THE LANGUAGE-GAME OF PRIVACY.

• Author: Fairfield, Joshua A.T.
• Position: Book review

PRIVACY REVISITED: A GLOBAL PERSPECTIVE ON THE RIGHT TO BE LEFT ALONE. By Ronald J. Krotoszynski, Jr. Oxford and New York: Oxford University Press. 2016. Pp. xx, 292. $90.

INTRODUCTION TO THE PROBLEM

Start with a thought experiment. Suppose someone asks you what the word "privacy" means. After thinking about it for a while, you realize you don't really know. You can do one of two things: you can go and ask people how they use the word privacy, or you can begin to internally construct your own definition. But because you don't know what privacy means, if you choose to construct your own definition, the only information that definition can contain is what you think privacy means. Worse: you may be aware that privacy could mean other things to some people, but those meanings don't fit the operational definition you're trying to reach here, and so you cut those meanings out, using the theorist's scalpel. Those things are not privacy, you argue, because they fall outside the very theoretical structure that you just built. But you soon realize your attempt to "pare down" language until it meets your exacting criteria entails literally putting words in others' mouths. You are claiming that what other people mean when they use the word "privacy" is precisely and only that which fits your theoretical structure. (You're also taking words out of their mouths: you're saying that when they use the word "privacy," they're often wrong.) And that's obviously not true. It's a dead-end street.

So instead you decide to take the richer path and find out how other people use the word privacy. You might read a few people's papers, most of them by privacy law scholars, and come to the conclusion that one of their definitions is the answer. But of course you wouldn't have found out what privacy means; you would have just gotten some data about what those scholars mean when they use the word privacy. There are more people in the world than Western privacy scholars trained in common law, and they use language too. Instead, it might be reasonable to ask how people use the term in daily life because the word privacy means exactly and only how people use it. In other words, meaning is use. In the grand debate about how the meaning of privacy should be explored, I fall into the "ask people" camp, whereas many thinkers about privacy, especially legal thinkers, fall into the "define it" camp.

The "define-ask" divide does not just exist in law. (1) It exists in logic, philosophy, and mathematics, and thus lies at the root of the scientific method. (2) This is not a distinction to do with the divide between empirical and theory work. Rather, it is a divide in empiricism itself. On both sides of the divide, people develop sensible hypotheses and then test them by comparison to theories of reality. The debate regards not whether we should test our conclusions, but which attribute of real experience should our hypothesis be tested against. In the case of privacy, should we test how person A uses the word "privacy" in any one instance against a core essential definition, or should we test the use of a word in one instance against how people actually use it?

One way to see this clearly is to observe the revolution that Ludwig Wittgenstein's theories worked in the social sciences. Traditionally, a scientist would develop a theory of reality, and would test whether results conformed with that theory. If the result was an outlier, then, that would point to a fault in the experiment or test. If it had lots of company, perhaps the theory would need revision. So before Wittgenstein, one could imagine (and this was the case) (3) that some cultural anthropologists might arrogantly judge technologically undeveloped cultures, because those cultures' usages fell short of the best theories the anthropologist had. But Wittgenstein's point was that in examining language (and other constructed cultural artifacts), the empirical reality against which theories of linguistic meaning must be tested is the use of language itself ? This shift worked serious changes in social science practice. Modern anthropologists would now find it alien to tell their subjects that their subjects' use of language failed to match some definition or theory. Rather, any attempt to determine meaning must lie in describing uses.

Even among the "ask people" crowd, the technique of asking people is often applied ad hoc, or without really understanding what's at stake. We read some case law to find out what those judges thought, or we read cases and theory from another legal tradition to try to get a handle on that. Rarely do we work out a system for how to ask--to ensure that we are gathering good-faith meanings of the words (and not lawsuit-motivated lies, for example)--or, more importantly, how to feed what people think "privacy" means back into an understanding of the term, without just letting the term mean whatever people say it means.

We are talking here about two views of how words fit together: Are words atomistic sounds that refer to clearly defined things? Or are they drifting and interlinked, while remaining anchored in context and community? Linguists take the latter approach, as does anyone who notices how words shift and flow into one another over time. (5) The question is how law and legal terms--and in particular legal terms like "privacy" that are also terms of common use--should be analyzed.

There is nothing wrong with carefully defined terms of legal art. The difficulty arises when lawyers and legal academics fail to signal that they are using a term of general use as a term of art--that they are using one explication of the term, not its full meaning. And this is particularly so when lawyers and legal academics make claims about the essential meaning of terms. For a legal thinker to say that privacy "is" something--by reference to case law, statute, and legal theory--is too often to say that privacy "is not" its everyday use. From a linguist's perspective, that cannot be correct. (6) And even lawyers must recognize that "privacy" is not a legal term of art. It can't be. We ask whether ordinary people have a reasonable expectation of privacy. The meaning of the term is dependent in significant part on what people expect when they consider something "private." Privacy is practiced in the bedroom and the bathroom, not the courtroom. So we must take the linguist's path.

I am, of course, not the first to see that law is language and that, like other languages, the language of law develops according to systems of negotiated meaning--grammars--and does not in fact refer to atomistic physical or defined conceptual "things." This is a straightforward application of some philosophical and linguistic observations made by Ludwig Wittgenstein, a little more about which below. (7) 1 think that the field of law is likely to provide many fruitful applications of Wittgenstein's theories. Wittgenstein's point is that the meaning of language is not essential to any term; it is negotiated in context and by community. (8) That is, to my mind, a pretty strong description of how law works. And understanding that law works in this way might help resolve some legal problems.

Dan Solove notices that law follows Wittgensteinian principles in his signal book Understanding Privacy, (9) He argues that the search for a definition of privacy is futile because privacy, as used by humans, simply means a range of related things. (10) And if that were all I needed to point out, that would be that. The difficulty is not with the choice of method, but its application. Solove starts well. He delves into case law, legislative histories, and a few recognizable other legal traditions and emerges with what he thinks privacy is, across a range of "privacy problems": a coherent-but-sometimes loose set of related difficulties that concern information, the individual, and society. (11) It is useful because it serves as a haystack, a collection of the thoughts from the immediately surrounding field of ideas, a one-stop shop for the conclusions of one legal academic mind that has done an enormous amount of reading about how other legal academic minds use the word "privacy." Solove emerges from this haystack with a taxonomy of privacy--a hierarchy with kingdom, phylum, class, order, family, genus, and species. (12) But it would be an enormous mistake to confound this taxonomy with the linguistic community, contexts, and tasks that gave rise to the range of cases Solove collects. Taxonomies are definitions in sheep's clothing.

Comparative law is precisely where Dan Solove's method diverges from his conclusions. He claims that the taxonomy he creates identifies generalities that cross all cultures. (13) He does this by reading a lot of case law, and sometimes the courts in these cases seemed to be working on similar problems despite cultural differences. (14) His method of inquiry--asking people what they mean when they use the word privacy--must certainly be a constant across cultures. It is how language works, and so will work for any language-using culture. But his narrow conclusions--the taxonomy he produced--would not transition across the Atlantic. No French scholar would have written that taxonomy. Indeed, no other person would have written that exact taxonomy. To mistake the taxonomy for the process of linguistic inquiry would be to commit a serious error. There are of course elements of Solove's taxonomy that resonate strongly across the Atlantic. But there are parts of it, including the most basic part of constructing a taxonomy in the first place, that may not translate at all.

If one must produce an accurate definition of privacy, it must be only this: privacy is the linguistic-conceptual network of ideas, constrained within a grammar, that is invoked in one human when another human honestly and with intent to communicate makes the set of noises "privacy." From this perspective the...