Thursday, December 05, 2019


In my corner of Twitter, I recently became obliquely aware of a kerfuffle about Michael Bloomberg saying that “Xi Jinping is not a dictator”. (Twitter is funny that way: sometimes all you get is a trace of some public person’s comment as reflected in memes or snark, and it takes some effort to find the original motivation for the jokes. It’s like a cloud chamber for political events). Not that this blog has any interest in current affairs, but I thought I actually had something to say about the question of the terms we use to talk about non-democracy, and I haven’t written here in a while.

At one level, this is an entirely unimportant and uninteresting controversy. The term “dictator”, like all political terms, can obviously be used as a weapon in the endless skirmishes for positional advantage in American politics and culture; and failure to label Xi Jinping as a “dictator” is easily equated with support for the crackdown in Hong Kong and repression in Tibet, Xinjiang, and elsewhere in China.

Similar controversies often arose when Chávez was alive (was he a dictator?), and still do occasionally with other people as well (is Orbán a dictator? Putin? Erdoğan?). They all have a similar structure: the person denying “dictatorship” status focuses on some attribute of the leader in question that seems incompatible with a standard “image” of dictatorship – e.g., the lack of violence against particular people, free election to the post of president or term limits, formal or informal accountability to selected constituencies, popularity – while those arguing for dictatorship tend to focus on the non-democratic aspects of the system over which the leader presides (lack of freedoms, use of violence, lack of formal accountability to other constituencies, life tenure). Though these disputes are ostensibly about correct descriptions – is this person a dictator or not? – their (usually obvious) point is coalitional – does the person making the descriptive claim belong in this or that group? (Unsurprisingly, Chinese leaders themselves prefer not to be called dictators, despite the fact that the Preamble of the Constitution of the People’s Republic still straightforwardly endorses the “people’s democratic dictatorship”).

This is unavoidable and to be expected; ostensibly descriptive terms have always been used in cultural and political struggles to increase or decrease the status of people and places, and their meaning is therefore “essentially contested.” Moreover, because they have long histories of use in these struggles, all such terms become sedimented with multiple valences that are impossible to “bracket off”. When we talk about “democracy” or “dictatorship” we invoke a plenitude of historically accumulated meanings and values that cannot simply be wished away by appeal to the dictionary. Yet alongside the political use of these words there is also an analytic use. (I’m a Weberian; I still see a distinction between Wissenschaft and politics, so sue me). From this point of view, the proper analytical use of a term like “dictatorship” should be informed by its historical roots, which in turn point the way to certain core ideas, but cannot be identical with its current usage in political struggles.

The history of the word “dictatorship” is probably familiar to readers of this blog; indeed, I sketched it here a long time ago, and then wrote about it at more length in chapter 2 of my book. The name of an ancient Roman republican magistracy, a person temporarily entrusted by the Senate to “save the republic” in emergency situations without being encumbered by procedural niceties, it already had a bad odour by the time of Sulla. These connotations were further cemented by Caesar’s assumption of the office in perpetuum, but none of this mattered for a very long time, as dictatorship was associated with republics, and republican forms of government basically ceased to exist for many centuries afterwards in the West. (The word for a bad ruler in non-republican contexts was tyrant or despot, not dictator; as Hobbes caustically put it, tyranny was “monarchy misliked”).

Accordingly, “dictatorship” came again into common usage in English and other European languages the mid-19th century, and more clearly at the beginning of the 20th century, when “democracy” (republics with universal suffrage) became more fully established; dictatorship is the pathology of republics, not of monarchy. And initially at least users of the word harked back to its “positive” or neutral sense; José Gaspar Rodríguez de Francia in Paraguay even declared himself “perpetual dictator” in 1816. Dictatorship is also associated with the invention of terms like “Caesarism” in the 19th century to describe regimes like those of Napoleon III or Bismarck, which combined appeals to “popular” legitimation, republican forms, and authoritarian features; these were the first people in a long time that could be described as “dictators” in European politics, and writers looked back to the analogy with Caesar to think about them.

When Marx and Engels appropriated and positively valorized the idea of the “dictatorship of the proletariat” (a term coined by Marx’s associate Joseph Weydemeyer), which they contrasted with the “dictatorship of the bourgeoisie”, they thus had already a rich cluster of meanings to draw on. There is the idea that dictatorship is rule without regard for law or traditional norms in special situations; that dictatorship is associated with republican forms, and can be extended, even “perpetual”, yet is not an inheritable office; and the idea that dictatorship (unlike tyranny) need not be, as a matter of definition, used for ill. To this I would say that they added the idea that dictatorship could be a system, not just an individual office, and the “dictator” could be a party or a class.

I’m guessing that this last conceptual addition made it possible for political scientists to use the term “dictatorship” to talk about all forms non-democracy – that is, to set up the modern opposition between democracy and dictatorship. In the age of democracy, even divine-right monarchy could be understood as a form of dictatorship, despite the conceptual opposition in classical political thought between these terms. But this came at a cost, insofar as the rich variety of non-democratic political forms became associated with the pre-existing connotations of the term “dictatorship.”

I think this is a bit unfortunate, not only because of the gap this opened between scholarly and popular uses of the term “dictator” (many people think of a “dictator” as the name for one particular kind of non-democratic ruler, not as a generic term for all such rulers) but also because it hides a number of important conceptual distinctions within non-democratic systems. (As you may have noted, I prefer the very general term “non-democracy” to talk about what many scholars call simply “dictatorships”).

In particular, we can distinguish, roughly, between how authoritarian control over society is (that is, how much of social life is patterned and directed by centralized uses of state power), and how much is the authority of a top leader divorced from control by norms, laws, or other elites more generally; and it is the latter dimension that many people tend to focus on when they talk about dictatorship. Implicit in Bloomberg’s comments, for example, is the idea of the dictator as the unaccountable or unbound leader; and having taught a course on dictatorships for many years I can testify that many of my students come in with the same idea. (I talk about this in my book at more length - this is just the capsule version).

These two “dimensions” of non-democracy often go together, but they can also be decoupled. Consider a place like Vietnam today (where I just visited, so it’s on my mind). A one-party state, most of my readers would be unlikely to call it a democracy, despite the fact that its constitution asserts the principle of democracy, Vietnamese media has been significantly liberalized since the early 2000s, and there’s even some competition in elections to the legislature and within the VCP – certainly more so than in China, and potentially with implications for economic and other outcomes. After all, this is still a country that tightly restricts competition for state power and where independent political organization is not allowed; it is authoritarian in the sense that a great number of activities – especially those connected with the competition for state power – are tightly restricted by the state.

Yet it would be difficult to name a single leader as the “dictator” of Vietnam; specialists in the politics of Vietnam talk instead about a “troika” (the General Secretary of the Communist Party, the President, and the Prime Minister) of mutually accountable and similarly powerful figures, all of whom are ultimately answerable to the Central Committee of the party, which has not been a rubber-stamp body since probably the death of Le Duan. While one could bite the bullet and say that Nguyen Phu Trong, the current General Secretary of the VCP, is Vietnam’s dictator, since Vietnam is a “dictatorship” (= “a non-democratic regime”) and he is by common consent the most powerful figure in the regime, this would sound a bit weird. Better, in my view, to say that Vietnam has an authoritarian regime where the leadership is more or less constrained by formal and informal rules, and so there’s no single “dictator”.

In this framework, what you get is a two-dimensional space where non-democratic regimes vary along the two dimensions of social control or “authoritarianism” and elite accountability or “personalism” (or “dictatorship”). In theory, you could have regimes where social control is fairly extreme (on the “totalitarian” side of the spectrum) but top leaders are tightly constrained by norms of collective leadership, for example, and vice-versa. We can even use data to get a rough picture of these regimes. Using a measure of “personalism” in non-democratic regimes from the latest book by Geddes, Wright, and Frantz (How Dictatorships Work, about which more in a later post perhaps) and a measure of civil society control from the Varieties of Democracy dataset we can get this picture (this is interactive, hover over a dot to see the name of the leader):

Now, this is a purely illustrative picture. I don’t have space to delve into the construction or quality of this data, and I certainly don’t believe all of it. Moreover, the “personalism” data only goes to 2010, so it does not allow us to see what Xi Jinping would look like here. The numbers do not have a real scale either - they represent an arbitrary index. But it does show that there is a lot of variation along these dimensions. Consider China:

According to this measure, the peak of “dictatorship” in China (Mao in the 1960s) is during the Cultural Revolution, which is also the peak of social control. But Mao did not always have this degree of power; earlier he was more of a first among equals, and as the cultural revolution dragged on he also lost prestige, though social control remained high throughout his tenure. In the 1990s, Deng and Hu represent lower levels of personal power – collective leadership eventually became more established at this time – as well as an opening of social control.

The graph's four quadrants represent four distinctive patterns. The “dictators” – leaders with high personal power, unbound by norms, institutions, or the elite – are at the top of the graph. These are the people we might expect: Trujillo, Mobutu, Qaddafi, Franco, the Duvaliers, the Kims, etc. Their regimes take their name from them; they are “dictatorships” in the sense that they can be identified with their dictator. But in some cases (e.g., Compaoré in Burkina Fasso, Kérékou in Benin, Eyadéma in Togo) they did not manage to get the state to exercise much social control (at least by this measure! results may vary on other measures!). Or consider Venezuela, which Geddes, Wright, and Frantz code as a non-democracy from 2006:

Here Chávez appears as a ruler with high personal power, more so than the military dictator Pérez Jiménez, who was constrained by a military Junta. But control over civil society was not extraordinarily high during his tenure, unlike during the military regime (note Pérez Jiménez was overthrown in 1958 - hence that “low social control” point in 1958, since V-Dem data is coded at the end of the year).

At the bottom of the graph, by contrast, we have non-democratic regimes that are not defined by their occupants, because their rulers are much more tightly constrained by norms, institutions, or other elites. In the lower-right corner, in particular, we have regimes that are sometimes confused with democracies. These are places like Taiwan in the early 1990s, or Mexico in the 1980s and 90s (where the President was term limited and constrained by the PRI, other parties existed and were able to organize and compete in elections). Some of these regimes engaged in targeted repression, but it would be a bit weird to say that Zedillo, for example, was the “dictator” of Mexico – though Mexico at the time was authoritarian (even if only in the mild “competitive authoritarian” sense coined by Levitsky and Way).

There is an interesting question here about how leaders come to have personal power – i.e., how they become full-fledged “dictators” in the popular sense of the term. This is a question that Geddes, Wright, and Frantz explore in their book in some detail, and about which I hope to say more later. For now, however, note only that a leader who is unbound by legal or traditional constraints either has what Weber called “charismatic” authority (e.g., Mao in the mid-1960s), or the kind of control over resources that allows them to bypass such legal and normative constraints (e.g., Mobutu in his later years). But that's a story for another post.