The last couple weeks have been rough, on a personal level. Nothing I can’t get through with a little help (and thankfully, there is help), but nothing particularly fun to get through, either. I’ve spent a lot of time going over and thinking about some old issues. Part of that has involved digging up a bunch of things I wrote years ago; a minor crisis occurred this past weekend when I tried to power up an old laptop — its hard drive is the only place certain archived emails and other bits of text still exist — and found that I couldn’t. As soon as I plugged it in, the adapter started beeping and failed to provide any power, a condition that Google tells me indicates a short or some other serious fault. Luckily I was able to find a supplier who provides replacement adapters, and one will arrive this week, at which point I’ll hopefully be able to turn this old machine on for long enough to finally copy everything I want from it.
This naturally provokes some thoughts on the persistence of information, and whether it’s something I really want in the long run.
Every once in a while I see someone writing articles and trying to rally support for some sort of digital archiving project. Ancient Sumerian tablets, we’re often reminded, are still readable millennia after they were inscribed; baked clay lasts a long time. But a lot of what we now produce has a lifetime of a handful of years, which can be a blow to history and to the people who lived it. Moving parts wear out, interfaces become obsolete, encryption formats are abandoned, or sometimes a cosmic ray just hits the wrong place at the wrong time and flips a bit, and that’s the ballgame. And that’s without considering things like services that shut down and just cause all their (or their users’) data to vanish in one fell swoop. Some of those old emails of mine met exactly that fate, which is why a single hard drive in my apartment is the only place they still exist.
There’s an article on Wikipedia that I stumbled across a while back, on the concept of information-theoretic death, which it defines as the destruction of the information within a human brain (or any cognitive structure capable of constituting a person) to such an extent that recovery of the original person is theoretically impossible by any physical means. While this may seem an awfully macabre thing to be thinking about, it’s actually pretty tame compared to some of the topics covered when I was doing my philosophy degree. And honestly, it’s a fascinating idea.
How far can we take that, though? Someday my brain will, of course, be just so much dust or ash. The same is true of the brain of every person who will ever meet me or have a memory of me. Could “information-theoretic death” be extended to include the point at which their brains have degraded sufficiently that no memory of me is recoverable? And — this is the tricky bit — what about information that’s not stored in human brains? Plenty of cultures have had a general belief that a person is never truly “dead” as long as their name is still spoken or written somewhere. Is there ever a point when all of that information is finally just gone? And, more to the point, do I have a right to that kind of information-theoretic death? Do I want that, and if I do how could I ensure it happens?
To take an example: a few years ago, a good friend of mine suddenly died. He’d been active on a certain IRC network, and had a distinctive nickname there. The network’s policy was and is that failing to identify to NickServ for a certain period of time will release your nickname. But his never released. I’ve never confirmed it, but I have reason to believe that one of the network ops kept an eye on it and made sure it stayed registered. It’s a nice gesture, though not one I’m sure I’d want for myself, and he didn’t get to make that decision. I don’t know what he would have wanted, had the choice been offered.
Quite a lot of the information I produce these days is of a technical nature. There’s code and blog posts and books and conference talks, and to be honest I don’t care whether it lasts another five years or another five hundred. On a personal level, it’s irrelevant. Some of it will continue to exist as long as various companies and services exist. Some of it would vanish if I shut down an account or stopped logging in or stopped paying a bill. I also don’t particularly care about my Twitter or Facebook accounts, since I never post anything to them that I’d have strong feelings about preserving or deleting. It doesn’t really matter whether they go out of business and it all vanishes, or they keep trying to monetize all of my tweets and funny cat pictures and social-graph data until the heat death of the universe.
So that’s not the kind of information I’ve been concerned with recently, or why I’ve been thinking about this. Rather, I’ve been confronted with the question of just what I need to keep and what I need to throw away and move on from, and that tends to involve information of a much more intensely personal nature. And what I’ve realized is that I don’t actually want most of it to be preserved indefinitely. In fact, I want basically the opposite: I wish there were some way to see that this stuff lasts only as long as I do and then goes when I go, whether that’s a car accident tomorrow, cancer a few decades from now or just the inevitable victory of entropy who-knows-how-many years down the road.
It’s not that there’s anything that needs to be hidden, of course, just that this is a strange sort of information. I am the source of its meaning, and when I’m gone, its meaning will also be gone. This is in the nature of anything truly personal; it needs the person to define it. I’m also at a point where, although there are lots of things I could do with my life, and varying legacies I could leave behind, I know which ones truly matter to me. If the sole informational residue of my existence is just a few words on a tombstone, I’ll rest quite peacefully as long as they’re the right ones. An information-theoretic death of that sort — a gentle fade from living memory, followed by the gradual erosion of some words from a piece of stone — is actually a rather appealing thought. But how to achieve it?
The first part, preserving that information for as long as I’m around, is not terribly hard.
For anything digital, there is a certain amount of effort involved. Storage has to be maintained, files occasionally have to be copied into some new form (not doing so was my mistake with the old laptop that gave me grief this weekend), some accounts have to be kept active and passwords have to be remembered or stored, but that’s about it. At most, it adds up to a few hours of work every few years.
For non-digital information, less effort is required. In a closet in my apartment, there’s a wooden box, given to me by my grandfather. At first I didn’t know what to do with it, and he didn’t give any hints, but slowly I started to put things into it, things I wanted to keep but didn’t have any other place for. Now it’s a sort of surrogate memory, a way to piece together the track of my life, which may have been what my grandfather had in mind when he gave it to me. There’s a clipping in there from when I was 12 and the local newspaper published my winning entry from an essay contest. There’s a music box I had as a child. There’s the tassel from my high-school graduation, my name tag from the first job I ever had, a couple of photos of my mother and my grandfather, and a lot of other odds and ends of that sort. There’s also a compartment — not really secret if you look, though it was exciting to my younger self to think of it that way — in the lid. In there are a few photographs and a letter.
The box itself has, currently, lasted over a hundred years. Its exterior bears a number of scuffs and scratches and cracks, but it is still fundamentally sound and requires no maintenance to keep it doing its job, which is the preservation of the objects it contains. As for those objects: the music box still works. The mechanism has lasted close to thirty years now, and it could be repaired if it ever broke. Time will eventually catch up to it and the other items, of course, but on a longer scale. The photos in the lid, for example, will eventually fade. The paper the letter is written on will someday start to break down, along with the ink and some other molecules that give it a certain scent. But they will, if simply left alone, almost certainly outlast me and my memory of what those images and those words and that scent mean. With careful preservation, they could even last for centuries beyond that.
So what about preventing preservation, or even actively destroying information, past a certain point?
On that side of the equation, anything that exists on some online service is potentially beyond my control already. But for the services which offer the ability to delete or close an account, and which actually follow through on that when asked, it wouldn’t be too difficult, provided I’d kept anything sufficiently personal exclusive to such services. Fortunately there’s not too much in this category that really needs to go when I do; mostly it’s just some emails and one very old, very private blog I used to keep as a form of therapy. For services which are less respectful of a deletion request, I suppose I could just leave instructions to someone to go log in and commit an egregious terms-of-service violation on my behalf.
As for the various computers I own, most of them will lose their informational content, or at least the ability to retrieve it (which is almost the same thing, depending on your philosophy), within a short span of years, much shorter than my current life expectancy. Solid-state drives actually make this easier, since they start degrading the moment they’re first used, and continue to do so, unstoppably, until one day there just aren’t enough floating gates still functioning to do anything useful. Rendering their contents permanently inaccessible is just a matter of applying enough write cycles. Similarly, that old laptop — which has an old-fashioned spinning-platters hard drive — likely is at or near the end of its useful life, and anything it holds could be destroyed with some repetitive writing of random noise, followed by application of a hammer.
And the box in my closet? Well, it’s made of wood, and a lot of what’s in it is made of paper. Our species mastered the making of fire quite some time ago, and that’s all it would take.
But notice that this is more complicated than just blind preservation. If I want that information to vanish someday after I’m gone, I’ll need to, at a minimum, leave a pretty long set of written instructions and credentials and probably even some documentation(!) explaining how to make it happen, and I’ll have to trust someone else to carry it all out. Some of this already has existing frameworks in place, courtesy of the long history of the legal system in the society I live in. Some of it is still very, very new, and involves problems that still don’t have tried-and-tested solutions, so setting up any kind of automatic dead-man’s switch isn’t possible.
I’ve read stories of spouses and other relatives trying to gain post-mortem access to an email account, with varying degrees of success. Facebook has a concept of death now, even if it’s implemented in a very Facebook way. Other services will probably get there someday. But I have a feeling, and I’m pretty sure this isn’t just me getting older, that it’s going to become surprisingly important, very soon, as the ways in which we produce information make information-theoretic “death” that much more complex. So maybe it’s something to be thinking about right now.