Virginia Appellate Court Rejects Dendrite / Cahill Approach to Online Anonymity

by Paul Alan Levy

In a decision issued yesterday morning, the Virginia Court of Appeals parted company with appellate decisions in eleven other states and held that the First Amendment allows a court to compel the identification of a company’s anonymous online critics even though the company has done no more than claim that it “suspects” that the statements were false, and then represent that the suspicion is based on an “investigation” of its customer database.  

Until recently, there was broad unanimity among state courts, following the analysis of New Jersey’s Appellate Division in Dendrite International v. Doe, or the analysis of the Delaware Supreme Court in Doe v. Cahill, about the substance of the analysis governing a would-be plaintiff that hoped to identify online critics so that it could sue them for defamation, or for other tortious speech.  These states recognize that the First Amendment protects the right to speak anonymously, that overcoming such First Amendment right requires a compelling state interest, and whether as a matter of First Amendment obligation or as a matter of state procedural rules, that the interests favoring enforcement of the subpoena cannot prevail unless the court is presented with evidence supporting the allegation that the speech was tortious or otherwise illegal.

Coupled with the somewhat conflicting results recently reached by two different panels of the Michigan Court of Appeals over the past few months (in the Thomas Cooley Law School and Ghanam cases), that unanimity no longer exists, although the majority rule still requires an evidentiary showing of merit before a Doe can be identified.

The Hadeed Court’s Decision

The appeal arose from a lawsuit filed by Hadeed Carpet Cleaning, alleging that seven Yelp reviews defamed it by accusing Hadeed of bait and switch tactics, advertising low prices for certain routine services which, when its staff arrived to clean the carpets, the range of reviews indicated were consistently determined not to be available.  In suing for defamation, Hadeed did not assert that the charges of bait and switch were false.  Instead, it claimed that it had conducted a review of its database of customers and been unable to identify customers whose names and cities matched the pseudonyms used for the Yelp reviews.  Consequently, it argued, it had a reason for believing that the very assertion that the reviewers were customers was false, and that, consequently, everything else in the reviews must have been false at least in the sense that the wrongs were not perpetrated on that specific reviewer.  Hadeed then issued a subpoena to Yelp, which objected to the subpoena.  A trial judge in Alexandria enforced the subpoena, and Yelp obtained appellate review by disobeying the enforcement order and being held in contempt.

In assessing the standard to be applied by Virginia courts in deciding whether to enforce subpoenas, the Court of Appeals focused on Section 8.01-407.1 of the Virginia Code, in which the Virginia legislature, acting well before the broad consensus developed in other states around the Dendrite / Cahill approach, adopted a standard applied by a Virginia trial judge in a case involving a subpoena to America Online in 2000.  The statute imposes excellent notice procedures to ensure that both the party being subpoenaed, and the anonymous speaker whose identity is sought, are informed about the reasons for the subpoena and have an opportunity to object to compelled disclosure.  Unfortunately, the excellent procedures are compromised by an ambiguous disclosure standard that could be construed as incorporating the Dendrite and Cahill approach, by requiring evidence that the speech is tortious, but can also be construed as weaker  – it requires a showing that the anonymous communications “are or may be tortious or illegal” or that the party requesting the subpoena has a “legitimate, good faith basis to contend” that it is the victim of actionable conduct.  “Are or may be” and “legitimate . . . basis” could be understood to mean that the plaintiff has evidence of wrongdoing, but the Hadeed panel's majority chose to impose a lower standard that apparently focused on the “good faith” terminology and, accordingly, allowed discovery based on no more than the plaintiff’s unsworn assertions about its “investigation” into whether the reviewers were customers.

A short dissenting opinion from Judge Haley makes clear the weakness of the majority’s requirements.  He accepted the majority’s decision to apply the Virginia statute rather than the full Dendrite or Cahill approach requiring evidence, but concluded that Hadeed has still not shown enough reason to believe that the reviews were false, considering the constitutional values at stake.  He noted that, during oral argument below and in the Court of Appeals, Hadeed’s counsel had admitted that it really didn’t know whether the Doe reviewers were customers, but only that “we suspect not,” and that it would not learn the truth until it identified them.  
    "This, I suggest, is a self-serving argument – one that proceeds from a premise the argument is supposed to prove. If Hadeed were an individual, he would be attempting to ‘lift himself by his own bootstraps.’ . . . A business subject to critical commentary, commentary here not even claimed to be false in substance, should not be permitted to force the disclosure of the identity of anonymous commentators simply by alleging that those commentators may not be customers because they cannot identify them in their database.”

Some Problems with the Virginia Court’s Reasoning

The Court of Appeals did not expressly address the “compelling state interest” analysis or explain how the mere filing of a complaint creates a compelling state interest in removing anonymity.  The Court of Appeals based its conclusions on two reasons that seem to me wrong — first, that it could not apply a different approach without finding the Virginia statute unconstitutional, and second, that it could not  follow a constitutional analysis that goes further than the statute without rejecting what it assumed to have been the legislative policy choice to reject the approach taken by courts in several other states, such as the New Jersey Appellate Division in Dendrite.  

The court’s error here was two-fold.  First, it is not at all uncommon for states to have statutes that co-exist with constitutional protections and hence provide alternate means for parties to assert their rights, as, for example, when a court considers a subpoena seeking to identity a reporter’s sources or outtakes, and the court decides first whether the statutory shield applies, and then decides whether the constitutional shield applies.  Court do not feel obligated to declare state shield laws unconstitutional just because the First Amendment provides more protection.

Second, apart from the fact that legislative “policy choices” cannot override the First Amendment, the court apparently rested its conclusion that the Legislature had considered, but rejected, Dendrite’s requirement of presentation of evidence, on a report from the Executive Secretary of the Virginia Supreme Court which, according to the court, cited the appellate decision in Dendrite as one of the outstanding legal standards before going on to recommend the language eventually adopted by the legislature. The report is available online, and it does not support this reasoning. 

To be sure, a footnote in the report mentions both the trial court decision in Dendrite and the appellate affirmance of that decision (at pages 26-27), but it misstates the New Jersey appellate court as having endorsed the Virginia trial court’s approach of requiring even less than a “full motion to dismiss standard.”  In fact, far from reporting a range of "factors that other states use in their unmasking standards," as the Hadeed majority opinion characterized it, the report commented on the paucity of authority from other jurisdiction, noting "only two 'tests' have been reported," neither of them involvking an evidence requirement; the text at page 24 of the report states, "no state or federal appellate court has yet endorsed a particular formulation of the level of scrutiny or balancing test to be applied."

Thus, the existence of this report, which did not tell the legislature what Dendrite actualy decided, and did not mention a possible test requiring evidence of the elements of a prima facie case as one possible way to decide such cases, provides no basis for concluding that the Virginia legislature made a policy choice to reject Dendrite’s evidence requirement as Virginia’s standard for deciding whether to enforce such subpoenas.

Whither From Here

Yelp has indicated that it is disappointed with the ruling because of the lack of privacy that it affords website users, the chilling effect that it may have on free speech in Virginia, and its potential to promote meritless litigation to silence critics without supporting evidence.  Yelp  is considering whether to seek further review from the Virginia Supreme Court.  Note that although Scott Michelman and I represented Yelp in this case, this post is my independent analysis as a public interest litigator, and does not speak for Yelp.