Gonzales v. Google Validates My Theory of Legislative Drafting — Be Really, Really Detailed and Longwinded.

Every now and then, I do some legislative drafting. I tend to get pushback on my habit of including a bunch of legislative findings and statements of policy, and what some consider my over-detailed definitions. The usual challenge I get is “everyone knows what we’re talking about.” My response: “I’m not writing for us. I’m writing for some judge 25 years from now with no idea what we’re talking about or trying to do.”


Which brings me to Gonzales v. Google, the Supreme Court case in which the Justices will take their first shot at interpreting Section 230 of the Communications Act. Distill down the thousands of pages of briefs, brush away the policy arguments, and it all boils down to one question: “What did Congress actually mean when it said don’t treat online services as the ‘publisher of speaker’ of third-party content”? Does it mean (a) the plain English ‘don’t treat the provider of the online service as if that provider actually said the thing’ – so you can’t sue a provider of an “interactive computer service” (to use the actual statutory term) for anything relating to third party content; or (b) does it mean ‘this section provides only protection from liability as a ‘publisher’ under the common law’ – but feel free to impose liability as a common law distributor of third party content (or possibly for any other kind of liability outside the rather narrow common law universe of defamation)?


Because this question comes a lot, and because I expect lots of folks to follow the Gonzales case, I decided to run through the type of analysis courts typically engage in when trying to interpret what Congress meant and why courts can come up with wildly divergent explanations.


Yes, policy issues and outcome determination matter. But good judges at least try to figure these things out, and even bad judges (by which I mean those determined to reach a specific outcome no matter what the statute actually says) need to couch their opinions in the form of legislative analysis. This is why lawyers and scholars spend so much time on the subject.


So if you want to understand how this game works to follow the arguments in Gonzales v. Google, see below. Along the way, I’ll highlight how W. VA v. EPA may complicate things with its stupid ‘let’s look at what Congress didn’t pass’ analysis.



I’ll start by making three points here. First, I’m not trying to predict an outcome. I personally think it makes better sense to read the statute as plain English rather than impute the common law definition of “publisher or speaker,” but that’s just me. As I explain below, this isn’t a case where I roll my eyes and say ‘how could anyone take this argument seriously?’ That’s what makes this a fun case and well worth analyzing.


Second, there is an interesting theory championed by Adam Candueb that Congress really intended to replicate traditional common law immunity. You can see his brief here. Enough folks have latched on to the “big tech platforms are/should be common carriers” argument that this might play spoiler. But it is something of an outlier so I’m not going to discuss it here. (But feel free to say in the comments if you want to see me tackle why – despite my love of common carrier – I don’t agree with this approach.)


My third caveat is that my employer Public Knowledge] submitted a brief arguing ‘even under a common law definition of “publisher or speaker,” Google’s recommendation system still falls under the definition of publishing, so Google should win either way.’ And, as Google argues in its own brief, this whole case should be moot if the Court decides in the companion case Taamneh v. Twiiter that you can’t hold a social media platform liable under the Antiterrorism Act for the conduct at issue here. So there are ways the Court could decide this without actually reaching the core statutory interpretation question (including a judgment against Google but no majority for an opinion).


With that out of the way, let’s begin.


Step 1: What Does the Statute Actually Say and Why Would I Think It Doesn’t Mean What it Says?


If you never had to engage in an exercise in statutory interpretation for a court or a federal agency, you probably think “Dude, how did common law get into this?” Sure, the common law has its own specialized definitions of stuff developed in specialized contexts over the course of 500 years or so. That’s why Black’s Law Dictionary exists and why you spend a stupidly ridiculous amount of money to go to law school. But shouldn’t we presume that Congress meant the usual plain English stuff? It’s not like Congress forgot to define other terms like “interactive computer service” or even “Internet.” If Congress wanted to use the more limited common law meaning, wouldn’t it have actually said so? That’s what the first major Court of Appeals decision held almost immediately after Congress passed Section 230, and what every subsequent Court of Appeals decision has held ever since.


Except that we know very clearly that the triggering event for getting Section 230 written and passed was the (in)famous Stratton Oakmont, Inc. v. Prodigy case, where a court found that ye olde ISP Prodigy should be treated as a publisher for defamation purposes since they held themselves out as trying to exercise some editorial control over the content. So, the argument goes, clearly Congress intended Section 230(c)(1) to address the rather narrow issue of publisher liability because “everyone knows that’s what we were talking about.” A walk through the legislative history can find a bunch of proof to support this view. Under this argument, for the past 25 years everyone has gotten it wrong. Therefore you – oh wise Supreme Court Justices – need to set everybody straight.


If that seems odd, well it happens. Another boost that this is how the Supreme Court is heading is that they actually took the case in the first place. The Supreme Court usually only hears these kinds of cases when circuit courts come up with conflicting interpretations, and courts have been remarkably consistent in this interpretation. So it’s a bit odd that SCOTUS wants to reexamine fairly unanimous holdings below. But people have been increasingly unhappy with the fairly broad grant of immunity courts have provided “interactive computer service providers” (ICSPs – the technical name in the statute for online platforms). We’ve had a few dissents (including one biggie from Justice Thomas in Malware Bytes v. Enigma Software dissenting from denial of cert) as well as some law review articles questioning the reasoning of the conventional judicial wisdom. And since just about every circuit decided relatively early on to follow the “plain language” interpretation rather than the “common law” interpretation, there isn’t likely to be a circuit split.


The Counter Argument: Congress Meant What It Said and Said What It Meant.


So does the fact that SCOTUS took the case mean it will ‘set things straight’ and adopt the common law meaning of “publisher and speaker.” Not necessarily. We also have a bunch of evidence – including a brief from the actual authors of Section 230 – that Congress did actually intend to give websites and other ICSPs broad immunity as incentive and protection to “curate content.” But if Prodigy dealt with defamation, why would Congress go further and grant broad immunity?


Because even though the Prodigy case dealt with defamation, it provided a signal of the kind of problems internet companies that hosted third-party content would encounter. Members of Congress who saw huge potential for the Internet to revolutionize commerce and make the universe generally more interactive (which in 1996 those people actually thinking about it saw as a positive – yeah, I was there and we were naïve, but I digress). And, as someone who was around at the time, 1996 was when everyone was deciding we hated liability and regulation for companies and wanted to unleash the magic of the private sector blah blah get government out of the way blah blah limit class action liability because lawyers greedy something about McDonalds hot coffee. A mere two years previous, Congress had seriously undermined the ability of the FTC to protect consumers from unfair practices with the FTC Amendments Act of 1994.


So it’s perfectly reasonable to believe, as courts have interpreted for the last 25+ years, that Congress intended to grant ICSPs very broad immunity for what other people said on their websites (or through their other services). It’s basically what Congress was into back then. Which brings me to the the “Congress could not have foreseen” argument.


“Congress Could Not Have Foreseen” The Thing I Don’t Like.


The basic argument sounds more in policy, even though we pretend it’s an idea of statutory interpretation. The idea here is that in 1996, Congress had no idea about services like YouTube, Facebook, or social media generally. Congress could not have predicted how incredibly important to, and intrusive in, our lives these “interactive computer services” would become. Had Congress but known, they never would have drafted such a broad grant of immunity.


The problem with this argument is that it basically says Congress only drafts laws that deal with a specific problem rather than trying to draft laws that deal with a type of problem. Whether or not Congress actually could predict specific things like Twitter and Twitch, it is entirely possible (I would argue likely) that Congress wanted to see things evolved and wanted them to evolve without services getting sued over 3rd party content.  We see support for this (IMO) in the findings and policy sections of Sections 230. Section 230(a), the findings, are all forward looking and describe a world evolving rapidly and unpredictably, but with a lot of potential for good stuff if it gets the chance to evolve. Thus Congress finds that “The rapidly developing array of Internet and other interactive computer services “ and about their “potential” to evolve further. Section 230(b), the policy section, basically talks about letting these interactive services continue to evolve going forward.


In other words, Congress in the statute freely acknowledged that it had no idea where thigs would go. That was rather the point of creating the immunity of Section 230(c). And yes, as many have pointed out, Section 230 was part of a larger amendment to the Telecommunications Act called the Communications Decency Act, which was also about requiring filtering of indecent content to keep smut away from innocent youth. Hence the provisions about providing parents with tools to better control what content they actually got (an approach used elsewhere in the 1996 Telecom Act for television, which gave us the TV ratings system). But there is no contradiction between a Congressional compromise that granted broad immunity to interactive computer services for everything in exchange for specific provisions designed to require filtering filth-flarn-filth because someone must think of the children. Or, to quote Hamilton the Musical: “No one really knows the game is played, the art of the trade, how the sausage gets made. We just assume that it happens.”


Which brings us back to next week’s Supreme Court argument.


The W. VA v. EPA Wild Card.


One of the more annoying points in West Virginia v. EPA had to do with interpreting Congressional intent by what Congress didn’t pass. Generally, as SCOTUS has said in other places, this is a bad move. Who the Hell knows why a bill does or doesn’t pass, let alone what that tells us about what an entirely different Congress meant? But in W.VA v. EPA, slip op pages 27-28, the Court buttresses its conclusion that Congress didn’t intend to give EPA the power to create a “cap-and-trade” system in carbon emissions because Congress had repeatedly failed to pass a cap-and-trade system. Mind you, as pointed out in US v. Southwestern Cable, Congress might have decided not to pass legislation because it felt the agency already had the authority. But whatevs.


Over the last several years, we have seen lots of attempts to amend Section 230 to limit the available immunity. The only one that passed Congress was FOSTA/SESTA, which was explicitly designated in the statute as a “clarification” that Section 230 was not meant to protect sex-trafficking. What, then, does that tell us about the meaning of Section 230? Does it mean Congress has endorsed the unanimous opinion of courts over the last 25 years that “speaker or publisher” means normal English, rather than the specialized meaning from defamation common law? Or does it mean that Congress wanted the courts to keep playing with it? Should we read anything into it at all?


Frankly, this is why I think the W.VA. v. EPA approach to introduced-but-not-passed legislation is pretty dumb. Members can always introduce a bill for whatever reason. Sometimes it is simply a “marker” bill, an effort to send a signal or start a debate. A lot of times something will take multiple sessions of Congress to pass. The Telecommunications Act of 1996 started making the rounds and getting hearings as early as 1989. Over the way it had lots of different hearings and markups before finally settling into the form Congress passed almost exactly 27 years before the Gonzales v. Google oral argument. Can the debate around things like the PACT Act or EARN IT actual inform what a bill written 27 years ago means? At best, it tells us that members of the current Congress may be unhappy with how courts have interpreted Section 230. But it is a leap to say that the failure of Congress to pass any amendment to Section 230 clarifying that “we really meant ‘speaker or publisher’ as used in defamation common law, not as you would mean it in plain English” sheds any light on what the actual drafters of the legislation actually intended.


But that’s what W. VA v. EPA seems to suggest. So maybe the Court will back off that part of W.VA v. EPA, or at least clarify it some. (I expect that if there is a majority to limit Section 230, the majority will just ignore it.)




I don’t expect others to care about the underlying question of “how do we interpret statutes.” But it comes up an awful lot, and it’s damn hard to anticipate every which way a court will think up on how to read things. Hopefully, this summary will help folks follow the argument and, ultimately, the opinion. But if it also inspires legislative drafters to greater clarity, even if it makes the statute somewhat longer and wordier, that would also be good.


Stay tuned . . . .

One Comment

  1. Nice analysis. Though I’d suggest that broad immunity is a good thing – and I hope that SCOTUS maintains the prevailing view. Allowing private publishers to arbitrarily censor material – in the name of protecting people, would truly be a slippery slope toward REQUIRING such censorship, with the resulting chilling effect on speech. (Something I worry about a LOT, given that I run a bunch of public policy forums of various sorts – and, so far, have imposed a “free speech for adults” policy – both as a matter of philosophy, and to maintain immunity from legal action.)

Comments are closed