The Copyright Act includes a “radio loophole” that permits radio stations to broadcast songs without paying royalties to sound recording copyright owners. In recent years, critics have attacked this loophole as being fundamentally unfair, depriving artists of revenue, hampering innovation in music, and contradicting international norms. This Note argues that Congress should close the loophole for another reason: it is out of line with historical trends.

Throughout U.S. copyright law history, Congress has routinely bestowed statutory carveouts and loopholes upon commercial and nonprofit entities alike. Some apply to copyright owners, and some apply to users of copyrighted works or entities that distribute copyrighted works. Although any given carveout is often the fruit of successful lobbying efforts, a broader look at the staying power of statutory carveouts reveals that they are more likely to persist when they serve a public interest, like many nonprofit entities do. Conversely, statutory copyright advantages tend to erode for commercial entities that lack strong public benefit rationales.

In the late 20th century, Congress could justify its codification of the radio loophole by pointing to the public benefits of radio—including its music discovery, public education, and emergency broadcasting functions. However, technological, legislative, and societal changes have combined to erode both the public benefits of radio and the public benefit justification for the radio loophole. Today, radio is less like public-benefitting entities whose statutory copyright advantages endure and more like commercial entities whose statutory advantages erode. Placing the radio loophole within this historical framework illustrates that the time has come for Congress to eliminate it. Without strong public benefits of radio as a justification, the radio loophole is now an antiquated result of bare lobbying efforts.

This Article introduces a novel empirical method for measuring judicial importance beyond traditional metrics, such as citation counts, which often undervalue decisions that shape legal practice without generating extensive case law. Drawing on techniques from computational text analysis, it proposes linguistic shifts—systematic changes in the length, language, and structure of legal documents—as an alternative proxy for a judicial opinion’s legal significance.

The Article applies this method to a case study involving Blackboard, Inc. v. Desire2Learn, Inc., 574 F.3d 1371 (Fed. Cir. 2009), a Federal Circuit decision that has received relatively little attention under conventional citation-based measures. By programmatically analyzing more than 300 patents filed before and after Blackboard, the Article demonstrates that Blackboard triggered a measurable linguistic shift in how patent prosecutors draft the very type of patents at issue. These findings demonstrate that Blackboard significantly shaped practitioner behavior despite its low citation count.

More broadly, the Article demonstrates that computational analysis of linguistic shifts can serve as a scalable, doctrinally agnostic, and empirically verifiable framework for assessing judicial influence. By capturing changes in legal practice beyond the litigation process, this approach expands the scope of judicial influence to include legal actors and behaviors that are otherwise invisible to conventional metrics.

Text is a vehicle to convey information that reflects the writer’s linguistic style and communication patterns. By studying these attributes, we can discover latent insights about the author and their underlying message. This article uses such an approach to better understand patent applications and their inventors.

While prior research focuses on patent metadata (i.e., filing year or gendered inventor names), we employ machine learning and natural language processing to extract hidden information from the words in patent applications. Through these methods, we find that inventor gender can often be identified from textual attributes—even without knowing the inventor’s name. This ability to discern gender through text suggests that anonymized patent examination—often proposed as a solution to mitigate disparities in patent grant rates—may not fully address gendered outcomes in securing a patent.

Our study also investigates whether objective features of a patent application can predict if it will be granted. Using a classifier algorithm, we correctly predicted whether a patent was granted over 60% of the time. Further analysis emphasized that writing style—like vocabulary and sentence complexity—disproportionately influenced grant predictions relative to other attributes such as inventor gender and subject matter keywords.

We also examine whether women disproportionately apply for patents in fields with higher rejection rates. Using a clustering algorithm, applications were allocated into groups with related subject matter. We found that 85% of women-dominated clusters (over 50% women inventors) have abnormally high rejection rates, compared to only 45% for male-dominated groupings.

These findings highlight complex interactions between textual choices, gender, subject matter, and success in securing a patent. They also raise questions about whether current proposals (e.g., anonymized examination) will be sufficient to achieve gender equity and efficiency in the patent system.