Si je me concentre sur l'écriture de critiques, j'ai eu l'occasion, à l'été 2011, de publier l'article ci-dessous sur une bataille entre japonais et chinois fin 1943. Le texte original est en anglais (il n'y a pas de version française) et a paru dans Battles Magazine, en accompagnement d'un wargame sur la bataille. Les cartes sont copyright Olivier Revenu.


In the literature of World War 2, the war between China and Japan is the least studied, the great unknown war that is hardly given a paragraph in history books. What happened on the continent beyond the initial campaigns of 1937-38 and the rape of Nanking is uncharted territory. From 1941 onwards, Western historians naturally focus on the Pacific theatre. Should China be mentioned, it is in relation to the British or the Americans, or what Westerners did there. Consider the level of detail we get on this sideshow of Burma compared to Ichigo, the greatest campaign ever conducted by the Japanese Army. Consider the controversies around the Stilwell - Chiang relationship compared to any investigation of Chiang - Mao diplomacy. Consider the multiples of books about the handful of Flying Tigers compared to the vacuum on field armies of millions of men. Even the Mandchurian operations opposing Japan and the URSS, both in 1938-39 and in 1945, are a wealth of documents compared to the Chinese theatre. Anything that relates to a third party has been decently investigated, yet not Japan-China war proper.

The war between Japan and China war is a controversial topic among its actors. The Chinese Army did not beat the Japan Imperial Army and did not liberate its occupied territory. Although allied with the winners, Chinese contribution to the final victory was not, to put it mildly, major compared to the contributions of Russia, America, Britain, Canada or Australia. A Chinese history of the war could not be glorious. Moreover, the bulk of the conventional fighting involved Chiang’s Nationalist troops, expulsed from China in 1948. Taiwanese historians may not have the best access to continental sources, while Communist Chinese may not want to praise fighting by the Nationalists too loudly. Japanese historians are well aware of the crimes committed by the IJA in China, and hence do not focus on their seldomly defeated military operations. The war also had its sequel - the 1945-48 civil war – and this dispersed documents and witnesses, so sources are scarce and incomplete. Military history of the Japan-China war still awaits an objective and fact-based review of the military operations, which would probably uncover countless of “forgotten battles”, just like David M. Glantz found on the Eastern Front.

And yet, so many questions, so many topics to explore! What was the actual performance of the Japanese and of both Chinese opponents? What was Japan’s actual commitment to China and its impact on the Pacific war (latest research suggests China did not divert any assets from the Pacific; quite the opposite happened: the Pacific war distracted assets from China)? What economic importance had China within Japan’s war effort? Were the operations actually led by Tokyo or triggered by local generals? How could Japan put together an operation involving half a million troops as late as 1944 and what did it actually achieve? What impact did American lend lease and bombers have? From how much “lack of will to attack” did Chiang suffer and what could he have really done differently? What lessons can be drawn from the guerilla war and how relevant could have they been for the wars of decolonization of the 50’s and the 60’s? How much has this war - its crimes as well as its military - shaped memory of Taiwan, China and Japan?

Before I focus on the battle, let me underline that, while the available documentation on the various campaigns of this war is thin, it borders on a total void for the 1943 Changde campaign. In the specialized English literature, the campaign, if mentioned at all, is hardly given a paragraph. Only in one book – 1971’s History of the Sino-Japanese War – can one find a map, an order of battle, and a general overview of the battle; but this Taiwanese source, written in broken English, has such a Chinese bias that any reported fact needs be taken with a grain of salt. I am not aware of Chinese books focusing on this campaign (but let me admit that my Chinese language skills could be improved), but I did spot a Japanese source, published in 1983 and long out-of-print. On the web, besides the contradicting stances in Chinese and Japanese Wikipedia articles, a deep search would give almost no additional data. The 2010 movie Death and Glory in Changde may trigger some serious historical work though.

The war of attrition in Central China


In 1943, the Japan-China war had been a war of attrition for a few years. The frontline had not significantly changed since 1939. In spite of all the territory ceded to Japan and all the military losses, the Chinese government has showed no inclination for surrender. The Japanese had hunted the Nationalist government in Nanking, then in Wuhan, but it eventually relocated to distant Chunking. In 1940, Japan tried a major bombing campaign against Chunking, only to realize it had had no impact on Chinese will to fight. So Japan tried another way. It put a puppet government in Nanking in 1940 to isolate Chiang Kai-shek politically and decided that “peace and order” should be its primary goal in China. Destruction of the Nationalist would come from “a long time siege” and the army would avoid unnecessary expansion of the frontline, since it lacked means to control the occupied territory against guerrillas.

In 1943, the closest frontline was 600 km and one mountain chain away from Chunking. The general staff of the Imperial Japanese Army (IJA) in China had studied what it would take to get to Chunking and concluded it did not have enough strength to get there: plans were shelved end of 1942. Moreover, the IJA had faced stiff Chinese resistance during its limited operations in central China since 1939. The Chinese even put together a large-scale counter-offensive in December 1939 that showed how vulnerable the widely dispersed Japanese could be.

It was also in 1943 that Japanese land assets first started being diverted from China to the Pacific. Most of the airforce had focused on the Pacific as soon as the war against the USA had broken out. Given the weakness of Chinese airforce, that did not create serious military issues in China, at least until the first American long-range bombers began operating from the country in late 1943. By the time the Changde operations kicked off, American-supported airforce were not a major nuisance yet, and Japan had yet to realize what threat they could eventually become. However, following Guadacanal, Japan started reinforcing Pacific islands with units from the continent. Some experienced divisions had been relocated from China to such distant place as the Mariannas. Priorities had changed: the Japanese army lost their reserve, and so its ability to conduct strategic penetrations into Chinese held territory.

The most powerful component of the Japanese in China, and the only operationally effective force, was the 11th Army, located in the central part of the country. Centered on Wuhan, it held a front facing south and west along the Yangtze river. The 11th Army has been relatively active since 1939. It had extended its grip along the Yangtze up to Yichang, the furthest point reachable by large ships, at the gates of the mountains leading to Chunking. It had also conducted a series of offensive operations, the most famous being the campaigns against Changsha. Since permanent occupation of conquered territory was not possible because of insufficient manpower and overextended supply lines, the primary aim of the 11th Army was to forestall any enemy counteroffensive and to keep up army morale. The 11th Army’s typical pattern was an advance to capture a limited objective, followed by a retreat to its starting positions. Even if it captured important points (such as Changsha), the 11th Army had to abandon them and return to its lines. The Chinese were quick to adapt, first by refusing fight in open terrain and retreating to fortified positions, and then by re-occupying territory as soon as Japanese left. The Changde campaign is a variation on this theme.

The Changde campaign: context and operating plans


Changde itself is located west of the large Dongting lake in the Hunnan province. The city was approximately 40km south-west of the frontline, where Japanese had secured a 10km deep foothold on the south bank of the Yangtze river. Changde belonged to Chinese 6th war zone, which also included Changsha, 170 km in the south-east, on the other side of Dongting lake. Although Chunking was right 400km west of Changde, no valley or road linked the cities. The terrain is rather flat but divided by sub rivers of the Yangtze (running north-south) and by numerous small lakes. Two large rivers that feed the Dongting lake ran west-east: the Li, north of Changde, that formed the main path of attack, and the Yuan through Changde and into the lake.

By late 1941, Changde was the site of an intense bacteriological attack. It is not clear why Japan picked this location to test its infamous weapons, but it certainly did not intend to follow-up with a land invasion. Airplanes dropped material infected with plague on the city, causing sizeable epidemics in March 1942. Chinese historians have conducted deep investigations into the matter, finding more than 7,600 confirmed cases of deaths linked to the bacteriological attack. The epidemic did not seem to have significant impact on military operations, however.

The Japanese 11th Army, led by general Yokoyama, launched two campaigns west on the Yangtse in 1943. A first operation involved six divisions in May 1943, between the Yangtse and the Li river. The 11th Army had previously attacked south (towards Changsha) and this move west surprised the Chinese. The operation seemed to be an attempt to test Chinese strength in the area and to disrupt any offensive preparation, rather than to secure any long-term objective. The fighting lasted a few weeks after which Japanese returned to their starting positions. During the Changde campaign, the IJA would fight along the Li river again, conquering positions it had occupied and evacuated in May.

General Yokoyama decided to mount another operation in the area in November 1943. This time, target was to destroy Chinese capabilities by seizing Changde. The plan was not to hold the city indefinitely, and some sources say Japanese had orders to leave the place after three days. Other sources mention another objective: to distract Chinese units from Burma. Although this could have been an argument to convince Tokyo, operational reality (time needed to move units from southern to central China and actual threat from the 11th army) made it irrelevant.

The Japanese force included more than eight divisions in a complex plan. From starting positions on the right bank of the Yangtse, the bulk of the divisions would advance west along the northern bank of the Li river, just as they had done in May, securing the bridges, from east to west, in Jinshi, Lixian, Shimen and Cili. In the north, a limited advance from the 39th division would secure the main thrust’s right flank. On the left, the 40th division would advance at a slower pace to the south-west, directly towards Changde. The overall idea was to deceive the enemy into thinking the main move to the west was targeted at Chunking. The second phase of the operations was to have the main force make a 90° turn to converge on Changde. The city would then be assaulted by 3 columns, coming from the north-east, the north, and the north-west. Except to secure the city itself, there was no plan to cross the Yuan river and move further south than Changde.

Attacking to the west along the Li river had one advantage: putting the weight of the assault at the junction of two Chinese war zones. However, the Japanese plan had two obvious shortcomings: the most advanced division had to run 70 km west up to Cili, then another 40km to get back to Changde, stretching the supply lines. And the Chinese could create choke points at each crossing on the Li river to delay or frustrate operations south to Changde.

Chinese were probably aware of a coming attack by having observed the Japanese reinforce their bridgeheads on the right bank of the Yangtze. Their defense had two main lines: first a “river defense force” holding a continuous line a few km west of the Yangtse; then two army groups (10th and 29th) for a deeper defense. The Changde area itself was under responsibility of another army group, the 27th, sometimes named after his general, Wang Yao-wu. It included the 57th division, specifically in charge of the defense of the city. As we will see, the Chinese reacted to the attack by a defensive plan seemingly like Stalingrad: luring the attacker into an urban battle to be able to encircle it.

Although the campaign took place deep inside China, navy services had some role to play. Some Japanese supplies were to be shipped from ports on the Yangtze river across the Lake Dongting, and eventually up the Yuan river to Changde. Meanwhile, a Chinese river force tried to mine the Japanese ports. It is not clear how many ships and barges were involved in the campaign, but US air force routinely strafed these exposed targets on their way to or from Change.

The advance to Changde (November 2nd – November 20th)

Japanese moved to their jump-off positions at the end of October and attacked on the night of November 2nd. After crossing the river defense line, the main Japanese force (3rd , 13th, 34th divisions, as well as the 17th mixed brigade) advanced in parallel columns to the west, a few km north of the Li river. They progressed 30km in 5 days until stopped by serious resistance from Chinese 79th corps around Nanchuisheh and Wangchiashang on November 8th. The 116th and 68th divisions, which were to seize the bridges across the river, could not advance quickly enough. On November 10th, they had secured the first bridge (in Jinshi), but were still 20km east of the main force.

Meanwhile, the right and left columns made little progress. In the north, a limited attack by the 39th division progressed 20km before moving back to its starting positions around November 10th. They had accomplished their mission: fixing Chinese units to protect the main force’s right flank. Further south, the 40th division, which was taking the shortest way to Changde, had a harder time against Chinese’s 29th Army group, and only got to Anxiang (15km into Chinese territory) on November 10th. Part of the Japanese actually crossed the lake on barges to speed up the attack.

At that point, the Japanese front showed a long salient that Chinese considered cutting. The Chinese 73rd and 44th corps, both from the 10th army group, attacked facing south east with the vague hope of linking with 29th Army group, but Japanese reacted with enough flexibility to check them.

But the Japanese had to divide their main force in 3. Some troops moved backwards to help the 116th and 68th divisions at their next crossing point (Lixian); others moved immediately south of Wangchiashang to seize the Shimen bridge; and the last division (13th) pushed further west to keep Chinese at bay. Since the Chinese 10th Army Group had seemingly exhausted its energy, that was a reasonable risk.

It took almost a week to secure the crossings (finally accomplished on November 15th). The Chinese army had a tendency to be routed and dispersed when pursued over flat terrain, but was at its best when entrenched in fortified positions, for example around a bridge or a city. Moreover, this was the time US planes engaged in harassing missions on the Japanese, targeting ground units and river shipping. The Japanese, who had insufficient air support, had feared enemy air would be a severe issue, but it did not become decisive. These delaying actions gave time for Chinese staff to prepare for a trap. Having guessed that Japanese may target Changde, they started pouring reinforcement in the city and calling support from the 9th war area. Fresh troops moved from the Changsha area along the southern shore of the lake to defend the Yuan river.

The battle of Changde (November 20th – December 9th)

The battle of Changde was actually two battles: the conquest of the city (November 20th – December 3rd) and the Chinese attempt to annihilate the Japanese locked up in the area (December 3rd-December 9th).

The first Japanese columns reached Changde on November 20th or 21st and an advance party of (walking) paratroopers were briefly surrounded in TaoYuan, a few km upstream from Changde. The Japanese assaulted the city but got bogged down into house-to-house fighting by the Chinese 57th division. After securing the area between Changde and Lake Dongting to allow for supply to get through by motorboats sailing up the Yuan river, the other Japanese columns joined the fight. By November 24th, 3 Japanese divisions were fighting in the city. The Chinese reinforced the besieged troops with their 10th corps, clearing the south bank of those Japanese that had crossed the Yuan river. The Japanese entered Changde fortress on November 28th. On the next day, they published a statement warning they would systematically burn the city (which proved difficult). The city was literary on fire, yet the 8,000 men of the 57th division, supplied by air drops from American A-29s, showed fanatic fighting spirit. Just like the Russians did at Stalingrad, the Chinese tried to cut off the attacking forces by surrounding the city. One source mentions that four Chinese corps did encircle the city by November 27th, yet this did not seem to decrease the fighting effectivenessof the Japanese or to trigger any rescue operation. It could be that the bulk of Japanese supplies came by the river – unhampered by strafing and river mines.

What exactly happened during the last days of the Changde siege is unclear. Chinese sources mention that the Japanese made liberal use of poison gas to reduce Chinese strong points. This would be realistic as Japanese frontline units routinely included gas specialists, and it demonstrates how tough the fight became. Japanese sources report that after the fortress had been invested on November 28th, an exit route was left open, through which the most senior Chinese officers escaped. The Chinese fail to confirm the anecdote. Similarly, it looks like no reinforcement were sent across the river to the defenders (contrary to Stalingrad). What is certain is that the 57th division defended until almost complete annihilation. When the last defenders eventually surrendered, they controlled an area no wider than 300m, and numbered about 500 survivors. Casualties amounted to 95% of officers and 90% of heavy weapons – again resembling to Stalingrad.

Yet the battle was not over. Chinese forces mopped up around the city attacked the Japanese. US aircraft could now bomb the city without hitting friendly units, and did so liberally. They repelled a weak interception attempt by Japanese fighters and kept control of the air until the Japanese temporally froze enemy activity by bombing airfields on December 15th. By December 9th, the Chinese had taken the city back from the Japanese. Whether the Japanese made a deliberate withdrawal (after all, their original plan was to keep control of the city for three days only) or were forced to do so because of enemy pressure has not been properly investigated. Similarly, what happened during the Japanese retreat and the casualties inflicted by the Chinese is not documented, but the fact the US air force was not at all called for support after December 9th suggests Japanese did not have a hard time getting back to their lines, both by walking or sailing across the lake.

The aftermath

The total campaign involved about 100 000 Japanese troops (included an unknown number of Chinese enrolled in four puppet divisions), facing 40-50 Chinese divisions (or 13-16 corps) that may have totalled some 200 000 troops. Casualties estimates are rather vague. Japanese records mention 2,977 dead and 1,274 wounded, of which a mere 148 dead and 72 wounded would be related to the siege of Changde siege itself. Number of prisoners in Chinese hands was not reported. Chinese losses are only estimates and range from 10 000 to 30 000 dead, to which one should add 15 000 prisoners. The 10:1 ratio between Chinese and Japanese losses would be surprising on other theatres, but is common on the China front.

So, who won? Both sides claimed victory, no one admitted defeat. The Japanese reached their target and destroyed the capabilities of the Chinese armies facing them by inflicting more than 25% casualties. Their own losses, although significant given lack of replacements, were not proportionally important, as the success of the same 11th Army in Ichigo proved six months later. Yet the Chinese kept control of the battlefield and showed such stubborn fighting spirit they even put the Japanese Army in embarrassing situations. Obviously their case was easier to publicize than Japanese’s: press reports and official communiqués claimed they frustrated the Japanese offensive. Their best argument could have been the famous photograph showing a demoralized, kneeled Japanese prisoner guarded by two Chinese soldiers. The Japanese could argue the whole operation was just a spoiling action, but, even if true, it would not be convincing.

The Changde campaign was the last “invade-destroy-return” campaign of the SIno-Japanese war. From 1944 on, strategic priorities changed: having lost control of the seas, Japan looked for a land link line from Indochina to Mandchuria; the threat of American bombers, which could reach Honk-Kong, Taiwan, and maybe Japan, had to be eliminated; and securing a major victory somewhere (be it Burma, Leyte or China) could give Japan a chance to get a conditional peace. Japan engaged its reserves and launched Ichigo.