Chris Harman


For Democratic Centralism

(July 1978)

From Socialist Review, No. 4, July–August 1978, pp. 37–39.
Transcribed & marked up by Einde O’Callaghan for the Marxists’ Internet Archive.

Richard Kuper’s case rests on six main arguments. Each seems at first sight to be almost unquestionable. Each is, in reality, fundamentally false and dangerous.

(1) Democracy and centralism are opposed to one another

This is an old red herring that has received a new lease of life in the most unexpected quarters with the international crisis of the revolutionary left. For instance, Alain Krivine of the French Trotskyist organisation the LCR said in an interview in the issue before last of this review, ‘democratic centralism – it’s two words which contradict one another’.

But the popularity of an idea does not prevent it being false. Democracy is a method by which an organisation takes decisions. Those decisions only make sense if they are binding on members of the organisation.

If they are not binding, there is no point in their being made. If a minority can ignore the will of the majority, why bother abut finding out the will of the majority? Why go to all the effort of having elections, counting votes and so on? You cannot have democracy without some means (moral or physical) of ensuring obedience to majority decision. When people enter a democratic organisation they necessarily surrender some of their freedom of action in favour of a centralised decision making process – whether or not the organisation calls itself ‘democratic centralist’.

It is absolute nonsense to pretend that there is something ‘undemocratic’ about this surrender of individual freedom: democracy depends upon limitation of individual freedom in the interests of majority decisions. You cannot have democracy without centralism. There are, of course, various forms this centralism can take. It can be the centralism of a ballot box into which the organisation’s members drop their voting papers; it can be the centralism of a mass meeting; it can be the centralism of a leadership elected to take day-to-day decisions. These various forms differ from one another significantly. But all are democratic and centralised.

The extent to which an organisation lacks this element of centralised decision making, it ceases to be democratic. The workers’ movement has a rich historical experience of non-centralised and therefore, in reality, undemocratic, forms of organisation.

For example, in Germany in the course of the first world war the old Social Democratic Party expelled those who objected to its collaboration with the government’s war effort. Nearly half the members broke away to form a new Independent Social Democratic Party (usually known by its initials as the USP). The new party reacted against the bureaucratic tendencies of the old by building a highly decentralised structure.

But this did not in practice result in a more democratic party than the old SPD. The MPs, trade union officials and newspaper editors of the new party, who tended to be on its right wing, were able to exploit the lack of a coherent centralised structure. Without it there was no discipline capable of making them obey the feelings of the majority of members (who veered rapidly towards the revolutionary left). They continued to publish anti-revolutionary articles and to make anti-revolutionary speeches.

Eighteen months after the foundation of the Independent Social Democracy a revolutionary party, the German Communist Party (Spartakus) was founded also on a decentralised basis. Again the result was the opposite of ‘democratic’! In this case the lack of centralism and discipline meant that the party was compromised on its ultra-left. Courageous, but foolhardy or inexperienced members were left to initiate armed actions, to call for localised seizures of power, and to form breakaway unions without any consultation with the organs of the party as a whole. The party could not control their actions, but still got the blame for them in the eyes of the rest of the working class. Without democratic centralism, minorities were able to take decisions for which the rest of the party was held responsible – and often paid for the responsibility with their lives.

These were not isolated .or accidental examples. Ultra-left adventurers and self-seeking careerists alike often relish in the joys of ‘decentralisation’ – because it means a movement they can exploit to their own advantage without being bound by its discipline. Today in Britain, for example, lack of a common discipline is one of the hallmarks of the Tribune group of Labour MPs. Why? Because it allows the members to enjoy an aura of ‘leftness’ without impeding their pursuit of careerist and opportunist policies. In the same way it is precisely the lack of centralism of the Broad Left in a number of unions that gives it such an appeal to aspiring bureaucrats. It can elect them but not control them.

Whatever the intentions of those who propagate it, the notion that ‘centralism contradicts democracy’ can only provide an ideological cover for the unprincipled and the self-important.

(2) Centralism may be needed in the revolutionary situation, but certainly not in the day-to-day struggles most of us are involved in now.

This assumes that we face a centralised enemy, capable of manoeuvre, of taking on and defeating us one at a time, on the day of the insurrection but not before. But we are always faced with enemies who are organised to manoeuvre against every struggle of the working class. Hardly a single strike takes place without management – and often the state – trying to turn workers in one place of work against workers elsewhere. Revolutionaries have to try to provide a single, unified, ‘centralised’ response in such situations. For, we know that if one group of workers returns to work while others stay out, the whole struggle can be smashed.

The struggle for power brings out in the starkest detail the need for centralisation. But it does not exhaust that need. Richard writes as if the Bolsheviks needed centralism only ‘between July and October’ 1917, when they were ‘concerned with the military preparations for the seizure of power’. Really Richard, have you not heard how in July they had to work, with a single, centrally determined will, to prevent a premature attempt to seize power.

(3) Centralism chiefly means hierarchy, ‘a massive concentration of authority’, ‘unbridled discipline, jumping to at a word’

The need for centralism flows out of the very character of the class struggle itself – of the way it is composed of battles, big and small. In any battle it is an advantage to an army to operate in a coordinated fashion, according to a single set of tactics.

All of us have experience of struggles where the most proficient fighters do not operate in this way, where militants make decisions of enormous importance in the heat of the moment without consulting their fellow militants, where demagogues arise who make speeches demanding action without having given a thought as to how the action is to be carried through, where no-one has bothered to assess the overall balance of forces or to seek out the weaknesses in the enemies ranks, where at best militants come together on an ad hoc basis, without any real knowledge of each others strengths and weaknesses, so that requiring the utmost reliability are given to those who are temperamentally unreliable and decisions requiring the coolest head are taken by those most easily carried away by events.

The party is necessary precisely to overcome these weaknesses at every level – to provide a pool of militants who have collectively trained themselves to intervene, whether in a sectional strike or a full blooded insurrection, to discuss with each other what to do, to assign tasks to one another on the basis of known abilities, to provide at least the crude elements of some centralised direction to the struggle. Such an organised network of experienced revolutionary socialists (often referred to as ‘cadres’) does not arise out of thin air. It takes many years for a party to develop the necessary traditions of coordinated, collective effort: otherwise it is all too easy in the heat of the struggle for members to forget the need for a cool head, for scientific appraisal and, above all, for coordination with the other members of the party.

So democratic centralism is essential, not just as an abstract national principle, but as the germ of party activity in each locality or factory. But once you accept the need for coordination and centralisation in this way, you also have to accept mechanisms to make it efficacious. You cannot organise a referendum of party members every time you want to decide whether to spread a strike or to respond to a racialist attack. The question is not whether there is going to be a centralised response (or non-response), but whether this response is properly organised.

The selection by the party of those of its members who are best able to take rapid decisions is, like the building of a tradition of collective intervention, a process that takes many, many years. The existing leadership has to be tested in struggle, its deficiencies discovered and corrected. Hence the importance of the democratic component of democratic centralism: it provides the mechanism by which the members discipline the leaders, feed the experience of the class struggle back into the centre of the party.

But none of this is possible unless leadership decisions are implemented. Does there then have to be blind obedience by the membership to every call from the leadership? There are all sorts of incidents in the class struggle which are not of a vital nature, which a centralised national leadership certainly cannot provide detailed guidance about. Here the unit of decision making is the branch, the workplace organisation of the party, or even the individual militant. The leadership has to try to coordinate these decisions by developing an overall theoretical and political perspective among the membership.

There are, however, many occasions on which the party has to move very quickly, as a single force. Then the leadership has to be able to demand immediate action from the membership and to be judged on its performance after the event, without debate beforehand. Otherwise the party reacts to emergencies in a disorganised way. The decisions of the leadership are not tested by the practice of the party as a whole, and so there is no easy way to tell whether the decisions were right or not.

(4) The real day to day activity of revolutionaries consists in ‘fostering and encouraging the self-confidence, self reliance, self activity of those among whom they work’, not in executing demands from the party centre.

This assumes that somehow ‘self activity’ and ‘self confidence’ arise independently of struggles in which centralised direction is necessary. But they do not. The ‘self activity’ of the working class develops through a struggle against the enemy class. As part of this ‘self activity’ revolutionary workers haveto be able to suggest ways of generalising the struggle, tactics that can produce victory. They can only do so successfully by suggesting tactics, by offering leadership, that fits in with the leadership offered by revolutionaries active in other parts of the class. The question of coordinated direction, of centralised leadership, necessarily arises again.

The existence of a centralised revolutionary party does not, therefore, form an obstacle to the self-activity of the masses – on the contrary, the latter is incomplete without it.

(5) ‘Under conditions of an ebbing revolutionary wave democratic centralist organisation has proved to be extremely harmful ... Once a monolithic party monopolises the loyalty of militants in a non-revolutionary situation, the construction of global alternatives (i.e. of revolutionary parties) is a non-starter’

Here Richard simply confuses effects with causes. It is true that in periods of defeat arid demoralisation workers parties with centralist notions have often declined into counter-revolutionary sects (whether big of small). But many workers organisations with non-centralist, notions have historically undergone similar degeneration. The cause of the degeneration lies in the period and the political response of the party leaders to it, not in the organisational form.

Again, once you have a party with mistaken views that ‘monopolises the loyalty of workers’, it makes the creation of new revolutionary parties very difficult, regardless of whether it is formally centralised or not. One of the most graphic examples of an organisation leading the class to defeat was that of the anarcho-syndicalist CNT in the Spain of the 1930s: its failure to see the need to build a workers state to wage the war against Franco left the room open for the liberals, social democrats and Stalinists to rebuild a bourgeois state structure that could only lose the war against Franco. The libertarian notions of the CNT did not stop it from ‘monopolising the loyalty’ of the most advanced workers.

For an existing revolutionary party, the answer to a declining revolutionary wave is not to do away with the whole notion of developing a leadership – it is to develop a leadership that knows how to retreat as well as how to advance.

(6) Democratic centralism leads to the party substituting itself for the class

You don’t have to be a democratic centralist to substitute yourself for the class. Trade union bureaucrats, councillors, MPs, do it every day. So do the proliferation of terrorist groups throughout the world with their ‘proletarian justice’ without the proletariat. So too do many ‘individual revolutionaries’.

In the case of Russia, the democratic-centralist Bolshevik party distinguished itself in the course of 1917 by its repeated refusal to substitute itself for the class. The ‘non centralist’ Mensheviks were prepared to go behind the backs of the workers to arrange governmental coalitions. The ‘non-centralist’ anarchists were prepared to substitute their own adventures for the mass action of the class. It was the Bolsheviks who insisted to waiting until, very late in the year, they had the majority support of the workers.

The problem after 1917 was not that the democratic centralism of the Bolsheviks led the party to substitute itself for the class – but that the decimation of the working class in the course of the civil war gave the party the crude choice between either substituting itself for the class and resisting the whites or seeing the victory of full blooded counter-revolution. As Trotsky noted, had that happened the word for fascism would not be an Italian word from 1922 but a Russian one from 1919.

Of course it is true that the Stalinists adopted the words ‘democratic centralism’ to describe their bureaucratic dictatorship, just as they took over the words ‘Communism’ and ‘Socialism’ to describe state capitalism. But that does not mean there was not another model at work in the revolution itself: after all, even as late as 1921 the platform of the workers opposition was printed in a quarter of a million copies on the official_party presses. No doubt that is why one of the early oppositions in the party referred to itself as the ‘democratic centralists’.

Incidentally, Richard is out of this world when he claims that the ‘Communist International was wrong when it affirmed at the Second Congress (1920) “the importance of the Communist Party does not diminish after the conquest of power by the working class, but on the contrary grows enormously”.’ Here Richard is playing with the ‘suicide’ view of the party, popular in semi-anarchist ‘council communist’ circles – the party is necessary to propagate a regime of workers councils, but must dissolve itself once this comes into existence.

What the arguments forgets is that the victory of the workers in one country is not the end of the struggle, but in many ways the beginning. It will be greeted with bitter and violent internal and external resistance, which will produce wavering away from the revolution among the lower middle classes and the less militant sections of workers. Only the most determined, organised, coordinated and centralised agitation by the advanced section of the workers – that is the most energetic activity of the revolutionary party – can lead the rest to defend the revolution. To believe otherwise is to substitute a pacifist dream for the reality of revolution and civil war.

(7) What matters is the correct politics and not the form of the organisational structure

Richard justifies this claim by writing, for example, that ‘no organisational structure was capable of compensating for the political inadequacy of the German left from 1919 to 1923’.

Rarely have the facts of history been so inverted for the purpose of argument. When the German revolution broke out in November 1918 the most experienced leaders of the revolutionary left – Rosa Luxemburg, Leo Jogiches, Karl Radek and Johannes Knief – had a very clear idea of what needed to be done. They all saw that it would take some months of partial struggles, especially of economic struggles, for the majority of workers to break with reformism and to support a dictatorship of workers councils. In the meantime, they insisted, it would be folly for the revolutionary minority of the class to try to seize power behind the backs of the rest.

Rosa Luxemburg’s articles for these months are absolutely clear on these matters. If the revolution went down to defeat it was not through ‘inadequacy of polities’ – it was because the politics of this leadership was not tied to a coherent ‘organisational structure’. There was not even the embryo of a party capable of transmitting the political analyses of Rosa into the key sections of the class. Indeed, such was the lack of a tradition of coordinated revolutionary activity that Karl Liekbnecht simply ignored the decisions of the rest of the leadership of the newly formed party and, in the heat of the moment, put his name to a call for the forcible overthrow of the Social Democratic government. The result was that the most advanced layer of militants blundered into a premature struggle for power, which led to the annihilation of much of the Communist leadership.

The tragedy in Gemany was that he democratic centralist party was not built until after the party had suffered major defeats and until after many of its best leaders had been murdered. Of course organisation is useless without the correct politics. But correct politics is impotent without organisation. To pretend otherwise is to guarantee in future a repetition of many of the massive defeats of the past. Yet this is effectively Richard’s conclusion.

Building a democratic and centralised revolutionary organisation is not an easy task. Our model cannot be the so-called ‘Marxist-Leninism’ that was elaborated after Lenin’s death by the new bureaucratic rulers of Russia. We have to develop forms of leadership that learn from the spontaneous struggles of workers, generalising the lessons, and feeding them back into the class. But that also means that our model cannot be the existing ritualised patters of bourgeois democracy. Few things are more stultifying for debate in a revolutionary organisation that a ‘government-opposition’ arrangement by which one section of the organisation feels that it is compelled as a matter of principle to oppose the elected leadership on every issue: this makes it extremely difficult for either the leadership or the opposition to learn from the concrete development of the class struggle.

Finally we have to remember that a small revolutionary organisation certainly is not the embryo of a new society. We do not exist as an island of socialism within capitalism, but as a voluntary organisation of militants whose task is to lead the class as a whole to construct the new society. So the aim of internal democracy is not to show ‘this is how things will work under socialism’, but to tie the development of the party to the concrete experiences of its militants in the workplaces.

But what then happens when the ‘democracy’ of the party fails to reflect the experiences of the most advanced sections of the class? When the party members have become routinised and cut off from new upsurges of spontaneous struggles, or when they come from milieus which have no real contact with the factories? In such cases, as Cliff argues in the first volume of his Lenin or as Trotsky argues in his Lessons of October the party leadership cannot simply sit back and reflect the ‘democratic will’ of a party that is lagging behind the class. It has to campaign vigorously for the sudden turns in the line of the party if necessary reaching to forces outside the party to pressurise the party members to shift their position.

This may not seem very ‘democratic’ to Richard, but the alternative is disastrous: it is to abandon the aim of building a vanguard organisation that can lead the class to power for the easier path of remaining a cosy, ‘democratic’ sect that coexists with the system. That is why the history of any serious revolutionary organisation – whether in the time of Marx and Engels, in the time of Luxemburg and Lenin, or today, is not just a history of linear growth from conference to conference, but also of ‘lurches’ in one direction and then in another, and on occasions of splits and expulsions.

It is not possible to tell from his article whether Richard has given up the perspective of building a revolutionary party for the easier option of friendly, ‘democratic’, non-centralised discussion circles. All he says is that ‘new organisational forms are urgently required’, without specifying what these are or how we are to get to them. But it does look as if those of us who remember the succession of massive defeats the working class has suffered because its most militant elements were not united in a revolutionary party are going to have to build without him.

Last updated on 13 September 2019