Delete: The Virtue of Forgetting in the Digital Age [Hardcover]

In this interesting but not always persuasive book, lawyer and policy analyst Viktor Mayer-Schonberger asserts that being able to forget stuff is a requirement for human social evolution.

For anyone who misplaces his spectacles or keys, this may seem surprising, but Mayer-Schonberger makes the case for it in at least some aspects of daily life. He concentrates on old resentments, which may cripple us if brooded over too long.

Maybe. Further, he claims that the digital revolution has made it impossible for us to usefully forget.

He presents a couple of examples: One is a Canadian psychologist who wrote a research paper in a journal mentioning his use of LSD in the '60s. American immigration officials, using Internet search, matched his name and - declaring him to be a dangerous drug user - denied him entrance.

This seems to me less a problem of too much remembering than of too stupid governors, but Mayer-Schonberger does explain in great detail about how much information the combination of digital speed and cheap memory can store. And even create, by data mining.

It doesn't have to be information you put on the Internet, either. Insurance companies routinely get records of most of the prescriptions pharmacies sell, and they can reconstruct much of your medical history - a history that is otherwise legally supposed to be private.

This part is plenty scary, whether there is a problem with not forgetting or not.

Mayer-Schonberger then leads us through various legal and technical fixes to the problem of too much memory too long. The Europeans have taken a hard-line view of privacy. This leads to absurd results: German universities are not allowed to reveal who they have awarded degrees to.

This much of "Delete" is must reading, unless you've lived in a cave the past 20 years.

The remainder, the frankly controversial part of "Delete," is only interesting if Mayer-Schonberger has already persuaded you that not forgetting is a problem.

He proposes, as a partial and initial defense, a policy of sunsetting or expiring digital data.

This is problematic. He uses the example of yesterday's newspaper. However, the uselessness of yesterday's paper resides in the fact that we have not yet had time to forget what was in it. A copy of a 100-year-old paper is worth more now than it was when fresh.

Imagine how useful it would be socially if 150 years ago the whole world had had as many newspapers as America or Europe, and if they had published daily temperatures. We could save billions in trying to reconstruct past climate and maybe trillions if the result showed that global warming has been overstated.

Similarly, in the United States, we consign census data to only temporary oblivion, keeping it secret for decades but then throwing it open for research useful to both sociologists and geneaologists.

So there is a throwing-the-baby-out-with-the-bathwater aspect to Mayer-Schonberger's solution.

Also, we now know that when we call up memories, we distort them when we restore them to our brains. Elizabeth Loftus demonstrated this conclusively 40 years ago (her book "Eyewitness" ought to be read by every person called to jury duty), but Mayer-Schonberger does not mention her. He does refer to a Harvard colleague who has made similar studies more recently, but they seem not to care about such things as accuracy of testimony about past events.

It may be well to forget past insults but then again, maybe not.

Mayer-Schonberger writes, "Forgetting is at least in part a constructive process of filtering information based on relevance."

Whether that's a bug or a feature depends on circumstances.

The first half of the book is dedicated to setting the stage. It is a rather detailed and rich account of the history of the contemporary information environment particularly print, evolution of the memory devices and information storage, and development of information governance institutions (defined in broader terms) such as copyright. While I was aware of some of the stories, many of them were rather new to me. For example, did you know that the subject index, as an alphabetical list of topics covered in a book, was introduced in thirteenth century, but the idea of adding page numbers to the index to ease the actual navigation was added only in the sixteenth century? Quite interesting.

Telling this history Mayer-Schönberger draws a picture of ever growing body of information about us, as individual members of society, and the way we may interact with it, even if in an indirect way. One of his favorite examples is the story of Stacy Snyder who was denied her teaching certificate because of a picture she had posted on MySpace of her dressed as a drunken pirate. The gist of the argument, if I read it correctly, is that while it becomes easier and cheaper to collect and store information about us and our behavior, we, as individuals, are losing more and more control over that information (once you or somebody else posts your picture online, you no longer have control over where it may appear, who may see it, and in what context). He labels it in terms of remembering and forgetting - if in the past it was difficult and costly to remember and easy and cheap to forget, this balance has reversed.

These days it is so easy and cheap to remember that we start losing our ability to forget. The repercussions of this development are that the accessible, durable, and comprehensive digital record of our past directly impacts the way we conduct and make decisions in the present. For example, I know that once this post will be published, it will become a permanent record of my take on "Delete". Knowing that, I should probably be very careful with what I say about it, because it may impact my future interaction not just with Viktor (with whom I am currently working), but also with other potential readers of this post. I may choose to self censor myself, to present a biased view, or abstain from publishing it altogether. The point is that my behavior today is guided by the uncertainty about the future uses of this information - on the one hand I know it is there to stay, probably attached to my name, but on the other hand, I have no idea who, when, and under what circumstances will use and interpret this post.

To better understand this idea, I think it is helpful to focus on some aspects of socio-psychological functioning of information, which Mayer-Schönberger discusses in length in the book. One of those aspects is interpretation. The bits and bytes in themselves do not mean much, unless we interpret them (similar to the idea of data in knowledge management). It is through interpretation that the information gains meaning and thus also social functions. This leads to another important aspect, which is context. In different contexts we will interpret the same information differently and this is one of the dangers of digitized memory - information is recorded in a certain time and in a given context, but when it gets retrieved at a different time and in a different context, it will likely have different meaning. Thus we are losing control over the interpretation and meaning of the digital information about us and our behavior. When we, as individuals, are losing control over the information, we are becoming powerless compared to other actors (like the state and the corporate world) who have the capacity to collect, store, and retrieve information about us, thus making them even more powerful (they know more about us than we know about them and they control the interpretation process of information about us). Another aspect of this is the negation of time, which threats our ability to make rational decision in the present. Instead of focusing on the big picture, we are focusing on managing the mundane details of our lives, because those are recorded and stored and will have impact on us in the future.

The shift of control over information and negation of time are at the heart of Mayer-Schönberger's concern with digital remembering. The rest the book is dedicated to analysis of potential responses to this concern and finally a proposal of an alternative solution. The book lists six different potential responses, each addresses either the power or the time aspect of digital remembering on one of the three levels: individual, law, and technology. The six solutions are digital abstinence, information privacy rights, digital privacy rights (sort of a DRM for personal information), cognitive adjustment, information ecology, and perfect contextualization. Each one of the approaches has its merits, but each one also has its drawbacks either at the conceptual or practical levels.

Mayer-Schönberger suggests expiration date for information as his solution to the negative effects of digital remembering. On the face of it, this is a rather straight forward idea - we need a piece of meta-data attached to each bit of information, which will determine how long this bit of information should be retained. Of course, his suggestion is much more nuanced and he goes into various scenarios of different ways in which information can be forgotten or partially forgotten, but I hope my one-line explanation carries over the gist of the argument. Mayer-Schönberger acknowledges in his book that expiration date addresses the time-related aspect of digital remembering, but it does little at the "power" front. In fact, the "power" is supposedly influenced indirectly, as by allowing automatic deletion of information the powerful side in the interaction is giving up some of its powers (if my power stems from having information about you and being able to mine it for my purposes, giving up the control over when this information is deleted, is equivalent to giving up part of my power).

I think that the main weakness of the expiration date argument lies not in the fact that it focuses primarily on the "time" aspect of the issue, but in the fact that it puts great hopes into the agency of the user. The idea of expiration dates gives user the power to decide for each and every piece of information how long they want to retain it. However, I am still slightly skeptical whether the user will use that power, because it comes with a cost. This idea assumes that (1) people want to make a decision about each bit of information they process and (2) they are capable of estimating the usable time span of each and every bit. I am not sure that people are that zealous about managing their information and are that thoughtful about the future prospects of its use. Just imagine if you had to decide for each one of the 300 pictures from your last trip, how long you want to retain it... wouldn't it be easier just to keep them all? ... just in case?

However, I think the main task of "Delete" is not offering a practical solution, but undertaking a rather ambitious conceptual and educational task - bringing the idea of "finitness of information" (p.171) into the public consciousness. There may be numerous socio-technical solution to the negative effects of digital remembering, but you need a well stated argument to start thinking in that direction. I think this is what "Delete" is trying to achieve.

We are all beginning to appreciate that digital information is changing our world. Casual posts to blogs, personal home pages, indiscreet pictures or video posted to social networking sites -- these will continue to live online and in worldwide databases long after we would otherwise have forgotten them. Viktor Mayer-Schoenberger's new book explores how the inability to forget one's past in the digital age is already changing our society, and may potentially change one of the most powerful human characteristics for personal and societal growth - the power to forget. Mayer-Schoenberger traces the history of humans efforts to preserve information, and the corresponding importance of some information being forgotten. He suggests that the ever-expanding and seemingly permanent storehouse of individual and societal remembering is something we might want to control before it reshapes our world in unpleasant ways. He proposes a variety of individual, technological and governmental mechanisms (from expiration dates on information to cognitive adjustments on how we look at data) for curbing our increasing reliance on digital remembering of things best left to drift into the shade of human editorial memory. Mayer=Schoenberger challenges us with both anecdote and statistics to determine for ourselves just how much remembering is too much.