What Copyright Law and Plane Crashes Have in Common
posted by Bruce Boyden
As others have already noted, the Atlantic Monthly is now making its articles available online, including browseable issues going ten years back and select articles through most of the twentieth century. I immediately checked it out to see if one of my favorite Atlantic articles was up, and it is: William Langewiesche’s The Lessons of ValuJet 592. Langewiesche’s article is a captivating look at a classic “system accident,” the 1996 crash of a ValuJet (now AirTran) plane due to the improper loading of unspent oxygen generators in the hold. I highly recommend it.
System accidents are fascinating events that have a sort of Rube Goldberg quality to them. They typically occur in highly complex organizations that have adopted systems and procedures to avoid simple accidents — such as planes flying into each other in mid-air or cargo exploding in flight. The airline industry, nuclear power plants, large modern military forces, NASA, and contractors that build and maintain large structures such as buildings and bridges are all examples of such organizations. Such organizations are complex, with highly detailed procedures that cover every aspect of their endeavors, because they are dealing with tasks that carry with them the possibility of catastrophic damage. These procedures tend to weed out the simple and easily understood accident causes. In the process, however, the very complexity of the organization and the procedures required tax the ability of the human participants to follow them. What sometimes results is accidents that do not stem from a single, obvious cause, but rather from a number of small errors, any one of which would not cause an accident by itself, but which together combine to produce a disaster. In ValuJet 592′s case, for example, confusion among contract workers about “expired” and “expended” generators, between “generators” and “canisters,” about whether caps were required, about what color tags to place on them, and about why they were placed in the shipping department, all led to the accident — which nevertheless still could have been prevented if either the ValuJet receiving clerk or the co-pilot had questioned why they were being loaded aboard the plane. It’s only because every single one of those things went the wrong way that the accident occurred.
One frequent element of a system accident is the way in which humans and machines fail to work together well. Air accident reports often attribute such failures to “pilot error,” but that usually does not capture the whole story. Beginning in the 1970s, accident investigators and aircraft designers started to go beyond a simple notation of “pilot error” and ask if there was anything about the design of the aircraft or the procedures that made such error more likely. In other words, could things have been designed better to handle predictable and likely mistakes? This research is referred to as “human factors” engineering — that is, considering the likely human response to various situations as part of the engineering design. An early example was the response to an Eastern Airlines crash in the Everglades in 1972. While both pilots were trying to figure out if the landing gear indicator light bulb had burned out (itself a waste of pilot resources), one of them accidentally hit the steering wheel, which silently disengaged the autopilot. By the time they figured out the autopilot was off and the plane was in a descent, it was too late. Sure, that was “pilot error,” but it was an entirely predictable one — people accidentally nudge stuff all the time, particularly when they’re focused on some other task. The solution was to add an audible alarm when the autopilot is turned off — e.g., a recorded voice saying “autopilot disengaged.”
Such research applies beyond hazardous environments such as airplanes. Everyday products are often poorly designed to interact with actual humans. Take glass doors in office buildings, one of my favorite examples of where simple design choices can make a task difficult or easy. It’s often not clear from looking at such a door whether it’s supposed to be opened by pulling on it or pushing. This can be cleared up with a simple visual cue: a horizontal bar across the door, or a steel plate on the side of the door, indicates the “push” side. A short vertical bar indicates pull. Other designs, however, may not indicate which is which; indeed, many doors are very poorly designed with “pull” bars that in fact are supposed to be pushed. (Check out the Bad Designs website for lots more examples.)
How does all of this relate to copyright? Copyright law is badly designed to relate to humans. It’s particularly maladapted to apply to the humans that, more and more, need to know what the rules of copyright are: non-lawyer individual consumers.
Copyright law’s original purpose was to govern the publishing industry, and later other commercial enterprises such as music publishing, plays, record production, movie-making, and television broadcasting. In theory it’s always applied by its terms to everyone, but in fact it only applied to businesses that have the capacity to hire lawyers to make sense of vague statutes and conflicting case law, or to morally obvious situations like the pirate who sells knock-off copies for profit. As Alan Latman argued to the Supreme Court in 1972, in other cases, such as noncommercial copying by individuals, “Nobody would sue…, because it’s impractical for anyone to sue.” It’s only relatively recently that this hidden assumption behind copyright law has been exposed, as suddenly the activities of ordinary individuals have started to have a real impact on content owner bottom lines. Increasingly, it’s falling upon individuals, and not just businesses, to understand copyright law.
That would not be such an onerous task if there were clear rules for non-lawyers to follow. But copyright is bereft of clear rules. Take one of the most frequent questions individuals might have about copyright law: is X a fair use?
Of course nonlawyers are often unaware of the complexities of any law, not just copyright law. But the basic elements of most laws are within the grasp of ordinary citizens. Most people understand what libel is, and how to avoid it. They may not realize that opinion is protected (or if they do, that sometimes even opinion can give rise to liability), but they understand the basics that writing something false about someone in a way that harms their reputation can subject someone to a libel suit.
Compare that to copyright law. Fair use is a defense, but since even the basics of what’s protected (e.g., idea vs. expression) and what constitutes a violation (hello, “substantial similarity”) are fairly fuzzy, the question of what noncommercial uses are permissible devolves quickly into a discussion of fair use. And as all copyright lawyers know, fair use is a rabbit warren of questions. It begins with the standard recitation of the “four factors” codified in Section 107, all of which are extremely vague or difficult to apply. It’s a balancing test, so you don’t need all four factors to be present to have a fair use; indeed, it’s not clear any more how much some of the factors (the second and third) even matter. In fact, one of the most important elements of a fair use claim — “transformativeness” — isn’t even in the statute. Nor is there any indication in the listing of the factors which side tilts toward liability and which doesn’t. For that, you need to read some cases. And of course, don’t forget the uses mentioned in the preamble to the factors — “criticism, comment, news reporting, teaching…, scholarship, or research” — not that the preamble is determinative either way. When you’re done compiling the factors and the preamble, take a guess on how a court will rule. That judgement may depend on whether you or your opponent has done anything that strikes the judge as underhanded, even if it’s not traditionally part of the fair use analysis. An appellate court could come out differently.
You can see the temptation for people to boil this down for nonlawyers as “never copy anything without permission” or “all personal use is acceptable” or “copying up to 500 words is fine.” These are attempts to change the design of copyright law to meet the cognitive capabilities of ordinary people. As descriptions of the actual content of the law, however, they all fail.
If fair use and all of the other elements of copyright law are going to depend on being internalized by citizens in order to be effective, they have to be reasonably internalizeable. That means one of two things. Either copyright law will need to find some other means of de facto enforcement — technology comes to mind — or the law will need to be changed to make it simpler. Note, however, that copyright law is complex for a good reason; it’s not very easy to come up with rules that slice neatly along the boundary between things we want to be protected and things we want to be used. That’s why the law essentially leaves it up to the parties to decide what to sue over, and courts to decide cases on a fact-by-fact basis. That works well if disputes are relatively rare. But that situation has changed, and any simple rule that attempts to divide users’ rights and owners’ rights will inevitably, like Solomon, threaten violence to both sides’ interests. Nevertheless, the status quo is unstable. Something will eventually have to give.