Книга: Code 2.0
In Chapter 7, I sketched briefly an argument for how the four modalities I described constrain differently. In this appendix, I want to extend that argument. My hope is to provide a richer sense of how these modalities — law, the market, norms, and architecture — interact as they regulate. Such an understanding is useful, but not necessary, to the argument of this book. I’ve therefore put it here, for those with an interest, and too much time. Elsewhere I have called this approach “the New Chicago School.”
Law is a command backed up by the threat of a sanction. It commands you not to commit murder and threatens a severe penalty if you do so anyway. Or it commands you not to trade in cocaine and threatens barbaric punishments if you do. In both cases, the picture of law is fairly simple and straightforward: Don’t do this, or else.
Obviously law is much more than a set of commands and threats. Law not only commands certain behaviors but expresses the values of a community (when, for example, it sets aside a day to celebrate the birth of Martin Luther King, Jr.); constitutes or regulates structures of government (when the Constitution, for example, establishes in Article I a House of Representatives distinct from a Senate); and establishes rights that individuals can invoke against their own government (the Bill of Rights). All these are examples of law, and by focusing on just one kind of law, I do not mean to diminish the significance of these other kinds. Still, this particular aspect of law provides a well-defined constraint on individuals within the jurisdiction of the law giver, or sovereign. That constraint — objectively — is the threat of punishment.
Social norms constrain differently. By social norms, I mean those normative constraints imposed not through the organized or centralized actions of a state, but through the many slight and sometimes forceful sanctions that members of a community impose on each other. I am not talking about patterns of behavior: It may be that most people drive to work between 7:00 and 8:00 a.m., but this is not a norm in the sense I mean. A norm governs socially salient behavior, deviation from which makes you socially abnormal.
Life is filled with, constituted by, and defined in relation to such norms — some of which are valuable, and many of which are not. It is a norm (and a good one) to thank others for service. Not thanking someone makes you “rude”, and being rude opens you up to a range of social sanctions, from ostracism to criticism. It is a norm to speak cautiously to a seatmate on an airplane, or to stay to the right while driving slowly. Norms discourage men from wearing dresses to work and encourage all of us to bathe regularly. Ordinary life is filled with such commands about how we are to behave. For the ordinarily socialized person, these commands constitute a significant portion of the constraints on individual behavior.
Norms, like law, then, are effective rules. What makes norms different is the mechanism and source of their sanction: They are imposed by a community, not a state. But they are similar to law in that, at least objectively, their constraint is imposed after a violation has occurred.
The constraints of the market are different again. The market constrains through price. A price signals the point at which a resource can be transferred from one person to another. If you want a Starbucks coffee, you must give the clerk four dollars. The constraint (the four dollars) is simultaneous with the benefit you want (the coffee). You may, of course, bargain to pay for the benefit later ( “I’d gladly pay you Tuesday for a hamburger today”), but the obligation is incurred at the time you receive the benefit. To the extent that you stay in the market, this simultaneity is preserved. The market constraint, unlike law and norms, does not kick in after you have taken the benefit you seek; it kicks in at the same time.
This is not to say that market transactions cannot be translated into law or norm transactions. Indeed, market transactions do not exist except within a context of law and norms. You must pay for your coffee; if you do not, the law of theft applies. Nothing in the market requires that you tip the waiter, but if you do not, norms kick in to regulate your stinginess. The constraints of the market exist because of an elaborate background of law and norms defining what is buyable and sellable, as well as rules of property and contract for how things may be bought and sold. But given these laws and norms, the market still constrains in a distinct way.
The constraint of our final modality is neither so contingent nor, in its full range, so dependent. This is the constraint of architecture — the way the world is, or the ways specific aspects of it are. Architects call it the built environment; those who don’t give out names just recognize it as the world around them.
Plainly some of the constraints of architecture are constraints we have made (hence the sense of “architecture”) and some are not. A door closes off a room. When locked, the door keeps you out. The constraint functions not as law or norms do — you cannot ignore the constraint and suffer the consequence later. Even if the constraint imposed by the door is one you can overcome — by breaking it down perhaps, or picking the lock — the door still constrains, just not absolutely.
Some architectural constraints, however, are absolute. Star Trek notwithstanding, we cannot travel at warp speed. We can travel fast, and technology has enabled us to travel faster than we used to. Nonetheless, we have good reason (or at least physicists do) for believing that there is a limit to the speed at which we can travel. As a T-shirt I saw at MIT put it, “186,282 miles per second. It’s not just a good idea. It’s the law.”
But whether absolute or not, or whether man-made or not, we can consider these constraints as a single class — as the constraints of architecture, or real-space code. What unites this class is the agency of the constraint: No individual or group imposes the constraint, or at least not directly. Individuals are no doubt ultimately responsible for much of the constraint, but in its actual execution the constraint takes care of itself. Laws need police, prosecutors, and courts to have an effect; a lock does not. Norms require that individuals take note of nonconforming behavior and respond accordingly; gravity does not. The constraints of architecture are self-executing in a way that the constraints of law, norms, and the market are not.
This feature of architecture — self-execution — is extremely important for understanding its role in regulation. It is particularly important for unseemly or unjust regulation. For example, to the extent that we can bring about effects through the automatic constraints of real-space code, we need not depend on the continued agency, loyalty, or reliability of individuals. If we can make the machine do it, we can be that much more confident that the unseemly will be done.
The launching of nuclear missiles is a nice example. In their original design, missiles were to be launched by individual crews located within missile launch silos. These men would have been ordered to launch their missiles, and the expectation was that they would do so. Laws, of course, backed up the order — disobeying the order to launch subjected the crew to court-martial.
But in testing the system, the army found it increasingly unreliable. Always the decision to launch was checked by a judgment made by an individual, and always that individual had to decide whether the order was to be obeyed. Plainly this system is less reliable than a system where all the missiles are wired, as it were, to a single button on the President’s desk. But we might believe that there is value in this second check, that the agency of the action by the soldier ensures some check on the decision to launch.
This is an important consequence of the automatic nature of the constraints of architecture. Law, norms, and the market are constraints checked by judgment. They are enacted only when some person or group chooses to do so. But once instituted, architectural constraints have their effect until someone stops them.
Agency, then, is one distinction between the four constraints. The temporality of the constraint — when it is imposed — is a second one.
Here I should distinguish between two different perspectives: that of someone observing when a constraint is imposed (the objective perspective), and that of the person who experiences the constraint (the subjective perspective). So far my description of the four constraints in this single model has been from the objective perspective. From that perspective they are quite different, but from a subjective perspective they need not differ at all.
From the objective perspective the difference is between constraints that demand payment up front and constraints that let you play and then pay. Architecture and the market constrain up front; law and norms let you play first. For example, think of the constraints blocking your access to the air-conditioned home of a neighbor who is gone for the weekend. Law constrains you — if you break in, you will be trespassing. Norms constrain you as well — it’s not neighborly to break into your neighbor’s house. Both of these constraints, however, would be imposed on you after you broke into the house. They are prices you might have to pay later. The architectural constraint is the lock on the door — it blocks you as you are trying to enter the house. The market constrains your ownership of an air conditioner in the same way — it demands money before it will give you one. From an objective perspective, what distinguishes these two classes of constraints is their temporality — when the sanction is imposed.
From a subjective perspective, however, all these differences may disappear. Subjectively, you may well feel a norm constraint long before you violate it. You may feel the constraint against breaking into your neighbor’s house just at the thought of doing so. A constraint may be objectively ex post, but experienced subjectively ex ante.
The point is not limited to norms. Think about a child and fire. Fire is a bit of real-space code: The consequences are felt as soon as the constraint it imposes is violated. A child learns this the first time he puts his hand near a flame. Thereafter, the child internalizes the constraint of fire before putting his hand in one. Burned once, the child knows not to put his hand so near the flame a second time.
We can describe this change as the development of a subjective constraint on the child’s behavior. We can then see how the idea extends to other constraints. Think about the stock market. For those who do not shop very much, the constraints of the market may indeed be only the objective constraint of the price demanded when they make a purchase. However, for those who experience the market regularly — who have, as it were, a sense of the market — the constraints of the market are quite different. Such people come to know them as a second nature, which guides or constrains their actions. Think of a stockbroker on the floor of an exchange. To be a great broker is to come to know the market “like the back of your hand”, to let it become second nature. In the terms that we’ve used, this broker has let the market become subjectively part of who she is.
Each constraint, then, has a subjective and an objective aspect. Laws are objectively ex post, but for most of us, the fact that a law directs us in a particular way is sufficient to make it a subjective constraint. (It is not the objective threat of jail that constrains me from cheating on my taxes; instead, I have made subjective the constraints of the law with respect to taxes. Honest, IRS. This is true.) As a subjective constraint, it constrains us before we act.
For those who are fully mature, or fully integrated, all objective constraints are subjectively effective prior to their actions. They feel the constraints of real-space code, of law, of norms, and of the market before they act. For the completely immature, or totally alienated, few objective constraints are subjectively effective. They step in the mud and only then learn about the constraint of mud; they steal bread and only then learn about the punishments of the law; they show up at a wedding in cut-offs and only then learn about the scorn of their friends; they spend all their money on candy and only then learn of the constraint of market scarcity. These two types mark out the extremes; most of us are somewhere in between.
The more subjective a constraint, then, the more effective it is in regulating behavior. It takes work to make a constraint subjective. An individual must choose to make it a part of who he or she is. To the extent that the norm is made subjective, it constrains simultaneously with the behavior it regulates.
This points to one final distinction between law and norms, on the one hand, and real-space code, on the other. Law and norms are more efficient the more subjective they are, but they need some minimal subjectivity to be effective at all. The person constrained must know of the constraint. A law that secretly punishes people for offenses they do not know exist would not be effective in regulating the behavior it punishes.
But this is not the case with architecture. Architecture can constrain without any subjectivity. A lock constrains the thief whether or not the thief knows that it is a lock blocking the door. The distance between two places constrains the intercourse between those two places whether or not anyone in those places understands that constraint. This point is a corollary of the point about agency: Just as a constraint need not be imposed by an agent, neither does the subject need to understand it.
Architectural constraints, then, work whether or not the subject knows they are working, while law and norms work only if the subject knows something about them. If the subject has internalized them, they can constrain whether or not the expected cost of complying exceeds the benefit of deviating. Law and norms can be made more code-like the more they are internalized, but internalization takes work.
Though I have used language invoking architects, my language is not the language of architects. It is instead stolen and bent. I am not a scholar of architecture, but I have taken from architecture its insight about the relationship between the built environment and the practices that environment creates. Neither architects nor I take this relationship to be determinative. Structure X does not determine behavior Y. Instead, these forms are always influences that can change, and when they are changed, they alter the affected behavior.
Like Michael Sorkin, I believe that “meanings inhere in forms, and that the settings for social life can aid its fulfillment. ” His book Local Code: The Constitution of a City at 42N Latitude suggests each feature of the model I am describing, including the ambiguity between law and architecture (building codes) and the constitution the two enable. Whatever the source of the content of these codes, he writes, “their consequences are built.” This is the feature to focus on.
My suggestion is that if we relativize regulators — if we understand how the different modalities regulate and how they are subject, in an important sense, to law — then we will see how liberty is constructed, not simply through the limits we place on law, but by structures that preserve a space for individual choice, however that choice may be constrained.
We are entering a time when our power to muck about with the structures that regulate is at an all-time high. It is imperative, then, that we understand just what to do with this power. And, more importantly, what not to do with it.
- Appendix A. Detailed explanations of special commands
- Appendix C. ICMP types
- Appendix D. TCP options
- Appendix E. Other resources and links
- Appendix F. Acknowledgments
- Appendix G. History
- Appendix H. GNU Free Documentation License
- Appendix B. Common problems and questions
- Appendix I. GNU General Public License
- Appendix J. Example scripts code-base
- Appendix D. SDRAM Interface Considerations
- APPENDIX A The History of Red Hat and Fedora