Why Experts Screw Up
Experts have a lot of knowledge most people don’t have. This includes both factual knowledge and knowledge of skills for learning new truths in their own field. When it comes to these matters, experts should usually be trusted. The rest of us don’t know these facts, and we can’t figure things out as well as they can in these areas. Despite this, though, experts really suck when they try to make predictions, and when they discuss issues they disagree about. Experts are often worse than random in these areas. Monkeys throwing darts are often better than experts at knowing what will happen. On average, we’re better off ignoring them than agreeing with them. How can it both be the case that experts should be trusted as a rule on certain questions, and ignored as a rule on other ones?
To help understand the answer, it’s good to focus on a fascinating example raised by James Surowiecki in The Wisdom of Crowds. He discusses a case where people wanted to know the location of a sunken ship. There were numerous factors that would have affected the location of the ship, such as its size and speed, the conditions of the water, the weather conditions, and so on. Those who wanted to know where the ship was asked experts in the relevant areas to each independently determine where the ship was. The experts were all wrong. However, the point of best fit between their answers was almost exactly right. The experts were wrong, but somehow, their errors, when combined, got the right results. How could this happen?
The reason the point of best fit was basically correct is that the experts could answer the following question very well: if your knowledge was all that mattered, then where would the ship be? They had the required knowledge and skills to know exactly how the information and its implications worked. If all that mattered was what they were good at, they would have been right. But it wasn’t all that mattered. Other factors that they didn’t know about also mattered, and they didn’t have the skills required to know how those factors changed that final answer. They weren’t incompetent, they were myopic. It required a broader perspective that took into account all the relevant factors to know the answer.
Numerous examples described by Surowiecki, Kahneman, and Tetlock, among others, fit this pattern. Experts are both wrong and very confident about a wide range of issues. They are justifiably confident in their own skills and how they relate to what they know about. They are wildly ignorant of the importance of additional factors outside their area of expertise. Because they know what would be true if all that mattered was what they knew about, they are sure they are right. Because other things matter, and they don’t know how to account for them, they are also usually wrong. Complexity breeds confidence and error. It’s very important, if we want to know the answer to complex questions, to avoid myopia as much as possible.
Philosophy and Expertise
Philosophy tries to answer the broadest possible questions. It is therefore more likely than other disciplines to be subject to problems arising from myopia. Even among experts, approaches are varied. Philosophy has included thousands of years of discussion all around the world. It has been done within a variety of methods, and there has been no accepted method that has limited the range of answers philosophers could endorse. Experts at doing philosophy following one approach might never seriously try to think within a different framework, and while there is some standard of what counts as “good” work from within a framework, there is very little consensus on what counts as “good” overall. Even the smartest people in history have, unsurprisingly, failed to resolve these big questions amidst this arena of inconsistent methods and vast complexity. If their method is no good, they won’t find any right answers, except by accident. If it’s good for some things, but not others, they’ll only find the answers limited to what their approach is good for, and this won’t include the answers to most philosophical problems due to their complexity. To actually make progress, philosophy needs an approach that allows us to find answers, which means it needs a method broad enough to bring into consideration a wide variety of types of knowledge without thereby sacrificing high standards that can keep out error both at the outset and along the way. The best method I know of to work with is common sense philosophy.
How does Common Sense Philosophy Work?
One of the immediate questions for this approach is how to understand common sense. If it just means common judgement, then it wouldn’t be a very good approach. Lots of people get lots of things wrong. Numerous common beliefs in one culture don’t exist in another one, and progress has demonstrated that a large number of false views were widely held in history. Fortunately, common sense philosophy doesn’t rest on these sorts of judgments. Instead, common sense consists of presuppositions of ordinary thought, language, and behavior. Whenever we act, we do so within a broad context of underlying assumptions. We assume we exist in a physical environment, that it has objects in it, that those objects have features, that others can detect the same objects, that we can discuss them, that they can understand things, that they know they shouldn’t kill us for no reason so it’s safe to approach them, that they have the ability to make choices and decisions, that they will usually act in ways that reflect their beliefs and desires, and so on.
These presuppositions are built into language, thought, and behavior. Language divides the world up into basic categories of noun phrases, adjectives, verbs, adverbs, tenses, and locations. We experience the world as having the things these correspond to, and we believe that these types of things exist. In addition, when we think, we do so taking these kinds of things for granted. We don’t have to remind ourselves that we have physical bodies that can move around our environment when we decide to go to the store. We instead think within a framework that includes these facts, and deliberate with them as a given. We do the same thing when we act and when we see the behavior of others. When we’re at the store, we assume the other people are there because they want to get food and they believe they can do so here. We take it for granted that there will be a rational connection between behavior, desires, and beliefs. We also take it for granted that they can make choices about their own behavior, and that they have the ability to control what they’re doing, just like we do.
There are frameworks of thought underlying every experience we have and every judgment we make. The most fundamental propositions in those frameworks are the basic principles of common sense.All other judgments are made within frameworks that take such judgments as presuppositions. Even when philosophers reject these views explicitly, they still rely on them. Skeptics about the external world may claim not to know there is anything outside their minds. They still act within the assumptions that there are objects they can use to write with and on, that their efforts will produce effects outside of them in the form of documents, that they can communicate their ideas to others, and so on. Their sensible behavior could not exist without these sorts of beliefs. Because these beliefs shape the other ones, they place limits on which views can be sensibly endorsed. Removing any substantial portion of these judgments cuts us off from accepting the opinions that rest on them, which include theoretical positions that reject them. In this way, common sense constrains which views we can rationally consider. This doesn’t guarantee that the truth is within our grasp. It does mean, though, that if it isn’t, we could never have good reason to believe what is actually true. Our knowledge is only possible if the truth lies within the scope of the judgments compatible with common sense. As a result, common sense can serve both as the starting point and as a constant constraint on which views we can rationally accept.
Common sense philosophy then proceeds in much the same way that an ideal science would proceed. It takes the data of common sense as the starting point. Because common sense typically works in the background, this requires us to unearth and articulate the commitments of common sense. In addition, though, exercising common sense produces immediate judgments on a number of simple issues. Most obviously, it produces beliefs in the things we see with our eyes or perceive with our other senses. In addition, it produces immediate judgments of thought, such as that identity is transitive, modus ponens is valid, and it’s wrong to toss babies into woodchippers. We don’t need to think about these things, we can just see that they are true. Introspection similarly delivers immediate judgments of our own minds. These immediate deliverances of exercising our abilities to learn also deliver common sense judgments. Jointly, they provide a rich set of initial beliefs to start from.
Once the data set is in place, we can start to try to build on it. We can identify similarities across cases and seek to develop general theories that might explain them. We can thereby build up our knowledge very slowly from where we started. This approach solves the problem of speculative belief that plagues rationalist approaches. Nothing that isn’t built in at the foundations gets added unless it is directly connected to these foundations. General theories about broad issues are the far away results of study, not the points from which to begin. Such theories are also incredibly unlikely to be true because every step towards them would have to have been accurately guessed by sheer luck. Big pictures, general worldviews, castles in the sky are all so implausible at the outset that beginning with them would be a huge mistake. By grounding beliefs in the foundations of common sense, we are discouraged from building and embracing the cool, beautiful, and almost certainly false pictures that draw us down the wrong roads.
Even when making relatively small theoretical judgments, though, they need to be tested before we consider accepting them. In science, which is often similar to natural philosophy done in a common sense way, testing involves the development of predictions, and of experiments that see whether or not those predictions hold. Other areas of thought don’t allow us to use empirical experimentation. They do, however, allow us to test theories against common sense by using thought experiments to test them. Thought experiments deliver immediate judgments in relatively simple scenarios where possibly confounding elements have been omitted. These sorts of judgments are more directly grounded in common sense than theories are, and should be trusted more than the theoretical conclusions. Theories can and should be regularly rejected when they conflict with these counter-examples. Common sense philosophy should look like science, with other areas of common sense as the data set, and with thought experiments in place of empirical ones as tests of any additions to it.
How does this relate to Myopia?
While it might not be initially obvious how, common sense philosophy also solves the problem of myopia. The reason for this is that common sense covers the widest possible range of beliefs. Since other beliefs grow out of common sense, every type of belief we can express can also be subjected to its judgment. Since its legitimacy holds over theoretical conclusions in general, implications for one view outside of its immediate area as also legitimate reasons to reject those theories. Even if a theory does a good job solving a particular puzzle, if that answer conflicts with common sense elsewhere, then we should reject that theory. The approach holds every addition to common sense subject to the entire basis of judgment in every case. It doesn’t allow limited areas of inquiry, or limited application of knowledge and expertise to a subset of the data to be authoritative. Therefore, even experts in an area need to be mindful of possible incompatibilities or relevant additional considerations from elsewhere. Rejection of a theory is always easier to justify than accepting it on this approach, including by considerations the creator didn’t have in mind at the time. This makes it less likely that myopic conclusions will stick around in the face of the range of possible objections.
At this point, I can more clearly explain some of my other posts as examples of the approach in application. Why do I hate physicalism? Because it tries to limit all knowledge to things that make the modern scientific worldview look good. It is the conclusion of a myopic approach that only cares about certain types of knowledge, and dismisses the obvious deliverances of things like introspection and basic modal judgments about psychological states. Why does Hume suck? Because he revels in views that elevate philosophers and degenerate the knowledge everyone actually has. He raises reason up as a judge of all else, thinks smart people must have used it correctly, and trusts the skepticism philosophers have been driven to over the fact that people can see tables. Why does Marx suck? Because he develops views about history and human nature that could only be justified if he had detailed knowledge of an array of subjects that he obviously lacks an understanding of. He trusts his speculations about economics, history, psychology, sociology, and other areas of study despite not even having sufficiently studied the limited knowledge of these subjects available at his time. He simplifies everything down to a dataset he likes, pretends it can somehow justify answers to questions that dwarf what he knows about, and draws sweeping conclusions that are subject to a myriad of counter-examples.
Above all, my animosity toward these views grows from the fact that advocates of them almost all suffer from a complete lack of humility about them, and a shocking certainty that they have the answer to questions no human is in a good position to have anything more than a hopeful inkling about. We don’t know the limits to which kinds of things can exist. We certainly don’t know it with such certainty that we can limit the number of things that can exist to an amount lower than the number common sense tells us about.
Reasoning about anything complex is incredibly hard. Very smart people throughout all of history have usually gotten it wrong when they tried. Trusting what they’ve been saying recently in your own area more than you trust your own eyes is ridiculous. Reason is great, but it’s nowhere close to good enough to justify anything Hume says. Nothing has such magical prowess.
People who have deeply studied one subject have usually only studied a small portion of it. They often fail to see problems only visible to those in a different part of their field. In the absence of lengthy study to replicate their skills, we can’t hope to know answers to the kinds of questions still left to be figured out by these experts, let alone be certain of answers to such questions without having obtained such expertise. Drawing from multiple areas of relative incompetence to create big answers about the fundamental nature of history and human nature, and then using these as the basis of your political philosophy is the absolute height of hubris and absurdity. Not backing down after a view whose nature guaranteed that it had no hope of being true in the first place leads to tens of millions of deaths is sheer insanity.
We need to stop pretending we know things that we don’t know. We certainly need to stop acting on the pretense that we do. Knowledge that goes beyond the obvious is hard to get. We need an approach that gives us the most tools to use that we can justly hold on to, and the strictest limits on how we use them as possible in order to prevent us from going astray. We also need to stop pretending we have knowledge when we don’t. We need to understand our limits, refuse to trust speculation, and recognize the likelihood of myopia and error even in areas where we know better than most. Common sense philosophy allows for this, which is one of the best reasons to accept it.
Leave a Reply