A Review of Matthew B. Crawford’s Why We Drive

Into this set of issues steps Matthew Crawford, the philosopher and author best known for his book Shop Class as Soulcraft. Motivated by the question “what is so special about driving?” Why We Drive attempts to “explore this one domain of skill, freedom, and individual responsibility…before it is too late, and make a case for defending it” (7). Crawford makes use of insights from Alexis de Tocqueville to examine two broad categories of phenomena that car culture produces. First, are the various car sub-​cultures from soap box derbies to hot rod clubs, all of which are, he argues, Tocquevillian associations. But second, and more important from a libertarian perspective, is the way in which driving and the human decisions and interactions it involves lead to the emergence of undesigned social norms and require us to exercise our facility for self-​governance in a number of important ways. Both of these Tocquevillian processes are, in Crawford’s view, important ways in which car culture pushes back at centralized power, both public and private.

I want to focus on the issues of knowledge and the undesigned emergence of rules and norms, as those are both particularly interesting to libertarians and are at the heart of my one major frustration with the book. Though thinkers like Jane Jacobs and James Scott make appearances, people like F. A. Hayek, Michael Polanyi, Vincent and Elinor Ostrom, and Don Lavoie do not, despite the fact that many of their ideas are very much part of Crawford’s argument. Even as he relies on the same sorts of arguments that libertarians make, he takes numerous shots at libertarianism, which he seems to see as a sort of unsophisticated rejection of rules of all sorts.

What he doesn’t see is that all of the things he’s talking about, and that he really likes about the freedom that comes with driving, are the very things that actual, real libertarians talk about all of the time. It’s an incredibly Hayekian book in places. If he thinks those things are normatively good, he should be more sympathetic to markets and libertarianism than he is. And that’s the frustrating part to a libertarian reader of this otherwise wonderful book. It’s just not clear from whom or where he got his understanding of what he thinks is libertarianism, but either he read the wrong people or misunderstood what he read. And it says something problematic about modern libertarianism that someone so smart and well-​read could make such good use of these ideas, but also not realize that they are at the core of libertarian political economy and social analysis.

There are a number of great discussions in the book (and his description of a soap-​box derby in an upper-​middle class section of Portland is a glorious piece of writing), but I want to focus on the ways in which his understanding of knowledge and the bottom-​up emergence of social norms play into his argument. In the Introduction, Crawford discusses the way in which psychologists believe that our motor skills and our ability to direct our own physical movements is tied to learning. Exploring our physical world is linked to memory. We are able to remember because we were able to move. Crawford points out that this is why we remember little to nothing of our infancy. We were unable to navigate in the world. We build up our memories through our development of the skills of locomotion. Driving is a great example of this process. It is a crucial way we explore the world, create memories, and thereby create the narratives about ourselves that constitute our identity. Crawford wonders what will happen if we turn even more of the driving over to the car and other automated processes.

This point about movement and learning is related to the idea of “embodied cognition.” In his critique of the ways in which cars have taken control away from the driver (e.g., cruise control or anti-​lock brakes), Crawford argues that this means drivers have lost the “feel” for their cars and that makes them less safe in situations where controlling the car is crucial. He points out that our relationship with our tools is such that when we become skilled tool users, the tool becomes an extension of our bodies. He uses the example of a hockey player and a hockey stick: the player’s attention “isn’t directed to his stick, it’s directed through his stick to the puck” (111). This is a point that Michael Polanyi raises in his book Personal Knowledge, where he distinguishes between “focal” and “subsidiary” awareness. Skilled use of a tool (including things like musical instruments) puts the tool in the background (“subsidiary”) and allows us to focus on the ultimate task. Crawford refers to this process as “embodied cognition” (without citing Polanyi), which is a way of saying that what we know is known through our physical movements even if we cannot explain exactly what we are doing.

Embodied cognition is one example of what Polanyi and others have discussed as “tacit knowledge.” There is much that we know, such as how to skillfully use a tool, that we cannot articulate through language or statistics. Crawford is interested in this knowledge because of the way it becomes part of a driver’s confidence and good judgment and cannot easily be captured by automation and artificial intelligence, or even articulated by the driver. He worries more broadly about a world in which other forms of skill-​based knowledge are undermined by automation, leaving us in a world that might be less safe or of lower quality because of the ways that automation cannot capture these forms of tacit knowledge. They might also undermine our ability to self-​govern as individuals, harming our capacity to do so at the social level.

His argument here is parallel to the criticisms of socialism, and defenses of markets, put forward by the Austrian economists. The argument for why governments will always fail in their attempts to plan economies is that planners cannot acquire the knowledge that is relevant for economic coordination because that knowledge is much more than statistics and preferences. In fact, a good deal of the knowledge that matters for deciding what to produce and how to produce it is “embodied” in the skills and experiences of both entrepreneurs and consumers. Tacit knowledge is ubiquitous in the market and those judgments and preferences can’t be captured in traditional notions of “data.” As Don Lavoie put it, “the data, in the relevant sense of the term, do not exist.” What does exist is the knowledge embodied in skill and experience that is made useful to others through the buying and selling of the marketplace, and reflected in the way that market prices serve as knowledge surrogates to guide our behavior.

Crawford further notes that the attempts to compensate for this loss of driver “feel” with various lights and chimes are ineffective because these “representations” end up weakening “the natural bonds between action and perception.” One could analogize this to the ways regulation and other interventions break the link between prices and knowledge in the marketplace. If he’s right that automation causes drivers to lose this important feel and thereby actually reduces automobile safety, how then can skillful drivers ensure that there is some sort of order on the roadways? What makes us safe if we put more human control back in driving?

His answer to that question is very much a Hayekian one. The short version is that faced with a minimal set of rules of the road, such has holding people accountable for the damage they do, drivers learn to engage in mutual adjustment with others on the road and that leads to the emergence of a variety of social norms that govern our behavior on the road. This requires, in his view, a return to a world with more individual responsibility and self-​governance. He offers a number of examples of how this might work.

The most powerful of those is his chapter on “Managing Traffic: Three Rival Versions of Rationality.” He starts with a contrast between what he calls “rule-​bound” traffic controls found in the US and the organically generated order of intersections in other parts of the world. By “rule-​bound” he means a system of traffic lights that is enforced by surveillance, both human and automated. Crawford points out that this takes away human judgment, as anyone who has sat at a red light late at night with no traffic coming the other way is well aware of. Intersections in Rome, for example, lack those externally imposed rules and are governed by what Crawford calls a “spectacle of improvisation.” He says “if there are rules being followed, they are not the simple kind that can be stated in a driving manual.” Once again, we see the previously noted tacit knowledge being made socially usable as Italian intersections are not significantly more dangerous than their US counterparts, thanks to the evolution of those unwritten rules. Crawford would have been on better ground here contrasting explicit, top-​down, imposed rules with tacit, bottom-​up, emergent rules, rather than suggesting that what guides the Italian intersection might not be rules at all.

He has a nice explanation for how those bottom-​up rules emerge. His preferred term for it, which he borrows from the philosopher and cognitive scientist Andy Clark, is “socially scaffolded mutual prediction.” As Crawford describes it, we are “continually updating our predictions of the world, including others’ behavior, and modify our own behavior so as to make it more easily predictable by others” (244). This is the exact same kind of process that Hayek describes at work in the human mind and that Adam Smith described with respect to human morality in The Theory of Moral Sentiments, and that libertarian social theorists have talked about in general terms as the “mutual adjustment” process that is central to emergent orders. Social coordination depends upon our ability to form correct expectations of others, and those expectations emerge unplanned out of human interaction. Hayek’s discussion of market coordination in his “Economics and Knowledge” from 1937 talked about this sort of process with respect to the spontaneous order of the market. We also see it in various discussions of the evolution of the common law. All of these emergent norms are no less rules than are the signals of a traffic light. The difference is that they have emerged out of the actual behavior and learning processes of drivers and have enough flexibility to allow human judgment to determine their applicability in any specific situation.

The way in which humans are able to discover rules for coordinating traffic at intersections without traffic lights is a smaller-​scale example of the ways in which free people can develop emergent rules to solve a whole variety of social problems. The reason that these bottom-​up emergent processes work so well is that they are able to take advantage of the knowledge, including the tacit knowledge, of those who are actually involved in the problem solving and not just the knowledge of a group of experts trying to impose a solution from the top down. Again, this is the argument against government planning of an economy, whether comprehensive or partial, that was been made by Ludwig von Mises and Hayek, and refined by more recent thinkers like Lavoie. Freely determined market prices enable the mutual adjustment required for economic coordination and the progress it produces. Emergent order through mutual adjustment is at the core of libertarian social analysis and also at the core of libertarianism’s normative arguments for why government intervention can’t achieve its stated goals.

This is what makes his occasional shots at libertarians so frustrating. Crawford seems to buy into the idea that libertarians reject all kinds of rules and constraints in favor of some metaphysical notion of freedom. Of course, that’s not the case at all. Thinkers like Hayek were adamant that there was no meaningful idea of freedom without rules, as rules make it possible for us to form expectations and thereby coordinate our behavior so that we can have mutually beneficial interactions. And libertarians have always been interested in where those rules come from. In other words, the libertarian argument for the importance of rules is exactly the case that Crawford is making for the bottom-​up, emergent rules of the road. Similarly, the libertarian emphasis on the responsibility and self-​governance that are required to live by the discipline of rules parallels Crawford’s similar emphasis on the character of drivers necessary for a world with less automation and fewer imposed rules. Finally, libertarians have also described the ways that the most effective rules are those that emerge from the bottom-​up, via actual human action, rather than being imposed from the top down. This is part of Hayek’s work the emergence of the common law as well as Elinor Ostrom’s work on community solutions to commons problems.

So the question remains, why does an author as smart as Crawford seem to miss all of this about libertarianism? I’m not sure what the answer is, but it’s a problem that should concern libertarians. Here is a well-​read philosophy PhD who probably knows something about libertarianism, but does not want to be perceived as a libertarian and doesn’t understand the ways in which his argument shares important analytical tools with contemporary libertarianism. This suggests that libertarians who use those tools need to think more about how their way of seeing the world might become more typical of the public perception of libertarianism, at least among intellectuals. It’s fine, of course, that a given person disagrees with libertarianism, but we have to take some responsibility for the fact that people like Crawford don’t understand our arguments and don’t want to associate with the name. There’s too much that could be gained by conversation with someone like Crawford to lose that opportunity for these sorts of reasons. That’s especially true given how much this particular book has to say that’s of interest to libertarians.

There’s also a lot more to this book than I could cover here and would be relevant for this particular readership. All I will says is that if you love cars and you love to drive, you will enjoy this book in ways beyond the social science Crawford deploys. His concerns about autonomous cars and the kind of power it would give organizations like Google are to be taken seriously, as is his more general concern with the loss of skills, judgment, and self-​governance that would result from losing our ability to drive. Why We Drive is well worth your time both for Crawford’s understanding of how social norms emerge and enable us to coordinate in the absence of top-​down processes and for the great joy he takes in the freedom of being behind the wheel on the open road.