Wishful Thinking and the Perversion of Political Knowledge
What "Deep Fakes" Mean for Our Political Order
In the Humean tradition that still largely infuses economics and decision theory, it is standard to account for choice behavior in terms of desires and beliefs.[1] A choice results from the conjunction of at least two intentional states, one that evaluates what is true or likely to be true about the world (a belief) and one that prescribes how the world should be (a desire). Decision theorists and economists typically assume on this basis that choices can be represented in terms of expected utility maximization where utility is a measure of desires and beliefs are represented by a probability measure. We don’t need however to commit to such a maximization framework and what follows is not based on this assumption. What is important however is that virtually all accounts of rational decision-making stipulate that desires and beliefs should be formed independently from each other.
To understand why this is so, we have to look at the respective nature of these two intentional states.[2] A belief is a cognitive state about how the world is. Philosophers say that it has a “mind-to-world direction of fit.” A belief aims at the truth and its true-value can be assessed by determining whether its “propositional content” (what the belief is about) fits with how the world actually is. For instance, if I have a belief with propositional content “The Denver Nuggets won the 2023 NBA Finals,” whether it is true depends on whether the Nuggets did indeed win the NBA Finals in 2023 (they did). A desire has the opposite direction of fit, i.e., a “world-to-mind direction fit.” A desire also has a propositional content, but this time the content expresses how one would like the world to be. A desire is a conative state that doesn’t aim at truth (you don’t assess a desire based on whether its propositional content is true). For instance, if I have a desire “The Boston Celtics win the 2024 NBA Finals,” there is no point assessing the truth-value of the propositional content – not because the relevant event hasn’t happened yet, but because it is not a proposition about how the world actually is.
The difference of nature between beliefs and desires doesn’t mean that there should not be any interaction between what one believes and what one desires. For instance, suppose that as a teenager, you strongly desire to become a professional basketball player and allocate most of your time and effort to fulfill your goal (i.e., you’re trying to change the world to make it fit with your mental state). As it happens, while you’ve some ability at this game, you’re small and not very athletic. Many people of the same age are way better than you. It is then reasonable to suggest that, provided you form correct beliefs about this state of affairs, these beliefs should induce you to revise your desire and the resulting allocation of your time and effort. This will help to reduce your frustration and increase the likelihood of other valuable achievements in your life.[3] However, the reverse relationship is more problematic. In general, we should not want our beliefs to be revised in light of our desires. Beliefs aim at truth, and how the world actually is does not depend on how we would want it to be. Allowing our beliefs to be influenced by our desires is to indulge in wishful thinking.
Wishful thinking leads to adverse consequences both at the individual and collective levels. In general, if beliefs are shaped by our desires, that will make them farther from the truth. It is hard to find cases where this can do any good.[4] As far as individual rationality is concerned, it’s clearly recommended to avoid succumbing to wishful thinking as it is highly likely to lead to bad decisions. This is also true for collective decision-making. As for individual decisions, we want collective choices to be based on accurate assessments of how the world actually is. Consider any election for instance. The social choice (who gets elected) results from individual decisions that are themselves produced by the conjunction of desires and beliefs. Beliefs about the current socioeconomic situation and the effects of various candidate policies are relevant to determining who is the best candidate. If desires influence beliefs, then the social choice is no longer based on any attempt to assess correctly what will happen if a given candidate is elected. This affects moreover everyone, not only those who indulge in wishful thinking. Persons trying to form correct beliefs about the state of affairs are also likely to be penalized, especially if they are in the minority.
A significant liberal tradition emphasizes the limits of our political knowledge. In political matters, when citizens vote and when officials make decisions on their behalf, we should acknowledge that there are many things that we don’t know about. This should make us reluctant to engage in vast programs that aim at remodeling society to pursue some ideals. Our beliefs should then reflect this caution. We should always leave open the possibility that things don’t turn out as we desire. It’s not hard to see how wishful thinking can make us oblivious to the limits of our political knowledge. Rather than admitting that we don’t know, we substitute our (alarming) uncertainty for confidence based on the conviction that we can shape the world as we want. There clearly is some psychological comfort in this.
Because it is blatantly irrational, wishful thinking is most of the time nonintentional. Nobody has reason to voluntarily indulge in wishful thinking, except to gain some psychological comfort. We can conjecture that we are even more prone to succumb to it the more the information we receive is ambiguous and difficult to interpret. This opens the door however for intentional political manipulation.
Decision theorists argue that rational agents should revise their beliefs, not in light of their desires, but in light of the information they receive. They generally suppose that the information received corresponds to chunks of knowledge, i.e., information that is assumed to be true.[5] Most of the time, however, the information we receive is ambiguous. If I hear a politician calling for more restrictions on immigration, I can interpret this in different ways, e.g., as cheap talk targeting the far-right electorate or a genuine commitment to implement new policies restricting entry into the country. This makes information more difficult to interpret (i.e., to turn into revised beliefs) and probably more prone to be distorted by desires. Consider the extreme case of deep fakes. As explained in a recent article in The Economist, AI will make it easier to fake videos, making it more difficult to determine if the event that is shown has really happened. Until now, images were (largely) conveying information easily turned into knowledge. This is no longer the case. The Economist tells us that, acknowledging that “images, audio or video of something do not prove that it happened,” people will turn on “reputation and provenance” to interpret and assess the information received. This is what should be rationally done indeed. But this is not what will happen very often. Rather, it will give even more space for indulging in wishful thinking. Many people will just interpret what they see as they desire it to be the case. Manipulators will use it to their advantage. More than limited, our political knowledge is perverted by manipulations taking advantage of cognitive biases. This perversion is nothing but part of the crisis of what Jonathan Rauch calls the “epistemic liberal order”[6] and another proof that what Harry Frankfurt[7] was calling “bullshit,” i.e., an (intentional or not) disregard for the value of truth, part and parcel of our political order.
[1] Economists and decision theorists talk of “preferences” rather than desires. This is not merely a difference of vocabulary as the concept of preference refers to properties that we would not ascribe intuitively to desires. For instance, preferences are comparative, not desires. Moreover, at least in situations of risk and uncertainty, preferences (over lotteries) are partially determined by beliefs. So, it is not correct to say that the standard choice-theoretic framework of economics and decision theory is Humean. What is nonetheless true however is that it is plausible to argue that preferences are partially determined by what philosophers call “desires.”
[2] I’m largely following John Searle. See for instance John R. Searle, Mind, Language and Society: Philosophy in the Real World (Basic Books, 1999).
[3] What I’m describing is close to what is sometimes called “adaptive preferences.” It consists in revising one’s preferences (desires) so that they are more likely to be satisfied. Amartya Sen, Jon Elster, and Cass Sunstein, among others, have shown that adaptive preferences can have socially detrimental effects. For instance, poor people get used to poverty, which contributes to keeping them in this state. That doesn’t make the mechanisms through which desires are revised irrational, however.
[4] There may be some exceptions. For instance, it may happen that holding false beliefs leads to actions turning into a self-fulfilling prophecy. Self-fulfilling prophecies can be disastrous (e.g., a bank run) but some may lead to good outcomes. These are very rare cases, however.
[5] This is implied by the use of Bayes’s rule. Bayes’s rule states that the probability that proposition A is true given proposition B should be conditioned by assuming that proposition B is true. The so-called Jeffrey’s rule extends Bayes’s rule to the cases where the information received is ambiguous.
[6] Jonathan Rauch, The Constitution of Knowledge: A Defense of Truth (Washington: Brookings Institution, 2021).
[7] Harry G. Frankfurt, “On Bullshit,” in On Bullshit (Princeton University Press, 2009), https://doi.org/10.1515/9781400826537.