A Japanese Perspective on the Nanjing Massacre

JAPAN ECHO Vol. 25, No. 4, August 1998


The Nanking Atrocities: Fact and Fable

HATA Ikuhiko

Last year marked the sixtieth anniversary of the start of the Sino-Japanese War of 1937-45. Sixty years is a period of two generations, a sufficiently long time, one would like to think, for academic scholarship to move ahead without the hindrance of rancor over old events. In fact, however, a number of aspects of the fighting between Japan and China continue to be the object of ideologically and politically motivated debate, with views sharply divided.

One episode in particular, the carnage that began when the Imperial Japanese Army captured the city of Nanking (Nanjing) in December 1937, is still being argued about as heatedly as if it had happened only yesterday.

A number of reasons may be adduced for this, but probably the biggest one is that the scale of the atrocities has not been pinned down. Nowadays some people seem to be saying, in effect, "God only knows how many people were killed," and suggesting that the actual numbers are, in any event, beside the point. In view of the nature of the incident, however, it strikes me as unnatural to brush off the question of scale when discussing what went on.

Last autumn a work of special note in this connection appeared. I refer to the Japanese translation of the wartime diary of John Rabe (Nankin no shinjitsu [The Truth About Nanking], trans. Hirano KyÙko, KÙdansha [German title, Der Gute Deutsche von Nanking]). Rabe, a German national who remained in Nanking after the leaders of the Nationalist government and military had fled the city, leaving their subordinates behind, was head of the International Committee for the Nanking Safety Zone. As such he was effectively "mayor" of Nanking at the time of the atrocities. His account is detailed and vivid.

In it is a letter he addressed to Adolf Hitler upon his return to Germany in June 1938, in which he describes the numerical extent of the killings as follows: "According to Chinese claims, 100,000 civilians were killed; this, however, is probably somewhat of an overstatement. We foreigners view the figure as having been from about 50,000 to 60,000. Reportedly the Red Swastika Society, which has buried the corpses, was unable to handle more than 200 bodies a day, and as of February 22, when I left Nanking, some 30,000 bodies were lying unburied in suburban Hsia-kuan [Xiaguan]."

Rabe's account may be interpreted in various ways, but it certainly bears a fair amount of weight, being the result of observations by a number of neutral eyewitnesses. Historians who have written about the massacre may have to reconsider their numbers.

Rabe's diary, incidentally, was published in the original German and in Chinese translation around the same time as the Japanese translation's publication. The work turned into a bestseller in Japan for a while. But the planned publication of the English translation has been delayed because of problems with the translation rights.

In the first part of this essay (based on a paper originally prepared for the "Nanking 1937" conference sponsored by the Alliance in Memory of Victims of the Nanjing Massacre and held at Princeton University on November 22 last year), I will present the various numerical estimates that have been advanced and offer my own observations on their reliability. In the second part (taken from an article I wrote for the April 1998 issue of Shokun), I will turn to the photographic record, considering problems with some of the pictures offered as evidence of the slaughter, and comment on some of the shortcomings of the recent bestseller on the atrocities written by Iris Chang.


Students of modern Chinese history know that four events in Nanking since the mid-nineteenth century are all commonly referred to as "the Nanking Incident" (Nankin jiken). The first occurred in 1853, when Taiping rebel forces occupied the city and conducted a massacre; figures for the number killed range from several thousand to 30,000. The second happened in September 1913 when military forces under Chang Hs¸n (Zhang Xun) killed 1,000 student revolutionaries. In his article "Nankin daigyakusatsu no shinwateki kÙzÙ" (The Mythical Structure of the Nanking Massacre) in Shokun, September 1994, Matsumoto Ken'ichi discusses these two cases, noting, "Nanking is a city where the defending soldiers are always cursed with the fate of being decimated."

The third "Nanking Incident" was the attack on foreign legations by the Nationalist Army under Chiang Kai-shek in March 1927. The Japanese consulate was one of the targets, and because the soldiers on guard did not resist, a Japanese naval officer attempted to ritually disembowel himself out of indignation. This event began the process that put an end to Foreign Minister Shidehara KijšrÙ's diplomacy of cooperation with the Western powers in dealing with China. When Japanese people talked about "the Nanking Incident" before World War II, it was this 1927 attack they had in mind.

The fourth event was the carnage left behind by the Imperial Japanese Army when it ran over the city in December 1937. This is commonly known as "the Nanking atrocities" or "the rape of Nanking." In this paper I will use the former term.

During its eight years of war with China from July 1937 on, the Japanese committed many other barbarous acts in mainland China. But the Nanking atrocities represent and symbolize all of this savagery. Huge volumes of material have been written about this in books and articles in many countries, especially Japan, China, Taiwan, and the English-speaking world. Because of this quantity and the variety of interpretations, the average reader may find it difficult to judge which accounts are reliable. The difficulty is compounded by the fact that political propaganda, commercialism, and humanism are all mixed together in the debate.

Those in Japan who discuss this issue fall into three categories based on political orientation and their assessment of the number of Chinese killed: radicals (the "massacre" faction), conservatives (the "illusion" faction), and moderates (the "in-between" faction). They have been heatedly debating this matter for a long time. Yoshida Takashi of Columbia University uses the terms conventionalist and revisionist for the radicals and conservatives, respectively, but he does not seem yet to have come up with a term for the moderates.

I am categorized as belonging to the moderate faction. One might think that this in-between position would be the most comfortable, but in fact it is the least so: For one thing, we are subject to attack from both the other factions. For another, our numbers are small among the historians concerned with this issue, most of whom espouse either the radical or the conservative viewpoint.

Now I will turn to an analysis of the figures for the number of Chinese killed in the Nanking atrocities.1 First let us specify the time and place. I define these in accordance with the proceedings of the International Military Tribunal for the Far East (commonly called the Tokyo Trial), where the prosecutors claimed that for about six weeks after Nanking fell on December 13, 1937, Japanese soldiers committed illegal killings of Chinese soldiers and civilians in the city of Nanking and adjacent suburban areas in violation of the Hague Convention of 1907. This definition includes soldiers who were executed after being captured but does not cover those killed in action.

This standard is not always used in the interpretations of the numbers killed in the Nanking atrocities. Table 1 shows the main numerical interpretations. The numbers vary widely, ranging from "very few" to a million. This wide variation is in part a reflection of the fact that I have deliberately included even some highly unreliable figures for reference.

The figure of a million, for example, appeared in the "Nanking Massacre" entry of the 1996 edition of a general reference work compiled under the supervision of a professor of European history at the University of Tokyo. When a newspaper reporter asked him what evidence he had for this figure, he replied, "I recall finding this number written in a document I saw when I went to China. . . . If it's wrong I'll correct it in the next edition." This number will therefore probably disappear from the next edition of this work, which is revised annually.2

The "very few" claimed by the deniers is also troubling. Whenever asked about the numbers killed, these people never give any specific figure. Perhaps they fear that, given the contemporary mind-set, an acknowledgement that 200 civilians, or even 20, had been slaughtered would be taken as meaning that a "massacre" had occurred.

Furthermore, even some of the relatively reliable-seeming figures appear to count the toll without considering the cause of death. In other words, they fail to distinguish between deaths actually attributable to the atrocities and those that merely happened to occur at the same time.

The wide variation in the numbers killed leads to the fundamental question: Is it impossible to come up with even an approximate count of the victims in and around Nanking? Before discussing this issue, I will examine the merits and demerits of four methods used to count the numbers killed.


Oral history: The first method relies on the oral history of the survivors of the Nanking atrocities. Interviews with survivors were hastily conducted by the Nationalist government when it returned to Nanking in the autumn of 1945, eight years after the event. The accounts were compiled for use in the Tokyo Trial.

For example, in describing the mass slaughter that occurred on the banks of the Yangtze (Yangzi) River on December 18, an eyewitness named Lu Su declared:

The Japanese Army encircled four or five villages located at the bottom of the mountain, confining 57,418 men and women, including refugees and Nationalist soldiers. . . . On the night of December 18, the captives were bound together in pairs with wire ropes, formed into four rows, and marched to the riverbank. There they were machine-gunned and bayoneted, their bodies doused with oil and burned, and the remains dumped into the river. I was hit in the leg by a stray bullet but managed to hide in a cave and saw the dreadful scene from close at hand.3

This account was accepted as evidence in the Tokyo Trial, and it always appears in Chinese documents as the largest mass slaughter during the Nanking atrocities. But no explanation has been offered of how the witness, while concealed in a cave, managed to come up with the number 57,418.

Other eyewitnesses offered accounts of massacres observed under similar conditions, and when these figures are added together they total 190,000, a figure that should probably be considered a substantial exaggeration. But the judges at the Tokyo Trial apparently accepted this total; possibly they rounded it up to produce their own conclusion that 200,000 people were killed in the atrocities.

Burial records: The Red Swastika Society, a charitable organization that was operating with the approval of both the Japanese occupiers and the International Safety Zone Committee, reported having buried 40,000 people. Another charitable group, which was called the Tsun-shan-tang but whose history is not well known, said it buried 110,000 bodies. The sum of these figures is 150,000. The average daily figure for the Red Swastika was 320 burials, and the average for the Tsun-shan-tang was 75 through March 1938. But in a three-week period of April, the latter society claimed to have buried an additional 105,000 corpses, or a staggering 5,000 per day; this is close to an impossible feat. I surmise that this group operated as a "subcontractor" of the Red Swastika and judge its count to be unreliable. Because the two charity organizations probably overlapped in their responsibilities at some of the burial sites, at least some of the corpses are likely to have been counted twice. Also, burials would have included those of soldiers killed in action and civilians who died either of illness or from being caught in the crossfire.

The Nanking war crimes trial of 1947 took the sum of the two organizations' burials as the total number of individual killings. This was added to the aforementioned death toll from mass slaughters, producing an estimate of 300,000-340,000 people killed. This dubious estimate is the basis of the figure of around 300,000 that is currently cited by the Chinese.

The Japanese historian Hora Tomio used the figure of 150,000 for burials and, adding numbers for other deaths, came up with a combined total of 200,000.4 Professor Hora's number seems to be accepted by most of those belonging to Japan's radical (massacre) faction. Ienaga SaburÙ has written "several hundred thousand." Recently Kasahara Tokushi has come out with an assessment of "at least one hundred and several tens of thousands," the vagueness of this wording perhaps being motivated by consideration of the position the Chinese government has staked out.5

Approximation based on data sampling: In June 1938 a report titled "War Damage in the Nanking Area, December 1937 to March 1938," was submitted to the International Safety Zone Committee by Professor Lewis S. C. Smythe of Nanking University. The weakness of this report is that it uses extrapolated figures. It was based on the result of a survey carried out with visits to one out of every 50 households in Nanking and one out of every 10 households in the countryside around the city.

The Smythe report concluded that 2,400 people died of either rape or assault in Nanking and 4,200 people were abducted. In the nearby Chiang-nin (Jiangnin) County, 9,160 died of either rape or assault, and in all five counties surveyed the total deaths from such causes came to 26,870. One issue with this report is the appropriateness of its geographical extent. But more to the point, the fact that it suggests fatalities of under 30,000 has caused it to be used as no more than an incidental reference by researchers.

Japanese Army field reports: If the central command of the Imperial Japanese Army or of the military police had ordered an investigation of the entire incident in Nanking, and if a report based on such an investigation had remained in existence, it would offer definitive evidence. But despite over 30 years of earnest searching by other researchers and me, no such report has been found. It appears likely that no investigation was conducted.

Both the veterans' group KaikÙsha and I accordingly decided to shift our attention to a search for the field reports of the units involved. We managed to find reports from 16 of the 56 battalions directly involved in the battle for Nanking--in other words, just under 30% of the total. These documents of course do not use the word "massacre." But they record, as part of their military operations, the "annihilation" of the remnants of the defeated army, including soldiers who had changed into civilian clothes (a common practice in the Nationalist Army), and the "execution of prisoners." One reason such records were kept was to serve as future reference for the granting of medals. If those keeping them had had any sense that these acts were illegal killings, they would naturally not have put them down in writing.

Fujiwara Akira has calculated that these field reports record the killing of 12,921 Chinese soldiers who were either prisoners or remnants of the defeated army.6 The figure for the Japanese Army as a whole can only be estimated by extrapolation. This is not such a simple task, however. A full 60% of the 12,921 killings recorded were carried out in two incidents involving just two units, namely, the execution of prisoners by the Yamada Detachment and the extirpation of those thought to be soldiers in civilian clothing in the International Safety Zone conducted by the Seventh Infantry Regiment. It is hard to reach a consensus on how the actions of the recorded battalions should be extrapolated to the battalions whose field reports cannot be found.

Using the existing reports and adding in various estimations, I have come up with a figure of 40,000 for the total of soldiers and civilians killed; the KaikÙsha group has estimated a total of 32,000.7 But using the same approach, Itakura Yoshiaki has produced a lower estimate of 13,000-19,000.8 Hora has also done work using the field reports, and he has come out with an interim report suggesting a much larger number in the range of 110,000-120,000,9 but he has not taken back his earlier assertion that over 200,000 people were massacred.

Apart from such estimates, we also have the estimate of "very few" that I have listed in the table for Tanaka Masaaki, who may be taken as representative of the conservative faction. Those belonging to this group focus most of their energy on criticizing or denying the figures released by the Chinese and by both the radical and the moderate camps in Japan. Their most distinctive feature is that they will never acknowledge any specific number of killings, even when pressed. But if one reads their works carefully, one discovers that they do not completely deny the occurrence of atrocities; rather they deny that it was a massacre, admitting only that some of the soldiers committed individual crimes. I have therefore listed not "zero" but "very few" in Tanaka's entry. But I will not dwell any further on this faction.


Each of these four methods provides only incomplete data, and even combined they cannot give a definitive answer on the numbers killed. I would like to pursue this issue further by taking into account two supplementary factors: Nanking's population and the number of casualties among the Chinese garrisoned in the city. These factors merit attention because they relate to two propositions, namely, that it is impossible for the number of people killed to have exceeded the city's population and that it is implausible that the number of dead civilians would have exceeded the number of dead soldiers.

According to the official figures released by the municipal authorities of Nanking, the population of the city was 1,006,968 in December 1936 and 308,546 in August 1938. No figures exist for December 1937, when Nanking fell. Rabe of the International Safety Zone Committee estimated the population in the zone at 200,000-250,000 but did not touch on the population of the city as a whole. In fact it is probably not possible to come up with a reasonably accurate number because the inflows and outflows of people around this time were so great. Outflows were produced by the flight from Nanking of most of the senior and middle-level bureaucrats of the Nationalist government, businessmen, and other citizens belonging to the middle and upper classes. Inflows were produced by the arrival of many refugees from nearby areas and Chinese soldiers either retreating from Shanghai or brought in to reinforce the Nanking garrison.

Based on two documents, the official military history compiled by the Nationalist government on Taiwan and the diary of Major General Iinuma Mamoru, who was chief of staff of the Shanghai Expeditionary Army, I estimate that the Nanking garrison consisted of 100,000 soldiers; I also estimate that 50,000 were killed in action and that 30,000, after being taken prisoner, were executed. Recently, however, Sun Zhaiwei (of the Institute of History at the Jiangsu Academy of Social Sciences) and Kasahara Tokushi have estimated the size of the garrison as 150,000, of whom 80,000-100,000 were killed.

For the moment, let us accept their estimate and say that up to 100,000 soldiers died, perhaps 50,000 in action and another 50,000 in violation of the Hague Convention. It follows that out of the 300,000 killings officially attributed to the Nanking atrocities by the Chinese, at least 200,000 must have been murders of civilians. To posit such a high civilian death toll is, to say the least, unnatural. It would imply that the Japanese Army exterminated virtually all the refugees in the International Safety Zone. Even if we discount the 300,000 total, setting it at 200,000, we are still left with a civilian toll of 100,000, meaning that the Japanese Army illegally slew twice as many civilians as soldiers.

If we instead estimate the civilian toll at 40,000-60,000, we can avoid such implausible conclusions. This estimate corresponds to the figures from three sources. One is the Red Army's official journal of the time, Hangdibao, though the sources for this are not known. Another is that of Miner Searle Bates of the International Safety Zone Committee. And the third is the aforementioned figure reported by Rabe in his letter to Hitler.

The judges at the Tokyo Trial rejected the death toll figure of 430,000 submitted by the Chinese prosecutors. Although they took the total toll to have been at least 200,000, the only breakdown they provided was of 50,000 soldiers and 12,000 civilians, or a total of 62,000; they did not attempt to account for the remaining 140,000. It thus seems that the court judged 62,000 to be the maximum number of killings for which there was strong evidence.

In sentencing General Matsui Iwane to death, the court stated that 100,000 were killed. Although this number obviously differs from 200,000, the court may have judged that 100,000 was enough to justify the sentence and hence did not see the need to pursue the number of victims any further.

We can also find discrepancies in some of the figures cited in Chinese works. For example, the book Qin-Hua Rigun baoxing zonglu (Complete Record of the Atrocities Committed by the Japanese Army Invading China), which was published in Beijing in 1995, gives the number killed in the atrocities as 300,000, but the figures for the eight mass slaughters that are detailed in the book and documented by sources in the Chinese National Archives (Nanjing) add up only to 110,000-130,000, far short even of the 190,000 figure produced by the Nationalist government in 1945. Among these eight cases, the largest massacre was 57,418, as noted earlier, and the smallest was 2,000. Even if smaller mass slaughters and individual killings are added to these cases, the total is nowhere near 300,000. Meanwhile, in the decision handed down by the Nanking war crimes trial, the figure of 190,000 was used as the total for the mass slaughters. As for individual killings, the court claimed that the Japanese Army had massacred 150,000 people. It seems to have taken this number from the aforementioned burial records.

Why do the Chinese stick to the 300,000 figure for the number of people killed in the atrocities? I surmise that, as Bruce Cumings has suggested, this number has become a political symbol or metaphor.10 People like Allen Whiting say that this number is used to give the impression that the Nanking atrocities were far worse than the atomic bombings of Hiroshima and Nagasaki.11 According to a report submitted to the United Nations secretary general by the mayors of Hiroshima and Nagasaki, the atomic explosions in these two cities killed 140,000 and 70,000 respectively. Perhaps Whiting is right.


Estimated numbers killed
Miner Searle Bates (member of the International Safety Zone Committee for Nanking), information provided to Harold Timperley, as reported to the International Military Tribunal for the Far East (Tokyo Trial)January 1938
Ku Wei-ch¸n (Speech of the Chinese delegation to the League of Nations)February 2, 1938Ý
Hangdibao (official journal of the Red Army), found in the Marco Polo Bridge MuseumApril 30, 1938ÝÝ
John Rabe (German head of the International Safety Zone Committee for Nanking), letter to Hitler after returning to GermanyJune 1938Ý
Why We Fight: Battle of China(Movie by Frank Capra)1943Ý
Nanking war crimes trial verdict (incl. 190,000 mass slaughter deaths and 150,000 individual killings)March 10, 1947ÝÝ
Tokyo Trial verdict for General MatsuiNovember 1948ÝÝ
Tokyo Trial final judgmentNovember 1948
History of the Sino-Japanese War(translation of official history produced by the Nationalist government in Taiwan; Taipei, p. 213)1971ÝÝ
History Department, Nanking University1979ÝÝ
Hora Tomio, Nankin daigyakusatsu (The Great Nanking Massacre), rev. ed.1982ÝÝ
Tanaka Masaaki, “Nankin gyakusatsu” no kyokÙ (“The Nanking Massacre” as Fabrication)1984ÝÝ
very few
Hata Ikuhiko, Nankin jiken (The Nanking Atrocities)1986
KaikÙsha, Nankin senshi (History of Nanking Military Operations)1989
Itakura Yoshiaki, “Nankin jiken no sšjiteki kenkyš (A Numerical Study of the Nanking Incident),”Gunji Shigaku, vol. 26, no. 11990
Unemoto Masami, as quoted by Itakura (above)1990ÝÝ
Kabayama KÙichi, Gendai yÙgo no kiso chishiki (Basic Knowledge About Contemporary Terms)1996ÝÝ
Œsaka Shoseki (publisher), Japanese middle school textbook1996ÝÝ
Memorial Hall of the Victims of the Nanjing Massacre (inscription at main entrance)ÝÝÝ


On December 13, 1997, the sixtieth anniversary of the fall of Nanking, events marking the occasion were held in cities around the world, including Nanjing (as Nanking is now called), Tokyo, Osaka, Taipei, Hong Kong, and San Francisco. In Tokyo a two-day international symposium was convened; the program included words of greeting from Ursula Reinhardt, granddaughter of John Rabe. As noted above, Rabe's German diary was translated into Japanese and published just last autumn, producing complex repercussions with its suggestion that the toll from the Nanking atrocities was 50,000-60,000.

Attending this symposium, I observed an interesting scene. Kasahara Tokushi, obviously considering the sentiments of the Chinese delegates, remarked: "Rabe gives a figure of fifty thousand to sixty thousand, but if we add the numbers for the victims in the areas outside the city that he couldn't observe and those killed after he left Nanking, the total will surely come to about three hundred thousand." To this, Sun Zhaiwei, the ranking Chinese delegate, retorted with words to this effect: "The figure of three hundred thousand refers only to victims within the walls of the city. Don't go arbitrarily expanding the area or time span."

At this point I directed a question to Sun: "What do you think of the story from the New China News Agency, as reported in this morning's daily Yomiuri Shimbun, that the toll was over five hundred thousand civilians and close to one hundred thousand disarmed Chinese soldiers?" "I don't know about that report," Sun replied.

The December 15 issue of the overseas edition of China's Renmin Ribao (People's Daily) carried a story from Jiang Zhi, a New China News Agency correspondent covering this symposium. Jiang reported that "participants honored the memory of the 300,000 Chinese soldiers and civilians who were butchered." Both in Japan and in China, the figures for the numbers of victims appear to be shaped by the political agenda of those who present them.

Let me cite a similar case of changing numbers: At a memorial hall commemorating the Chinese fight against Japan near the Marco Polo Bridge, the figure displayed for the number of victims of the war was initially 21.69 million, including 9.32 million killed. But in 1995 President Jiang Zemin gave an address in Moscow in which the number jumped suddenly to 35 million, and subsequently the memorial hall adopted it. The figure seems to have been produced without consulting experts; nobody has been able to offer a breakdown. It is certainly quite a remarkable jump from the figure announced immediately after the war by the Chinese defense minister, who reported that 1.75 million soldiers had died. Some China watchers tell us with a smile that the reason for such inflation is simply to make the point that "a lot of people were killed."


John Rabe's wartime diary was reportedly discovered by Iris Chang, a 29-year-old American journalist of Chinese extraction. I say "reportedly" because according to a senior member of the Alliance in Memory of Victims of the Nanjing Massacre (a political action group run by Chinese Americans and the host of the conference at Princeton I mentioned at the start), somebody else actually found it. But I will not speculate further on this issue here.

Chang has written a 304-page book titled The Rape of Nanking: The Forgotten Holocaust of World War II (Basic Books, 1997). As she explains in this work, her grandparents had been residents of Nanking, but they found a refuge in the countryside a few months before the city fell. In 1949 her father escaped from the mainland to Taiwan; he moved to the United States, studied at Harvard, and became a professor of physics at the University of Illinois. Chang was born in Illinois and majored in journalism at her father's university. She says she learned about the Nanking atrocities as a small child.

Just what led to the discovery of John Rabe's diary is unclear, but this is how Chang explains the course of events, which read like a spy thriller: Chang got in contact with Ursula Reinhardt, Rabe's granddaughter, and Reinhardt borrowed the diary from her uncle Otto, the current head of the Rabe family, and read it for the first time. After discussing the matter with aPeople's Daily reporter, Chang, and others, Reinhardt decided to make the contents public. Because of concern that Japanese rightists might break in and destroy the document or else try to buy it for a large sum, she made a photocopy and spirited it to the United States, where she entrusted it to the Yale Divinity School library and revealed the diary's existence in a press conference.

Chang commuted to Yale to do research for her book. But evidently she does not read German herself; she used translations provided by Barbara Masin (a friend) and Jeff Heynen (of theAsahi Shimbun). For translations from the Japanese she relied on Sugiyama Satoko, who volunteered her time. It is hardly surprising that the results of Chang's efforts--treatment of a major topic by a woman who reads neither Japanese nor German and who has no connections with any scholar of modern history capable of checking her work--are seriously flawed.

Table 2 lists some of the basic errors that caught my eye as I skimmed through the book. To give one example, on page 40 she refers to "Taisa Isamo"; this is clearly a mistake for what should be "Lieutenant Colonel ChÙ Isamu," a key staff officer who is said to have ordered the large-scale slaughter of Chinese prisoners in defiance of a restraining command from General Matsui Iwane.

The confusion about ChÙ's actual rank of chšsa (lieutenant colonel), which is rendered as taisa (colonel), is a trivial affair. But the assumption that taisa is his surname--analogous to identifying President Bill Clinton as "President Bill"--is not a mistake that anybody born and raised in Japan could possibly make. Somewhere along the line it seems that taisa was taken as the surname, and the given name Isamu got distorted into the nonexistent Isamo. It also appears that the translator was not familiar with military matters in general. For example, the termhohei rentai, "infantry regiment," gets turned into both "battalion" and "wing."

Item 12 in Table 2 concerns the number of victims of the atrocities, which Chang estimates at 260,000-350,000. On page 102 she even goes to the trouble of showing how the 260,000 figure was calculated. But it turns out that this is merely a list of disparate numbers collected from various sources and preserved in the records of the Tokyo Trial; she has reproduced it without a thought to the duplication within it. I have seen similar misuse of statistics before, but still I was astounded at her cavalier approach.

As for the 350,000 figure, she initially attributes this to "some experts" (p. 4). Later she lets us know that the source may be "the authors James Yin and Shi Young," who in a telephone interview gave her the figure of approximately 355,000 (pp. 103, 254). But since she fails to cite any relevant work by them, the validity of the estimate cannot be verified. Meanwhile, she kindly deigned to refer in passing to my own estimate of 38,000-42,000 in the same breath as Rabe's 50,000-60,000 (p. 100). But she does not bother to consider the possible veracity of these figures, even though she took Rabe practically as her guardian angel. I imagine this happened because the numbers struck her as unacceptably low.

For the number of rapes Chang presents an estimate of 20,000-80,000 (p. 6). This is also a very rough set of statistics. When she appeared on a U.S. TV broadcast on January 11 this year, her interviewer asked her the reason for the big gap between the high and low figures. She made no serious attempt to explain this, merely asserting that the nature of the events made it impossible to arrive at an exact figure.

The final item in the table has to do with the photos used to illustrate her book. I shall return to this later.

What surprised me, though, was the warm reception the work received from reviewers and the book-reading public in the United States. I personally came across three reviews, inNewsweek (December 1, 1997; translated in the December 10 Japanese edition), the Washington Post (December 11), and the New York Times (December 14). All were long, with sensational headlines. Thanks to this prominent coverage, the book promptly turned into a bestseller.


Let me introduce the review that appeared in the New York Times, said to be the journalistic guardian of America's conscience. The reviewer, Orville Schell, is identified as dean of the Graduate School of Journalism at the prestigious University of California, Berkeley.

Schell starts out with the bold assertion that the Nanking atrocities have "virtually no parallel in recent history." According to him, "Many experts now believe the number to be over 350,000." He continues, "The carnage was the result of a secret order sent to Japanese forces in China under the seal of Prince Asaka, uncle of Emperor Hirohito: 'Kill all captives.' Soon competitions arose among soldiers to see who could kill most efficiently."

Here Schell has subtly replaced Chang's "some experts" with "many experts." And while Chang herself withholds judgment on Prince Asaka's responsibility, writing, "What is not clear is whether Asaka himself issued the orders [that all captives be killed]" (p. 40), Schell arbitrarily labels him the culprit. Ordinarily one expects a reviewer to discount some of the excesses of the author, but Schell has done the opposite. And the distortion of Chang's message seems to have been not an oversight but deliberate.

Book reviewers often include excerpts that highlight the work. Here is the passage that Schell chose as representative of Chang's book:

Many soldiers went beyond rape to disembowel women, slice off their breasts, nail them alive to walls. Fathers were forced to rape their daughters, and sons their mothers, as other family members watched. Not only did live burials, castration, the carving of organs, and the roasting of people become routine, but more diabolical tortures were practiced, such as hanging people by their tongues on iron hooks or burying people to their waists and watching them get torn apart by German shepherds. (p. 6)

I was curious where the author got the materials for this passage, which details scenes worse than a medieval witch trial. The only documentation she offers, however, is interviews with people said to be survivors.

Chang seems to have a strong taste for the grotesque. In a passage just before the one quoted by Schell, she reproduces these bizarre calculations by "one historian" without thinking the results through:

If the dead from Nanking were to link hands, they would stretch from Nanking to the city of Hangchow, spanning a distance of some two hundred miles. Their blood would weigh twelve hundred tons, and their bodies would fill twenty-five hundred railroad cars. Stacked on top of each other, these bodies would reach the height of a seventy-four-story building. (p. 5)

I pulled out pencil and paper and attempted to verify these calculations. The figure for stretching from Nanking to Hangchow (Hangzhou) seems reasonable if one accepts her estimate of the total toll. But the reference to a "seventy-four-story building" is actually far short of the mark. If we conservatively take the thickness of an average corpse to be 10 centimeters, it would take only 1,000 bodies to reach a height of 100 meters. Am I miscalculating? I wondered, but I was forced to abandon my arithmetic by visions of iron hooks and German shepherds. Even so, I find it amazing that the dean of a graduate school at a top-rate university like Berkeley could take this sort of preposterous fable seriously.

Schell was not the only one to accept Chang's work at face value. The reviews in Newsweek and the Washington Post were similar in their overall appraisal. The Amazon.com on-line bookstore (http://www.amazon.com) has also uploaded a set of reviews from people including well-known university professors and Pulitzer-prize-winning authors, most of them full of comments like "outstanding," "powerful," "meticulously researched," and even "one of the most important books of the twentieth century." Only a couple of critical notes appeared, such as "while rigorous in its moral earnestness, the book is inadequate as a history."

I wondered why so many such luminaries were being so soft on Chang. The explanation that occurred to me was that they may have been swayed by her charisma, which caused a number of people to refer to her presentation as "powerful," and by the force of her feminist rhetoric.

As noted above, last November I attended the "Nanking 1937" conference, and while in Princeton I had the chance to see Iris Chang close up. As I looked around at an audience moved to absolute silence by the overwhelming impact of her speech, it occurred to me that she is rather like Ueno Chizuko, an energetic Japanese feminist scholar.

I thought of asking a couple of questions after her speech, but decided not to. I feared that I might get snapped at for "sexual harassment." Nor did Chang approach me, so we missed having a direct encounter. When "holocaust" is coupled with "rape," neither Americans nor Japanese of any sensitivity can help lapsing into silence. Skillfully using this psychology, Chang has built up a strong anti-Japanese presentation and turned her book into a long-running bestseller.


Next let us consider the 40-odd photographs that Chang selected to illustrate her book, especially the 7 in the "beheading" category and the 4 in the "sexual crimes" category. All have appeared previously in various places and will not surprise experts on the subject, but they are sure to have a considerable impact on those seeing them for the first time.

The sources of the photographs are given in the captions, but in no case is the date given, and in few cases are the photographers identified. In other words, all we have to go on is the names of the organizations from which the pictures (presumably copies) were borrowed, without knowing where the originals came from.

There are some cases, however, in which we can pin down the source of photos that have been in circulation for a long time. For example, one used by Chang shows severed heads in a row on the ground. Chang identifies the heads as belonging to victims of the Nanking atrocities, but it is known that the photo is actually of bandits executed by the Chinese police in 1930.

All in all, the 11 photos in the beheading and sex-crime categories are a combination of fakes, forgeries, and composites; not one of them can be clearly identified as showing Nanking or its environs in 1937. Allow me to explain this assertion with respect to a few of the pictures.

Photograph1Reproduced here as Photograph 1 is a scene labeled in Chang's book as showing a Japanese bayonet drill in Nanking after the capture of the Chinese capital, with Chinese prisoners being gored to death. A group of Japanese soldiers can be seen looking on. "As to the authenticity of the photo," Chang writes, "it was sent to Look by W. A. Farmer of Hankow, who says the photograph was made by a Japanese soldier. The film was sent to Shanghai to be developed. Chinese employees in the Japanese-owned shop made extra prints and smuggled them out (UPI/ Bettmann)."

The journalist Watada Susumu heard a similar story from Wu Shi, a former official of the government Japan installed in Nanking, when he interviewed him in 1987. According to Wu, the photo is one of seven that were developed in Nanking (not Shanghai) and submitted to the Nationalist government shortly after the war; copies of them are now on display in Nanjing's Massacre Memorial Hall. Chang's book reproduces not just Photograph 1 but also another photo from the set of seven, and Wu's story may account for both of them. But do they provide proof of the Nanking atrocities? To conclude that, Watada asserts, "would be very dangerous."12

Looking more closely at this photograph, one sees that 10 or more of the Japanese soldiers standing by are in white shirts with their jackets off. It can also be seen that the shadows cast by the bayonet-wielding soldiers are too short for midwinter. This is obviously not a photo taken in Nanking in December or January. Even if the photo itself is authentic, it clearly shows a scene from a time and place other than Nanking after its fall. But despite its lack of credibility as evidence for the Nanking atrocities, the New York Times, the Washington Post, andNewsweek all used it to illustrate their reviews of Chang's book, perhaps because its impact is so strong.

Consider next Photograph 2-a, which appears in Chang's Rape of Nanking with this caption: "Japanese soldiers sometimes forced their victims to pose in pornographic pictures, which were kept as souvenirs of rape." It is said to have come from the family of George A. Fitch, an American on the committee running the International Safety Zone. I have seen several versions of this photo. One was in Shashinshš: Nankin daigyakusatsu (An Album: The Great Nanking Massacre), a collection of photos put out by Japan's Elpis Co. in 1995. It and most of the other photos in the same album are identified only as having come from the Massacre Memorial Hall.

Looking over several reproductions of this photo, I noticed some interesting discrepancies. Photograph 2-b, which I came across in a book I acquired in Taiwan, differs from the others in that it shows a Chinese-looking man standing on the right.13 Also, if we closely examine the soldier in the center, we find that he seems to be wearing a civilian jacket and that neither his cap nor his facial features look very Japanese.

This reminded me of the memoirs of Sasaki Motokatsu, a high official of the field postal services assigned to the Nanking operation. Sasaki wrote that on November 22, 1937, he was told by a military policeman assigned to the censoring of homebound mail that sadistic and pornographic photos were in wide circulation and that while their origin was unknown, some believed they were fakes created in Japan using people wearing wigs. Sasaki writes that he had a look, and "one of them showed a young Chinese woman standing crying, her body naked below the waist. But it is not known whether it was taken by a Chinese or Japanese soldier."14

The military policeman informed Sasaki that while the censors had instructed them to burn all the lewd photos they found in mail for Japan, he could help himself to any that he wanted. We cannot be sure that the photos shown here as 2-a and 2-b correspond exactly to the one Sasaki mentions being shown. But it seems apparent that such pictures were in wide circulation as souvenirs at the time, and it would not be the least bit extraordinary for a convenient caption to be attached after the fact.

Photograph3-cA more serious misrepresentation can be found in the photographs numbered 3-a, 3-b, and 3-c. Fortunately the details of the trickery involved are known, so I shall explain them. First of all, here is Chang's caption for Photograph 3-a: "The Japanese rounded up thousands of women. Most of them were gang raped or forced into military prostitution." Her photo credit reads "Politburo of Military Committee, Taipei."

At least three other works include this same scene: (1) the book from Taiwan in which I found Photograph 2-b, (2) the mainland Chinese record of the war with Japan I have mentioned (see note 3), and (3) Kasahara Tokushi's Nankin jiken (The Nanking Incident) (Tokyo: Iwanami Shoten, 1997). They all describe it in similar terms, but only Kasahara explicitly identifies the source. His caption reads: "Chinese women in the Chiang-nan [Jiangnan, south of the Yangtze River] region being taken away by Japanese soldiers," and he says the photo is from a work published in 1938 by the politburo of the Nationalist military. From this it appears that Chang's photo and his, shown here as Photograph 3-b, are of the same provenance. What I found suspicious, though, is that the trimming and resolution of the two versions differ.

As it turns out, what is probably the original version of both photos was carried in the November 10, 1937, issue of the illustrated magazine Asahi Gurafu. I present it here as Photograph 3-c. Kasahara used the same trimming, and so did the book I picked up in Taiwan. But the right side of the picture, which shows an elderly woman pulling a cart, was cropped out of the mainland Chinese record, and it is also absent in the photo in Chang's book (3-a). The more recent versions show reduced resolution; this is only natural. But what I found peculiar is that Kasahara's version alone shows blurring as if by halation in the spot that appears to show a young woman next to the girl at the head of the group on the left-hand side.

The general mood of the group seems altogether too relaxed for a body of women being hauled away to be military prostitutes. One woman is carrying a baby, and at least four children are present, including a boy with a smile on his face. The soldiers are carrying rifles on their shoulders, as they would when on the march, but no bayonets are attached.

Actually it was only last January that I found a solution to this set of photographs, which had puzzled me for quite a while. It was then that I discovered that the original had appeared in theAsahi Gurafu (a sister journal to the daily Asahi Shimbun). It was also carried in a three-volume work titled Shina jihen shashin zenshš (Pictures of the China Incident) published by the Asahi organization in March 1938. It is one of a set of four photos credited to correspondent Kumasaki Tamaki; in the magazine they appeared under the headline "Utopia Amidst the Gunpowder: The 'Rising Sun' Village South of the Yangtze."

The text notes that this photo and the others in the same set were taken in the province of Paoshan (Baoshan), near Shanghai, and offers the following overall explanation: "There are two villages where peace returned thanks to protection by our army. One is Sheng-chia-chiao, known as the 'Rising Sun Village.' . . . The approximately 400 villagers of this 'utopia' are free from the depredations of remnants of the defeated Chinese soldiery and commune with the earth, their minds at ease, as a result of the protection provided by our army."

Cotton was a major local crop, and one photo in the series shows a group of villagers, young and old, singing as they picked cotton under the autumn sun, escorted by Japanese soldiers. The original photo, whose later versions were used by Chang and others to illustrate the rounding up of women by the Japanese military, carries this caption: "A group of women and children from the Rising Sun Village returning from the fields, guarded by our soldiers."

Reexamining the scene after reading this caption, I found that, sure enough, the little girl at the front and the boy shown walking next to a soldier are smiling. Another soldier has his hand on the boy's shoulder. We can also see that the cart which a plump woman is pulling at the far right of the picture contains a load of cotton. When I tried contacting the photographer Kumasaki through the Asahi's association of retired employees, I learned that, unfortunately, he passed away three years ago at the age of 80. But I am certain that the scene was exactly as he described it.

Who, then, is the culprit who rewrote the caption, turning it into the complete opposite of the original? As far as Photograph 3-b is concerned, Kasahara made a public admission of "deep regret" for misusing the original to the Asahi organization and the late cameraman Kumasaki; this apology was carried in the April 1998 issue of Tosho, a public relations journal of the publisher Iwanami Shoten. On the same page Iwanami, in a message to its readers, promised to replace the photo with another one in its next edition of Kasahara's book on the Nanking atrocities. Sad to say, though, trickery with wartime photos of China is seen even in publicly funded museums around the country. And the media, which is so quick to pounce on errors of other sorts, is strangely quiet about these distortions.

David Bergamini, author of Japan's Imperial Conspiracy, laid out a fanciful scheme of a Japanese blueprint for world conquest. Chang remarked critically on this work in her book even as she appeared to trust it. Seen from a distance, she seems a second Bergamini.


1. “Taisa Isamo” (p. 40)Lieutenant Colonel ChÙ Isamu (See text.)
2. “A hidden Chinese artillery emplacement opened fire and killed several hundred men, including a cousin of the Empress Nagako [in August 1937].” (p. 33)No such fact
3. “[General Nakajima was formerly] chief of the Japanese secret police for Emperor Hirohito.” (p. 37)Former head of the military police
4. “Tani Hisao, .Ý.Ý. commander of the Japanese 6th Division, was later found guilty of raping some twenty women in Nanking.” (p. 50)Not mentioned in the indictment or verdict
5. “[Doctor] Nagatomi Hakudo spoke candidly about his emotions in the fallen capital.Ý.Ý.Ý. Patients can watch videotapes of his trial in Nanking and a full confession of his crimes.” (p. 59)Nagatomi’s given name is “Hiromichi.” A student at the junior college Kokushikan, he visited Nanking in early 1938. He is now an acupuncturist.
6. “There is even compelling evidence that the Japanese themselves believed [that] the death toll at Nanking may have been as high as 300,000” (referring to a January 17, 1938, message to Washington from Foreign Minister Hirota Koki). (p. 103)The message only quoted (in English) a censored cable to the home office of the Manchester Guardian from the correspondent Harold Timperley.
7. “For three days starting on December 15, 1937, Ohta’s army unit dumped . . . a total of some 150,000 bodies [of Chinese victims into the Nanking River]” (according to Ohta’s confession at a war criminal camp in 1954). (p. 101)The confession of Major Œta was forced; he was assigned to Nanking on December 25, 1937.
8. “[Japanese children in cram elementary schools] study from 9:00 P.M. to 6:00 A.M.” (p. 205)Meaning unclear
9. “Screeners at the ministry [of Education] ordered textbook authors to eliminate all references to the numbers of Chinese killed.” (p. 208)Of the seven currently approved middle school history texts, six give figures from 150,000 to 300,000 in acknowledgement of Chinese assertions; one says “many” were killed.
10. “official denials of wartime atrocities [by the Japanese government]” (p. 224)The Japanese government has never denied the atrocities in Nanking and other places. In the San Francisco Peace Treaty, it accepted the judgment of the Tokyo Trial.
11. “The Japanese government needs to issue an official apology to the victims.” (p. 225)The emperor, the prime minister, and other representatives of the Japanese government have issued apologies to victims of the war in China, South Korea, and elsewhere.
12. Civilian deaths in the Rape of Nanking totaled 260,000‚350,000; 20,000‚80,000 women were raped. (pp. 4, 6, 102)See text.
13. Eleven photos show acts of sadism and sex crimes by the Japanese.See text.

1. This presentation is based in part on my book Nankin jiken (The Nanking Atrocities), published by ChšÙ KÙron Sha in 1986; a Chinese translation of this work was published in Hong Kong in 1995 as Nanjing da tusha zhenxiang by Shangwu Yinshukuan.
2. The entry in question has been deleted from the following years' editions.--Ed.
3. Qin-Hua Rigun baoxing zonglu (Complete Record of the Atrocities Committed by the Japanese Army Invading China) (Beijing: Hebei Renmin Chubanshe, 1995), p. 645.
4. Hora Tomio, Ketteiban Nankin daigyakusatsu (The Definitive Account of the Great Nanking Massacre) (Tokyo: Tokuma Shoten, 1982), p. 150. The March 1995 issue of the journal Rekishi Chiri KyÙiku carried an article by Fujiwara Akira in total support of Hora's conclusion (p. 72). This opinion is also shared by Kasahara Tokushi; see his article in the September 1987 issue of Rekishigaku Kenkyš.
5. Kasahara Tokushi, "Hakken sareta Nankin gyakusatsu no shÙgen" (Testimony to the Nanking Massacre Discovered), Sekai, September 1997, p. 159.
6. Fujiwara, op. cit., p. 67.
7. KaikÙsha, Nankin senshi (History of Nanking Military Operations) (Tokyo, 1989) pp. 366, 374.
8. Itakura Yoshiaki, "Nankin jiken no sšjiteki kenkyš (A Numerical Study of the Nanking Incident)," Gunji Shigaku, vol. 26, no. 1 (1990), p. 68.
9. In the third edition of his Nankin daigyakusatsu no shÙmei (Proving the Great Nanking Massacre) (Tokyo: Asahi Shimbun, 1988), Hora gives the figure 110,000 (see the table on pp. 336-37). In the forthcoming fourth edition he gives 120,000.
10. See Cumings's comment on p. 270 of Kojima Noboru, Jinrui wa sensÙ o fusegeru ka (Can Humankind Prevent War?) (Tokyo: Bungei Shunjš, 1996).
11. Allen S. Whiting, China Eyes Japan (Berkeley: University of California Press, 1989), p. 187
12. Hora Tomio, Fujiwara Akira, Honda Katsuichi, eds., Nankin daigyakusatsu no genba e (Visiting the Site of the Great Nanking Massacre) (Tokyo: Asahi Shimbun, 1988), pp. 217-22.
13. T'ieh-cheng ju shan (A Mountain of Ironclad Evidence) (Taipei: Hsien-tai Chung-kuo She, 1982), p. 15.
14. Sasaki Motokatsu, Yasen yšbinki (The Military Field Postal Banner) (Tokyo: Gendaishi Shuppankai, 1973), p. 178.

Part I is a slightly revised version of a paper presented at the "Nanking 1937" conference sponsored by the Alliance in Memory of Victims of the Nanjing Massacre, held at Princeton University, November 22, 1997. Part II is translated from "Nankin gyakusatsu 'shÙko shashin' o kantei suru," Shokun, April 1998, pp. 82-93; abridged by about one-third, with minor additions from the author. (Courtesy of Bungei Shunjš)

© 1998 Japan Echo Inc.

HATA Ikuhiko
Born in 1932. Received his doctorate in modern history from the University of Tokyo. Has been a research associate at Harvard University, a senior fellow at Columbia University, and a visiting scholar at the University of Michigan and Princeton University. Is now a professor at Nihon University. Author of Nankin jiken (The Nanking Atrocities), Shôwashi no nazo o ou (The Riddles of Shôwa-era History), and other works.