Severed Knowledge: Negative Trade Secrets Through the Lens of Compartmentalized Memory

Compartmentalized Knowledge and the Boundaries of Memory

Thomas Edison is credited with saying "I have not failed. I've just found 10,000 ways that won't work." If the details behind those 10,000 ways were disclosed to Joseph Swan, one of the scientists racing with Edison to create a commercially viable light bulb, Swan might have overcome Edison and become a household name. Negative trade secrets, often referred to as "negative know-how," embody the institutional memory of failure—the catalog of what does not work, built through repeated research, experimentation, and error, to (ideally) ultimately reveal a solution that does work. Before an invention is fully realized, knowing what does not work is critically important, but it is not always as easy to protect against disclosure.

In the popular television show Severance, Lumon Industries protects the secrecy of its research and development by surgically altering its employees' brains so that their professional knowledge is secreted away when they leave the building. As far as we know, that technology is pure fiction, and real companies engaged in research and development must enforce strict boundaries around both their breakthroughs and their missteps. Failed experiments, discarded strategies, and lessons learned from missteps can be as significant as the knowledge about what efforts led to a successful outcome, but because they are not always as well-documented, they can also be harder to protect.

Evolution of Negative Trade Secret Law: From Exclusion to Recognition

The 1939 Restatement (First) of Torts required that trade secrets be in active use, effectively ruling out negative trade secrets, which definitionally were not used at all. The adoption of the Uniform Trade Secrets Act ("UTSA") in 1979 marked a shift, explicitly including information with negative commercial value in the UTSA commentary. 

In response, courts began to view negative know-how as part-and-parcel of the affirmative trade secret. For example, in Metallurgical Indus., Inc. v. Fourtek, Inc., the Fifth Circuit rejected the notion that a negative trade secret is never enforceable, noting that the trade secrets there resulted "from realizations of what not to do; but so does every human process: the selection of one action at a given moment involves the rejection of every other conceivable one that might have been chosen … This characterization, however, can always describe the invention or modification of a device. Knowing what not to do often leads automatically to knowing what to do." 790 F.2d 1195, 1203 (5th Cir. 1986). Other courts have suggested that negative information that informs positive trade secrets can be more explicitly protectable. See Courtesy Temp. Serv., Inc. v. Camacho, 222 Cal. App. 3d 1278, 1287 (Ct. App. 1990) (defendants had acquired valuable negative information when they received plaintiffs' customer list because they could discern uninterested customers and avoid the time and effort of calling those people).

In a more extreme example, the court in Avery Dennison Corp. v. Finkle applied the Connecticut Trade Secrets Act (based on the UTSA) to prevent an employee from starting a new position based in part on negative know-how theory. No. CV010757706, 2002 WL 241284, at *2 (Conn. Super. Ct. Feb. 1, 2002). There, the court found that, among other things, the defendant's "knowledge of the procedures and information relating to concept testing data, including 'dead ends,'" relating to writing implements was sufficient to enjoin his employment with a competitor. Although such injunctions, sometimes premised on inevitable disclosure, are rare, the court noted that "it seems virtually impossible for Avery Dennison's trade secret information, particularly as to 'dead ends,' to not affect" his relationship with his new employer.

Today, courts that recognize negative trade secrets treat them the same as positive trade secrets, considering whether: (i) reasonable measures were taken to keep the information secret; and (ii) the information has independent economic value. Trade secret claims always require some level of specificity, but courts often require more detail when asked to recognize negative trade secrets. Compare Alta Devices, Inc. v. LG Elecs., Inc., 343 F. Supp. 3d 868 (N.D. Cal 2018) (allegations allowed the defendant "to ascertain at least the boundaries within which the secret lies"), with AlterG, Inc. v. Boost Treadmills LLC, 388 F. Supp. 3d 1133 (N.D. Cal. 2019) (broad statements alleging misappropriation of "positive and negative learnings" were too vague). 

The Federal Circuit has affirmed judgments of misappropriation of negative trade secrets. In CardiAQ Valve Techs., Inc. v. Neovasc Inc., the court affirmed a $70 million jury verdict in favor of CardiAQ. In that case, the parties entered an NDA for Neovasc to assist in the development of a prosthetic mitral heart valve that could be implanted via a catheter entering the body through a small incision in the patient's leg, making open-heart surgery unnecessary. During the collaboration, Neovasc applied for a patent on the technology, and did not list any of CardiAQ's doctors as inventors. After Neovasc's patent application was published, CardiAQ sued for trade secret misappropriation relating to several revisions of its device (none of which became the ultimate product) and sought to have its doctors added as inventors on the patent. 

The jury awarded damages on the trade secret claim based on a reasonable royalty, and the court found that CardiAQ's doctors were co-inventors. The Federal Circuit affirmed, holding that "the jury had a sufficient basis to conclude that CardiAQ continues to use [the trade secret] in the operation of its business because, [a]s it develops its TMVI device and works towards FDA approval, CardiAQ continues to use its past discoveries about what works and what does not, and [i]ts development history has influenced the current design of its [product], and will likely influence any future modifications to it." 708 F. App'x 654, 665 (Fed. Cir. 2017) (internal quotations omitted). Among the evidence the jury considered was a shareholder presentation by Neovasc stating that its "[i]ntimate understanding of what has and has not worked so far in the development of (percutaneous) valves" gave it a competitive advantage. Id.

In a negative trade secret case, it may be advantageous to develop damages theories specifically about "use" and also "acquisition" of the know-how. Judge Alsup highlighted the distinction between acquisition and use in Waymo LLC v. Uber Techs., Inc., concluding that Waymo "did not preserve an acquisition-only damages theory by alleging that defendants misappropriated negative know-how," but finding nonetheless that "a defendant might use a negative know-how trade secret by taking its lesson to avoid developing apparently fruitless technology." No. C 17-00939 WHA, 2018 WL 466510, at *4 (N.D. Cal. Jan. 18, 2018).

Conclusion

Recognizing and protecting negative trade secrets can require a careful balancing act, one that echoes the central tension of Severance and the true purpose of Macrodata Refinement: who controls the boundaries of knowledge, and what are the consequences when those boundaries are breached? As courts continue to refine the standards for pleading, protecting, and valuing negative know-how, the law increasingly mirrors the challenges of managing compartmentalized memory in the modern workplace. The evolving treatment of negative trade secrets underscores the importance of clear boundaries, meticulous documentation, and robust confidentiality measures in a world where the value of what is forgotten is finally being recognized. That said, courts' increased willingness to recognize the value in negative know-how suggests that future litigation will further delineate the boundaries of this concept from pleadings to damages.