2002 Copyright Law Exam Sample Answer — Eric Goldman

UC Berkeley, Boalt Hall School of Law
Copyright Law – Spring 2002
“Sample” Answer
Eric Goldman (ericgoldman@onebox.com)


In general, the exams were very good and a pleasure to read.  Compared to my Cyberspace Law exams, I had a strikingly small number of incorrect statements of the law.  I’d like to think this was due to my teaching brilliance, but I suspect it had more to do with the general skill level and interest of the class, as well as the foundation many of you got in the Intro class.

Usually, students distinguished themselves at the high end of the curve through deeper issue spotting, accurate application of legal tests, and more insightful arguments and creative applications of tests to the facts.  But most exams covered most core issues, so the distinctions I had to make were finer-grained than normal.

One hot button to note, although I trust this was clear from class: I HATE circular/tautological reasoning.  You may need to use circular reasoning where it is legally adopted, but it’s always a good idea to acknowledge the circularity.  Those of you who parroted circular reasoning without realizing it or, even worse, created your own circular reasoning that went beyond the circularity already baked into copyright law usually lost some points.

There were 4 HHs, 16 Hs and 28 Ps.  I apologize for the relatively low number of HHs but I didn’t really have a choice based on the mandatory class curve and the way I graded the exam.

Question 1

This was a racehorse question with too many issues to address fully in the allotted time.  I got feedback that there wasn’t enough time to answer this question the way you wanted to, so you were not alone.  Organization, time-management and efficient analysis were keys to high-scoring responses.

As usual, this question is based on a real-life question I encountered in practice.  Indeed, this fact pattern plays out constantly in client/law firm relationships.  I have seen IP issues addressed in engagement letters only twice in my career.  Note this is not necessarily a problem—I didn’t add IP provisions to the engagement letters I formed for Epinions—but it does raise interesting issues when the client wants to exploit the attorney’s work!

Is the agreement copyrightable?

Initially, we should assess to what extent the agreement constitutes copyrightable subject matter.  We need to determine if we have an original work of authorship fixed in a tangible medium of expression.

  • Fixation.  The agreement was published.
  • Work of authorship.  An agreement can qualify as a literary work.  Alternatively, because the agreement represents an aggregation of multiple elements from multiple sources, the agreement could be a compilation.
  • Originality.  This is a harder question.  To have originality, the agreement needs independent creation and a minimum degree of creativity.  I’ll discuss independent creation below in the “ownership” section.  As for the minimum degree of creativity, the idea/expression dichotomy limits copyrightability to some degree:

–         clever solutions are not protectable as such, although their expression can be protectable.

–         the merger doctrine limits copyrightability when there are a limited number of ways to express an idea.  In this case, the statutory-driven provisions are likely subject to the merger doctrine and not protectable.  Morrisey.

–         blank forms are also subject to limits on copyrightability (Baker v. Selden; 37 CFR 202.1), but a form agreement is different from a blank form or system incorporated in a form.

In this case, we have evidence that the agreement is complicated and required substantial judgment from multiple attorneys about what to include and exclude and how to phrase it.  The agreement should have the minimum degree of creativity required to support copyrightability.

Who owns the copyright?

Next, we need to figure out who owns the copyright.

  • Yabe marketing language.  Yabe probably owns this language as a work for hire of its employees.
  • Statutory-driven provisions.  These are probably public domain given the merger doctrine and Morrisey.
  • Provisions cut ‘n’ paste from the Internet.  These provisions could be public domain either because of the merger doctrine or because they are otherwise too lightweight to qualify as an original work of authorship.  Otherwise, the copyright is owned by the respective authors.

–         This is a crucial point.  Many of you identified that Yabe may not own the copyright in these provisions, but a subset of you identified that Yabe in turn may be infringing.  To properly answer the call of the question—advising Handy—you needed to reach this issue.

  • Provisions drafted by Yabe’s previous attorney.  We don’t have enough facts to determine who owns these.
  • Provisions I drafted from scratch. There are 3 possible owners of these provisions: me, the firm, or Yabe.

–         Me.  Copyright vests in the author (Sec. 201(a)), so the starting point for our analysis is that I own the copyright in this material.

–         The firm.  I could have assigned ownership to the firm contractually, but I don’t have any agreement with the firm that addresses IP.  Nevertheless, if my drafting constitutes a work for hire in the scope of my employment, then the agreement’s copyright vests in the firm.  Sec. 201(b).  I am an employee of the law firm and the agreement should be within the scope of employment.  Thus, the copyright should vest with the firm.

–         Yabe.  We know that there is no written agreement between the firm and Yabe that either transfers ownership or an exclusive license.  Therefore, Yabe doesn’t have either.  Sec. 204(a).

Further, Yabe can’t take advantage of the specially commissioned work for hire doctrine in Sec. 201(b).  While the agreement could qualify as a compilation or perhaps a supplementary work or a contribution to a collective work, a specially commissioned work for hire requires a written agreement and we don’t have that here.  The Second Circuit caselaw enabling a work for hire to be codified in a subsequent written agreement if it was the subject of an oral agreement prior to commencement isn’t all that helpful; it still requires a written agreement to implement the work for hire, and an assignment can be done in that written agreement just as easily.

–         Some of you argued that I was an employee of Yabe or that the law firm was the employee of Yabe.  I didn’t give you enough facts to really analyze this issue properly under the CCNV factors, but a law firm isn’t an employee of another company; and in this context it doesn’t make sense for me to be both an employee of the law firm and an employee of the client at the same time for the same activity.  While a multi-factor test like CCNV can be manipulated to reach any result, be careful of letting your desire to achieve a client’s desired result trump your common sense.

On balance, I would conclude that the law firm owns the provisions I created and the client has an undefined nonexclusive license to those provisions.  This of course creates a significant standing issue for Yabe.

Some of you tried to overcome the standing issue by suggesting that the firm could assign or grant an exclusive license to the work.  Note that, given the amount of recycling inherent in law firm life, most law firms won’t grant such rights.  Therefore, Yabe may very well not be able to get standing to enforce the copyrights in these provisions.

A better “solution” would be for the law firm to join Yabe as a co-plaintiff.  The law firm may be willing to do this but probably would want Yabe to pay the firm’s expenses, substantially increasing the costs of enforcement.

  • Edits from the starting point agreement and provisions from other agreements you drafted.  You created a derivative work of the starting point agreement, so we need to know who owned it.  If your analysis concludes that you/your firm are an employee of Yabe such that Yabe owns the resulting agreement, then you have a dilemma—the same analysis should lead you to suspect that the starting point agreement was a work for hire for that client, meaning the new agreement is an infringing derivative work of the starting point agreement.  However, if you conclude that the firm owns all agreements it creates, then the derivative work is permissible because it’s made by the owner of the underlying work.
  • Edits and integration.  The ownership of these tracks the analysis on the provisions drafted from scratch.  To the extent the underlying content is not copyrightable, these contribute to the selection/arrangement/coordination to create the compilation.
  • Comments from other attorneys in the firm.  Anything copyrightable in these should be owned by the firm as works for hire within the scope of employment.
  • Comments from Handy.  Yabe should own anything copyrightable here as a work for hire within the scope of employment.  Although it’s possible, it’s unlikely comments on their own represent copyrightable works.  However, these, combined with the Yabe marketing language contribution, create the possibility that the agreement is a joint work of authorship.  To have a joint work, there has to be an intent to merge contributions into a unified whole.  The agreement might qualify.  See Thomson v. Larson.  Of course, in that case Yabe would have a duty of accounting to the law firm, which depending on how the revenue stream is defined could be enormously expensive.
  • “Taxonomies.”  Some of you discussed the agreement’s taxonomy.  This is a bit oxymoronic, as a taxonomy is a classification scheme and thus doesn’t really apply to agreements as such.  The agreement will have an organizational structure or outline, and this is protected if at all as part of a compilation copyright.

In sum, there are 3 ways Yabe could have ownership rights sufficient to give it standing to sue:

  • You or your firm is an employee of Yabe.  This is dubious.
  • The agreement is a joint work, giving Yabe the right to sue as a co-owner but creating the duty to account.
  • Yabe owns only the marketing language (and Handy comment contributions), but it can sue to protect those to the extent they are copyrightable.

This section was a big mover in many scores.  Some of you did a very careful job in this section, admirable under the time pressure, and you usually were rewarded for your meticulousness.  Others of you breezed through this section in a couple of sentences, catching few of the nuances, which usually nudged you towards a P.


Now that we’ve analyzed what Yabe owns, we can determine if it was infringed.  To establish infringement, we need to establish copying and violation of an exclusive right.

  • Copying.  Copying can be shown directly, or it can be shown indirectly by showing access to the work and substantial similarity.  The facts specified copying, but even if they didn’t, indirect copying can be shown.  The copyrighted work is published in the marketing collateral, so defendants should have access.  The facts specify substantial similarity—the agreements are the same except for the name change—and you might even find “striking similarity” such that access could be inferred.
  • Violation of exclusive rights:

–         The competitors are reproducing and distributing the agreement.  Even if the agreement is only subject to a compilation copyright, we’d still find infringement because the entire compilation was taken.

–         Arguably, the competitors are making a derivative work by doing the global name change.  This depends on whether minor nonsubstantive modifications qualify as a derivative work.  See the ART cases.

–         If the agreement is a literary work, then the competitors are also publicly displaying the agreement

–         By replacing Yabe’s name, Yabe could have a separate action for the removal of its copyright management information under Sec. 1202.  This was a subtle point picked up by maybe 25% of you.

Fair Use

Any defendants would claim that their use was fair.  Although the analysis is similar, a different fair use analysis could be done depending on whether the infringed property was just the Yabe marketing language or the agreement in total (excluding the portions not able to be owned by Yabe).

  • Purpose of use.  It’s reasonable to assume that most defendants would be commercial enterprises, so many of you then concluded that this factor favored the plaintiffs.  However, this factor looks at the nature of the use, not the nature of the user.  Nevertheless, this factor probably weighs against the defendants in context as the use was made to facilitate commercial trade.  Some of you argued that the use was commercial because the defendants saved money by not having to draft the agreement themselves.  While this argument has been accepted by courts (e.g., Texaco), it’s a circular argument and should be used advisedly.

As for whether the use was transformative, I think the light degree of changes (just changing the name) and its use for exactly the same purpose makes it not transformative.

  • Nature of the work.  On a scale between fact and fiction, the agreement is much closer to fact than fiction.  Some of you argued that its status as a literary work made it more like fiction.  I think this factor distinguishes if the work is more like core copyrightable material such as books or plays or more like factual information such as factual databases.  In this case, a functional, legal liability-driven document is more like a factual database than a core copyrighted work even if it is a literary work under Sec. 101.  Also, the agreement was published.  Under either approach to this factor, it favors the defendants.
  • Amount taken.  Effectively 100% was taken.  This factor favors the plaintiff.
  • Effect on the market.  This is a difficult factor to analyze, because we can’t easily define the market.

We could define the market as the market for licensing user agreements, but this market doesn’t currently exist and little suggests that there is a “traditional, reasonable, or likely to be developed market.”  (Texaco).  On the other hand, there is a small market for the license of agreements and forms generally, so Yabe could in theory enter this market.  If this is the correct market, the disposition of this factor entirely rests on the judge’s willingness to believe that Yabe might want to pursue such a licensing program or should have the right not to do so, and thus is entirely up to the judge’s biases.

If Yabe is suing solely on the marketing language, then we could conclude that there is no market for that.  Sony v. Bleem.

Some of you argued that the market is Yabe’s core business generally.  This is not totally consistent with the statute, which talks about “the potential market for or value of the copyrighted work,” but it’s an interesting argument nonetheless.  However, this argument doesn’t buy you much, as it’s not clear that competitors all using the same agreement hinders or reduces their respective competitive positions.  Certainly copyright law can’t be used to preclude competitors from establishing the same business practices, so the harm from articulating the business terms in the same way is unclear.

Similarly, a few of you looked at the market as the market for your legal services, which would be reduced if a lawyer’s work were recycled without compensation.  This would be relevant only if you or the firm were the plaintiff, but this wasn’t the call of the question.

On balance, I believe that the use isn’t fair because it’s not transformative and 100% is taken.  It’s not a slam dunk either way.

I was fascinated by the disparity of the ways in which you analyzed and concluded on fair use.  As you might expect, collectively your arguments were all over the map, illustrating the complete failure of the 4 factor test to provide any rigor or predictability to the results, at least among 48 bright Boalt students.  As a result, my grading of this section usually depended on depth of thought and accuracy of applying the factors.  Those of you who misapplied the factors or, worse, applied circular reasoning without acknowledging its circularity tended to lose some point.

Advice for Handy

Many of you stopped with fair use and didn’t actually answer the call of the question—what should Handy do?  Those of you who allocated your time to provide a thoughtful response to this question usually picked up some points.  In this situation, I advised the client to let the issue go for the following reasons:

  • The agreement is subject to a thin copyright at most.
  • Yabe’s ownership may be limited to the marketing language it contributed.
  • If Yabe establishes joint ownership, it will have an uncertain duty to account.
  • Yabe could be deemed an infringer itself.  Calling attention to the agreement increases that risk.  Even if it has little legal risk of being sued for infringement, it does have some risk of looking hypocritical (“we can steal portions from other people, but they can’t take from us!”).  Press attention to this hypocrisy could be embarrassing and counterproductive to long term corporate objectives.
  • Each lawsuit will be expensive, and the defendants can quickly mitigate their liability by rewriting the agreement.
  • The damages that Yabe could receive are limited (no statutory damages unless it’s registered on time; and limited actual damages).
  • Therefore, the cost/benefit analysis doesn’t favor litigation.
  • Plus, there may be strategic benefits to competitors using the same agreement.  Consider a situation where Company A does thoughtful legal work and competitors B-Z are small and not likely to have good legal representation.  A may very well want B-Z to use a solid legal agreement if they (as opposed to A) are possibly going to establish legal precedent.  So it could be counterproductive to Yabe’s long-term competitive interests to force its competitors to use a worse agreement.


Because of all the detail in this question, this sample answer doesn’t illustrate what you could have done in the time allotted.  Indeed, the best scoring answers touched on less than 75% of the detail above.

High scoring answers worked through the copyrightability and ownership issues with some detail, reaching points like the breakdown of ownership by source and the identification of the agreement as a compilation.  They also had a thoughtful fair use analysis.  Alternatively, answers that skipped the fair use analysis entirely lost points, as did answers that treated the agreement as a monolithic whole for the ownership analysis.  Extra points were given for identifying the 1202 issue.

Question 2

As some of you determined, this question is loosely based on the behavior of a well-known search engine.  I chose this question to illustrate the way in which commercially available services can be launched—and used by many students in class!—without even conceiving of it as a copyright violation.  This highlights that one should “just do what everyone else is doing” advisedly.


  • Robot accessing the web page.  This triggers a copy into RAM and violates the copyright owner’s reproduction right.  This is the same argument that concludes browsing is infringement, so we might question its general validity.
  • Robot making a copy of the PDF for the central database.  This also violates the reproduction right.
  • Converting the PDF to HTML.  This is a derivative work.
  • Displaying a portion of the page as a search result.  This is either a copy or a derivative work (as a redaction of the entire page or by putting the content into a new context per the ART case).  This also violates the owner’s distribution and public display rights.
  • Linking to the PDF.  Unlike Kelly, this is not a public display because there’s no frame or in-line linking.  However, linking can be contributory infringement if browsing is infringement (Utah Lighthouse) or per Napster.
  • Distributing the HTML version.  Most search engines include a link to a “cache” version or some other version of the page they indexed.  It’s hard to reconcile this practice with copyright law, despite its ubiquity.  (512(b) provides a caching safe harbor but, despite the nomenclature, I don’t think this behavior is caching as contemplated by 512(b)).  In this case, distributing the HTML version violates the owner’s reproduction, distribution and public display rights.  In addition, the added highlights and the display reformatting arguably violates the owner’s derivative work rights. The additions/changes aren’t particularly substantive but may be enough to constitute a derivative work.
  • Criminal liability.  If one concludes that any of the above activities is willful, then Toggle likely has criminal liability because it does so for commercial advantage.
  • Secondary liability.  Even if Toggle is permitted to engage in any of the above activities under copyright law, Toggle could have secondary liability if any of the files it picks up were themselves infringing.  This is basically why Ditto.com couldn’t stop indexing Kelly’s photos—it stopped visiting Kelly’s site but picked up copies from people who infringed Kelly.
  • Anti-circumvention.  The system of converting the PDF to HTML and then displaying the HTML version arguably violates Sec. 1201 as circumvention or the trafficking/provision of an anti-circumvention system.

Presumably Toggle doesn’t violate 1201(a)(1) because PDF doesn’t attempt to control access.  However, the PDF bypass could violate 1201(b) because it violates the copyright owner’s control over a 106 right (either reproduction or the creation of a derivative work, take your pick) if it meets one of the three tests:

–         it was “primarily” designed or produced for the bypass.  Toggle will argue the bypass is primarily for indexing purposes, but I think the statute could be read to apply to this case

–         it has limited commercially significant purposes other than the bypass.  I think the same arguments apply.

–         marketing.  We don’t have any facts about the marketing.

Also, if Toggle links to any anti-circumvention devices, it could be trafficking in the files.  Corley.

Not Really at Issue

  • User Editing.  A number of you argued that Toggle could have secondary liability because users were editing the HTML files.  This is a very hard argument to support because it doesn’t really pass the common sense test.  Applying this argument to a Napster extreme, it’s like arguing that Napster is liable for any subsequent sampling done by users who download MP3s through their system. The causality connection is just too strained.

However, applying the law, the test for contributory infringement is “one who, with knowledge of the infringing activity, induces, causes or materially contributes to the infringing conduct of another.”  Many of you immediately jumped to the Napster bastardization of this test, which should be done only by acknowledging that it is a bastardization.  (Remember how many times in class I suggested that Napster may be explainable as precedent only by limiting its applicability to Napster?).  Here, the facts don’t give you a lot to support knowledge of infringing activity.  Toggle’s contribution to the infringement is debatable.  True, they provide the capacity for users to infringe by bypassing the PDF lock, but Toggle doesn’t actually do anything with respect to the editing decision by users.   I think Toggle’s behavior doesn’t meet the test unless you use the Napster bastardization.

Vicarious liability is dubious because Toggle doesn’t derive any financial benefit from the infringing activity and has no ability to control what users do after the files have left their servers.  Of course, you could reach a contrary result using the Napster bastardizations, but again you need to acknowledge the bastardization.  In particular, a number of you argued that Toggle had the ability to control the infringements because they could just stop the service.  At minimum, this is circular/tautological, and I think it’s an exaggeration of current copyright law.

  • Copying URLs.  Unless they are aggregated into a protectable databases, URLs on their own likely don’t qualify for copyright protection.  (Tickets.com)
  • Deep Linking.  Deep linking isn’t a cause of action by itself, and it’s hard to articulate why deep linking is any more infringing than regular linking.
  • Toggle’s Display of PDFs.  Toggle is not displaying the PDFs itself.  I tried to make the facts clear that Toggle also isn’t displaying the PDFs as Ditto.com was, but I think the facts confused some of you anyway.
  • 1202.  There was no facts suggesting CMI omissions, and the Kelly court found there wasn’t a CMI violation in its case.  Indeed, given that Toggle takes the entire PDF, which is usually intended to be a self-contained file (unlike a graphic, which can go either way), there’s even less likelihood of a CMI issue.  The only CMI issue that could arise is with the search results listing, where only a portion of the work is displayed.  If this is considered an infringing copy or derivative work, it may also exclude some CMI that was in the main file.  Of course, this type of CMI violation will occur in many types of derivative works, and it’s not clear the law will be read that broadly.


The Campbell test says transformation occurs when the new work “adds something new, with a further purpose or different character, altering the first with new expression, meaning or message.”  The Kelly court omitted the requirement of something new being added, meaning that transformation occurs just if you can show a different purpose.

It’s important to distinguish the two tests, because they may lead to different conclusions.  If you apply the Campbell test, not much new was added.  True, Toggle adds the highlights and puts the data in a new format, but these are not really substantive additions.  So under the Campbell test, the conversion of PDF to HTML is probably not transformative because it lacks something new.

Applying the Kelly test, one could argue that the HTML files serve a different purpose.  Specifically, the conversion is an integral step in allowing PDF files to be indexed in the search engine, this serving “search engine purposes” much like the thumbnails in Ditto.com served that purpose.  On the other hand, one could argue that this is just a different form of redistribution and therefore it’s not a different purpose.  In our hearts, I suspect most of us feel that search engine behavior like Toggle or Ditto.com is not really transformative but we may want to protect it and encourage it nonetheless because of its socially beneficial effects.  This, of course, puts judges in the awkward position of following the law or following common sense, an unfortunately all-too-common tension in copyright law.

Those of you who failed to answer this part of the question lost points.