a return
ticket in her pocket and she left no written testament. She was nonetheless unafraid to die: she
had attempted to kill herself in prison, as a protest against force-feeding during a hunger strike.
Indeed, she intimated that “a life would have to be given before the vote was won” (quoted in
Colmore 1913, p. 49). Although the movement embraced her as a martyr, her death did not
inspire others to die.
Two episodes from the end of the Second World War are worth recounting. After Japan
surrendered, many military leaders died by seppuku (ritual disembowelling) to atone for their
sins. Closer to self-immolation as defined here, two groups of civilian nationalists—numbering
26 in total—and several members of the Imperial Guards Division committed seppuku before the
Imperial Palace, in separate incidents (Morris 1960, pp. 25-9). More surprising, perhaps, is what
happened in Europe when American and British troops repatriated former Soviet citizens who
had been living under Nazi rule (Bethell 1987). As they were transported back to the Soviet
Union, hundreds killed themselves, and sometimes killed their children. Knowing that their fate
was execution or life in the Gulag, suicide represented a preferable option. To some extent,
however, death was a protest against the actions of their captors. Although many soldiers were
disturbed by the suicides, they never disobeyed orders. Indeed, the military eventually designed
procedures and modified trains to prevent people from taking their own lives. With the onset of
the Cold War, the American and British governments ended the repatriations; the mass suicides
counted for little, if anything, in the decision. As the operation was carried out in strict secrecy,
these suicides had no wider impact.
These episodes belong to the prehistory of self-immolation, as it were. Its history dates from
1963. The subsequent waves of self-immolation attracted the attention of psychologists and
psychiatrists, who defined their object as suicides by fire (Bostic 1973; Bourgeois 1969; Crosby,
Rhee, and Holland 1977). Medical professionals contribute to a specialized literature on suicide,
which includes two diagnostic studies of Indians who attempted self-immolation in 1990, and
survived (Mahla et al. 1992; Singh et al. 1998). Pacifists have evaluated self-immolation from a
normative standpoint (King 2000; Ryan 1994). Film-makers have documented and dramatized
individual acts (Puhovski 1997; Park 1995). Yet with two recent exceptions (Kim forthcoming;
Uehling 2000), self-immolation has not attracted the attention of social scientists—which is
surprising given the vast literature on collective protest and the classical importance of suicide
for sociologists.
Objectives
This research investigates patterns of self-immolation by compiling a database of 306 individual
acts, all over the world, since 1963. Global scope is necessary in order to trace the diffusion of
self-immolation across different countries and from one cause to another. It also enables
comparison of differences among cultures. Gathering information on every individual act, no
matter how inconsequential, avoids the fallacy of selecting on the dependent variable. To
understand how an individual act of self-immolation could have a huge political impact, it is
necessary to also examine those ordinary cases which had no meaningful consequences.
Whatever their biases and limitations, news reports are the only practical source of information.
For the period from 1963 to 1976, the New York Times and The Times are used, because they are
comprehensively indexed.1 The superior search capability of the Nexis database becomes
available from 1977 onwards. This provides access to global newswires: the Associated Press
(AP), United Press International (UPI) from 1981, and Agence France Press (AFP) from 1991.2
Needless to say, not every act of self-immolation is reported in the news media. The two barriers
are lack of information and lack of interest. Totalitarian states can effectively suppress
information on protest, including self-immolation. For example, Western tourists in Moscow’s
Red Square—visiting for the 1980 Olympic Games—saw someone on fire. One was knocked
down as he attempted to take a photograph; others had their film exposed by KGB agents. The
authorities insisted that it was merely a burning cigarette or a “garbage fire” (AP, 2 Aug. 1980;
17 June 1981). How many others committed suicide in more remote locations, out of sight of
Western visitors?3 There is an important offsetting factor, however. Self-immolation is intended
as a public act, and so individuals have good reason to orchestrate publicity—by traveling to a
public place where there will be witnesses or by alerting journalists to be on hand. Several Falun
Gong believers traveled 350 miles from Kaifeng to Tiananmen Square to set fire to themselves—
before cameras from the Cable News Network.4
The second barrier is lack of interest. Self-immolation is rare and spectacular, and so it is
exceptionally newsworthy—far more than conventional protests like strikes or demonstrations
(McCarthy, McPhail, Smith 1996). Nevertheless, the space for news is limited.5 Journalists and
their audiences have different levels of interest in different parts of the world. During the mid
1960s, because Americans were focused on Vietnam, the New York Times reported every selfimmolation there (ascertained by checking against more detailed accounts). But it never
mentioned some incidents in South Asia—which we know about from reports in the Times. As
this suggests, the bias can be partially offset by including news media from different countries.
Nevertheless, this is still the world viewed from New York, London, and Paris.
Methods
These news reports have been used to compile a ‘sample’ of each individual act of selfimmolation: where someone apparently attempted to kill him or herself, without harming others,
1
In the index to the New York Times, this requires searching entries under the heading ‘suicide’. The index to The Times
can be searched electronically, under the heading ‘suicide’ until 1965 and thereafter ‘self-immolation.’
2
Other newswires, from different parts of the world, have recently been added to the Nexis database. These are not
included because this would destroy comparability of coverage over time. The keywords utilized are ‘immolate’,
‘immolated’, ‘immolating’, and ‘immolation’. Some reports, unfortunately, may not include these words, but this is the
only feasible method of searching.
3
One example is Musa Mamut, who set fire to himself in a village in the Crimea (Uehling 2000). Curiously, this was
reported by the official Chinese news agency (Xinhua News Agency, 1 Sept. 1978)!
4
The Cable News Network denied that it was tipped off beforehand; its producer just guessed that some sort of protest
was likely (AP, 8 Feb. 2001). The film was confiscated by the police.
5
This is brought home by the Times index, which includes articles dropped from later editions of the day’s newspaper.
Some tantalizing reports of self-immolation were driven off the page by late-breaking news. Because only the day’s final
edition is microfilmed and archived, these reports effectively vanish from the record. (Even the Times itself does not
hold these early editions.)
as an act of public protest on behalf of a collective cause.6 (The term ‘sample’ is used not to
imply statistical sampling, but to emphasize its partial character.) Note that an act is counted
whether or not the person actually died (which is not always reported). Thwarted attempts at
suicide are included, but threats are not.7 In many cases, only the most basic information was
reported. This is an inevitable consequence of choosing not to select only the most notable or
successful acts. A minimal amount of specific information is required (at least two of the
following: name, date, location); extremely vague reports are discounted. For this reason, selfkilling by prison inmates has been excluded. Prisons resemble totalitarian states, and so it is
often difficult to know whether ‘suicide’ is actually murder.8 Moreover, if an inmate does
actually kill him or herself, it is not always possible to ascertain whether this was an act of
protest. Admittedly, even outside prison the distinction between personal suicide and selfimmolation is not always clear. My judgment depends on whether the cause is potentially
generalizable. I include a man in Pakistan who set fire to himself to protest against continual
police harassment—on the grounds that police corruption is a widespread problem in the
country. I exclude a Pakistani man who set fire to himself in Australia after his family were
refused asylum, because this was a personal grievance (there was no declaration against
immigration policies which mistreated refugees in general).9
Results
The results permit a synoptic overview of self-immolation over the four decades since 1963. The
sample includes instances from three dozen countries, from Argentina to Yugoslavia. Table 1
shows the distribution by country. Three countries stand out: Vietnam (South Vietnam until
1975), India, and South Korea. Lithuanians are counted separately, even though their nation was
subsumed by the USSR (and one man traveled to Moscow to kill himself). Kurdish refugees
living in Europe also deserve special treatment. Their acts were part of a struggle against the
Turkish state. This should be considered when interpreting the figure for Turkey. That also
excludes (for reasons explained above) at least two dozen prison inmates—most of whom were
Kurdish—who have set themselves on fire since 1996.10 The sample greatly underestimates the
huge wave in India from mid September to the end of October 1990. This occurred after V. P.
Singh’s government proposed to set aside an additional 27 percent of places in universities and
government employment for Other Backward Castes. Western newswires reported the wave in
some detail, but provide specific information—warranting inclusion in the sample—on only
twenty individuals. According to detailed analysis of two major Indian newspapers (The Times of
India and The Hindu), a total of 239 people tried to kill themselves in protest over this period.
With this addition, India heads the list, contributing almost half the total of 525.
This augmented total still underestimates the extent of the phenomenon. We can gauge the
magnitude of the omission by comparing more comprehensive totals for particular times and
places. The sample excludes slightly over half of the fatal immolations in South Korea from
1970 to 1997 (Kim forthcoming). More surprisingly, perhaps, it omits over two-thirds of the
6
The sample includes an additional eighteen individuals identified by other sources—because they provided specific and
detailed information (Chân Không 1993, pp. 206-8; Human Rights Watch 1995; Lee and Bishop 1998, p. 171; Puhovski
1997; Radio Praha n.d.; Times of India, 22 Sept. 1990; Uehling 1990). This includes a group of twelve which was actually
reported in the Times (8 Sept. 1976), though not indexed under ‘immolation’; the newspaper report alone does not
provide sufficient information to warrant inclusion in the sample. A variety of sources have been used to provide
supplementary information on other cases.
7
A thwarted attempt is included only if it was physically prevented at the last moment; this excludes arrests for
apparently planning self-immolation. The sample excludes self-mutilation, like cutting off limbs or fingers.
8
An examples will suffice. In South Korea in 1991, a labor activist (Park Chang-soo) in custody fell to his death. The
government claimed that he took his own life; the opposition believed that he was killed.
9
I would tend to exclude suicides on the illness or death of a public figure—like M. G. Ramachandran in Tamil Nadu
(Pandian 1982, pp. 17-18)—though the point is moot because there is not sufficient information on individual cases,
even in Indian newspapers.
10
The newswires reported 21 in total, but this is obviously an underestimate. For November 1998 alone, six appeared in
news reports, but official Turkish figures counted 22 (Ergil 2001, p. 115).
immolations in Europe that were committed by Kurdish refugees (Olivier Grojean, personal
communication). For India, it is possible to tabulate reports in two major national newspapers
(The Times of India and The Statesman) for 2000. They reported eighteen instances of selfimmolation—but only two filtered through to the Western newswires. Thus we may hazard a
guess that the real total could be anywhere from two times to nine times larger than the sample.
The lower bound could hardly be less than a thousand; it is hard to imagine that the upper bound
could exceed three thousand.11
To compare countries, it is necessary to calculate rates. The denominator should be urban rather
total population because virtually all the cases in the sample occur in cities. Indeed, there are
only four ‘farmers’ (only 2 percent of recorded occupations) in the sample. Estimates of urban
population refer to 1985, approximately the mean year in the sample (United Nations Population
Division 2001, table A.3).12 When rates are calculated from this denominator, the figures yield a
rather different story. Vietnam and South Korea remain prominent, but two small countries from
the former Soviet bloc enter the top four: Lithuania and Romania. The Kurdish diaspora in
Europe has by the highest rate by far. Correcting the rate for India to include the wave of
September and October 1990 yields 1.5, just below South Korea. We will postpone the challenge
of explaining this variation among countries until section 6.
Comparison across time is rather less meaningful than comparison across space, because of the
different sources used before and after 1977.13 Figure 1 (see “Dying without Killing”) shows the
numbers by year. There was a noticeable lull in the late 1970s and early 1980s. In 1990 came the
unprecedented outbreak of self-immolation in India, followed by another wave in South Korea in
1991. The last few years have witnessed another outburst: Kurds protesting against Turkey’s
capture of Abdullah Ocalan in 1999; unemployed coal miners demanding government assistance
in Romania in 1999; Falun Gong adherents protesting against persecution in China in 2001. Of
course the world’s urban population multiplied about two and a half times over these four
decades, and so an increasing number of cases is not surprising.
Self-immolation is naturally clustered in space and time: it occurs in waves. Defining the
boundaries of a wave is inevitably arbitrary. Some notion can be gained by counting how long it
takes between one self-immolation and the next in the same country (classifying Kurds outside
Turkey and Lithuanians separately, as above). Over a quarter occurred on the very same day;
most of these are group actions. Those aside, 14 percent occurred from one to six days after the
previous self-immolation, and a further 12 percent from a week to a month. These figures are
based on the sample. Taking into account the full extent of the wave in India in 1990 would
show that the great majority of self-immolations occurred within one or two days of another.
To some extent, waves can be explained by exogenous shifts in political and economic
conditions. The endogenous process of inspiration is also clearly important (Biggs 2003). An
Indian student, Monica Chadha, provides an illustration. After reading the morning newspaper,
she casually announced to her family that she would kill herself. She was particularly impressed
by the first woman to die. “If all these other boys and girls can sacrifice themselves like this to
shout down the policies of this prime minister, then I will sacrifice my life as well” (Los Angeles
Times, 20 October 1990). She then left the room and set herself alight on the terrace outside.
There are several instances of very direct connections between one case and another: a Tamil
labourer burned himself exactly one year after Chinnasamy’s death; an Indian student set fire to
herself outside the hospital where the wave’s first victim lay; a Vietnamese monk burned himself
11
Note that the figure for the wave in India in 1990 is unlikely to be much of an underestimate, and so this should not
be multiplied.
12
Unfortunately, these estimates are not based on a uniform definition of ‘urban’; they reflect the different definitions
adopted by each country’s statistical office. The United Nations eliminates Taiwan from its statistics; the figure is
estimated from Tsai 1996. The Kurdish population in Europe is estimated as a million (following Grojean). Obviously
the rate of population growth varies greatly among countries, but this will not distort the comparison—though
differential growth before and after 1985 will.
13
The newswires appear to be more comprehensive, at least judging from 1977: the Associated Press reported two
cases, but neither appeared in the New York Times or the Times.
to death to protest against the confiscation of a previous victim’s body by the police; a Korean
student jumped off the morgue containing the body of the wave’s first victim. Inspiration is not
always confined to those who believe in the same cause. In 1990 one lower-caste man followed
the example of the Indian students—he killed himself in protest against their protest. At a greater
remove, two Turks burned themselves in 2000 in protest against their government’s decision to
delay executing Ocalan. They were surely inspired by their enemy: a year before, Kurds had
used the same technique to protest against the capture of Ocalan.
By far the most common method of self-immolation was Quang Duc’s: nine out of ten
individuals chose to set themselves on fire. This is spectacular and painful. Even if it does not
prove fatal, it is likely to leave permanent disfigurement. “To burn oneself by fire is to prove that
what one is saying is of the utmost importance,” as Nhat Hanh (1967, p. 118) observed. “There
is nothing more painful than burning oneself.” In this respect, it is akin to seppuku. As Yukio
Mishima, the celebrated Japanese writer explained, “everybody knew that this was the most
painful way to die … it proved the courage of the Samurai” (quoted in Scott-Stokes 1975, pp.
16-17). In actuality, the agony was quickly terminated by beheading. Similarly, individuals could
combine burning with a less painful or more certain means of death: either by taking poison or
jumping from a building. The latter combination was especially common in South Korea.
As for individual demographic characteristics, three quarters of the sample were male. This
certainly fits the pattern of suicide deaths, in which men almost invariably outnumber women
(the exception is China). Figure 2 (see “Dying without Killing”) shows the age distribution. It is
remarkably similar for men and women: it peaks in the range from 15 to 24 (the mode is actually
25), and then declines with increasing age. We can view the same fact from another perspective:
almost a third (where occupation was recorded) of the sample were students of one sort or
another. Taking into account the full extent of the wave in India in 1990 would further
accentuate this youthful profile. It is not straightforward to interpret the age distribution without
calculating age-specific rates, but there is no suitable denominator. Nevertheless, we can be sure
that individuals in their late teens and early twenties were disproportionately prone to selfimmolation. This provides a point of comparison with suicide deaths in general. Unfortunately,
statistics are not available for Vietnam or India. Comparison can be made with three countries
which feature large in the sample: South Korea, the United States, and Romania (all for 1997:
World Health Organization n.d.). (For comparability with Figure 2, we examine the number of
suicides in each age-sex group as a fraction of the total number of suicides, and not the agespecific rate.) In all three countries, for males the distribution peaks in the 35-44 age group. For
females it peaks somewhere in the 35-55 range except in South Korea, where it peaks between
15 and 24—this is the only distribution which resembles that of self-immolation. We may
suspect, then, that self-immolation bears greater resemblance to the age profile of participation in
less extreme forms of protest.
Table 1: Self-immolation by country, 1963-2002
Number
Fraction
of total
Rate per
million
South Vietnam/Vietnam
70
23%
6.06
India
50
16%
0.27
South Korea
43
14%
1.62
USA
29
9%
0.16
Kurds outside Turkey
14
5%
14.00
Romania
13
4%
1.11
USSR/ex-USSR (except Lithuania)
12
4%
0.08
Pakistan
9
3%
0.32
China
9
3%
0.04
Japan
5
2%
0.05
Lithuanians
5
2%
2.17
France
5
2%
0.53
Czechoslovakia
4
1%
0.47
Chile
4
1%
0.04
Turkey
4
1%
0.15
East/West Germany
4
1%
0.06
Taiwan
3
1%
0.23
Bulgaria
3
1%
0.52
Thailand
2
1%
0.22
Malaysia
2
1%
0.28
UK
2
1%
0.04
14
5%
0.01
Total
306
100%
0.15
India plus all September/October 1990
269
51%
1.45
Country
elsewhere
The results of the research allow us to answer questions about the relation between intention and
fatality, and the role of audiences and organizations.
Threats of self-immolation are too common to warrant any sustained attention. A hunger strike
gradually demonstrates a willingness to suffer; a threat to kill oneself costs nothing and so is
incredible. One gains the impression that threats are almost invariably idle words. This holds for
Vietnamese Buddhists in the mid 1960s, even though so many proved willing to die. In July
1963, monks at one pagoda staged a press conference to announce that a nun—from a
prestigious family—had requested to be allowed to burn herself. Foreign journalists found this
spectacle repugnant (Schecter 1967, pp. 191-2). And the nun never did kill herself. In May 1966,
monks in Danang declared that three of their number would kill themselves if government troops
attacked the city (NYT, 17 May 1966). Dozens were killed in the attack, but the pagoda
surrendered a week later without any self-immolations (Schecter 1969, pp. 230-1). Although
threats seem to be rarely carried out, there are some examples of success. A Sikh bus conductor
in Birmingham threatened to burn himself unless beards and turbans were permitted; the local
authority eventually relaxed these senseless restrictions (Times, 10 April 1969).
Confining attention to actions rather than words, it is still necessary to ask whether the
individuals really intended to die. Studies of personal suicides which treat death as a contingent
outcome of deliberate risk-taking—more or less dangerous—have yielded great insight (Firth
1961; Stengel 1969). For this reason, the sample includes attempts that did not prove fatal.
Thwarted attempts, when someone douses him or herself with flammable liquid, but is prevented
from setting it aflame by the intervention of fellow protesters or police, account for 9 percent of
the sample. This could be faked, of course, though deception is alleged in only one case. When
Rajeev Goswami was arrested in Delhi in 1993, he was ridiculed by fellow students. “It was all
stage-managed,” said one (AFP, 21 Sep. 1993). Ironically, Goswami was the same man who
sparked the wave of 1990, suffering serious burns after his attempt. (He is the only individual
entered in the sample twice.) Even when someone actually set fire to him or herself, we cannot
assume that death was intended. The only way of investigating this systematically is to examine
how many actually died. Although it is not always reported whether an individual eventually
died (especially when the immediate outcome was hospitalization), we can estimate the overall
death rate in the sample. The ex ante chance of surviving is only one in four—or one in five if
the event is not thwarted. Disaggregated figures reveal considerable variation. In Vietnam,
virtually everyone died. In India, a third survived. And comprehensive figures on the wave in
1990 (which is greatly underestimated in the sample) indicate that three out of five survived.
Among Kurdish refugees, almost two-thirds survived. In Romania there seem to have been no
deaths at all. These differences cannot be explained away by variations in the quality of medical
care or the actions of bystanders, because the United States also has a high death rate.
In a minority of cases, then, self-immolation is a dramatic and risky gesture (perhaps akin to selfmutilation)—rather than a determined effort to die. But those who survive setting themselves
alight will probably suffer permanent disfigurement. Moreover, the overall death rate is still
remarkably high. Even adjusting the figures to fully incorporate the Indian wave of 1990, three
out of five acts of self-immolation end in death. This is in stark contrast to suicide attempts,
where the survival rate can be as high as nine out of ten (Stengel 1969). It also contrasts with
hunger strikes. Although systematic figures are not available, one has the impression that hunger
strikers rarely choose death if they fail to win concessions. In 1923, for example, about 8000
Republican prisoners in Ireland went on hunger strike to demand release; two died (Sweeney
1993).
In the vast majority of cases, self-immolation occurred in a public space, where others would
witness the death. More to the point is whether the act was performed in front of a crowd
gathered to further the cause—an audience of potential sympathizers. Over a quarter of the
individuals had such an audience. By itself that does not indicate that the audience was complicit
in the act. In most situations, protesters attempted to intervene to help the individual, typically
trying to douse or quench the flames—sometimes being burned themselves. In an extreme case,
Chan Yih-hua set himself alight in the midst of a packed memorial for a dissident in Taiwan,
causing severe injuries to those standing beside him (AP, 19 May 1989). Much less typical were
situations where the crowd effectively assisted the suicide, for example by preventing the police
from intervening; one in six actions were given such assistance. In this respect, Quang Duc’s act
was exceptional. Other examples come from Vietnam and from India and Pakistan. Insinuations
and accusations that an individual’s death was not entirely voluntary are very common. The
Archbishop of Hué claimed that monks bludgeoned one of their fellows to death before setting
him alight. Such claims can be discounted as absurd propaganda, with a single exception. A
Pakistani journalist recalls a self-immolation in protest against Zulfikar Ali Bhutto’s death
sentence (UPI, 25 March 2002). According to the man’s relatives, others in the People’s Party
had assured him that it would be symbolic—they would immediately extinguish the fire. But as
he ran around in flames, screaming for help, the assembled crowd watched impassively.14
Unlike suicidal attacks, self-immolation is quite easy to accomplish without the help of any
collective organization. From the perspective of an organization, it may be quite dangerous to be
seen to authorise or encourage self-immolation, because it implies a callous disregard for its
members. This is illustrated by the fate of Falun Gong (discussed below in section 5).
Nevertheless, the organization of self-immolation has been particularly clear in Vietnam. In 1963
Buddhist leaders explicitly sanctioned two deaths: those of Quang Duc and Thich Tieu Dieu.
Both were elderly (as was the nun at the press conference), while there is evidence that younger
novices were refused permission. This is understandable: the elderly had less life to sacrifice and
had presumably attained sufficient wisdom to make a responsible choice. In 1966, the militant
wing of the Unified Buddhist Church apparently staged at least one self-immolation—that of
Thich Nu Thanh Quang (55 years’ old)—though officially it was announced that she had not
received permission. “Burning oneself to death is the noblest form of struggle” declared the
militant leader, Thich Tri Quang, “which symbolizes the spirit of nonviolence of Buddhism”
(quoted in Schecter 1967, p. 233). After five deaths over the next few days, Tri Quang called on
followers to halt (NYT, 31 May 1966). After the Communist invasion, when Buddhism was
facing renewed persecution, the Central Executive Council in 1977 called on monks and nuns
“to be ready to act, to sacrifice ourselves if necessary” (quoted in Forest 1978, p. 41). By the late
1990s, however, the Church was apparently dissuading members from killing themselves.
Supporters of the Kurdistan Workers’ Party (PKK) coordinated protest against the capture of
their leader Abdullah Ocalan. On the 16 February 1999, demonstrators attacked or occupied
Greek and Kenyan embassies across Western Europe. In many cases they threatened to burn the
embassies down; in the event, seven Kurds set themselves alight. It is frustrating that there is
little detailed evidence on these acts: whether they were spontaneous or assisted. They certainly
did not accord with the wishes of Ocalan. Only months before, after 22 PKK prisoners in Turkey
had set themselves alight, he broadcast the following message: “I categorically reject selfimmolation. I strongly suggest that they should refrain from setting themselves on fire” (quoted
in Ergil 2001, p. 115). Following Ocalan’s capture, his brother Osman repeated the same
message: he called on Kurds to end self-immolation and “burn the enemy” instead (AP, 19 Feb.
1999).
Activities
1. Workshop in Madrid with other contributors to Making Sense of Suicide Missions, November
2002.
2. Panel at the American Sociological Association, August 2003.
Outputs
14
A recollection so many years after the event—in 1978—cannot be treated uncritically, but it obviously seared into his
memory. The event is in the sample, but the Associated Press did not report sufficient detail to verify his account.
1. “Dying without Killing: Protest by Self-Immolation, 1963-2002,” Diego Gambetta (ed.),
Making Sense of Suicide Missions, Oxford University Press, forthcoming, 2003
2. “Protest by Suicide: Self-Immolation in the Global Repertoire, 1963-2002,” to be presented at
the annual meeting of the American Sociological Association, August 2003
Impacts
The notion of commercial exploitation is clearly inappropriate.
Future Research Priorities
I am continuing to add more cases to the database, utilizing the extensive newspaper collections
at the University of Illinois. I am particularly concerned to trace some of the earliest examples in
the 1960s, especially in Tamil Nadu. In addition, I am developing some theoretical insights
which emerge from this research on the benefits of ‘suffering’ in political protest. These will be
developed in a paper entitled “When Costs are Beneficial: Signaling and Suffering in Political
Protest,’ to be presented in May 2003 at the University of Chicago.