Stuart Shapiro, Categorical Denial: Deconstructing Personally Identifiable Information
Comment by: Lance Hoffman
Workshop draft abstract:
The concept of personally identifiable information (PII) has been drawing increased attention of late, sparked by problems with de-‐identification. Depending on who you talk to, PII is a distinction without meaning or an inescapable necessity for bounding regulation. Irrespective of their conclusions, one trait common to all these analyses of the “PII problem” is their failure to look at PII as the construct it fundamentally is: a category.
Categories constitute a basic conceptual building block of human thought, related to but distinct from other building blocks such as analogies. They are both socially and cognitively constructed and grounded in both objective and subjective perceptions. Most importantly for the purpose of analyzing the PII problem, they exhibit structure and a host of associated properties which continue to be investigated by researchers in a variety of fields, including cognitive science.
To fully grasp the problems and possibilities of PII, we must take it seriously as a category, as opposed to a legal or technical label. This paper aims to rigorously analyze PII as a category, leveraging the substantial body of existing research on how humans construct and use categories. This includes situating PII with respect to different types of categories, inferring its internal structure and related characteristics, and drawing out the implications of differential sorting behavior among subject matter experts and laypersons.
The problems with PII will not be resolved by abandoning the concept or by introducing ad hoc constructions. Ensuring the viability of PII as a category requires more explicit understanding and treatment of it as such. Doing so reveals new avenues both for understanding current difficulties and for addressing them in a coherent fashion consistent with what we know about categories qua categories.