« Metaphors, Models, and Meaning in Contract Law | Main | How is a Blizzard Like Unemployment Insurance?, or, Should States Save for Disasters? »

Tuesday, February 15, 2011

The Persistence of Legal Error

When I was in my first semester of law school, I was given a short memo assignment involving some principle of Connecticut contract law. I quickly found a case stating the relevant rule of law--every contract needs consideration, or something. But it quoted an earlier case. Being a good historian, I knew I couldn't just use the more recent case--I had to track this down to its source. So I looked up the earlier case. But that in turn cited an even earlier one for the same rule. So I looked up that one. After about nine or ten iterations of this, I was in the 18th century, and courts were still citing earlier cases, now from English reporters that I couldn't look up as easily. I gave up, and concluded that legal authority worked differently than historical authority--if an earlier court said it, that's good enough, no matter where it originated or what the original context was.

The upside of this is that rules can get transmitted from case to case much more efficiently. The downside is that errors can spread just as easily. Take the idea from copyright law that contributory infringement liability is derived from the tort law concept of enterprise liability. This explanation is widespread in the case law. See, e.g., Perfect 10, Inc. v. Visa Int'l Serv. Ass’n, 494 F.3d 788, 794-95 (9th Cir. 2007); Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259, 264 (9th Cir. 1996); Intellectual Reserve, Inc. v. Utah Lighthouse Ministry, Inc., 75 F. Supp. 2d 1290, 1293 (D. Utah 1999); Polygram Int'l Pub., Inc. v. Nevada/TIG, Inc., 855 F. Supp. 1314, 1320 (D. Mass. 1994). It's also featured in the influential Nimmer treatise: "A separate avenue for third-party liability in the copyright sphere is contributory infringement, which forms an outgrowth of the tort concept of enterprise liability," Nimmer § 12.04[A][3]. And, it's taught in law schools. The textbook I used to teach copyrights from 2007 through last year used to have only a one-paragraph introduction to secondary liability, followed by cases such as Fonovisa, which included the "enterprise liability" explanation. So, I dutifully repeated it to my students in both copyright and Internet Law, even though I was not really sure what "enterprise liability" was.

It turns out that it is flat wrong. Contributory infringement liability has nothing whatsoever to do with enterprise liability.

The earliest citation for the enterprise liability suggestion, and its apparent origin, is Demetriades v. Kaufmann, 690 F. Supp. 289 (S.D.N.Y. 1988). Demetriades was a case involving copyrighted architectural plans; the plaintiff claimed that the real estate agent and realty firm that had sold an empty lot were contributorily liable for the house built using infringing plans on that lot. The court had to determine whether those two defendants were indirectly liable for the infringement.

The Demetriades court began by correctly noting that "[i]n delineating the contours of this third-party liability, and because copyright is analogous to a species of tort, 'common law concepts of tort liability are relevant in fixing the scope of the statutory copyright remedy . . . .'" But then the court moved from that premise to a surprising conclusion: "Guided, therefore, by well-established precepts of tort liability, it appears that two avenues of third-party liability in copyright have grown up in the law--'vicarious liability' (grounded in the tort concept of respondeat superior) and 'contributory infringement' (founded on the tort concept of enterprise liability)."

Contributory infringement as defined in copyright law is liability for knowingly providing encouragement or assistance to an infringer; it's thus the analog of contributory tortfeasor liability for knowingly aiding a tortfeasor. See Rest. 2d Torts §§ 876, 877. It's not a form of strict liability. Enterprise liability is something completely different. Enterprise liability is a theory of tort liability that emerged in order to hold an entire industry responsible for harm caused by an industry-wide standard when identifying the particular tortfeasor in a given case is impossible. See Black's Law Dictionary, 9th ed. The classic case involves injuries caused by blasting caps to children. It can be impossible to determine after the fact who made a blasting cap that exploded. If the entire industry manufactures blasting caps that are unreasonably dangerous, enterprise liability permits a suit against everyone for their share in the harm. See Hall v. E. I. Du Pont De Nemours & Co., Inc., 345 F. Supp. 353, 376 (E.D.N.Y. 1972).

Bill Patry, in his recent copyright treatise, spends an entire section refuting the "enterprise liability" connection. See 6 Patry on Copyright § 21:44. He goes out of his way to point out that the Demetriades decision was issued by a court that, although within the confines of the Southern District of New York, was in fact located in White Plains. I take it the suggestion is that once you leave the sophistication of Manhattan, knowledge of copyright law decreases proportionately. But in fact the Demetriades court seems to have made its mistake out of an excess of caution. Not simply satisfied to repeat earlier claims about contributory liability's origins in tort law, it attempted to ground that claim. The problem is that it did just enough research to be dangerous.

The court didn't provide any direct support for the problematic sentence, but a sentence later it dropped a footnote that indicates what may have happened. First, the court complained that the relevant authorities were unhelpful in distinguishing between contributory and vicarious infringement. As to the canonical decision in Gershwin Publishing Corp. v. Columbia Artists Management, Inc., 443 F.2d 1159 (2d Cir. 1971), this criticism seems unfair, but it is certainly accurate as to the Supreme Court's then-recent decision in Sony v. Universal, the "Betamax" case, and may have been accurate as to the Nimmer treatise as it existed at the time. The Demetriades court next noted "the plain distinctions between vicarious and enterprise liability in tort" and cited the 4th edition of William Prosser's Handbook on the Law of Torts §§ 69 & 72, published in 1971, even though an updated 5th edition was released in 1984. It was probably what the judge had available.

Prosser's 4th edition does not mention enterprise liability at all. For theories of imputed liability, it covers, in § 69, vicarious liability, and in § 72, liability among members of a "joint enterprise." That is, if two people or two businesses engage in a joint enterprise, one member might be held liable for the tortious acts of the other. This is just close enough to the rules of liability for contributory tortfeasors to be confusing. Somehow the word "joint" dropped out, and the Demetriades court referred only to "enterprise liability."

For whatever reason--probably because most federal judges, like me, didn't know what enterprise liability was either--the Demetriades court's suggestion was picked up and repeated. Indeed, the popularity of the Demetriades court's suggestion shows that subsequent courts not only didn't know what enterprise liability was, but they also didn't know where contributory infringement liability originated. First, the enterprise liability suggestion was picked up by the influential district court decision in Polygram International v. Nevada/TIG. At some point, it also made its way into the Nimmer treatise, which cited Demetriades. Both Demetriades and Nimmer were cited by the Ninth Circuit in its even more influential decision in Fonovisa v. Cherry Auction in 1997. From there the enterprise liability suggestion has continued to spread, including in the Ninth Circuit's recent Perfect 10 v. Amazon decision.

Fortunately, in none of these cases did the reference to "enterprise liability" actually matter. In Demetriades itself, the court later went on to cite Section 876 of the Second Restatement of Torts--which provides for contributory tortfeasor liability. No court has held that contributory liability in copyright law is a form of strict liability that requires apportioning damages among defendants according to market share. But the repeated citation of contributory liability as originating in enterprise liability poses the danger of such a result, and in the meantime obscures contributory infringement's actual origins. Despite its wrongness, and despite the efforts of people such as Bill Patry to remove it, it's got a lot of case law inertia behind it now. I expect "contributory liability as enterprise liability" to have a long life to come.

[Cross-posted on the Marquette Law Faculty Blog.]

Posted by Bruce Boyden on February 15, 2011 at 11:59 AM in Intellectual Property, Judicial Process | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference The Persistence of Legal Error:


Interesting, Peter, thanks!

Posted by: Bruce Boyden | Feb 17, 2011 12:54:29 PM

Great post. I think the most interesting part of this post is the observation that the reference to enterprise liability didn't matter in any of the cases in which it came up. Your post reminds me of an article by my colleague Alan Watson on the supposed origin of trade secrets law in the Roman law of acto servi corrupti or the action for making a slave worse. Complete rubbish it turns out. Watson's article, which is at 11 Tul. Eur. & Civ. L. 19 (1996), inspired me to do an extended treatment of the claim that the procedural device of intervention--now in our modern Fed. R. Civ. P. 24 and its analogues--has its origins in the Roman law of civil procedure. Again, complete rubbish. My article is at 31 Ga. J. Int. & Comp. L. 33 (2002). It's fascinating that these false claims get repeated in quite reputable sources and that they never make a difference to the argument that the author is trying to advance. I think it's because lawyers are always searching for authority and are often trying to make themselves look erudite (which I guest they think burnishes the rest of their argument).

Posted by: Peter Appel | Feb 17, 2011 11:39:20 AM


That's a good one.

Posted by: Orin Kerr | Feb 15, 2011 3:01:00 PM

Thanks Orin; that post in turn reminds me of some formulaic practices in civil pleading that drive me nuts, e.g. the practice of, in each claim, beginning each claim with Paragraph X: "Plaintiff hereby realleges and incorporates by reference Paragraphs 1 through [X-1] as if fully set forth herein." This is a waste of time and paper. The complaint is one document, which is supposed to contain only a "short and plain statement of the claims" for which "no technical form is required" and which is required to be "construed so as to do justice."

But I bet complaints have been dismissed because Count 4 didn't "fully set forth" all of the allegations that came before, and so was construed to stand alone. So you can bet that all of my complaints followed the standard format.

Posted by: Bruce Boyden | Feb 15, 2011 1:56:44 PM

Interesting. I'm reminded a bit of this post on The Strange Practice of Indicting in the Conjunctive:

Posted by: Orin Kerr | Feb 15, 2011 1:22:21 PM

The comments to this entry are closed.