The Journal of Things We Like (Lots)
Select Page
Tim Samples, Katherine Ireland, and Caroline Kraczon, TL;DR: The Law and Linguistics of Social Platform Terms-of-Use, __ Berkeley Tech. L. J. __ (forthcoming 2023), available at SSRN.

Much has been written about ubiquitous online terms of service or terms of use (TOUs). But, as Samples, et. al. write in their forthcoming article, TL;DR: The Law and Linguistics of Social Platform Terms-of-Use, TOUs are poorly understood. Their interdisciplinary study examines the “law and linguistics” of 196 agreements for 75 smartphone-based social platforms. Most other studies of TOUs have a law and economics vantage point, but their study combines legal analysis with “natural language processing, data science, and corpus linguistics.” (P.5.) Corpus linguistics is the “scientific study of naturally-occurring language in the aggregate, often in large datasets, so-called corpa.” (P. 4.) All this means that their study focuses on what matters when thinking about consumer contracts: the language and readability of contracts.

The article begins with a summary of what most contract scholars know by now (nobody reads adhesive form contracts), and then proceeds with an overview of the law in this area. The authors note that social platform TOUs have characteristics that make them particularly problematic. First, they operate on an unprecedented scale. The largest platforms, such as Facebook, affect billions of users.

Second, the platforms have attention surveillance business models and are “positioned to harvest particularly intimate and sensitive data about their users.” (P. 14.) The attention-surveillance business model “generates problematic incentives” and tempts platforms to “deploy addictive interfaces (also known as ‘dark patterns’) to maximize user engagement.” (P. 17.) These addictive designs combined with “extraordinary scale and data sensitivities…add another layer of differentiation between general consumer contracting and social platform TOUs.” (P. 17.)

Third, because these platforms “mediate almost every aspect of modern human life,” they are hard to resist and using them becomes “almost inevitable.” (P. 18.) The authors note that “(t)heories of rational behavior falter in these conditions. Because the top social platforms play such an essential role in everyday life, users are hardly facing a real choice when they click the ‘I agree’ button.” (P. 18.)

Finally, the “ownership and management of data is elemental to governance” and “TOUs play a significant role in digital governance, especially in jurisdictions that have weaker data and consumer protection laws.” (P. 19.) They note that these platforms take a quasi-governmental role and “play outsized roles in shaping privacy and speech rights at the global scale.” The TOUs are “central in defining the relationships between technology and society” and through creating “governance frameworks for the users of digital platforms, TOUs shape basic human rights such as privacy, personal security, and political participation.” (P. 19.) For example, they note, when the former President Trump sued Twitter for removing him from the platform, a judge looked to Twitter’s TOU to decide whether to grant the motion to transfer. (P. 19.)

Having persuasively summarized what makes platform TOUs unique, the authors then assess the readability, linguistic complexity, and key tendencies of social platform TOUs. (P. 20.) They use several metrics, including the Flesch Reading Ease test and the Flesch-Kincaid test, that measure various factors, including text length and paragraph size, as well as lexical and syntactic structures and complexity. (In the article, they provide a comprehensive discussion of the datasets they used and the data processing methods for those who want more information).

The Flesch Reading Ease scores indicate that TOUs are “incomprehensible to most consumers.” (P. 32.)  The authors note that arbitration clauses are particularly unreadable. The Flesch-Kincaid score similarly shows the unreadability of TOUs. Both scores indicate that at least some undergraduate coursework is required to understand the average TOU.

They compare their dataset with earlier studies conducted by Michael Rustad and Thomas Koening and Uri Benoliel and Shmuel Becher. These prior studies also found TOUs unreadable although not to the same extent which the authors interpret to mean that TOUs are getting worse, not better. (P. 33.) As for syntactic complexity, TOUs across the board exhibited a “high degree of overall difficulty and lack of reading ease,” and arbitration clauses were particularly unwieldy. (Pp. 34-35.)

Finally, except for the platform Telegram, which was unusually laconic, the TOUs were quite wordy.  Venmo’s TOU had 20,505 words, which the authors note, is the length of a law review article. They observe that the benefits of such wordiness are one-sided because platforms “incur very little cost in adding terms to online contracts” while “snowballing length poses enormous transaction and opportunity costs for consumers.” (P. 40.) To make matters worse for consumers, these documents are typically read (or supposed to be read) on mobile phones, an impracticable if not impossible feat.

These unread clauses do not simply recite harmless company policies. They deprive adherents of valuable rights. Samples et. al. note that 64% of the TOUs (48 of 75) contain arbitration clauses (P. 42), and nearly all of those contain class or collective action waivers. Virtually all of them (71 of 75) contain unilateral modification clauses which platforms take advantage of by modifying their TOUs frequently.

The authors conclude that there is a “yawning gap” (P. 49) between classic contract doctrine and modern contracting and that “contract law remains static” (P. 49) despite profound marketplace changes. I disagree; I think that contract doctrine has (d)evolved dramatically, at least as applied to TOUs. Traditional contract law doctrines require more in the way of contract formation than what is required under the deviant strain of wrap contract law unleashed by ProCD v. Zeidenberg and its progeny. Samples et. al’s study does an excellent job of demonstrating just how far the law of TOUs is from the classic conceptions of mutual assent, intent, and reasonable expectations that are foundational to contract law.

Download PDF
Cite as: Nancy Kim, Click To Agree That Terms of Use are Incomprehensible, JOTWELL (March 24, 2023) (reviewing Tim Samples, Katherine Ireland, and Caroline Kraczon, TL;DR: The Law and Linguistics of Social Platform Terms-of-Use, __ Berkeley Tech. L. J. __ (forthcoming 2023), available at SSRN), https://contracts.jotwell.com/click-to-agree-that-terms-of-use-are-incomprehensible/.