My thanks to Eugene for the opportunity to address the Volokh Conspiracy audience and get some feedback on my article "Copyright and the World's Most Popular Song."

Supporters of copyright will no doubt like some of the things I have to say in the piece; questioners will like others. I'm going to start off with an aspect of the piece that's relatively pro-copyright.

The melody of "Happy Birthday to You" is quite simple and folksy. No one would confuse it with a song by Schubert or Cole Porter. Thus, one might think that it was probably created by the accretion of incremental contributions so small that none of them would qualify for copyright protection. One might also think that copyright protection wasn't needed to motivate its composition or dissemination, and that, in any event, it is not the kind of sophisticated music that copyright is really meant to promote.

What I have learned about the history of "Happy Birthday to You" has led me to question all of those thoughts. The melody of that song -- originally published with different words as "Good Morning to All" -- was the product of an intensely focused and extended creative process. Patty Smith Hill and Mildred Jane Hill, the two sisters who composed it, started with a specific goal in mind. They wanted a melody that could be easily sung and remembered by kindergarten students, yet would also have an emotional punch.

Because Patty was the principal of a kindergarten, the Hill sisters had a laboratory in which they could test melodies. And that's what they did. They would compose a song; bring it into the kindergarten; see how easily and enthusiastically the children would learn and sing it; and then go home in the evening, make changes, and bring the next draft into the kindergarten the following morning.

Out of many such rounds of testing came the "Happy Birthday to You" melody. It turns out that it's a melody that not only children but also adults with little musical talent can remember and sing a few times a year, building to that satisfying catharsis of the high note in the song's third line, celebrating a milestone of life.

The Hill sisters did not invent the melodic, harmonic, or rhythmic vocabulary they were using. But as far as I can tell, they were not just tweaking an existing melody either. Moreover, though they were not professional composers, they were quite aware of copyright. Patty told the kindergarten teachers that they could sing the song all they wanted, but that they should not write it down, because that would endanger copyright protection. (At the time, publishing a work before registering it with the Copyright Office would forfeit copyright.)

Where does this all lead? First, even simple melodies don't necessarily spring up without people devoting time and care to them. Many "folk songs" likely have authors whose identity has just been lost to our collective memory. Second, even nonprofessional composers who aren't seeking to get rich from a song may desire copyright protection, because they understand that it is important to publishers who invest in disseminating that song.

Finally, lowbrow domestic music may be worth promoting with copyright protection just as much as highbrow concert hall music. As much as I love Schubert and Cole Porter, I'm willing to concede that their entire output has probably not contributed as much to the happiness of humanity as a simple song by two unknowns, as sung in groups large and small by millions of people around the world every day.

Yesterday, my post to this fine forum argued that there were good policy reasons for granting Patty and Mildred Hill copyright protection for the melody of "Happy Birthday to You." Don't worry, they got it. Their book "Song Stories for the Kindergarten," which contained the "Happy Birthday" melody with different words under the name "Good Morning to All," was duly registered with the Copyright Office before publication in 1893. At that time, the Hill sisters would have been looking forward to a total of 42 years of copyright protection. The Copyright Act of 1909, however, gave them (and every other copyright holder) an additional 14 years of copyright -- by modern standards, a modest windfall. That brings us to 1949, when "Song Stories for the Kindergarten" took in its last royalty payment and expired, that is, entered the public domain.

Fast forward to 2008. Warner Music Group is pulling in about $2 million a year from "Happy Birthday to You." It's counting on that stream of income to keep flowing until 2030. What happened?!?

The very short answer is that WMG is not claiming copyright in the melody, but only in the combination of the melody and the words, which it claims was first copyrighted in 1935. These days, a properly renewed 1935 copyright is good for 95 years; 1935 plus 95 equals 2030.

Granting a new copyright for combining an old melody with new words, or even an old melody with old words, is not necessarily a bad thing. For example, a hymn called "Materna" and a poem called "America" both remained relatively obscure until the day that someone thought to combine them, and then voila!, "America the Beautiful" appeared. The right combinations can be of enormous cultural value, and to my mind there's nothing wrong with offering the incentive of copyright protection to experiment with new combinations.

In this case, however, the story's not that simple. There is very good evidence that the "Happy Birthday to You" words were already being sung to the Hill sisters' melody back in the 1890s. Moreover, those words and melody appeared together in many published songbooks in the 1910s and 1920s. Thus, WMG's claim that the song is still under copyright has to be much more complicated. We don't know exactly what WMG's litigation position would be, because there hasn't been any litigation about the combination of words and melody. (There was some litigation in the 1930s and 1940s, but that was about the melody alone, which was still under copyright at the time.) Here's my best guess about what WMG would have to assert:

(1) Patty and Mildred Hill actually wrote the "Happy Birthday to You" words back in the 1890s, but they did not authorize anyone to publish them until 1935. All of the songbooks in which the song appeared in the 1910s and 1920s were infringing.

(2) In 1935, the combination of "Happy Birthday" words and music were published in an authorized version with proper copyright notice (which was the necessary formality for gaining federal copyright protection at the time).

(3) In 1962, copyright in the "Happy Birthday" words and music was properly renewed.

I've spent a lot of time investigating whether these assertions are true. (I also take a lot of pages in my article draft to evaluate the assertions, which may be one reason why I haven't yet placed the article -- law review editors take note, I'll accept an offer conditioned on cuts!) It turns out that all three assertions have serious problems.

I'll discuss those problems in a post tomorrow. For now, let me just say that it seems quite clear that, back in 1935, no one associated with "Good Morning to All" was thinking that publishing it with the "Happy Birthday" words would extend copyright past 1949. For example, the registrations for the versions of the song published in 1935 claimed copyright only in the arrangements made by the publisher's employees-- piano accompaniments and the like -- not in the song itself. And those 1935 versions credit only Mildred Hill, the musician sister who wrote music for dozens of published songs but never once wrote lyrics; they fail to credit lyricist Patty Hill.

Yet, as Dr. Johnson would have said, when a music publisher knows that his successful song will soon slip into the public domain, it concentrates his mind wonderfully. As 1949 approached, someone came up with a colorable theory under which royalties for the use of "Happy Birthday to You" could be demanded until the distant year of 1991 -- which, thanks to repeated Congressional largesse, has now receded to the still-distant year of 2030.

In yesterday's post, I argued that to get a court to uphold the 1935-2030 copyright in "Happy Birthday to You," Warner Music Group would likely have to convince the court of three things. It turns out that WMG would face difficulties with all three. The details are in my article ( http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1111624), but here's a summary:

1. There's little or no evidence that Patty or Mildred Hill wrote the "Happy Birthday" lyrics. Indeed, Patty Hill testified in 1937 that she had written the original "Good Morning to All" lyrics, and that she had used the "Happy Birthday" lyrics, but she stopped conspicuously short of testifying that she had written the "Happy Birthday" lyrics. Over a hundred years after those words were first used with the Hill sisters' melody, no one is alive who could testify about their origin, and I have found no relevant documentary evidence — no drafts, no letters, nothing.

2. The 1935 publications of "Happy Birthday to You" bore the copyright notice "Copyright 1935 by Clayton F. Summy Co." It is very likely that the Summy Company did not own copyright in the song at the time (it probably had an implied license for the song, and owned only the musical arrangments that its employees had made). Under then-prevailing precedent, and for several decades thereafter, if you published a work with copyright notice naming someone other than the work's owner, you forfeited copyright. However, here WMG might be able to take advantage of later changes in judicial attitude.

3. The renewal registrations filed in 1962 — necessary to maintain copyright beyond 1963 — are only for the arrangements, and do not claim to renew the song itself. This is probably the point of greatest weakness in the copyright, and this issue could also likely be decided early in litigation, because the facts are clear.

So if there are these weaknesses, and if a $2 million per year income stream is at stake, why hasn't anyone challenged the copyright's validity? The short answer is that no single user is paying enough of that $2 million to make a challenge worthwhile, and it's extremely difficult for users to organize a collective effort.

In recent years, about 35% of the "Happy Birthday" income has come from performance rights licensing through ASCAP. (I can make this estimate because I discovered that there has been litigation over the alleged mishandling of a trust funded by "Happy Birthday" royalties, and I got access to the court files, which include income reports.) ASCAP collects money from thousands of restaurants, bars, and radio and TV stations for "blanket licenses" covering all of the millions of songs that are in its repertoire. The price of the licenses don't change when individual songs go in or out of copyright. That means that the licensees don't have an incentive to challenge the copyright on one song (and a court might even rule that they couldn't). Other music publishers that receive royalties through ASCAP might be able to mount a challenge, but there may be too much glass in their own houses to start throwing stones.

The remaning 65% comes from all other licensing — for uses in movies, on TV, in ads, and so on. Here's one small example I just learned of from a woman who e-mailed me yesterday from Australia. She was involved with an Australian movie called "Annie's Coming Out". It's about a social worker who works at a hospital for mentally disabled children. During one scene in the movie, a group of children with multiple sclerosis sing "Happy Birthday" to another child with MS. When the movie was released (on a very small scale) in the US in 1984, its producers had to pay WMG's predecessor $5000 to use the song.

Licenses for bigger US releases probably cost more (IMDB lists 176 movies that feature "Happy Birthday to You" but it misses "Annie's Coming Out"). But even, say, $30,000 is not nearly enough to consider funding copyright litigation. And in my article, I consider coordination problems that make it unlikely that a group of "Happy Birthday" users will ever find each other and jointly finance litigation. The result is that the copyright in the song will probably never be tested.

Ultimately, this is not just about one song. There are almost certainly other works out there generating significant licensing income in spite of serious copyright weaknesses. However, I don't see any easy fix.

Taking a breather from arcane copyright doctrine, here's an issue that's not specific to copyright, or even to law at all. What is the proper role of anecdotes in making policy arguments?

Suppose that I'm arguing that the term of copyright is too long, and I say, "because of Congressional pandering to special interests in Hollywood, even 'Happy Birthday to You' is still under copyright!!!" Why is this example likely to carry particular persuasive force? In part, because the audience probably makes false assumptions. "Happy Birthday to You" is one of the few songs that most people still learn as children, at home, from family and friends, far from the commercial world of iTunes and Vh1. That will cause many people to lump it with songs like "Mary Had a Little Lamb" and "Twinkle, Twinkle, Little Star," though the latter are considerably older. And many people may also assume that if "Happy Birthday to You" is under copyright, the copyright owner could demand a license for singing it at a family birthday party — outrageous! — though private performances have never been regulated by copyright. I discuss this sort of problem with the use of anecdotes in my article, I can't say that I have the answer.

Taking a hard line, one might say that all anecdotes and examples should be accurate and typical (and not foreseeably misleading) or they shouldn't be used. The requirement of being typical may be impossible to meet, however. Members of a group may be diverse enough that none of them is really "typical," and anecdotes can be misleading in very subtle ways.

Meanwhile, we human beings may be built in such a way that statistics often don't move us enough to take appropriate action. Specific examples, as supplements to numbers and percentages, may actually be necessary to engage us to respond adequately. So it's not clear that minimizing the use of anecdotal argument would even theoretically be optimal, given that we are who we are.

Here's a possible rule of thumb: small deviations of examples from the typical are inevitable, and just need to be accepted; medium-sized deviations should be accompanied with a proviso ("not all cases are like hers"); atypical examples should be avoided; and there should be some effort to figure out how typical an example is deploying it. I'm not sure how far that gets us, however. I have to confess that in yesterday's post, I mentioned the use of "Happy Birthday to You" in the movie "Annie's Coming Out" in part because the scene sounded striking: kids with multiple sclerosis were singing it to another kid with multiple sclerosis. Is that a typical use? Is the fact that that scene is touching something we should really take into account when formulating copyright policy in general, or in a fair use analysis?

Perhaps Volokh Conspiracy readers have insights to share.

Suppose you owned a piece of land and took no legal action while others used it for more than two decades. When you (or your successor) eventually tried to evict those users, would you prevail? For centuries, the doctrines of adverse possession and easements by prescription have stood in the way of such long-delayed action. These doctrines rest on a notoriously mixed bag of rationales, including protecting the expectations of users, who may or may not have been acting in good faith; promoting productive use of land; keeping title free of complications that may hinder transfer and investment; and obviating the need to decide cases on thin, stale evidence of ownership claims. But the doctrines themselves have remained a remarkably durable part of land law.

Copyright law has never had similar doctrines (nor has patent law). As long as the term of copyright remained relatively short, it arguably had little need for them. Copyrighted works were going to enter the public domain soon anyway -- after, at most, 28 or 42 or 56 years -- and they would enter the public domain even sooner if copyright holders didn't assert their continuing interest in the works by filing a renewal registration after 14 or 28 years. Now, however, copyright can last for 120 years or more, and has no renewal requirements. Thus, it is possible for a copyright owner to neglect to enforce copyright for decades, and then return and reassert ownership for decades more.

The history of "Happy Birthday to You" gives us a glimpse of what can happen when long copyright has no adverse-possession-like doctrine. The song was distributed widely for about 20 years from the early 1910s to the early 1930s without any permission from or enforcement by the Hill sisters or their publishers. It was during that period -- a period when the song was de facto in the public domain -- that it became THE standard birthday song in the United States. It is even conceivable that the lack of copyright enforcement contributed to the song's assumption of its central place in American culture. Then the putative copyright owners resurfaced and started claiming royalties, not for another decade or so until 1949, but, it now appears, for another 95 years until 2030.

The copyright community has started discussing the related and partially overlapping problem of "orphan works," but so far there's been no legislative progress on that front. The doctrine of adverse possession was actually a judicial creation, piggybacking on top of statutes of limitations. The copyright statute of limitations, however, is so short -- 3 years -- that it is unlikely that judges will ever read doctrines of adverse possession or prescription into it. The result is that copyright will likely increasingly face all of the problems -- clashing expectations, clouded title, stale evidence, use-inhibiting uncertainty -- that land ownership would have without those doctrines.

One comment earlier this week noted the evidentiary problems that arise as works under copyright get older and older. Some of those are inevitable. Others could be avoided with better Copyright Office recordkeeping policy, which has not kept up with the increasing copyright term. As I detail in my article, the Copyright Office has maintained a policy of discarding much correspondence and many deposits after a decade or two. In addition, there has never been a way for the public to learn when the Copyright Office has denied registration of a work, unless the dispute gets into the courts, because only successful registrations are assigned searchable numbers. I heard a rumor that there had been an unsuccessful attempt to register the "new" melody of "Happy Birthday to You," different from the melody of "Good Morning to All" only in that the quarter note to which "Good" was sung was split into two eighth notes to accommodate the word "Happy." The rumor is probably false, but since there's no way to search denied registrations . . . who knows?

With that, I'll conclude my series of posts about "the World's Most Popular Song." Thanks again to Eugene for giving me this opportunity, and to all of you who responded for your comments and discussion.