The Custom

The path outlives the errand that first wore it into the field. The earlier Collected plate on the origin of norms began with feet taking the way that other feet had taken, each passage making the next easier. But the destination may move while the path remains. A market keeps the day once fixed by a ferry that no longer runs; an office preserves a sequence devised for papers it no longer handles; a ceremony survives the danger it was meant to avert. These are imagined instances of the question, not histories of particular customs. Their puzzle is that a practice need not retain its original purpose to retain a present cause. The reason people first adopted it and the reason their successors continue it may be entirely different. What began as a useful response can acquire a second life in trained hands, coordinated expectations, and the conviction that this is simply how things ought to be done.

Habit preserves a sequence by relieving us of the need to choose it afresh. William James described the saving with a musician's hands and the betrayal with a familiar gesture made in the wrong place: practice can set an action going before attention has reconsidered its occasion. The economy is real. A person who had to deliberate over every movement would have little attention left for anything else. Yet what saves attention can also spare an obsolete routine its examination. The clerk may continue copying a figure because copying it follows entering it, although nobody now uses the copy. No watching colleague is needed for this small survival; the sequence can persist in one person's conduct. That makes habit different from a convention sustained by other people's choices. The two can reinforce one another, but they offer different answers to the same question. One practice is difficult to stop because the hand begins it; another because stopping would leave the hand alone.¹

Mutual expectation gives the custom a present advantage that its forgotten origin cannot explain away. Hume's account of convention turns on conduct adjusted to the expected conduct of others; repeated experience strengthens confidence that the adjustment will continue. Imagine, then, two traders choosing independently between an old market day and a new one. Each receives four if both attend on the old day, six if both attend on the new, and nothing if they miss one another. The ferry that once favored the old day has vanished; the newer arrangement would now serve both better. Nevertheless, if each assigns only a three-in-ten chance to the other's changing, changing yields an expected 1.8 while staying yields 2.8. Both stay. In this deliberately invented example, choosing the new day becomes preferable only when confidence in the other's change exceeds two-fifths. The custom survives without anyone mistaking four for six. Each can see the better destination and still lack a profitable route to it alone.²

A history of use can also change the things available to be chosen. Brian Arthur's model of competing technologies gives this claim a precise setting: adoption improves a technology, improvement attracts further adoption, and an early advantage can become difficult to dislodge. The later chooser encounters the consequences of earlier choices in the quality of the alternatives themselves. A custom can acquire analogous supports in skills, equipment, schedules, and the arrangements that have grown around it. The analogy does not make every custom a technology, or every old technology inferior. Arthur's conclusion depends on the returns process; increasing benefits that eventually level off do not by themselves guarantee that one technology takes the whole market. What the argument defeats is the easy inference from survival to superiority. An arrangement can prevail because it has been developed, connected, and made convenient by its own prevalence. The present may keep paying a return on a lead that the past bestowed for quite another reason.³

The custom becomes harder to examine when familiarity acquires the voice of conscience. The earlier Margin plate called this the given: a rule so thoroughly absorbed that it is experienced as the shape of the world. Montaigne had already noticed how practices received in childhood can appear universal, and departures from them appear unreasonable merely because they are unfamiliar. This adds something the market-day game leaves out. The trader may now regard the proposed change as disloyal, indecent, or unworthy of his station, even if everyone could coordinate it successfully. Custom has entered the standard by which its replacement would be judged. Yet the discovery of this history does not dispose of that judgment. A duty learned from one's neighbors may be a sound duty. Montaigne himself coupled his scrutiny of custom with a formidable distrust of innovation and the disorder it could release. To denaturalize is not to debunk. It is to recover the distinction between a practice's authority over us and the reasons that would justify giving it that authority.⁴

Changing a custom therefore requires an account of the crossing as well as the destination. In the market-day example, a credible agreement to move together would remove the coordination loss; another speech about the superiority of six over four would not. Even agreement need not make a transition immediately profitable. Suppose each trader must incur a one-time moving cost of three. The first new market brings a net three, below the old market's four; over two markets, with no discounting or other costs, the new arrangement brings nine against eight. A short horizon can reverse the judgment without reversing the ranking of the settled arrangements. Actual transitions may also have losers: the owner of the old market's best site need not welcome a change that helps everyone else. These distinctions matter because they identify what the reform must accomplish—credible coordination, financing a transition, or confronting a distribution of gains. Longevity alone cannot tell us which obstacle stands in the way.⁵

The custom endures because the past has become part of the present's machinery, not because time has issued a verdict on its merits. Sometimes that machinery preserves knowledge its users could not readily reconstruct; sometimes it preserves an inconvenience nobody can escape separately; sometimes it shelters an advantage its beneficiaries prefer to call tradition. The inquiry must distinguish these possibilities without pretending that everything old has been tested against every alternative, or that everything newly designed has been tested at all. Ask what the practice now accomplishes, what keeps it in place, and what those who bear it would face in changing it. The worn path remains useful as evidence of passage. It is no warrant that the destination still deserves the journey. And even a community that chooses a better road must find a way to rely on its fellow travelers. After the custom comes the faith.