Prof. Joseph LiuTrademark Seminar (Fall 2005)Yu Lu

Review Standard
for Likelihood of Confusion

I.Introduction...... page 1

II.The Likelihood of Confusion Tests
and Review Standards...... page xxx

A.The Likelihood of Confusion Tests...... page xxx

B.Review Standards and Rationale...... page xxx

III.Non-Traditional Infringement Use...... page xxx

A.Dilution (FTDA)...... page xxx

B.Anticybersquatting Consumer
Protection Act (ACPA)...... page xxx

C.Internet Domain Arbitration...... page xxx

D.First Amendment...... page xxx

IV.Analysis of the Issue and A ProposedAnswer...... page xxx

A.Question of Fact or Question of Law?...... page xxx

B.Two Ways to View Likelihood of Confusion...... page xxx

C.Comparisons to Copyright and Patent...... page xxx

D.Does the same analysis fits the other cases?..page xxx

V.Conclusion...... page xxx

Review Standardfor Likelihood of Confusion

Part I:Introduction

Federal laws should be applied consistently in all federal circuits, with a unified standard of review (to discourage forum shopping, predictability, etc.). While this being the ideal, many times, ideal remains an ideal. For example, in the case of traditional federal trademark infringement, confusions frequently arise surrounding the standard of review for the likelihood of confusion test. This could become even more probmatic, in view of the growing trend of protecting trademark as “property” in recent trademark practice and law. Questions remain as to what should be the proper review standard, and whether the same standard applies in all cases outside the realm of traditional trademark infringement, when the courts also use a likelihood of confusion-type test.The answers may depend on where and how this test is used.

Part II below summarizes the likelihood of confusion test, and the different standards of review adopted by the thirteen federal circuits, to illustrate the unsettled issue of review standards for likelihood of confusion. Part III introduces the relatively new problem of using likelihood of confusion-type tests in non-traditional trademark infringement cases, and asks the question of whether the same review standards should be applied in these cases. Part IV analyzes the issue and provides a proposed answer, by first discussing the two ways to view the likelihood of confusing issue, followed by how the author thinks the standard of review issue could be resolved thereunder, in both traditional trademark infringement and non-traditional uses. Analogies to related / similar issues in other intellectual property laws are drawn to aid the discussion. Part V concludes with a summary.

Part II:The Likelihood of Confusion Test and Review Standards

Different federal circuits adopted largely similar Likelihood of Confusion tests, but apply very different review standards, and the U.S. Supreme Court has chosen to remain silent (or did it?).

  • Test for traditional federal trademark infringement – Likelihood of confusion (LoC) (Lanham Act)
  • Different federal circuits have developed largely similar “LoC tests” with slightvariations (from 1960s to mid-1980s)
  • But different federal circuits adopt quite different standards of review for district court decisions (from mid-1970s to mid-1980s)
    – most treat LoC as question of fact (QoF), reviewed deferentially under the clearly erroneous standard
    - minority treat LoC as question of law (QoL) or mixed question of fact and law (QoF&L), reviewed de novo at the appeal court level
    - the U.S. Supreme Court has repeatedly denied cert.
    - although a couple of S. Ct.casesand the 1985 Amendment to FRCP 52(a) help to sway the most federal circuits to adopt the clearly erroneous review standard
  • Unresolved question: what should be the review standard?

Part III:Non-traditional Use of the Likelihood of Confusion Test

Recent trademark practice and trademark law have seen a gradual expansion from the trademark core value of consumer protection, towards protecting trademark as “property.” Along the process, some courts have used the likelihood of confusion-type test to assess the similarity of marks. This presents the question of whether the same review standard also applies in these cases.

  • Federal Trademark Dilution Act (1995): not based on protecting consumers from confusion, but on protecting famous-mark holders from a possible diminution in the value of their marks
    - no need to prove LoC
    - protection not limited to identical marks, also similar marks.
    - some courts used LoC type of test / analysis to test similarity
    - similarity in dilution likely more stringent (if no LoC, no dilution)
  • ACPA (Nov. 1999):
    - traditional TM infringement and dilution not a good fit for this type of cybersquatting activity
    - some courts use LoC test to see if confusingly similar
    - although 2nd circuit held that Wella’s test is for confusingly similar
  • Domain Name Dispute (2000):
    - WIPO case used LoC test to decide mark similarity
  • First Amendment:
    - the Rogersbalancing test – LoC and First Amendment interest
  • Most of these new developments occur after the different federal circuits have settled in their respective review standards
    - does these review standards still apply in these new settings outside the realhm of traditional trademark infringement?

Part IV:Analysis of the Issue and A Proposed Answer

A.Question of Fact or Question of Law?

  • QoF and QoL
    - QoF: historical fact (what happened), predictive facts (what will happen), etc.
    - QoL: pure legal question
  • Mixed Law and Fact
    - the analysis framework

B.Two Ways to View Likelihood of Confusion

  • LoC: on the face of it, seems like a predictive fact – decision-maker is asked to predict, based on a give set of circumstances, whether it is likely that consumers will be confused?
  • LoC: but it can also be casted differently as a judgment call, where the decision-maker is called upon to decide whether allowing the use of the accused mark so similar to another would violate the policies of trademark law.
  • casted this way, LoC does not askthe decision-maker to prediction what will likely happen (but is merely a policy choice of how closely two marks can be without violating TM policies)
  • S.Ct held that this type of policy decision should be made by the decision-maker (judge or jury) that is better positioned to decide a particular issue.Miller v. Fenton, 474 U.S. 104, 114 (1985) (“The fact/law distinction,...at times has turned on a determination that, as a matter of the sound administration of justice, one judicial actor is in a better position than another to decide the issue in question.")
  • Examples:
    - negligence (jury – community standard)
    - probable cause (judge – need to "unify precedent" and ensure consistent application of a constitutional requirement)
  • back to LoC: is jury or judge better positioned to decide the issue of how close we can allow one mark to another without causing consumer confusion?
    - seems to me that jury is, since consumers are ultimately the ones that may be confused.

C.Comparison to Copyright and Patent

  • compare to copyright law
    -copyright infringement uses the “substantially similar” test to determine whether there is “improperly appropriation.”
    - most federal circuits says this is QoF, since the ultimate assessor of substantial similarity are the consumers of copyrighted work.
    - only 2nd Cir reviews substantial similarity de novo, if it is a bench trial.
  • compare to patent law
    - differentiate TM rights (positive) and patent rights (negative)
    - patentability not co-extensive with freedom to operate, while Lanham Act says no TM registration if LoC
    - more proper comparison is to patentability: the obviousness issue in the eye of person of ordinary skill in the art (QoL based on fact)
    - the policy choice here is unified standard of patent protection (Congression intent in creating CAFC)
  • LoC in trademarkis more like substantial similarity in copyright (both viewed in the eye of consumer)
  • QoF & Clearly Erroneous Standard seem to be the better choice in the traditional TM infringement case in both views.

D.Does the same analysis fits the other cases?

  • Dilution:if used, like traditional LoC
  • ACPA and domain name dispute:
    - if use LoC test, then like traditional LoC.
  • First Amendment:
    - Constitution rights involved, judge seems to be better positioned.

Part V:Conclusion

  • Summary: must first decide in what context the LoC test is used before chosing a standard of review.

Bibliography:

1.Randall H. Warner, All Mixed Up About Mixed Questions, 7 J. App. Prac. & Process 101, 136 (2005).

2.RICHARD L. KIRKPATRICK, Likelihood of Confusion Issues: the Federal Circuit’s Standard of Review. 40 Am. U.L. Rev. 1221, at note 3 (1991).

3.Pullman-Standard v. Swint, 456 U.S. 273, 287 (1982).

4.Scandia Down Corp. v. Euroquilt, Inc., 772 F.2d 1423, 1429 (7th Cir. 1985).

5.J. Thomas McCarthy, 4 McCarthy on Trademarks and Unfair Competition § 24:90.2 (4th ed.) (Database updated September 2005).

6.Victoria's Cyber Secret Ltd. Partnership v. V Secret Catalogue, Inc., 161 F.Supp.2d 1339, 1351 (S.D.Fla. 2001).

7.Sporty's Farm L.L.C. v. Sportsman's Market, Inc., 202 F.3d 489, 496 (2ndCir.2000).

8. (last visited on October 30, 2005).

9.Diageo plc v. John Zuccarini, Case No. D2000-0996 (WIPO October 22, 2000).

10.Cliffs Notes, Inc. v. Bantam Doubleday Dell Pub. Group, Inc., 886 F.2d 490 (2nd Cir. 1989).

11.Miller v. Fenton, 474 U.S. 104, 114 (1985).

12.JOYCE et al., Copyright Law, sixth Ed., pp.690-92 (2003).

[1]