First off, I wanted to say that I very much appreciate everyone who has posted a review of our book, positive or negative, on Amazon or Goodreads (or both!). With that in mind, we had our longest, and easily most thoughtful negative review on Amazon yesterday, which echoes some of the most frequent criticisms of the book. Because we do get these criticisms frequently, and because this particular reviewer articulated them in a very coherent manner, I thought I would respond to a couple of his points over the next couple of days.
The reviewer begins by noting that our definition of democracy is pretty underwhelming, and then spends a lot of time pondering exactly what a democracy is, and is not–and noting that he doesn’t fully believe if Japan, where he lives, is democratic. And, to be fair, he is absolutely correct, in that our definition of democracy in the book is not very precise. However, as an author, sometimes you can’t do things as precisely as you would like; and this is one of those cases where I would have loved to have spent more time and energy delving deeper into the definition of democracy in the book. Unfortunately, the definition of democracy is extremely complicated, and that discussion wasn’t really relevant to the larger purpose of the book. I know you may find that sentence surprising, given that the book is, in fact, about democracy, so allow me to explain. But first, allow me digress momentarily into a discussion of pornography.
Some concepts are extremely difficult to define precisely–although just because we lack a good definition, doesn’t mean that we can’t hold a conversation about the topic. The classic example of this in law and politics is pornography. We all know what pornography is. We can hold political debates about pornography. Most Americans have admitted in polls to having purchased pornography or seen pornography at some point–clearly, if they’ve bought it, they must know what it is. And yet when it comes down to precisely defining pornography, as a distinct category from more legitimate artistic forms of expression, or just from a three year old running butt-naked down the hallway because he doesn’t know better and mom taking a picture because she thinks it’s funny… well, in practice it can be extremely difficult to come up with a precise definition that incorporates everything we want it to (e.g. naked people doing erotic things), and excludes everything that we don’t want it to (e.g. legitimate art or nudity that’s amusing but not erotic). Hence the classic phrase, from an actual Supreme Court ruling on the subject, “I know it when I see it.”
This is a common problem that applies to many different concepts. (Just for kicks some time, try to come up with a definition of “chair,” that successfully differentiates chairs from tables, benches, or ottomans, and that doesn’t rely on user or creator intent.) To get around this problem of discussing difficult to define ideas, social scientists often will lay out two separate definitions for the same concept: a theoretical definition and an operational definition.
First, they will lay out a theoretical definition, which usually hews closely to a standard dictionary definition or comes out of a philosophical debate on a topic. The purpose of this definition is to allow the author to build a theoretical framework, and to participate in a philosophical debate on a topic, without having to go into the nitty-gritty grey-area cases of what does and does not fit within the definition. We all know that pornography exists, we all know that it encompasses roughly a set of pictures of naked people with some sort of erotic intent, and that’s enough to have a conversation about the topic. We can agree on the general scope of what pornography is, while keeping our conversation about whether or not Michelangelo’s David is pornographic separate from the broader context of whether our theory of pornography is appropriate.
But a theoretical definition like that is just not good enough to collect data or run a statistical analysis. For that you need an Operational Definition, the purpose of which is to give clear labels to everything within the grey areas. The operational definition is what gets picked apart, where people get to argue whether or not a particular case ought to fall in one category or the other. Of course, a strong theory will be robust to slight changes in the operational definition–in other words, if the theory is strong, reclassifying a small number of cases should not have any impact on the results of your study. If the proof of your theory relies on having a few grey area examples all classified in very particular ways, then that is a strong indication of a weak theory.
Alright, so enough with the lecture on social science methodology. Why is this relevant to our book and to this gentleman’s critique of our book?
Democracy is like pornography, in that it is very difficult to come up with an adequate operational definition. In our book, we define democracy as a system of government that relies on regularly held free, fair and meaningful elections with near universal suffrage. This is a pretty standard definition of a modern democracy, as it is generally practiced and recognized around the world. But of course, what exactly is a “free election”, a “fair election” or a “meaningful election”? How regularly must they be held? How near-universal must suffrage be? These are all huge problems for any definition of democracy, but explaining them adequately would have required easily a chapter, and possibly a book, all by itself. So instead we give cursory definitions of each term, we basically just acknowledge that there are massive grey areas that we have no intention of addressing, and move on. But that’s okay, because we are largely giving a theoretical definition. We don’t put it in those terms–after all, we were writing a book aimed at a general audience, not a social scientific one. Our arguments didn’t require us to be worried about the grey areas. Once again, we were building a theoretical definition so that we could hold a conversation about the topic, not constructing an operational definition so that we could test a theory against specific data.
And trust me, operational definitions of democracy are never pretty. If you don’t classify a country as democratic until both sexes have the right to vote, then most countries don’t become democratic until well into the 20th century, including the United States. The Jim Crow laws functionally disenfranchised something like 15% of the American population… were we a democracy before 1965? In Japan, the same party one election every time for more than 50 years. Most observers believed that the elections were free and fair, but… I mean, come on, that does look pretty fishy, and how would we really KNOW that there wasn’t some kind of behind the scenes tampering? What about Mexico, where some amount of corruption and ballot-box stuffing has been proven, but how much corruption needs to be proven to make a country non-democratic? Or Israel, which is perfectly democratic if you are a Jew in Tel Aviv, but not so much if you are an Arab living in the Israeli-occupied West Bank (until the recent establishment of a Palestinian Authority)? And I haven’t even gotten to the really hard cases yet.
Finally, there is one other issue I want to deal with. One confusion about common operational definitions of democracy is that the data sets mix-up the concepts of “democracy” and “human rights”. Philosophical discussions about democracy revolve around who makes the laws and who selects those people–and yet the best and most reputable data base out there that social scientists use to determine how democratic a country is (Freedom House) includes within it measures of respect for religious freedom, free speech, free press, etc. I would argue that those things may correlate with democracy–in fact, they do correlate strongly with democracy–but they are not inherently part of democracy. Democracy is about the rules for choosing who governs. Freedom of speech comes from the decisions made by those who govern.
So yes, our definition of democracy in the book isn’t especially precise, but that’s because the alternative was to write an entire chapter to come up with a precise and controversial definition of democracy that would have been completely unnecessary to make the point that we wanted to make.