On Stipulations and Expressiveness

As the reaction to the Supreme Court’s adventure in First Amendment Lochnerism in 303 Creative v. Elenis begins to unfold, I want to make a brief intervention on the issue of stipulated facts. In my view, the majority strips the context of some very banal stipulations by the Colorado defendants, unfairly playing “gotcha” to transform the stipulations into legal concessions and conceal the breadth of its holding.

Stipulations in the Majority Opinion

Ken White lays out in a Bluesky thread an argument for the significant role of Colorado’s stipulations in the 303 Creative majority opinion, authored by Justice Gorsuch:

I think the majority’s best critique of the dissent is that the dissent largely ignores what the parties in this case stipulated was true (especially that the web sites and their custom content are expressive).

Now, it’s not clear to me why Colorado stipulated as easily as it did to the facts it did . . . .

Indeed, the 303 Creative majority makes a big deal about stipulated facts:

The Tenth Circuit held that the wedding websites Ms. Smith seeks to create qualify as “pure speech” under this Court’s precedents. . . . We agree. It is a conclusion that flows directly from the parties’ stipulations.

. . .

We further agree with the Tenth Circuit that the wedding websites Ms. Smith seeks to create involve her speech. . . . Again, the parties’ stipulations lead the way to that conclusion.

According to the majority, this blows a hole in Colorado’s arguments:

Colorado says, this case involves only the sale of an ordinary commercial product and any burden on Ms. Smith’s speech is purely “incidental.” . . . On the State’s telling, then, speech more or less vanishes from the picture—and, with it, any need for First Amendment scrutiny. . . .

[Colorado’s] theory . . . is difficult to square with the parties’ stipulations. . . . As the case comes to us . . . Colorado seeks to compel just the sort of speech that it tacitly concedes lies beyond the reach of its powers.

Chastising the dissent for overblowing the holding, the Court again turns to the stipulations:

[The dissent] reimagines the facts of this case from top to bottom. The dissent claims that Colorado wishes to regulate Ms. Smith’s “conduct,” not her speech. . . . Forget Colorado’s stipulation that Ms. Smith’s activities are “expressive” . . .

Instead of addressing the parties’ stipulations about the case actually before us, the dissent spends much of its time adrift on a sea of hypotheticals . . . .

[T]he moment one acknowledges the parties’ stipulations . . . the First Amendment implications be come obvious. As does the fact that our case is nothing like a typical application of a public accommodations law requiring an ordinary, non-expressive business to serve all customers or consider all applicants.

Finally, seeking to cabin the breadth of the holding, the majority again turns to the stipulations:

Doubtless, determining what qualifies as expressive activity protected by the First Amendment can sometimes raise difficult questions. But this case presents no complication of that kind. The parties have stipulated that Ms. Smith seeks to engage in expressive activity.

(emphasis original)

Colorado Stipulated to What Now?

These stipulations sound like they must be pretty damning. What exactly did Colorado stipulate to? Here’s how the majority frames the joint stipulations:

They have stipulated that Ms. Smith’s websites promise to contain “images, words, symbols, and other modes of expression.” . . . They have stipulated that every website will be her “original, customized” creation. . . . And they have stipulated that Ms. Smith will create these websites to communicate ideas—namely, to “celebrate and promote the couple’s wedding and unique love story” and to “celebrat[e] and promot[e]” what Ms. Smith understands to be a true marriage.

. . .

As the parties have described it, Ms. Smith intends to “ve[t]” each prospective project to determine whether it is one she is willing to endorse. . . . She will consult with clients to discuss “their unique stories as source material.” . . . And she will produce a final story for each couple using her own words and her own “original artwork.”

Reading this in parallel with the Court’s analysis, the parties supposedly agree here that (a) the designer’s speech is the kind of expression cognizable under the First Amendment, and is (b) attributable to the designer. And we’re in compelled-speech land, so no means-ends analysis or other interminable balancing tests; it’s game over for Colorado.

But wait a minute. Did the litigators at the Colorado AG’s office really just offhandedly concede the heart of a really important case? White posits that maybe “Colorado WANTED to pull the trigger on a test case of the core issue rather than try to avoid it with factual or procedural arguments.” Setting aside what exactly the “core issue” would be if not the First Amendment implications of the plaintiff’s conduct, this notion doesn’t square with Colorado’s brief, which centers on the non-expressiveness of the regulated conduct:

[The underlying law does not] regulate expressive conduct. The mere act of selling something— even something that may contain expressive elements—is not itself expressive conduct. Routine commercial transactions do not become expressive conduct just because the business believes a sale would convey approval of the buyer.

What Really Happened Here?

Taking a look at the stipulations themselves (scroll down to page 173a if you’re playing along at home), these explanations become even less satisfying. I think the majority is playing fast and loose, stripping context from uncontroversial statements about the basics of websites and web design to characterize them as substantial legal concessions that Colorado didn’t make that about the heart of the First Amendment issues in the case.

First, there’s all the stuff about expressiveness:

46. All of Plaintiffs’ graphic designs are expressive in nature, as they contain images, words, symbols, and other modes of expression that Plaintiffs use to communicate a particular message.

47. All of Plaintiffs’ website designs are expressive in nature, as they contain images, words, symbols, and other modes of expression that Plaintiffs use to communicate a particular message.

. . .

81. Plaintiffs’ custom wedding websites will be expressive in nature, using text, graphics, and in some cases videos to celebrate and promote the couple’s wedding and unique love story.

82. All of these expressive elements will be customized and tailored to the individual couple and their unique love story.

In context, these are banal facts about the nature of wedding websites. 46 and 47 are basically just saying that graphic designs and websites are lower-case-e “expressive,” in the sense that images, words, and symbols express meaning. And 81 and 82 are likewise just noting that wedding websites generally use words, images, and videos to communicate something about the couple getting married.

🤷🏻‍♂️ (Does WordPress support the shrugging guy emoji? In case it doesn’t, just imagine me shrugging.) None of this has anything to do with whether a web designer is doing anything “expressive” in the sense of “speech” under the First Amendment. Dale Carpenter offers a sophisticated analysis of First Amendment “expressiveness” at the Volokh Conspiracy, but I think we’re talking about something much more basic here. As Merriam-Webster’s first definition puts it, all these stipulations do is establish that websites are a mode of “expression” qua “an act, process, or instance of representing in a medium (such as words).”

In other words, Colorado is just politely stipulating the obvious: websites contain words and pictures. (Hindsight being 20/20, I probably wouldn’t have stipulated to the word “expressive” given its overloaded connection to the First Amendment, but it’s clear from the context that isn’t what Colorado meant.)

Next, there are the stipulations purportedly about the website’s “expressions” being attributable to the web designer, as opposed to her clients:

50. Each website 303 Creative designs and creates is an original, customized creation for each client.

. . .

53. Every aspect of the websites and graphics Plaintiffs design contributes to the overall messages that Plaintiffs convey through the websites and graphics and the efficacy of those messages.

. . .

83. Viewers of the wedding websites will know that the websites are Plaintiffs’ original artwork because all of the wedding websites will say “Designed by 303Creative.com.”

(there are more along these lines, but just sticking to the ones that the majority focuses on)

Again, in context, these are just banal facts about web design. Yes, web design involves customizing words and images at your client’s behest. Yes, websites express messages. Yes, websites often credit the designer. (Note, though, that the literal attribution here is at the designer’s behest; if you don’t want to put your imprimatur on your client’s message, you could just not put your name on it!)

And again, Colorado is just politely stipulating the obvious: web designers design websites. None of this adds up to a concession from Colorado that the message of the hypothetical websites at issue here is actually attributable to the designer in a First-Amendment-cognizable way.

The Non-Stipulated Future of Anti-Discrimination Law

So where does this get us? None of the foregoing addresses the substance of the broader First Amendment questions that swirl around this case. But the majority’s emphasis on the supposed significance of these stipulations is troubling. I see at least a couple of possibilities here:

  1. The majority is disingenuously mischaracterizing the stipulations as legal concessions so it can avoid articulating rules around the boundaries of expressiveness and attributability, thereby opening the door for litigants to flood the courts with a mess of new fact patterns challenging anti-discrimination laws under the First Amendment;
  2. The majority is implying that the banal facts underlying the stipulations that Colorado made actually add up to expressive, attributable speech, thereby suggesting that anyone in a vocation nominally involving activity on behalf of a client that nominally involves lower-case-e expressiveness (like, words) can discriminate under the protection of the First Amendment.

Either possibility (and why not both?) suggests that the holding could have a wide impact highly inconsistent with the majority’s superficially narrow framing.

Coda #1

I guess there’s a third possibility, which is that the majority really thought that Colorado inadvertently conceded the whole case with these stipulations and just wanted to reach a narrow holding with little or no impact on future cases. But does this really warrant a 26-page opinion closing with a dramatic quotation from George Orwell and waxing about the “opportunity to think for ourselves and express those thoughts freely?” Again: 🤷🏻‍♂️

Coda #2

Raffi Melkonian offers some thoughtful rejoinders on Twitter. Some of the discussion gets into speculation about Colorado’s tactics about which I don’t have much to say. But I’ll add a few thoughts on the rest of the discussion, which helps highlight the confusion this case creates between terminology around the use of “speech” and “expression” as factual descriptors on the one hand and terms of varying legal significance on the other.

First, Raffi notes:

Justice Sotomayor does not say so. In fact, she *agrees* the speech is expressive – she just thinks it doesn’t matter. See slip op. at 29 (“It would”).

I’m not sure I agree. Justice Sotomayor’s next sentence—”These questions do not resolve the First Amendment inquiry . . . .”—suggests that she is not invoking the term “speech” to reach a legal conclusion. (Note that the word “expression” doesn’t appear until later on in the paragraph.) And though I feel like I’m chasing my tail a little bit here, remember where we started: the majority opinion specifically criticizes the dissent for misunderstanding the significance of the stipulations, characterizing the dissent as erroneously “claim[ing] that Colorado wishes to regulate Ms. Smith’s ‘conduct,’ not her speech” and “[f]orget[ting] Colorado’s stipulation that Ms. Smith’s activities are ‘expressive.'” If Justice Sotomayor agrees with the majority opinion about the significance of the stipulations, the majority doesn’t seem to think so.

Relatedly, Raffi notes:

If in fact it is true that Colorado was just conceding that “websites contain words and pictures” – and was not conceding that those words and pictures were expressive in the First Amendment sense, then why did they not dispute this latter point?

To the contrary …

When the district judge specifically asked them whether he should have an evidentiary hearing, *Colorado* said no.

While I agree with Raffi that Colorado stipulated the fact of lower-case-e “expression” (in the “words and pictures” sense), I’m not sure why it would have made sense to contest the legal significance of that fact at an evidentiary hearing. Indeed, the screenshotted blurb attached to Raffi’s thread highlights Colorado saying merely that it doesn’t “believe that any facts are in dispute in this matter” (emphasis added).

Relatedly: Colorado’s response to the plaintiffs’ summary judgement motion at the trial court, filed right after the joint statement of stipulated facts, does full-heartedly dispute the legal point. The response includes a dedicated subsection on the expression issue, arguing that the law did not require the plaintiffs to convey “a message in support of same-sex marriage or any message.” The next subsection addresses the attribution issue, arguing that “to the extent any message is conveyed at all, reasonable observers would attribute that message to the individuals being married, not Plaintiffs.” Again, without weighing in on the merits of these arguments, it seems to me that Colorado indeed disputed that the “words and pictures” at issue were “expressive in the First Amendment sense.”