Collingwood's New Leviathan and classical elite theory

ABSTRACT R. G. Collingwood's New Leviathan (1942) presents an account of two ‘dialectical’ political processes that are ongoing in any body politic. Existing scholarship has already covered the first: a dialectic between a ‘social’ and a ‘non-social’ element, which Collingwood identifies in Hobbes. This essay elucidates a second: a dialectic between Liberals and Conservatives, which regulates the ‘percolation’ of liberty and the rate of recruitment into what Collingwood calls ‘the ruling class’. The details of this second dialectic are to be found not in Hobbes, but in the work of Vilfredo Pareto and Gaetano Mosca, yet Collingwood's connections to these fathers of ‘classical elite theory’ have not previously been discussed.

Interpreted as such, The New Leviathan earned early praise from readers including H. H. Price, W. D. Ross, and J. L. Austin, 9 and today it is beginning to be recognized again as a serious achievement among political theorists beyond the circle of Collingwood specialists. Robin Douglass has furnished the History of European Ideas with a close study of The New Leviathan, focusing on Collingwood's reading of Hobbes, which Douglass finds 'strikingly original'. 10 This originality is due in part to Collingwood's recognition that, for Hobbes, 'a body politic is a dialectical thing'. 11 As Douglass explains, Collingwood is following a distinction found in Plato between two types of discussion, 'eristic' and 'dialectic'. 12 In 'eristic' discussion 'each party tries to show that he was right and the other wrong', whereas in 'dialectical' discussion 'you aim at showing that your own view is one with which your opponent really agrees'. 13 Douglass also outlines the dialectic that Collingwood posits in a body politic between a negative element of 'non-sociality' and a positive element of 'sociality', which, according to Collingwood, is what Hobbes means by 'nature' and 'commonwealth'. 14 A further achievement of Hobbes', as Collingwood reads him, is the recognition that this dialectic can go either way. 15 A great deal more can be added about the manifold roles of 'dialectic' in Collingwood's philosophy of man, politics, and civilization beyond what he finds in Hobbes. Much of this has been amply covered by existing literature, and I will not attempt an overview here. 16 Rather, I would like to pay special attention to a second political dialectic, which narrates the correct relationship as Collingwood sees it between the Liberal and Conservative parties that dominated the Westminster system until 1922, 17 and between the timeless principles of democracy and aristocracy that he thinks those parties embody. Scholarly attention to this second political dialectic, and particularly to what it claims about the political role of Conservatives and 'aristocracy', has been very fleeting. 18 It is perhaps a natural neglect, since Collingwood tells us in his Autobiography that his attitude towards politics 'had always been what in England is called democratic and on the Continent liberal'. 19 Specialists have therefore discussed at length Collingwood's reception and Andrew Vincent, British Idealism and Political Theory (Edinburgh: Edinburgh University Press, 2000), 185; Nicholson, 'Collingwood's New Leviathan'; Gary Browning, 'New Leviathans for Old ', Collingwood Studies 2 (1995): [89][90][91][92][93][94][95][96][97][98][99][100][101][102][103][104][105][106]and Douglass,'Leviathans Old and New'. 9 See Connelly and Johnson,'Collingwood's The New Leviathan',[10][11] See Douglass, 'Leviathans Old and New', 529. 11 Collingwood,New Leviathan,24.68. See Douglass,'Leviathans Old and New',538. 12 See Plato, Meno, in John M. Cooper (ed.), Complete Works (Indianapolis/Cambridge: Hackett, 1997), 875 (75 c-d). 13 Collingwood, New Leviathan, 24.57-9. 14 Douglass, 'Leviathans Old and New', 538-9. 15 Douglass, 'Leviathans Old and New ', 539. 16 See especially Richard Murphy, Collingwood and the Crisis of Western Civilisation (Exeter: Imprint, 2008), 183-69; Boucher,Social and Political Thought,; James Connelly, Metaphysics, Method and Politics: The Political Philosophy of R. G. Collingwood (Exeter: Imprint, 2003), 50-2, 73-4, 94-5; Jan van der Dussen, 'Collingwood on the Ideas of Process, Progress and Civilization', in Philosophy, History and Civilization: Interdisciplinary Perspectives on R. G. Collingwood, ed. D. Boucher, J. Connelly, and T. Modood (Cardiff: University of Wales Press, 1995), 246-68; and Gary K. Browning, Rethinking R. G. Collingwood: Philosophy, Politics and the Unity of Theory and Practice (Basingstoke: Pallgrave Macmillan, 2004), 120-43. 17 It seems initially strange that Collingwood does not name the Labour party, since The New Leviathan is a thoroughly reconciliatory text from the parliamentary point of view, and by 1941 Labour was the second largest part of the National Government, with Labour MPs outnumbering those of the combined Liberal parties in the House of Commons three to one. Noel O'Sullivan notes that this omission was quite deliberate, but identifies it all the same as a weakness: 'The development of political consciousness in the working class … was one phenomenon which Collingwood found impossible to accommodate within his political theory'. N. K. O'Sullivan, 'Irrationalism in Politics: A Critique of R. G. Collingwood's New Leviathan', Political Studies 20, no. 2 (1972): 141-51, 147-8. Collingwood is, I think, allowing readers to think of Labour as another, newer, representative of 'democracy', rather than of non-dialectical class war. For some of Collingwood's criticisms of Marx and socialism,see New Leviathan,12.95,25.33,37.58. See also Collingwood,Essays in Political Philosophy, and Peter Johnson, A Philosopher and Appeasement: R. G. Collingwood and the Second World War: A Philosopher at War Volume 2 (Exeter: Imprint Academic, 2013), 142. 18 Although other authors have mentioned it in passing, commentary on this political dialectic specifically has been limited to work by David Boucher, in Social and Political Thought, 163-6;in Collingwood, Essays in Political Philosophy, 14-15;in Collingwood, New Leviathan, xliv-xlv;and in R. G. Collingwood, An Autobiography and Other Writings [1939]  reformulation of a liberal and democratic tradition of political thought, and have always acknowledged the largely Italian flavour of Collingwood's brand of idealism, which draws especially on Benedetto Croce, Giovanni Gentile, and Guido de Ruggiero. 20 It is not my intention here to reclassify Collingwood's political thought or to overturn any of this 21though I have explained elsewhere how his liberal-conservative dialectic might be useful in today's studies of conservatism. 22 It is true that Collingwood has to offer some sort of defence of the Conservative party because part of his purpose is to defend the 'talking-shop at Westminster' against the one-party, non-dialectical politics of Fascism. 23 But if we recognize that this dialectic in which conservatism and 'aristocracy' play an important role is also intended as a serious contribution to political philosophy, then hitherto under-explored aspects of Collingwood's thought are opened up.
First we can see how thoroughly Collingwood revises a dialectic that has in fact been long recognized in Western political thought. And second, we discover striking parallels with the work of two further Italian authors who have not been discussed before in relation to Collingwood: Vilfredo Pareto and Gaetano Mosca, the two leading figures of what has since been called 'classical elite theory'. 24 (I will continue to use the term for convenience).

Collingwood's democratic-aristocratic dialectic
In chapters fifteen to seventeen of The New Leviathan, Collingwood explains that the difference between the Liberal and Conservative parties concerns what he calls 'the percolation of liberty throughout every part of the body politic'. 25 It is the avowed aim of Liberals to 'hasten' that process, and the avowed aim of Conservatives to 'retard' it. 26 However the relationship is not eristical, but dialectical, owing to two important points of agreement. First, 'Both [parties] held it as an axiom that the process of percolation must go on'. 27 And second, both parties recognize that there exists an 'optimum velocity' for this processthat 'if it went too fast, and equally if it went too slow, the whole political life of the country would suffer'. 28 For Collingwood, this process of percolating liberty operates by 'recruiting' people into what he calls the 'ruling class' from the 'ruled' class, 29 or 'non-social community'. 30 The fact that a body politic is 'divided into a ruling class and a ruled class' Collingwood calls 'the first law of politics'. 31 There are three of these 'laws'. The second is that 'the barrier between the two classes is permeable in an upward sense', 32 and the third is that 'there is a correspondence between the ruler and the ruled, whereby the former become adapted to ruling these as distinct from other persons, and the latter to being ruled by these as distinct from other persons '. 33 contemporaries that he had 'converted to Marxism, or 32 Collingwood, New Leviathan, 25.8. 33 Collingwood, New Leviathan, 25.9. However, Collingwood does not provide a very detailed explanation of why percolating liberty and recruiting into the ruling class are in fact the same process. The logic seems to be that the liberty of an individual is furthered in and through the social liberties and duties that are bestowed by his being recruited into a social position that involves 'ruling' in some sense. But this equation of the percolation of liberty and the recruitment of people from the non-social community is a little awkward, since it seems to imply that the only people who are free in a body politic are the members of its ruling class, while being 'ruled' means being unfree. The awkwardness is mitigated by Collingwood's imprecision about what exactly the 'ruling class' is. It does not seem to mean central government, or any specific institutions. The awkwardness is also somewhat eased by Collingwood's claim that freedom is a matter of degree. 34 People are not simply free or unfree; rather, from less free they grow freer as they are made increasingly 'mentally adult', 35 and are 'assimilated in psychological character' to their rulers. 36 Accordingly, they are able to rule and be ruled by consent (or 'authority') rather than by force. 37 But as well as being liberating for individuals, this process of recruitment is also necessary for the maintenance of the ruling class. Comprising mortal beings, the ruling class must be perpetually 'replenished', or it will grow too small or too weak to solve new political problems. The perpetual replenishment of the ruling class is therefore 'the process which is the life of the body politic'. 38 Collingwood does not explain how and when the decision to recruit is made, or by whom, since such particulars will vary. But the ruling classany ruling classalways faces the question of how quickly it must replenish itself in order to be as 'strong' as possible. 39 And to this question, Collingwood says, two answers are always given: Democracy answers: "By enlarging it [the ruling class] so far as is possible. By recruiting into it, to discharge one or other function, every member of the ruled class who may constitute an addition to its strength." Aristocracy answers: "By restricting it so far as is needful. By excluding from its membership everyone who does not or would not increase its strength." 40 The liberal-conservative dialectic, then, is really a democratic-aristocratic dialectic. But Collingwood is deploying these terms, 'democracy' and 'aristocracy', rather idiosyncratically. They do not denote the relative size of the sovereign power, or 'citizen-body', as they do for Aristotle, 41 but instead stand for two different principles of ongoing recruitment into it. So long as both principles are allowed to influence the rate of percolation and recruitment, by means of a relationship that is dialectical rather than eristical, that rate will not exceed 'optimum velocity'. 42 Collingwood intends this democratic-aristocratic dialectic as universal, and he traces its operation through Western political history, from ancient Greece and Rome, through the Middle Ages, the English Civil Wars, and the American and French Revolutions, to 'the arch-Fascists and arch-Nazis of to-day'. 43 Each 'revolution' was indeed democratic, Collingwood says, but each was also aristocratic.
In England, he explains, 'democracy' and 'aristocracy' are embedded in the Liberal and Conservative parties respectively. Liberals, as 'democrats', ensure that the rate of recruitment does not fall 34 Collingwood, New Leviathan, 20.62, 21.8, 25.41. 35 Collingwood, New Leviathan, 20.23. 36 Collingwood, New Leviathan, 27.74. 37 Collingwood, New Leviathan, 20.45-8. 38 Collingwood, New Leviathan, 26.16. 39 Collingwood, New Leviathan, 26.13. 40 Collingwood, New Leviathan, 26.14-15. This chapter was probably complete by January 1941 (see Connelly and Johnson, 'Collingwood's The New Leviathan', 7). It is interesting that Collingwood considers part two of The New Leviathan to be the part that 'covers what is called social and political theory', while part three, in which this discussion falls, 'is about Civilization'. See Connelly and Johnson, 'Collingwood's The New Leviathan', 9. 41 Aristotle, The Politics, trans. T. A. Sinclair, ed. Trevor J. Saunders (London: Penguin, 1992), 189-90 (III, vii). 42 Collingwood, New Leviathan, 26.19, 27.89. 43 Collingwood, New Leviathan, 26.4-96. below the 'optimum velocity', while Conservatives, as 'aristocrats', ensure that the optimum velocity is not exceeded. For Collingwood it is true, then, that serving the interests of 'aristocracy' is a Conservative principle, as critics have often said. But of course by 'aristocracy' Collingwood does not mean what Aristotle means, and he certainly does not mean to indicate the sort of social class that Paine sought to excoriate in his Rights of Man. 44 If either principle ceased to exert its force, Collingwood warns, 'the whole political life of the country would suffer'. 45 It is because both parties are allowed to realize their objectives, and neither is allowed to defeat the other once and for all, that the 'optimum velocity' of percolation is maintained. Collingwood doubts that the parties were ever fully aware of their proper dialectical function. 46 But on occasion, he adds, they agreed on their joint task so completely that 'one party could steal the other's thunder'as when Disraeli's 1867 Reform Bill 'dished the Whigs'. 47 Understanding the dialectical rather than eristical nature of the two-party system is, for Collingwood, indispensable, and he places the blame squarely upon the Liberals for their own demise. They 'did not understand the dialectic of English politics'. 48

Sources of dialectical politics
Today the idea of 'dialectic' is associated above all with Hegel, 49 and indeed some readers have sought to highlight 'Hegelian' themes in Collingwood's writings. 50 It is Gary Browning's view, for example, that The New Leviathan 'highlights the profundity of the impact of Hegel on Collingwood's thought', 51 and Andrew Vincent's that Collingwood's peculiar reworking of social contract theory in The New Leviathan owes to Hegel's critique of it. 52 These judgements may be accurate, but it is worth pointing out that Collingwood's democraticaristocratic dialectic is not to be found in Hegel. Hegel does posit a mediation between the 'particular' interests of the professions and estates on the one side, and the 'higher viewpoints' of the state on the other; 53 he explains that the estates mediate between individual selfishness and universality, 54 and civil servants mediate between them and the state; 55 and these relationships are 'dialectical' insofar as, while the interests of the corporation may conflict with those of the state, the state also supports the legal recognition of the corporation. 56 This process, which loosely speaking pertains between something like an economic class and something like a political class, might be characterized as what Browning calls a 'constant process of accommodation between the rulers and the ruled', 57 and in that sense it could serve as an example of Collingwood's third law of politics. But what Hegel is describing is not a relationship between political parties or factions, or between 'democracy' and 'aristocracy' as principles of recruitment, and where Hegel does discuss the 'rise' of individuals from lower ranks, he says only that this occurs 'in a contingent manner and as particular  54 Hegel, Philosophy of Right § 201A. 55 Hegel, Philosophy of Right § 289. 56 Hegel, Philosophy of Right § 289. 57 Browning, Rethinking R. G. Collingwood,137. circumstances require'. 58 Even Browning, who has presented the most sophisticated case for reading The New Leviathan as 'Hegelian', recognizes that 'the possibility of a transition between being ruled and ruling' is a reworking more of Plato than of Hegel. 59 And although Browning also points to that 'dialectical' character of politics which, for Collingwood, 'provides the spirit of reasonableness that allows for reasonable argument over the extent and manner of democratic reform', he rightly observes that Collingwood is repurposing Plato here as well. 60 What, then, are the sources of Collingwood's democratic-aristocratic dialectic, if not Hegel? Although Collingwood barely points it out, 61 there is in fact a very long history in Western political thought of the notion that there should be a complementary relationship between the characters, factions, or principles of something like innovation and something like conservationenergy and moderation, progress and tradition, novelty and order, action and reaction, and so on. We find versions of this mysterious partnership, which is varyingly supposed to produce something like peace, harmony, or progress, in Plato, Machiavelli, Burke, Tocqueville, Coleridge, and in Milland even among some early-twentieth century politicians. 62 Coleridge's formulation bears an extra similarity to Collingwood's, in that it contains a distinction between 'contrary' and 'opposite powers' that echoes the Platonic formulation favoured by Collingwood. Unlike 'contrary' powers, Coleridge thinks, 'opposite' powers 'tend to union' and, in the specific case of 'permanence' and 'progressiveness', 'so far from being contrary interests … [they] suppose and require each other'. 63 A very similar dialectic is also narrated in Guido de Ruggiero's History of European Liberalism, first published in English in 1927. De Ruggiero's formulation is worth extra attention, as Collingwood was its translator, and not a disinterested one. He knew de Ruggiero personally,64 and in a letter of 1926 had written that 'The political principles expounded and implied [in de Ruggiero's History] are at every point my own, and expressed with a justness and completeness that leave me nothing to do but express my complete agreement'. 65 The dialectic that de Ruggiero describes is somewhat similar to that which Collingwood would later outline in The New Leviathan. It is a dialectic, de Ruggiero writes, 'of continuity and antithesis', 66 in which the 'ideal value of the principle which lies at the root of Conservatism … nourishes the dialectical antithesis of the Liberal thesis'. 67 It operates by 'the alternate rule of parties', 68 and unites 'the principle of conservation with that of progress, Radical initiative with historical tradition'. 69 Like Collingwood's, this dialectic is also about propagating freedom: together the two forces 'are aiming at a democracy of free men: at instilling a sense of autonomy into the masses, fostering a spirit of spontaneous association and co-operation'. 70 And from this perspective, de Ruggiero writes, 'the development of the struggle between Conservatives and Liberals, and their alternation in power, represent not an alternation of freedom and unfreedom, light and darkness, but the rhythm of an 58 Hegel, Philosophy of Right § 279. 59 Browning, Rethinking R. G. Collingwood, 135-6. 60 Browning, Rethinking R. G. Collingwood,137;  uninterrupted movement'. 71 All this seems very familiar to readers of The New Leviathan, though of course de Ruggiero's expression is more ornamented than the punchy decimalized assertions of Collingwood's 'late style'.
However, it is not at all the case that Collingwood's revisions of de Ruggiero's dialectic are merely stylistic. Upon closer inspection, de Ruggiero's dialectic is not actually between liberalism and conservatism, and it is not between democracy and aristocracy either. Rather, it is between democracy on the one side and liberalism on the other. 'Democracy' for de Ruggiero means the political force, against the status quo, of popular equality and unity/uniformity. 72 'Liberalism' on the other hand is the force, also against the status quo, for personal liberties, 73 the 'free play of individual forces', and of 'leaving men so far as possible to act for themselves'. 74 Both democracy and liberty, in de Ruggiero's understanding, are principles of self-government, but for different entities: democracy is the self-government of the people, liberalism the self-government of individual persons.
Collingwood's dialectic is significantly different in that, as we have seen, it regulates the percolation of liberty by the concrete process of recruiting into the 'ruling class'. Recruitment forms no part of de Ruggiero's explanation. Collingwood therefore redeploys 'democracy' as a recruitment principle, and aligns it with the Liberal party; and, on the other side, introduces 'aristocracy' as a different but equally necessary recruitment principle, and offers it as the political function of Conservatives. Further, for de Ruggiero conservatism plays no consistent dialectical role: it is simply resistance to either 'democracy' or 'liberalism', which, when active, work against 'passive mechanism and the conservatism of routine'. 75 Collingwood's version, then, evinces significant revisions of the account with which he had once wished to express his 'complete agreement': namely the mechanism of recruitment, and the dialectical function of restricting access to the ruling class.

Collingwood and classical elite theory
Collingwood's innovation is to revise an old theory of something like a liberal-conservative dialectic such that it is transformed it into an account of how power and freedom are bestowedand indeed must be bestowed. That is to say, rather than using the old dialectic to explain an abstract and general state of something like peace, harmony, or stability, he uses it to explain the process by which power is distributed, restricted, and perpetually diffused in a heterogeneous and evolving body politic.
Collingwood achieves this by accommodating into the structure of The New Leviathan a certain theory of the ruling class: (1) what it is and how it rules, (2) how it replenishes itself, and (3) the dangers of excessively restricted access to it. These key points are answered by Collingwood's 'three laws of politics', and his explanation of each reveals a surprisingly close correspondence with Pareto's Les systèmes socialistes (1902) and Trattato di sociologia generale (1916), and especially with Gaetano Mosca's Elementi di scienza politica (first published 1896, second edition 1923). 76 I do not wish to suggest that Collingwood could not have made his modifications without any other authors' work in mind, and it is in precisely such a case as this that we should heed Quentin Skinner's warning against identifying 'influence' simply because 'An argument in one work may happen to remind the historian of a similar argument in another and earlier work'. 77 Collingwood has already suffered his share of that sort of historiography, 78 and it is not true that he must have read Pareto or Mosca. Peters (in his History as Thought and Action) provides a detailed demonstration that Collingwood's thought is significantly independent of the Italian thinkers who some readers have assumed to be his 'masters'.
Nevertheless historians must sometimes contribute to their field by making the strongest possible case for an ambitious verdict, if only so that any shortcomings can be explicitly documented. With this in mind, I think we have good reason to say that chapters fifteen to seventeen of The New Leviathan present Collingwood's use of Italian classical elite theory, and that, although not certain, it is very likely that Collingwood had read Pareto and Mosca. Collingwood's account of the ruling class and its operation is more concise, but he seems to have incorporated into it the chief ideas and formulations that are to be found in the classical elite theorists, and the terminology that he uses is often identical with theirsespecially with Mosca's. It seems that the New Leviathan project gave Collingwood the perfect occasion not only for reading the classical elite theorists (perhaps again), but also for synthesising, adapting, and responding to ideas and language that seemed to be gaining new currency, and for shaping the discussions of new and future Anglophone readers of Pareto and Mosca. The case for this verdict proceeds as follows.
The English translations of Pareto's Trattato di sociologia generale (The Mind and Society) and Mosca's Elementi di scienza politica (The Ruling Class) appeared in 1935 and 1939 respectively. The New Leviathan was begun in the autumn of 1939, 79 and chapters fifteen to seventeen were probably complete by January 1941. 80 It is true that neither Pareto nor Mosca is named in it, but this fact is unremarkable, as it was Collingwood's principle 'never to name a man except honoris causa', 81 and he would have had good reason at this time not to advertise his agreement with Italian 'elitists'. 82 Of course Collingwood would not have had to wait for English translations. He read Italian and French very proficiently, so he could have been aware earlier than almost any of his British peers of Pareto's Trattato and Les systèmes socialistes, and of Mosca's Elementi. Indeed, he ought to have been aware of them, as all three appear in the bibliography of his own translation of de Ruggiero's History of European Liberalism. 83 He might also have been alerted to the nature of Mosca's work by a very favourable review by Croce, published in La critica in 1923, 84 two years before Croce and Mosca were made representatives of the Italian Liberal Party in the Senate 85or (later) by a discussion in The Review of Politics. 86 Either way, it seems very unlikely that Collingwood would have completely missed the appearance of the English translations, given his interest in recent developments in Italian thought, his reading history, and his personal connections in Italy.
Collingwood's knowledge of Mosca is also suggested in private correspondence. In 1935 Humphrey Milford, head of Oxford University Press's London operations, had forwarded to Collingwood a memorandum from OUP's New York office containing some sort of proposition in relation to 'the work of Gaetano Mosca', and asking for Collingwood's opinion. 'You no doubt know all about Mosca', Milford explains. 87 Collingwood's reply is lost, but he seems to have been interested, and three days later Milford writes again: 'Thanks very much. If the MS. [manuscript] comes over here, I will ask you to look at it, if you will'. 88 The manuscript Milford is talking about is presumably that which, in the end, came out through a different New York publisher, edited by Arthur Livingston . There is no hard evidence that Livingston was known to Collingwood, but there are suggestive personal connections beside this manuscript. Like Collingwood, Livingston had devoted much of his career to introducing Italian writers to Anglophone readers, including Croce, whose The Conduct of Life he had translated and published in 1915. 89 (Collingwood's translation of Croce's Autobiography was published in 1927.) 90 Both men were also vociferous anti-Fascists, corresponding with other anti-Fascists in Italy and elsewhere.
These, then, are some of the suggestive extra-textual connections between Collingwood and the classical elite theorists. It is my intention in what remains to provide an overview of the textual connections. The relevant points concern specifically (1) the nature of the ruling class; (2) how it rules; (3) its renovation and the meaning of 'democracy', and (4) the meaning of 'aristocracy' and the danger of its excessive strength.

The ruling class
The first point of agreement is obvious from even a superficial reading. For Collingwood, as for Pareto and Mosca, there is in all societies an organized ruling minority which directs or controls the less organized majority. 91 As we have already seen, Collingwood makes the universal existence of a 'ruling class' and a 'ruled class' his 'first law of politics'. 92 Though Pareto is usually credited with pioneering the theory of 'governing élites' (his preferred term), 93 there has been some debate about whether Mosca's formulation is the earlier, and partly the source of the other's. 94 There is, Mosca explains, a class that rules and a class that is ruled. The first class, always the less numerous, performs all political functions, monopolizes power and enjoys the advantages that power brings, whereas the second, the more numerous class, is directed and controlled by the first. 95 This is one of the 'constant facts and tendencies that are to be found in all political organisms … from societies that are very meagrely developed and have barely attained the dawnings of civilization, down to the most advanced and powerful societies'. 96 Mosca does not consider this to be a remarkable or original discovery. He thinks he is simply applying Taine's historical method to societies in general, he recognizes formulations of the same theory in 97 and he thinks that the existence of 'political' or 'ruling' classes is 'so obvious that it is apparent to the most casual eye'. 98 At any rate, it is Mosca's term, 'ruling class', rather than Pareto's 'governing élite', that we also find in Collingwood's formulation. 99 In relation to The New Leviathan, it is especially interesting that the basic claim of classical elite theory is actually a departure from Hobbes, whom of course Collingwood is all the time claiming as his model. Mosca details two widespread 'habits' of political analysis that, he thinks, distort the true nature of politics. The first is the assumption that in every political organism 'there is one individual who is chief among the leaders of the ruling class as a whole and stands, as we say, at the helm of  95 Mosca, Ruling Class, 50. 96 Ibid. 97 Mosca, Ruling Class, 329. 98 Mosca, Ruling Class, 50. 99 One of Pareto's formulations (Mind and Society, 2047), however, runs as follows: 'The least we can do is to divide society into two strata: a higher stratum, which usually contains the rulers, and a lower stratum, which usually contains the ruled. That fact is so obvious that it has always forced itself even upon the most casual observation, and so for the circulation of individuals between the two strata'. state'. 100 But, Mosca points out, it is hard to imagine how 'all men would be directly subject to a single person without relationships of superiority or subordination'. 101 Such an individual … would certainly not be able to govern without the support of a numerous class to enforce respect for his orders and to have them carried out … he certainly cannot be at odds with the class as a whole or do away with it. Even if that were possible, he would at once be forced to create another class, without the support of which action on his part would be completely paralyzed. 102 Collingwood does not provide as much detail on the internal structure of this class as Pareto or Moscaor in fact Weber. 103 But the basic notion is nevertheless there in The New Leviathan, where it is explained that 'The ruling class may … be subdivided into a multiplicity of graded subclasses … Thus the ruling class as a whole becomes a hierarchy of ruling subclasses'. 104 Here, then, by asserting as his 'first law of politics' the basic thesis of classical elite theory, Collingwood has not followed Hobbes, but has instead adopted something very close to Mosca's implicit correction of Hobbes.

Organization, force, and authority
For Collingwood, the ruling class can only rule because it is organized. It is a society, possessing social consciousness and 'joint will'. 105 (He also uses the term 'corporate will'.) 106 Relations between its members are 'civil'that is, based upon mutual and self-respect and agreement. 107 Political rule therefore 'is exercised by authority of the society'. 108 The ruled, on the other hand, are those who have not awoken to 'consciousness of their own and each other's freedom', 109 and who are therefore not sufficiently able to will jointly. This is why it is always also necessary to some degree to rule by 'force' 110not physical force, but 'force' as superior moral or mental strength. 111 Thus, for Collingwood, the ruling class rules by authority and force simultaneously, 112 and can do so only because it is socially organized.
These terms, 'force' and 'authority', are somewhat Hobbesian, 113 even if the distinction itself is not. But we can find the same distinction in Weber, who distinguishes domination by 'control' from domination by 'authority', 114 and in Pareto, who distinguishes between 'lion' elites that rule by 'force', and 'fox' elites that use 'cunning'. 115 Pareto is obviously drawing on Machiavelli's political methods and animal imagery. But unlike Collingwood, he also divides 'force' and 'cunning' between two essential types of elite that, as he understands it, attain power alternately in any history of successive revolutions. An elite that rules by 'force' is replaced by a new elite that operates by 'cunning', Pareto thinks, but as the people grow tired of political 'cunning', they again become attracted to a more 'forceful' elite, which attains power with the help of popular support. It has been noticed however that, for all this, Pareto's Mind and Society does not contain the observation that political elites are cohesively organized, that they have a 'group' character, and that, because of this lacuna, Pareto cannot really explain how it is that elites get their way. 116 Mosca however, like Collingwood, recognizes that being able to 'act in concert' 117 is the ruling class's great political advantage: 'A hundred men acting uniformly in concert, with a common understanding', he explains, 'will triumph over a thousand men who are not in accord'. 118 More strikingly, and also like Collingwood, he observes that 'Any political organization is both voluntary and coercive at one and the same time'. 119 The terminological difference here -'coercion' instead of 'force', volition instead of 'authority'should not distract us from the shared recognition that voluntary and coercive organization are not mutually exclusive, and in fact operate simultaneously. There is always some degree of consent in any political organization, Mosca points out, and the amount of consent depends upon the extent to which the ruled class 'believes in the political formula by which the ruling class justifies its rule'. 120 But there are also always those who do not consent, or cannot yet consent, but who are nevertheless coerced by nature and necessity into political organization, 'the human being finding himself unable to live otherwise'. 121

Democracy and replenishment
Of course for Collingwood, as we have seen, the aim should not be to rule the non-social community by force for its own sake, but rather to 'civilize' its members individually, so that they can be recruited into the ruling class, thereby replenishing and strengthening it. For Pareto too the governing elite is 'restored' when it loses 'its more degenerate members' and when 'vigorous' families rise from the lower classes. 122 The governing elite is therefore 'always in a state of slow and continuous transformation' 123though the 'velocity in circulation has to be considered not only absolutely but also in relation to the supply of and the demand for certain social elements'. 124 As for Collingwood, there is an optimum velocity, which is determined not absolutely, but relative to timely requirements or contingently, as Hegel puts it. 125 Collingwood does not use Pareto's term 'circulation', 126 and neither does he explain the continuous transformation of the ruling class in terms of rising 'families'. His explanation of what opens access to the ruling class, 'rule-worthiness', is also more prosaic than Pareto's: it is not 'vigour', but competence and willingness to do the required work that is, or should be, the criterion of recruitment into the ruling class. 127 Collingwood's version is, again, much closer to that of Mosca, who describes the ruling class being 'renovated' or 'restocked' 128 when 'individual energies' and 'certain individuals … force their way from the bottom of the social ladder to the topmost rungs'. 129 The relevant criterion for recruitment into the ruling class 'will be different according as the system is autocratic or liberal, or as the democratic or aristocratic tendency prevails'. 130 Pareto, Mind and Society, 2055. 125 Hegel, Philosophy of Right § 279. 126 Pareto, Mind and Society, 2025. 127 Collingwood, New Leviathan, 26.28. 128 Mosca, Ruling Class, 68. 129 Mosca, Ruling Class, 67. 130 Mosca, Ruling Class, 402. merit'. 131 Ruling classes are 'continually replenished through the admission of new elements who have inborn talents for leadership and a will to lead, and so prevents that exhaustion of aristocracies of birth which usually paves the way for great social cataclysms'. 132 It is a process, he later adds, that is 'indispensable to what is called "progress"'. 133 Like Collingwood, Mosca posits two 'tendencies' that prevail alternately in the renovation process. His phrasing is in fact rather exquisite: the history of civilized mankind, he writes, has witnessed 'an unending ferment of endosmosis and exosmosis between the upper classes and certain portions of the lower'. 134 For the first tendency, 'which aims to replenish the ruling class with elements deriving from the lower classes', 135 Mosca considers the term 'liberal', before explaining why, in his view, 'the term 'democratic' seems more suitable'. 136 'Democracy' is the tendency which 'results in a more or less rapid renovation of ruling classes'. 137 Collingwood, as we have seen, uses 'democracy' in exactly the same way.

Aristocracy and class isolation
Also like Collingwood, Mosca identifies 'aristocracy' as what 'we would call the opposite tendency', which, when it prevails, 'produces closed, stationary, crystallized ruling classes'. 138 What Collingwood shares with Mosca on this point is also what differentiates his dialectic from that of de Ruggiero. More so than de Ruggiero, Mosca and Collingwood are implicitly dismantling the Aristotelean classification of constitutionsmonarchies, aristocracies, democracies. (By radically 'updating' Aristotle, Collingwood, as no doubt he realized, is also following Hobbes.) 139 But in his redeployment of Aristotelian constitutional terminology, and especially in his recognition that democracy and aristocracy are not mutually exclusive, but are instead two essential 'tendencies', Collingwood is again much closer to Mosca. 'What Aristotle called a democracy', Mosca observes, 'was simply an aristocracy of fairly broad membership', and although 'the doctrine of popular sovereignty still holds sway over many minds', nevertheless 'modern scholarship is making it increasingly clear that democratic, monarchical and aristocratic principles function side by side in every political organism'. 140 For Collingwood, as we have seen, the principle of aristocracy is to slow the rate of recruitment/ percolation when it threatens to go too fast. But still, if the rate were allowed to fall below the 'optimum velocity'that is, if aristocracy were to become too strong -'the whole political life of the country would suffer'. 141 Collingwood does not really specify the exact effects of a suboptimal velocity of recruitment, and the reader is left to elaborate or piece this together for himself.
Pareto describes the dangers of a 'slowing-down of class-circulation' in more detail. He explains that 'decadent elements' accumulate in the higher classes, 142 while 'superior elements' build up in the lower classes. 143 Such a 'disturbance in the equilibrium' is likely to produce a revolt, during which a new elite, made up of more vigorous 'leaders from the higher strata', tactically uses the 'combative' attitude of the lower strata to overthrow the incumbent elite. 144 This is why, as Pareto's celebrated remark has it, 'History is a graveyard of aristocracies'. 145 The essence of 'aristocracy' according to Mosca is, however, notably differentand here we can see that Collingwood is not simply incorporating Mosca's account wholesale either. Aristocracy for Mosca means the tendency 'to stabilize social control and political power in the descendants of the class that happens to hold possession of it at the given historical moment'. 146 Accordingly, Mosca's version of the democratic-aristocratic dialectic is slightly different from Collingwood's. It is 'a conflict between the tendency of dominant elements to monopolize political power and transmit possession of it by inheritance, and the tendency towards a dislocation of old forces and an insurgence of new forces'. 147 Thus, when the aristocratic tendency exerts too much influence, a 'group spirit, a sense of caste, arises and asserts itself, so that the members of the aristocracy come to think of themselves as infinitely superior to the rest of men'. 148 'Class isolation' begins to set in, 149 and this prevents the rulers from understanding, and therefore from sympathizing with, the sorrows and tribulations of those who live on the lower rungs of the social ladder; and they are equally insensitive to the toils and efforts of those who have managed to climb a rung or two by their own achievement. 150 Further, because the rulers avoid contact with 'the lower strata of society', they 'are left in complete ignorance of real psychological conditions in the lower classes', with the overall result of 'depriving the ruling classes of any influence whatever on mental and sentimental developments in the masses, and so of unfitting the ruling classes for managing them'. 151 Collingwood does not understand aristocracy as a specifically hereditary principle of bestowing power. Neither does he include Mosca's specific prognosis of 'class isolation' in his discussion of aristocracy. But Mosca's account is very easily accommodated by Collingwood's third law of politics 152that is, as an example of its breakdown. The danger of the rate of recruitment dropping below the optimum velocity, then, may be that the third law of politics ceases to apply. The correspondence between rulers and ruled is eroded, as the ruling class, perhaps in much the same way that Mosca describes, begins habitually to resist infiltration from members of the non-social community, who are now understood to be essentially different from themselvesor, rather, misunderstood to be so.
Finally, in building into his third law of politics the acknowledgement that the law of adaptation works both ways, Collingwood is again in agreement with Mosca. The second of Mosca's widespread 'habits' of political analysis which, he thinks, obscure the truth about the ruling class, is the myth of de facto popular sovereignty. Mosca explains that 'pressures arising from the discontent of the masses who are governed … exert a certain amount of influence on the policies of the ruling, the political, class'. 153 But it does not follow, he adds, that the ostensibly 'ruling' minority is therefore actually ruled by the subject majorityspecifically, by the threat of its discontent and revolution. Instead, 'granting that the discontent of the masses might succeed in deposing a ruling class, inevitably … there would have to be another organized minority within the masses themselves to discharge the functions of a ruling class'. 154 Collingwood offers a concrete example of precisely this. The 'citoyens' upon whom the French Revolution aimed to bestow power, he observes, were not the whole population, not a 'rabble', but the bourgeoisie, and 'the bourgeoisie was already an organized body corporately possessed of economic power'; so 'The problem of the revolutionaries', he concludes, 'was to bestow political power where economic power already lay'. 155 As for Mosca, the habit of attending to the majority in a political movement blinds analysts and historians to the consistent applicability of the first law of politics: there is always, even in instances of popular revolt, an organized minority that is already in control.

Conclusion
This essay has sought to complement existing literature on Collingwood's 'dialectical' politics, and particularly a recent study of what I have called Collingwood's 'first' political dialectic between non-sociality and sociality, by providing an equivalent study of a 'second', and thereby to throw some light on what has been hitherto an overshadowed corner of Collingwood's political thought. This second dialectic is between a principle of democracy, and a principle of aristocracy, which operate together to regulate the percolation of liberty by way of recruitment into the 'ruling class', and which for Collingwood has traditionally been embedded in the Liberal and Conservative parties. Although Collingwood is drawing on a long-recognized dialectic between something like innovation and conservationan echo of which can be found in de Ruggiero's History of European Liberalismhis revision, accommodating a particular theory of what the ruling class is, how it rules, how it replenishes itself, and how it can weaken, shows him engaging with the chief themes of classical elite theory as it was developed in the late-nineteenth and early-twentieth centuries, in major texts that became available to English readers shortly before he began work on The New Leviathan. Conscious of Skinner's warning, I have not claimed that Collingwood must have been 'influenced' by these texts. But by surveying the relevant connections, textual and extra-textual, I have sought to offer a case for the verdict that in The New Leviathan we find Collingwood engaging with classical elite theory, and shaping in his own small way its reception in Anglophone political thought.