Two Notions of Ideology

Ideology is more like a narrative that follows actions, rather than a set of principles that precede them

---

Men follow their sentiments and their self-interest, but it pleases them to imagine that they follow reason. And so they look for, and always find, some theory which, a posteriori, makes their actions appear to be logical. If that theory could be demolished scientifically, the only result would be that another theory would be substituted for the first one, and for the same purpose.

—Vilfredo Pareto

---

Take one: Ideology as a set of principles that govern action

Suppose I produce the following sequence of numbers: 2,4 6, 8. Someone asks me which rule governs my number sequence, and I reply “+2”.

But then I continue the sequence with 10, 13, 16, 19.

An observer concludes that either I have changed the rule that I’m following, or that I’m following a rule that is different from the one I articulated. My rule could have been, say, “+2 to 10, then +3 to 20, then +4 to 30, etc.” instead of “+2”. The two rules are coextensive up to a point (number 10)—that is, they produce the same output—but they are not the same rule and precisely because of that, they start diverging at a certain point.

Suppose that I’m convinced that I was, indeed, following the same rule through the whole sequence, and I’m trying to assure the observer of this. Does my own sincere conviction constitute sufficient evidence regarding the rule that was governing my number sequence?

Now let’s switch to another example. Suppose I consider myself an egalitarian. An observer notices that although I say that I care about equality, I seem to care about it only up to a point— if it works against group A. If, on the other hand, on a particular outcome members of group B show a disadvantage, I show no interest in closing that gap.

This leads the observer to object:

If you claim that you have an “aversion to inequality” but show interest in reducing inequality only when it puts group A at a disadvantage while ignoring those inequalities when it puts group B at a disadvantage, is “aversion to inequality” really an accurate description of your behavior?

If you object to “inequality” only when it puts yourself or your ingroups at a disadvantage, is it really true to say that you care about inequality? It seems more accurate to say that you care about yourself and your ingroup’s interests. Saying that you care about inequality does sound better, which perhaps is the whole point of saying it.

How do I know that your actions were governed by the principle “there should be no inequality between groups” rather than “we should work in group A’s interests”?

The observer writes the following three points in his notebook:

1. “Acting in accordance with a rule” is not the same as “following a rule”. In the first case, the rule is merely correlated with the behavior in question while in the second case, the rule is causal with regard to that behavior.

2. The illusion that a person’s behavior is the product of following a certain rule or principle may arise from the fact that we are operating with limited data about that person’s behavior. If we observe someone's behavior long enough, we often reach a point where his behavior is inconsistent with the rule/principle he says he followed. Therefore, sufficient time and data enable us to distinguish between the rule or principle that was causally relevant for producing certain actions and the rule or principle that was merely in accordance with those actions.

3. What one says—or even sincerely says— about oneself, regarding the rule or principles he is following, is not sufficient to conclude whether or not his actions were in fact the outcome of following that rule or principle.

Of course, the mentioned situation could equally work with someone who claims he cares about “liberty” or any other political value or principle.

---

A common idea in the social sciences is that people's political behavior is explained by something called ideology which is usually conceived as a kind of doctrine, a set of political or moral principles that guide one's actions.

However, there are theoretical and empirical reasons to doubt that people are ideological in that sense of the word.

The theoretical reason is that being ideological in the sense of consistently following principles doesn't sound very adaptive either from an evolutionary perspective or from a rational choice perspective. Rather, we should expect inconsistency and selectivity to be the norm (exceptions are cases where one can incur reputational cost by appearing inconsistent).

For example, those principles that enable you to obtain power or resources are not the same principles that enable you to maintain them. As an illustration, consider the case of a former World Chess Champion, Anatoly Karpov. At one point in his chess career, Karpov was opposed to the rule that the world champion keeps his title if he draws against the challenger. When asked why he didn’t oppose the rule while he was the world champion, he replied rather bluntly: "Although the problem came up again and again it never bothered me. I didn't see any point in bringing it up. As long as I was world champion the rules were in my favor. My opponents never brought the matter up, so I’d have been crazy to open my mouth, wouldn't I?"

Karpov, in short, noticed that with changed circumstances a different rule favors his interests. Given that it is a fact of life that conditions tend to change, adaptive behavior would be the one that is dynamically inconsistent (with, again, the need to be reliable as social partners constraining this inconsistency).

This line of reasoning is supported by empirical evidence: instead of using moral principles to guide their actions, people often strategically use those principles to justify actions that exist independently of these principles.

Most people are not “ideological” in the sense of internalizing a set of general—moral or political—principles that they follow in practice. This is backed by political science research that shows that the median voter is not very ideological:

most voters have an incomplete grasp of what “liberal” and “conservative” mean. The exception is the most highly informed, highly partisan voters. The more engaged one is in politics, the more ideological one is.

But wait a second! Aren’t there many historical cases that convincingly demonstrate the power of ideology? I would argue that in most cases once you scratch the surface, less principled causes start to emerge.

Consider the case of a 13-year-old boy Pavlik Morozov, praised by the Soviet propaganda as a martyr, because he reported his father to the authorities, and was subsequently murdered by his own family as an act of revenge. Wasn’t he a clear example of how far ideological blindness can take a person? Both communists and anticommunists seemed to agree that the answer was yes but with a different spin on it.

For the Soviet regime, Morozov was an “upstanding young Pioneer who, in a situation of conflicting family and state loyalties, nobly put the interests of the state first” while for the anticommunists he was “indicative of the moral decay of totalitarianism, whereby ideological control undermined and destroyed even family bonds”. However, a more thorough investigation uncovered different motivations behind Morozov’s actions:

his father, the chairman of the local rural soviet, had abandoned his wife and children and moved in with a younger woman from the same village. Pavlik either denounced his father out of personal resentment (as the eldest child, at 13 or 14, he had to take care of his family) or was prompted by his mother out of revenge, or by a cousin who wanted to become chairman of the rural soviet.

To say that a person is “ideological” most of the time is just another way of saying that a person is stubborn in maintaining a belief despite new information that contradicts it. But this “stubbornness” is not intellectual consistency: when one's position on the grand chessboard of life changes, being “inconsistent” is what pays off. The time of change is the real test of whether one's behavior was the product of following general principles or whether one was merely invoking general principles in order to justify one's behavior.

Principles are underdetermined by evidence and they underdetermine actions

Going back to the initial examples, we can say that up till number 10, whether I was following the rule “+2” or “+2 to 10, then +3 to 20, then +4 to 30, etc.” was—from the observer’s perspective—underdetermined by data, with further data (going beyond number 10) giving him a reason to distinguish between the two hypotheses. But the fact that beliefs about which rules were followed are underdetermined by data available to us, raises an important question of how rules or principles determine actions.

How do rules or principles “regulate”, “direct” or “determine” future behavior? One answer might be: look at past applications of the principle! The problem with this answer is that past behavior is consistent with numerous, mutually incompatible, principles. And if two principles are compatible with the same behavioral evidence, how do we know which of the two principles to continue following? What necessarily follows from one rule does not necessarily follow from another rule. How can past behavior determine future behavior if it is by itself insufficient to distinguish which rule was followed?

Furthermore, in practice, it is very easy to rationalize any apparent deviation from a principle. Any belief can be protected from refutation, and any action is defensible, so long as you make sufficient adjustments somewhere in your web of beliefs.

One can even justify deviating from one principle by appealing to another principle. What determines giving more weight to one principle instead of another? Are principles “all the way down”, or is there, perhaps, something else that regulates decisions such as when to stop following a principle or when to switch between different principles?

What constitutes the correct application of a rule is not clear even in a “simple” continuation of a number sequence, let alone in the application of political or moral principles in all the complexity of real life. This ambiguity about what constitutes the correct application of a principle provides the actor with sufficient degrees of freedom for his interests and preferences to kick in and influence action. And if you stop following a principle, or switch between them (like Karpov) when they start to conflict with your interests, is it not correct to say that you followed your interests all along?

Take two: ideologies are more like narratives, rather than principles

Suppose you want to get from point A to point Z, say from your house to the pharmacy on the town’s main square. On the way there, your eyes were caught by the shop window, after which you entered that shop (point B). Maybe you visited more shops in a similar manner (points C, D...). Maybe you met an acquaintance at point E, and after you reached the square and did the work you set out for, you didn't immediately return the same way but stopped at a nearby park that was recently landscaped (point F). In the end, you took a taxi and returned home.

The agent that was assigned to report on your activities might be tempted to treat all points as relevant data for interpreting the intention behind you leaving the house, creating a narrative that “explains” all the points. This would be wrong—many of the points you stopped on were a matter of chance. They were not part of your plan. One point was independent of the next one (although it may have influenced it in a spatial sense). You didn't leave your house because of point B or C—the plan was to go to the pharmacy and come back. Everything else was incidental.

The problem is that the agent cannot enter your mind, so information about your intentions is unavailable to him. The mistake he would be making in interpreting your trip consists of trying to run the same explanatory thread through all of the points, i.e. from assuming that all of the data points are connected, part of a coherent whole, and that what explains one point also explains another one.

Now let’s tweak the example. The agent that has to report on your activities is yourself. You are writing your own autobiography and the points from A to Z represent important choices you made during your lifetime. Now although at the time you made those choices you were familiar with the intentions that lay behind those choices, looking back on your life after so many years leads you to unconsciously distort your intentions and impose coherence on your life as a whole, much in the same way the agent did while reporting on your trip.

These distortions occur not only because memory is imperfect but also because you have to sell your autobiography. The need to sell your story creates an incentive to misremember things, to recall them in a more positive light, or to impose coherence on the various and often incoherent life choices you made during your life. Your story needs to be believable, relatable, and interesting.

To get back to the example of chess champions, one of the things that Karpov and Kasparov, another World Chess Champion, competed over—apart from chess—was who was the biggest anti-communist at a time when it was not popular, prior to the collapse of the Soviet Union. But as Dutch grandmaster and writer, Hans Ree, points out, they “both played more than their required role in the party and related organizations.”

So when Karpov wrote his autobiography, was he consciously and intentionally distorting his past to fit with the requirements of the new age ensuing after the fall of Communism? Perhaps, but this assumption is not needed. Belief-formation process is influenced by the consequences of belief acquisition—he simply had the incentive to misremember and reinterpret things from his past in a way that was beneficial to him. This tendency of yesterday’s apparatchiks to reinvent themselves into tomorrow’s dissidents will be all too familiar to people living in former socialist countries.

The six dots in each figure represent data points. The data points are compatible with many possible figures. Which is the correct way to connect the dots is underdetermined by the data (evidence). One’s past provides many such data points enabling a person to cherry-pick for different narratives thus creating different possible autobiographies.

What does all this mean?

People have the need to create coherence between different life choices and preferences. That requires a narrative. When choices and preferences are shared by a larger group of people, we call that narrative an “ideology”. When they are idiosyncratic, they are called “rationalizations”.

Hence, while rationalizations are an individual-level phenomenon, ideologies are best understood as a group-level phenomenon.

It is not simply the case that narrative “unites” people into the same group. For an ideology to arise, and in order to exert a unifying effect, what is needed is sufficient similarity between people that enables them to use and share the same narrative for their own private purposes. A crazy person cannot be united with other people into a coalition insofar as his rationalizations are too idiosyncratic. Likewise, a narrative that is unable to provide coherence and justifications to any individual, will not spread and become an ideology.

Imagine moving your pieces across a chessboard while simultaneously trying to provide your opponent with a coherent story (that explains your every move) and persuade him that you're not really trying to checkmate him. Good-enough approximation. Now do the same for group vs. group.

But the coherence is an illusion we create for PR purposes in order to hide the dynamic inconsistency of our choices that actually wins the game.

---

So to conclude, if we think of “ideology” as a set of principles that governs people’s actions, then we are forced to conclude that either most people are not “ideological”, or that we operate with a wrong notion of “ideology”.

Ideology, properly understood, is more like a narrative that follows actions, rather than a set of principles that precede them.

And lastly, ideologies are group phenomena; they are to a group what rationalizations are to an individual.

I should add that said considerations apply to most people. That doesn’t mean that there aren’t people who are more principled, care more about consistency, and are more able to notice inconsistency in their actions. There are such people, but they are a minority. These people are closer to being “ideological” according to the first conception of ideology examined here.

In one of my previous posts, I wrote that:

many intellectuals tend to overestimate the power of ideas, arguments, and reasoning to bring about social change—it is a self-serving bias of the profession, after all. They tend to see social conflicts as a kind of intellectual disagreement; the controversy surrounding the legality of abortion is called the “abortion debate”, for example.

And they have every incentive to believe that! If social conflicts really do lie on the level of ideas and arguments (instead of e.g. interests or personalities), ideas and arguments can resolve them, which means the demand for their services will increase— together with their social status.

I think there’s something similar to be said of the tendency to over-attribute ideological motivations to actors of social change. It is a bias of the intellectual class that writes books about social change by inadvertently projecting their own mentality onto participants.

Political scientist Stathis Kalyvas noticed something similar in his own area of research, which is civil war. He writes that:

the urban bias is present in explanations of motivations that are heavily biased toward ideology. There is a clear epistemic bias, at least in the sociological and historical traditions, in favor of the assumption that all (or most) participants in conflicts are motivated by ideological concerns. Because “urban” scholars tend to be primarily motivated by ideology themselves, they often assign unambiguous ideological motives to participants, even if this is not the case.