For most of the period of the Cold War, the term ‘warlord’ was unpopular among scholars of politics. Only seven journal articles using warlord in the title can be found in Worldwide Political Abstracts between 1966 and 1993. One studied Lyndon Johnson, another examined Winston Churchill, and the remaining five dealt with China after 1911. The Oxford English Dictionary reflects this limited definition by specifying two particular uses for ‘warlord’, the first simply for a military commander (and a translation of the German Kriegsherr), and the second a translation of the Chinese junfa – ‘In China, a military commander who had a regional power base and ruled independently of the central government, esp. in the period 1916–1928.’ Despite the frequent association of the term with events in early-twentieth-century China, it was even problematic to use ‘warlord’ in that historiographical context. This ambivalence about the term was most evident in the work of Hsi-Sheng-Chi, who decided, on the one hand, to entitle his book Warlord Politics in China, 1916–1928 and, on the other, to declare in the first footnote that he would avoid using the term ‘warlord’ in the text of the volume because of the negative connotation of the word.Footnote ¹

The period after the Cold War, however, brought a demand for new political terms to describe unfamiliar situations. ‘Warlord’ became more popular, first to describe commanders in Somalia and then to describe similar rulers across the globe. More than fifty articles in the Worldwide Political Abstracts database have ‘warlord’ in their title since 1993, not just with reference to Somalia and China, but also to the Andes region, Afghanistan, Bosnia and Herzegovina, Liberia, Macedonia, Morocco, Philippines, Russia, Sierra Leone and Tajikistan. Clearly warlordism is now worth studying.

Russia in the period of war and revolution is an obvious place to undertake such a study. In the first place, the Romanov empire splintered into a mosaic of undergoverned statelets awash with violence and dominated by men in uniform. As we shall see, the preconditions for state collapse and the emergence of warlordism were present even before the revolution in the period of the First World War, when the imposition of martial law and the introduction of radical new projects of social engineering on the part of the army leadership began blurring the lines between civilian and military authority. In 1917, regime change and military defeat led to a full-blown collapse of the state and a near-total erasure of the distinction between civilian and military power. By 1918, uniformed men with guns held power throughout the former empire, and they were only partly organised into hierarchies of power with a centralised chain of command. They ruled through terror, and on occasion with charisma, but they ruled locally and tyrannically virtually everywhere. The decade from 1915 to 1925 was an era of warlords all across Eurasia, not only in China.

Second, the fact that similar phenomena were displayed at the same time in neighbouring Russia and China allowed the discourse of warlordism to be created in a kind of dialogue between the two crippled imperial spaces. The Bolsheviks appropriated the term ‘warlord’ to describe their enemies, and Chinese visual propaganda relied heavily on models generated in Soviet Russia. Contemporaries clearly saw the parallels.Footnote ² The parallels between Russia and China – and those between these early-twentieth-century examples and those that emerged at the end of the century – are worth pursuing. What accounts for warlordism? How does it begin, and how does it end? What happens to societies and economies subjected to extended periods of warlord rule? These are all questions that the Russian case can help us understand. They are also questions that require much more than a single article to answer. The aim of this essay is more limited. I aim to illuminate the processes that developed over the course of the First World War which led to the genesis of Russian warlordism and to indicate some of the ways in which warlordism manifested itself in Civil War Russia by sketching the experiences of two of the most notable Russian warlords – General Lavr Kornilov and Baron Roman von Ungern Shternberg.Footnote ³

The figure of the warlord – whom we define here as ‘a military commander who autonomously exercises political power through the threatening use of force’Footnote ⁴ – did not originate in China and Russia in the second decade of the twentieth century. Historically speaking, a great many political communities had their genesis in precisely this way. But the phenomenon of twentieth-century warlordism was somewhat surprising, since it occurred after the rise of the modern state, not before. The long, much-studied process by which civilian bureaucracies replaced earlier structures of authority based mainly on military force had played out centuries before, and warlords had not returned. There were good reasons why states had supplanted warlords, and a lot of good reasons why both rulers and the ruled preferred state rule. State bureaucracies, no matter how corrupt or inefficient, were less violent, more predictable and more likely to support prosperity and security than were strongmen and their armed bands. In Russia the Romanov dynasty had been around for more than three hundred years, and although several Romanov monarchs liked to fancy themselves as military men, none of them derived their legitimacy from their skills in violence. The governors-general installed by the autocracy in newly conquered or particularly troublesome areas such as Poland or Turkestan came closer to fitting the bill, but even they, on a day-to-day basis, were more governors than generals.

Several preconditions would have to be met before the political system could change from statism to warlordism. The fundamental precondition was state failure. In some fashion, civilian bureaucratic institutions would have to lose legitimacy, authority and effectiveness – lose power, in other words – for warlords to emerge. This was a process that would-be warlords were normally unable to instigate on their own. Rebel commanders tried, of course, but their history was a history of failure. In Russia, figures such as Stepan Razin and Emelian Pugachev had formed bands to oppose the state during the seventeenth and eighteenth centuries, but they were thoroughly defeated. When states collapsed, it was usually the result of other pressures, often external ones.

The second precondition is that there must be candidates to become warlords – men with military experience, sufficient individual authority to inspire obedience on the part of their men at arms, political ambition, interest in civilian affairs and a high tolerance for risk. Again, Russia had social groups who trained men of this type throughout the imperial period – most obviously the Cossacks, but also the groups who frequently clashed with those Cossacks in such places as Chechnya, Dagestan and Turkestan. All these militarised groups had clashed with the imperial state and the standing Russian army, and all had lost. As elsewhere, a functioning state was able to defeat warlord formations. For this balance of power to shift, something would have to occur that would both cripple the existing state and invigorate (or generate) a social group psychologically and experientially prepared to become warlords.

State collapse

In Russia, that event was the First World War. The process began at the very beginning of the war and was not simply the result of military defeat and revolution. Instead, the wounds that would eventually kill the Russian Empire were self-inflicted. State failure was mandated from above, beginning with the institution of martial law itself. The regulations issued by Nicholas II on the first day of the war were breathtaking in their scope. According to these new rules, a broadly defined front-line zone was established that came under the direct control of the leaders of the active army.Footnote ⁵ This decree in essence imposed martial law on all territories west of the Dnepr river and as far east as St Petersburg itself. Although the tsar apparently envisaged assuming the role of Supreme Commander-in-Chief himself, he was eventually persuaded to stay in St Petersburg and to appoint his cousin, Grand Duke Nikolai Nikolaevich, to the new position. It took two weeks for the Supreme Headquarters (Stavka) to be up and running, leaving a window of time in which civilian officials had to report to the commanders of regional military districts rather than to their ministry chiefs.Footnote ⁶

Even after Stavka began functioning regularly, the lines of authority were not particularly clear. The grand duke and his chief of staff, General Ianushkevich, were in charge, but at first they were mainly occupied with combat operations rather than civilian administration. In September, Nikolai Nikolaevich finally got around to appointing Prince N. L. Obolenskii as the point person for civilian affairs and then creating a Chancery for Civil Administration for him in October. He was assigned two other officials, an enlisted clerk, and two orderlies. This office of six men was in charge of civilian life across a territory larger than Germany.Footnote ⁷

It therefore goes without saying that civilian administration did not end at the doors of Stavka. In the first place, employees of civilian ministries were supposed to remain in place and continue performing their jobs as they had done before the war. This expectation was not always met, but enough stayed on to provide some continuity. In addition, a wide variety of military officers now enjoyed governance rights in the areas where their troops were stationed. Front commanders, army commanders, corps and divisional commanders could all issue edicts, as could the chiefs of supply and fortress commanders. In enemy territory occupied by Russian officers, depot officers exercised similar rights. These military men created regulations to enforce security in the zones of their authority – imposing curfews, conducting searches of homes and businesses, and deporting undesirables. They sought to control economic life by fixing prices, forbidding trade in certain goods and requisitioning labour. They endeavoured to suppress political life by imposing censorship, dismissing local officials and occasionally harassing civilian relief agencies. All this was done by decree, with no right of remonstrance on the part of affected civilians or outraged civilian bureaucrats.Footnote ⁸ The result was confusion, indeed anarchy. The chain of command looked clear enough, as each of the men in uniform worked in a hierarchical system headed by the grand duke, who in turn answered only to the emperor. But the fact that civilian administration was at best an afterthought meant that communication, authority and responsibility for these affairs during the war were unclear. Exigency – and its twin, arbitrariness – was the only common feature of political life in the front-line zones.

The defining feature of life for most inhabitants of those zones, meanwhile, was insecurity. In regions removed from the front line itself, columns of men on foot, horseback or train regularly travelled through, hungry and immune from local prosecution. Marauding enlisted men could be (and occasionally were) court-martialled and shot, but only at negligible rates.Footnote ⁹ In villages and towns closer to the front, there was the ever-present danger of becoming collateral damage. In Poland, Ukraine and Armenia, large swathes of territory saw fighting in the very first year of the war, and civilians were not immune to the shells, the bullets or the poison gas. Several areas even underwent multiple changes of military occupation over the course of a few months.Footnote ¹⁰

In September 1914, German troops got as far as the gates of Warsaw, occupying such places as the town of Pruszków, a mere 16 kilometres from the Polish capital. The Germans were forced to retreat soon thereafter, leaving the town ablaze with accusations, denunciations and recriminations. The case of Shmul Abramov Shraiberg provides us with an example of the insecurity produced by conditions of war. Shraiberg was the proprietor of a joiner's shop in the town, and two other Pruszków residents, Stanislav Smolenskii and Stanislav Furman, denounced him to a Russian counter-intelligence agent. According to testimony, when the Germans entered the town Shraiberg greeted them warmly and immediately presented them with his documents, which were examined and approved by an officer. After this vetting, Shraiberg allegedly led German troops around the town, showing them not only where food reserves were located but also where local residents were hiding. These residents were taken into custody, and many of them were taken away by the Germans on suspicion of espionage when German troops retreated. Finally, Shraiberg was accused of ‘railing at Poles’, in particular at Smolenskii's wife, already wracked with grief because her husband had been rounded up by the Germans. Shraiberg's story, needless to say, was much different. In his account, he claimed he had been unable to evacuate Pruszków when enemy forces neared because his wife was ill and bedridden. He claimed that he stayed at home for three days before venturing out to buy bread on 1 October 1914, the day the Germans retreated. On that day, he claimed, Germans were going from house to house, and he was swept up and sent with one hundred other Poles and Jews to Helenów and then to Grodzisk. A week later, with the German retreat continuing, all of them were abandoned in Grodzisk and made their way home. He claimed that he had never given assistance to the Germans or pointed out food stores or hiding townspeople. His defence was useless. In accordance with directives from the military, he was sentenced to exile in Siberia for the duration of the war on ‘suspicion of military espionage’.Footnote ¹¹

Shraiberg's case was typical of dozens of others still preserved in the archives, and it testifies to several levels of insecurity. One level was judicial. The documents included in his file were hardly conclusive and bore the marks of a high-speed prosecution. Smolenskii's testimony was the more extensive of the two, and the corroboration by Furman was limited to his having witnessed Shraiberg ‘talking about something with German soldiers’ as civilians were being rounded up. No work was apparently done to find other deportees who might have confirmed that Shraiberg had been the target rather than the author of the arrests. By the same token, Shraiberg's deposition was also uncorroborated and understandably exculpatory. The military mandated that the gendarmes’ exile anyone even suspected of treason or collaboration, however, and that low threshold was enough to send Shraiberg away indefinitely.

The second level of insecurity was physical. All the parties agreed that the Germans had indeed grabbed civilians as prisoners or hostages as they retreated, though they were eventually released. The experience of forced deportation in these first weeks of war was enough to cause significant anxiety for most civilians, many of whom had hidden themselves away as foreign troops arrived. Death was present throughout the war zone, but there were other dangers that weighed on the minds of civilians.

The third level was ethnic. Shraiberg was Jewish, and this was a fact of considerable significance in wartime Poland. As several recent studies have demonstrated, the war dramatically ‘ethnicised’ or ‘nationalised’ political and social relationships in the Russian empire.Footnote ¹² Relationships between Jews and Christians hardly needed ethnicising by 1914, of course. The Russian state was openly antisemitic, and even important liberals and progressive socialists had difficulty navigating what they called the ‘Jewish Question’ because of its political explosiveness. Still, there was a dramatic change in the years of the war. The worst pogroms of the pre-war period were relatively small affairs. The infamous Kishinev pogrom in 1903, which caused a worldwide outcry, killed roughly fifty people. As terrifying as these events were, they did not compare with the level of terror experienced by Jews during the war. Soldiers made sport of robbing, killing and humiliating Jewish townspeople from the very first days of the war. Thousands of Jews fled for their lives, and many more were forcibly deported, either on suspicion of spying (as with Shraiberg) or en masse, as occurred frequently during the Great Retreat of 1915. It is difficult to say with precision how many Jews were displaced or killed during the war, but the scale was substantially higher than had been the case in the pre-war pogroms. One aid worker counted more than 50,000 Jewish refugees in Warsaw by the end of 1914, before the worst even came.Footnote ¹³ In total, most estimates fall between 500,000 and 1,000,000 deportees.Footnote ¹⁴

In one sense, the insecurity described here was not that unusual. Deliberately or not, military occupation regularly lays waste judicial systems, economies and ethnic harmony. What began occurring on the Eastern Front in the First World War, however, was that the military officials who faced these issues began to devise long-term as well as short-term political strategies for dealing with them, and they used the conditions of martial law and military occupation to facilitate their political objectives.Footnote ¹⁵

Again, it was ethnic cleansing that became the focus of these politico-military strategies. Russian authorities took energetic action to arrest ethnic Germans and expropriate their property throughout the empire, shutting down factories and seizing an astonishing amount of land. By early 1917, more than 5 million acres of land had been taken, with another 10 million scheduled for auction in the near future.Footnote ¹⁶ These were steps far in excess of any ‘wartime necessity’, and they were taken in order to transform fundamentally ethnic settlement in large swathes of eastern Europe. The assault on Jews was even more dramatic. The mass deportations mentioned above were undertaken with great cruelty and humiliation, and the scale was so extensive that it forced the tsarist government to dissolve the boundaries of the Pale of Settlement that had restricted Jewish movement since the empire had annexed these concentrations of Jewish settlement in the partitions of Poland in the late eighteenth century. No one seemed to know for sure how the ‘Jewish Question’ would be resolved after the war, but, again, these were steps that were far more ambitious than was necessary to preserve order in the zones of martial law.

The genesis of the warlords

This new ambition to conduct political programmes with the aim of social engineering took root at the highest levels of the Russian military. General Ianushkevich, the most senior staff officer in the army for most of the first year of the war, was not the only general with grand schemes, but he was the most important one. On 3 October 1914, in the wake of the German offensive that had provided the context for Shraiberg's alleged actions, amidst a wave of similar allegations of treason, Ianushkevich ordered all enemy subjects to be deported from front-line zones.Footnote ¹⁷ Over the next few months this edict was put in force not only near the trenches but throughout the zone of military occupation. By January 1915, Ianushkevich was ready to extend the application of these orders beyond foreigners to Russia's Jewish population as well. The 50,000 Jews who had fled the early, anarchic period of this wave of antisemitic terror were about to be joined by hundreds of thousands more. On 25 January 1915 he sent a circular to army commanders condoning the deportation of ‘all Jews and suspect individuals’ from combat areas.Footnote ¹⁸ By March he was ordering the deportation of entire communities from occupied Galicia.Footnote ¹⁹ By May he had issued orders saying that all civilians living near troops who were members of any suspect ethnic group were to be exiled.Footnote ²⁰ The zones of military occupation were to become ‘Russian’ territories, with all the political and demographic upheaval that suggested. As Ianushkevich wrote to Count Bobrinskii, the satrap of the Galician territories just annexed in the battles of 1914, ‘we are established in that territory sufficiently firmly to feel that when we are in Eastern Galicia we are on Russian land, and we can allow Russian Galicians to think that as well’.Footnote ²¹

Ianushkevich was taking on many of the attributes of a warlord. He controlled both military and political life in the territories where Russian armies stood, and he saw little distinction between his military and political ambitions. Just as importantly, many of the restraining forces that might have served to retard the development of warlordism were weakening or had snapped already. Well-dressed ministerial bureaucrats and the fancy salons of Petrograd were far away, and men in uniform were becoming accustomed to making decisions on a much wider range of issues than they had done before the war. But Ianushkevich was not a warlord. Indeed, he was well known as a lively and entertaining guest, one ‘gentle to the core’, at those very salons whose importance was fading away.Footnote ²² He killed people with his pen, not his sword. Just as importantly, he still fully embraced the discipline and chain of command that linked him to the rest of the military organism and that linked that organism to the rest of the Russian political and social system. When he and his boss Nikolai Nikolaevich were abruptly dismissed from their positions after the disasters of 1915 and reassigned to the Caucasian front, they both left quietly and without open recrimination. His farewell speech to his staff consisted of his telling them that ‘I am to blame for all that has occurred.’Footnote ²³ And although the idea of a military coup to wash away all the blathering and betrayal in the capital certainly appealed to a wide range of generals as the war progressed, no one ever attempted one. Indeed, it was only after the revolutionaries in Petrograd forced a decision that the top military leadership even came out in favour of the abdication of the tsar. There were no lone wolves in the army pack in early 1917.

The revolution changed that situation. Most obviously, it completed the process of state destruction that had begun in the borderlands in 1914. ‘Russia’ remained, but institutions of governance suffered from a lack of legitimacy, a lack of personnel, and a lack of a coherent political programme other than the vague (and likely contradictory) goals of ‘revolution’ and success in the war. The revolution also buckled the knees of the armed forces. Soon after the fall of the monarchy, the Petrograd Soviet issued orders encouraging the formation of elected military committees and the questioning of officer authority, and it upended the traditional system of military discipline by requiring officers to treat soldiers as equals and by allowing soldiers to disobey any orders that contradicted those issued by the soviet. The Provisional Government furthered the transformation in military discipline by forbidding the death penalty for military crimes only days after the February Revolution.Footnote ²⁴

The immediate outcome of these edicts was a sharp decline in the legitimacy of the officer corps and a noticeable increase in the incidence of insubordination. One might have expected that this crippling of the social group most likely to become warlords would have retarded the growth of warlordism, but in fact it served to accelerate the process. As the army broke apart, opportunities arose for ambitious leaders to take command of parts of it on their own authority. As it turned out, the process of vetting, lynching, dismissing or reaffirming one's commander served to bolster the personal power of particular commanders even as it weakened the centralised army structure as a whole. It had become possible to gather groups of armed men more loyal to their commander than to the high command or indeed to fellow soldiers from other units. These commanders, precisely to the extent that they challenged central authority, required support from their soldiers and the means to supply those soldiers with civilian goods. If for no other reason than survival, nascent warlords would have to engage in civilian affairs, although most had political desires that would have led them to take the path of civilian rule in any case.

The first person to combine all these features and ambitions – the first true warlord of the period – was Lavr Kornilov. Kornilov was born into a Siberian Cossack family in 1870 and had rapidly ascended the promotion ladder during his military career. He earned a reputation for personal bravery, intelligence and strong leadership, both when engaged in military espionage in Persia at the turn of the century and in combat during the Russo-Japanese War, when he was decorated with the St George's Cross. In the First World War he had led a division at the forefront of the Carpathian campaign in 1914–15, but had been taken prisoner when the Carpathian positions were exposed by the successful German–Austrian offensive near Gorlice. A year later he escaped from imprisonment and returned to assume command of the XXV Corps on the South-Western Front under Brusilov's command. In the course of late 1916–17 he ascended rapidly, taking over from Brusilov as commander of the Eighth Army and then, briefly, as commander of the South-Western Front.Footnote ²⁵

Kornilov's talent for command and his extreme intolerance of military indiscipline were well known. Those who loved him, loved him. Those who hated him did so with a passion. He apparently recognised the dangers and possibilities of his situation quite quickly, and he began forming cadres of those of his men who were loyal to him rather than to his military or political superiors, most notably an elite corps of Tekke Turkomans called the ‘Tekintsy’. These men served him personally and well. So did other several other trusted units. But the polarisation produced by the revolution and by Kornilov's personality also led many men under his command to resist. The major showdown in this regard was the June 1917 offensive, when Kornilov's Eighth Army succeeded in punching a hole in the Austrian line but failed to develop the attack when many units refused to enter the battle. The ensuing counteroffensive broke many of those hesitating units. Soldiers fled to the rear, looting and pillaging as they went.

It was this disaster that led to Kornilov's transformation. There is no indication that Kornilov had harboured either a desire or a plan for striking out on his own prior to June 1917, but the choices he made over the ensuing month were fateful ones. Three choices were particularly important for his maturation into a warlord. First, he began using terror as a governing strategy. Second, he began to assert his independence from central authorities. Finally, he began to envision himself combining military and political authority in his own person.

As soon as the troops began to break, Kornilov turned to drastic measures. Soldiers evacuating Galicia turned the region into a platform for atrocities. The so-called ‘Savage Division’ was particularly active. The division had been detailed to the front for explicitly political work. On the eve of the offensive Colonel Chavachadze, the divisional commander, began his briefing by saying ‘Gentlemen, I am sorry indeed that the young officers who joined our colours recently will have to start their fighting career by doing a rather repulsive sort of police work.’Footnote ²⁶ That work was to attack any soldiers or units engaged in mutiny or desertion, and it was welcomed by many members of the unit. One of the officers in the division, Sergei Kurnakov, admitted that he ‘hoped to be able to avenge some of the insults the mutineers had inflicted on my brother-officers’.Footnote ²⁷ Instead, when the offensive collapsed, Kurnakov reported undertaking a far different kind of police work. Coming upon the town of Kalush, he encountered a monstrous scene. Driving down the main street, Kurnakov thought he saw snow in July. Instead, it was the down of feather beds floating through the air. He saw looting soldiers, ‘the body of an old Jew . . . hanging out of one of the windows, his arms nailed to the window sill’, and finally the gang rape of a pregnant woman, which he dispersed with machine-gun fire from his car.Footnote ²⁸

Other reports confirm the brutality of the retreating Russian army in Galicia that July, but most indicate that, far from acting to stop the carnage, Kornilov's favourite division was one of the main instigators of atrocity, not only in Kalush, but in towns such as Ternopil’ and Brody as well. As one official reported as late as August, ‘we are continually receiving petitions regarding robbery and violence’ by members of the Savage Division.Footnote ²⁹ One such petition from landowners in Edinets county complained that the Caucasian troops thieved vegetables, robbed men of their money and killed several people. ‘The local people’, they wrote, ‘are in a state of panic.’Footnote ³⁰ While Kornilov did not officially condone this behaviour, he did exacerbate it by issuing a scorched-earth policy similar to that which was implemented in 1915. Mass deportations were ordered once more, with all military-age men targeted for removal along with their goods and horses.Footnote ³¹ He also pursued a terror campaign against his own soldiers by summarily executing deserters and others whom he blamed for the military defeat.

Far from recognising that ordering his troops to conduct campaigns against civilians and shirkers was contributing to indiscipline, he came to believe that only further terror could end the problem of military collapse. He, like most former officers, was outraged by the insubordination, the desertion and, above all, the soldier committees and political commissars that had ruined his beloved army. In his opinion, the only way to reverse the collapse of the Russian military was to strengthen and unify the authority of officers that had been ruined by the Revolution. This was a non-negotiable issue for Kornilov. Furious that enlisted men would issue demands and ultimatums to their officers, he was nevertheless the first general to demonstrate insubordination to the political leaders above him in the chain of command. The pattern of conflict with the Provisional Government began early. When Kornilov was offered the position of commander-in-chief, he immediately issued a set of demands to the Provisional Government that included the controversial measure of reinstituting the death penalty for soldiers, and insisted on complete autonomy in choosing military commanders. Kerensky urged his colleagues to rescind the offer because Kornilov had demonstrated an unwillingness to subordinate himself to civilian authority, but he was rebuffed.Footnote ³² Other battles soon followed. Kornilov was head of the South-Western Front when promoted, and the Provisional Government had appointed General Cheremisov to command the front at the same time as it moved Kornilov up the chain of command. Kornilov took this as a violation of his ultimatum that politicians must not interfere with command decisions, and he named his own favourite, General Baluev, instead. Cheremisov did not yield willingly, and it took all the wiles of the Provisional Government's commissar M. M. Filonenko to resolve the situation without a major conflict in the high command erupting into the open.Footnote ³³

By August, Kornilov was convinced that saving the army required not only the restoration of ‘harsh measures’Footnote ³⁴ in the army but also dictatorial methods in the rest of Russian government. Other top generals worked together with the Union of Officers to insist that the Provisional Government admit its mistake in distrusting military commanders, stop sticking its nose into army affairs, render the declaration of soldier rights ‘null and void’ and, again, legalise the death penalty.Footnote ³⁵ Kerensky and other political leaders attempted over the course of July and August to calm the generals by explaining the lengthy period necessary for legislation and, when being totally honest, by indicating how suicidal such a legislative programme would be for the Provisional Government itself. These conversations led to a simple logical conclusion for Kornilov: the army could not succeed without discipline, discipline could not be restored without firing squads, and firing squads could not be implemented by a weak Provisional Government in tandem with a seditious Petrograd Soviet. Hence ‘dual power’ had to be eliminated to save the army. Dictatorship was necessary.

Kornilov and his later supporters claimed that although he had desired a dictatorship, he had not insisted that he himself be the dictator. It is true that Kornilov never openly avowed a desire to overthrow Kerensky's government, but he was, of course, far too experienced to issue such openly treasonous proclamations. His actions suggest otherwise. He soaked up the adulation of the political elite at the State Conference in Moscow, where he gave a speech that displayed political ambition as well as a strong military presence. More importantly, he began moving troops loyal to him close to Petrograd, most notably the Savage Division. The Kronstadt fortress and naval base, one of the cradles of revolution, was to be emptied, its troops dispersed to other units. His own chief of staff, General Lukomskii, threatened resignation on hearing of these moves before being reassured by the explanation that the repositioning was to defend the Provisional Government in the event of a Bolshevik coup. Kerensky was hardly Kornilov's friend, however, a fact highlighted during Kornilov's visits to the Winter Palace, visits always made with the Tekintsy by his side.

In late August, Kornilov made his move. He sent troops under the command of General Krymov from the front to Petrograd with the intention of disbanding the soviet, arresting the Bolsheviks, and instituting a military dictatorship with himself at the head. These troops, shunted into sidings by railway workers, were literally derailed. Similarly, Kornilov's expectation (fed by an extremely high level of miscommunication) that Kerensky would support the coup was dashed when Kerensky ordered him to be dismissed and arrested. The historical controversy that has swirled around the ‘Kornilov Affair’ ever since has focused on Kerensky's intentions and to a certain degree on whether Kornilov intended to allow civilian ministers into his circle of leadership, but these issues have only limited importance here. What is unmistakable is that Kornilov intended to institute military rule instead of civilian rule and that authority would be concentrated in a single man – himself.

Thus although the Kornilov revolt in August is normally understood as the major counter-revolutionary moment in 1917, in a certain respect it was also the moment when revolution (in the form of insubordination to the established hierarchy of authority) reached its apogee. The coup failed. Kornilov's bid to become Russia's supreme warlord was stopped, not, as most such attempts had been in the past, by defeat by the modern state and its army, but by the opposite phenomenon. Power had fractured to such a significant degree that military commanders could no longer expect to have their rule accepted over large territories. The army was divided against itself, as was the country as a whole. Kornilov had discovered an important truth about warlordism – it works better on a small regional scale than over a large territory.

The coup failure itself helped to bring about the final disintegration of the army and the country. A new wave of hostility between officers and enlisted men erupted. As General Gromyko of the South-Western Front admitted in mid-September, the Kornilov uprising had destroyed whatever trust remained between the ranks. ‘The general mood is satisfactory’, he reported, ‘but in connection with recent events it is angry with the command staff.’Footnote ³⁶ So, too, did ethnic tension build up, this time between Cossacks and non-Cossacks. Peasants emboldened by Bolshevik successes began refusing to sell grain or forage to Cossack units and fellow soldiers began to ostracise Cossack units. ‘The mass of soldiers’, one officer morosely reported, ‘hate Cossacks.’Footnote ³⁷ By the end of the year, the bands of armed men and the territories they covered would be small enough for aspiring warlords to have success.

Playground of warlords: the Russian Civil War

The importance of local power was a lesson that many others would learn over the next four years, starting with Kornilov. Released from prison by sympathisers during the October Revolution, Kornilov made his way to what he and many others took to be the best recruiting grounds for a warlord army – the Cossack territory on the Don. Much of the rest of the army high command would join him there, each plotting to overthrow the Bolsheviks, and each jostling for power in a situation of open entrepreneurial activity. These men banded together in an odd ‘Volunteer Army’ composed almost entirely of officers. Despite the fact that other top generals were present, including Mikhail Alekseev, the longest-serving and most successful chief of staff in the war, Kornilov took charge. His ‘charismatic presence’, his courage in combat and his reputation for armed decisiveness all played in his favour. The territory they would occupy, especially the city of Novocherkassk, soon took on the political attributes and feel of its occupiers. It became ‘St Petersburg on the steppe’, right down to changing its clocks to the time of the capital city.Footnote ³⁸

Kornilov's exploits were legendary, though short-lived. During his flight to the Don and subsequent ‘Ice March’ from the Don to the Kuban in winter 1918, he had his horse shot from under him, escaped his Red pursuers, and launched futile attacks at well-defended cities. In March that year he launched yet another doomed assault on the city of Ekaterinodar. It was his last decision. His hut was targeted by Red artillery and destroyed, crushing him in the ruins. General Anton Denikin, his successor, tried to keep the news of his death secret, since he feared that his long-suffering men had invested as much in Kornilov personally as in the idea of a Volunteer Army.

Small, territorially based armed formations marked by systems of personalised loyalties marked Kornilov as a warlord and the Volunteer Army as a warlord formation. It, too, deployed terror. During the Ice March, the Volunteer Army despoiled the villages it encountered, not only stealing supplies necessary for survival but also engaging in mass executions. In the village of Lezhanka, officers stripped peasants naked and whipped them before executing sixty of their number as a measure of collective punishment of a community that had a ‘Red’ reputation. This would be a pattern that later ‘White’ warlords would repeat ad nauseam. As one of the most prominent of those warlords, Petr Wrangel, would admit later, ‘we had not brought pardon and peace with us, but only the cruel sword of vengeance’.Footnote ³⁹

The terror campaigns launched by warlords were massive in scale. They targeted political groups. Anyone suspected of sympathy or support for the other side's cause was liable to be killed. But targets were also ethnically defined. The legacy of the anti-Jewish campaigns of the First World War was evident. Between 1918 and 1920 in Ukraine alone, there were more than 1,500 pogroms in approximately 1,300 towns, with total deaths amounting to somewhere between 50,000 and 200,000 people. This does not count the tens of thousands of additional wounded, including a great many victims of mass rape. As the most recent and comprehensive work on the subject puts it, ‘Jews were killed on the roads, in the fields, on trains; sometimes whole families perished, and there was no one left to report on their fate.’Footnote ⁴⁰ This catastrophe was not the result of a co-ordinated and centralised policy. A combination of toxic military antisemitism and political warlordism played out in similar forms throughout the country during the Civil War. To take the terror against Jews in the Ukraine again as an example, one researcher has estimated that in the course of 1919, 40 per cent of the assaults were conducted by troops associated with Semen Petliura's Ukrainian Directory, 17 per cent by those associated with the ‘White’ cause, 9 per cent by ‘Reds’, 4 per cent by those linked to Ataman Grigoriev, 3 per cent by Polish troops, and a full 28 per cent by bands that were difficult to characterise politically.Footnote ⁴¹

This breadth of bands in Ukraine is indicative of the spread of warlordism in the Civil War, but it was not limited to the west. All across the land, from the Golden Ring to the Urals to the far east, local strongmen combined military and political power and deployed terror against civilians. Political forms other than warlordism also developed during the war, as attempts at civilian governments, local clan-based independence movements and other innovations sprang up. In summer 1918 alone, nineteen separate governments emerged to claim power and territory in Asian Russia.Footnote ⁴²

It is, of course, impossible to describe the phenomenon of warlordism in each of these cases, so let us turn to the most obvious warlord of the era: Baron Roman von Ungern-Shternberg. This was another man with a distinguished career as a warrior, a St George's Cross, and a penchant for violence so remarkable that even other hard men backed down in his presence.Footnote ⁴³ Ungern-Shternberg, like so many other warlords, spent much of his time during the Great War fighting in a Cossack regiment. Although not a Cossack himself (the army high command liberally (and wisely) diluted the officer corps of Cossack troops with non-Cossacks), he fought in a Cossack manner – bravely on the battlefield and savagely in relation to civilians. The February Revolution and the collapse of military discipline so dismayed him that he travelled to the Mongolian border, the place of one of his earlier postings, to try to recruit a Buriat ethnic regiment.Footnote ⁴⁴ Ungern and a half-Buriat Cossack fellow officer from his unit named Grigorii Semenov had already attempted this sort of volunteer recruitment/mobilisation earlier in 1917 in the Caucasus, and they were ready to raise their own troops again, this time with the blessing of a Provisional Government that was very far away and extremely weak. Semenov organised his efforts in the city of Chita, while Ungern stayed in the remote outpost of Dauriia. By spring 1920 he commanded nearly 2000 men and much of the frontier zone between what was left of Russia and what was left of China. Bolshevik victories in that year prompted Ungern to cross the border. He moved his troops into Mongolia, defeated the Chinese troops garrisoned in Urga in February 1921, reinstated the corpulent Bogd Khan as the nominal ruler of Mongolia and unleashed a reign of terror.

Terror as a governing strategy had developed organically as part of Ungern's military life. Just like Kornilov, Ungern believed that fear was the linchpin of military discipline. He neither regretted violence nor took pity upon miscreants in his units, despite the fact that he (as again with Kornilov) had a history of insubordination that extended all the way to slashing a superior officer with a sabre in a drunken rage in 1916. Ungern's view as commander was that his brutal punishments, such as beatings, forced exposure on winter nights and summary execution, were in line with a true military tradition that ‘began with Frederick the Great and ended with [Grand Duke] Nikolai Nikolaevich’,Footnote ⁴⁵ commander-in-chief of the Russian forces in 1914–15 and then Ungern's commander in the Caucasian theatre.

Ungern apparently saw civilian governance in much the same light. When he commanded the Dauriia region he regularly set his troops on passing merchants to loot and pummel, and he had a very sensitive nose for potential enemies. He pulled passengers off trains (Dauriia was the last stop before Manchuria) and executed them on suspicion of spying or being ‘Red’. As with many of his anti-Bolshevik comrades, Ungern thought that Jews were evil. In typically extreme fashion, he did not limit himself or his troops to periodic pogroms, but pursued an eliminationist strategy, giving orders to his men to kill all Jews they came across, as well as their families.

All these aspects of the Ungern terror campaign were conceived on the Eastern Front, developed when he controlled Dauriia and came into full flower when he led Mongolia into the abyss in 1921. The slaughter started immediately. After defeating the Chinese troops in Urga, his men savaged the city, grabbing what they could, setting sections of the town on fire, roasting ‘Red’ children alive in bakery ovens, strangling old women, raping young girls and even killing one another in a spree of bloodshed.Footnote ⁴⁶ After three days Ungern ordered the killing to stop and began the task of governance. He issued new money, tried to stabilise the merchant economy in the town, and even pressed successfully to get the local power station up and running again. Through it all, terror remained a constant, as he set up interrogation squads that eventually arrested and killed 10 per cent of the Russian expatriate community in the city. As might have been expected, his policies of violence were more effective than his economic policies. Tradespeople and other civilians fled, and Ungern was reduced to property seizures and other forms of thuggery just to feed his troops. Desperate for action and new sources of plunder, Ungern gathered his men again to launch a raid to the north, into Soviet territory. In June 1921 he launched a failed offensive on Kiatkha, then escaped with some of his men to the steppe. Urga was soon occupied by the Reds, and Ungern was finally deposed and arrested by his own Mongolian troops in August and given over to the Bolsheviks, who interrogated, tried and executed him.

Kornilov and Ungern-Shternberg were among the most notable warlords of the period, but there were many others. Hundreds more aspired to warlord status. The files of the young Soviet state are filled with reports of ‘bandit gangs’ with political agendas. Only a few were able to exercise territorial control, but many were able to cause trouble for the Bolsheviks and local civilians alike. In the Altai in June 1920, for instance, military commanders reported that ‘armed bands of deserters under the control of their chiefs Dem'iachenko and Plotnikov are operating in Barnaul district. These bands are robbing and attacking civilians; they are conducting open agitation against Soviet power and communists.’Footnote ⁴⁷ Not all these groups were small, minor and easily dispatched; ‘In the Western Siberian district . . . the character of the location (mountains and forests), and also the kulak composition of the population has called forth a significant development of banditry and insurgent movements . . . There are a large number of so-called partisans, who earlier made up entire armies who fought against Kolchak . . . The insurgency has more than once threatened district cities like Slavgorod and Pavlodar and Barnaul itself; it has also been able to seize the district town of Zvenogorsk.’Footnote ⁴⁸

This type of localised power was not just the characteristic of anti-Bolshevik forces. On the contrary, Red military commanders built their own loyal bands and used them as they wished, frequently angering their ‘commanders’ in Moscow. This Red warlordism was tolerated by the Bolshevik leadership in the Civil War out of necessity, and it began immediately. After the Bolshevik coup there was a period of several months during which there was effectively no Red Army at all, just a constellation of soviets hoping to stave off attack from local opponents. Local soviets raised local militias (which had already sprung up during 1917 in many areas) and used them for self-defence. Once raised, these groups were reluctant to be sent away from their home towns. As Nikolai Podvoiskii, one of the top Bolshevik military authorities noted, each soviet had ‘ignorantly inculcated in them the idea that they were created above all to defend that soviet and its territory, for the preservation of order in its region and, at most, for activity at the nearest front’.Footnote ⁴⁹ Over the course of the entire Civil War, the Bolsheviks had enormous difficulty in running a centralised draft, and they were continually forced to ask local leaders to mobilise soldiers for them. Sometimes this was done by civilian leaders in the Soviet. More often it was done by military commanders via a process close to kidnapping. Article 12 of the Regulations for Army Commanders allowed armies to impress soldiers in their battle region on their own authority, and they did so officially 354,373 times during the war, with many more actions no doubt going unrecorded.Footnote ⁵⁰ The memorable image from Boris Pasternak's Doctor Zhivago of cavalrymen sweeping a needed doctor off the road and into the war was based on a rather widespread real phenomenon.

This local latitude expressed itself organisationally in many ways, not simply through the formal Red Army command but also by the expansion of the Red Terror and the daily use of terror by local officials. V. I. Shishkin, the noted historian of Siberia, has called this development ‘Red Banditry’, and the title is apt. ‘Over the course of 1920 in Siberia’, he writes, ‘a swift escalation of state violence sanctioned by Soviet law occurred.’Footnote ⁵¹ Sometimes this movement was directed against former enemies under the slogan ‘Respond to the White Terror with Red Terror!’ Just as often it was a violent struggle between two local communists, each seeking power.Footnote ⁵² This was not always warlordism. The campaigns were often conducted by specific institutions, such as the local militia or the local communist party, rather than by charismatic individuals. But it was always frightening and destabilising for local residents, who feared for their lives and well-being. The spread of warlordism hurt civilian populations most of all. Beaten, raped, robbed, murdered and abused, civilians saw little benefit from warlordism. Indeed, by early in the civil war most civilians expressed their main political desire as a desire for stability more than a preference for one warring party over another. Their slogan was ‘Down with the Civil War’.Footnote ⁵³