Delete? How could I?

Viktor Mayer-Shoenberger's book raises a fascinating topic - the impact of ever more powerful and pervasive digital memory on our psychology and society. But while his concerns are thought provoking and merit further discussion, his proposed medicine is for me worse than the disease.

Given the foreboding message of the book, namely an impending upheaval of the mental  landscape that humans have inhabited, and been shaped by, for millenia, there is a reassuring calmness in Viktor Mayer-Schoenberger's discussion of the issue. This is no futurist's fantastical rant about how the machines are about to take control, but a cool and collected presentation of what he sees as a real threat to our cognitive environment, one with tangible negative consequences for our society.

The basic premise of the book is that although mankind has always striven to remember more efficiently and completely, reliable records were always more the exception than the norm, and this is something being inverted by modern technology, resulting in a paradigm shift which neither our communities nor our minds are prepared for.  The first point is something we are all aware of : throughout history new mechanisms have constantly been developed to better preserve information, first via oral traditions, then with writing (inscription and printing), and finally with modern media for sound and image. Indeed the speed with which such a new technology spreads is clear evidence of how this is an important drive within humans; photographs, video recorders, hard discs etc., all have caught on like wildfire as soon as they became available. What was always the case until very recently however, was that this always involved some extra effort or cost. Of course the 20th century marked a fundamental change to previous ones in which such recording was only even an option for a wealthy elite, but, to take the example of photography,  even when cameras and film became commonplace in the 80s and 90s, every roll of film had a non-negligible cost, as well as the time and effort required to get the pictures developed. It was, quite simply, always easier to forget than to remember. But now this has all changed. Digital cameras have no tangible running costs, nor is there any cumulative work involved in obtaining and viewing the resulting pictures. Indeed, as Mayer-Schoenberger points out, even the cost of storing each picture data file has come down so much, that if calculated using the minimum wage, it is actually less than the value of the time required to actively choose and delete those pictures. Hence preservation has become the default, and deleting requires effort, which is a complete reversal of how things have been until now. And it is not just in the domain of photography that this applies; the content explosion of the internet, with online archives and social networks, means that vast swathes of our lives are documented and preserved, from the largest news event to the birthday wishes we email a friend. Furthermore, a lot of this is being done without our knowledge, by everyone from health services to Amazon.  

Of course one major, and well known, problem with this mass of retained information, is the issue of privacy. Digital data can be maintained indefinitely, anything stored for one purpose, personal or official, may at a later time be misused for another, and due to the unique ease with which computer information can be copied and distributed, such leakage from one domain into another is much more likely than ever before. Mayer-Schoenberger discusses this privacy consequence at length, and it is practically the most tangible issue raised by the book, but it's not actually the one I want to focus on here. While it is a complicated and important issue, it is one that is already being discussed at various levels.

What I find more interesting is his suggestion that beyond the misuse of information, the sheer quantity and quality of that information poses problems even for its correct usage. While having a better memory has always been regarded as a good and desirable thing, he points out how perfect recall can actually be more a curse than a blessing. One cultural warning to this effect is Borges' story "Funes the memorious" , where the protagonist, although having perfect memory, seems to lack basic understanding; being snowed under by the massive amount of trivial detail, he is blinded by the particular and cannot grasp the general, which is of course what is needed to truly know something. The point is that facts are not knowledge, no matter how many of them there are. He even presents a real world example, a woman known in the literature as AJ, but who revealed herself as Jill Price,  who despite her amazing memory powers obtains no great advantage from them, and in fact finds this powerful memory causes her to 'spend an unusual amount of time immersed in the past rather than enjoying the present'.

What these examples highlight is that information is highly dependent on context, and is worthless on its own. Mayer-Schoenberger points out that this feature is intrinsically linked to the way our human memory normally operates. Far from being a dry database of information, it is a network of relationships, and further more something which is dynamic rather than static. New information gets linked to old, influencing the surrounding context of both, and changing what was there before. Even recalling something is not a simple retrieval, but a reconstruction, a piecing together that is also heavily influenced by the mind's current state. This is of course massively different from the error free duplication of digital memory, and he points out that the more this digital variant becomes prevalent, the more it might conflict with our own biological memories. This might not have been so psychologically relevant when the amount and type of information being stored was limited, and confined to areas such as dates and events, or just to do with official or government business, but with the advent of personal electronic communication and social networking, this will now apply to personal and emotional areas as well. Mayer-Schoenberger illustrates this with a hypothetical example, that of two friends, Jane and John, who have been out of contact for a considerable time, but having bumped into each other by chance, arrange to meet again. Jane thinks it would be nice to go again to a cafe they used frequent, but in searching her email for the name of it, stumbles upon an angry message from John which she had forgotten about. Suddenly her present opinion of John is poisoned by this blast from the past. Her biological memory had filtered out this single event, and instead had preserved a more holistic conception of him, an aggregated sum of all his actions., but now this is thrown into conflict with perfect digital memory, which forces this single incident to the fore, without any subtle modification for its original context or its overall relevance.  The point is our minds and feelings have evolved to be adapted to imperfect memory,  it is an important element in how we handle our social relations, and digital memory disrupts this.
And of course this can apply to the rational as well as the emotional domain,as Mayer-Schoenberger shows with an other example, that of his friend who wanted to maintain a library of only 200 'best' books. What his friend did not realise is that his subjective judgement varied with time, and hence his library was constantly changing, as his values and priorities changed. There was no objective external 'best 200' which could be achieved, only the best 200 relative to any particular phase of his life, This is something inherently human and very distinct from machine organized information, and we need to bear this in mind when lured by the digital dream of perfect, objective, memory.  Unfortunately,  the long search for better external memory means we are culturally disposed to value it more highly, blinded by its error free perfection, and hence we are normally not aware that its remembered facts lack context and are qualitatively different from our own.

Mayer-Schoenberger discusses various options for dealing with this new state of affairs. Most of them again relate to controlling information abuse, and the problems of  'information privacy'. There are some very interesting ideas, from a kind of Digital Rights Data Management for personal information to the idea to allow people to declare 'Social Bankruptcy', so they can wipe their online slate clean. Mayer-Schoenberger's own proposal, adding 'expiry dates' to data, is interesting, and apart from its privacy potential in restricting how data can be endlessly passed on and hence misused, it has relevance also in the psychological domain which I am focusing on here.

The basic idea behind 'expiry date' is that when creating a file, or recording data, users need also to select (if only from a drop down range of options) a date after which the information is no longer valid, and should be automatically removed from a system. Whatever the practical problems, it does at least also try to help with what is perhaps the most important problem, our lack of awareness of the durability of data, and the related consequences. While I fear even a few extra clicks in any process might easily be viewed as too much effort, and prevent it from catching on, even the very fact that one should choose, or even can choose, the expected relevance time for information would make us think about it. The windows message "do you want to discard the changes to this file?", may in the future be followed by another pop up, "are you also sure you want to save them?"

Unfortunately considering how we might choose, reveals I think the fundamental problem, which is ironically related to the very contextual nature of human memory that it is supposed to support.  Whenever we are throwing something away, the first thing we ask ourselves is whether we will need it sometime in the future, and we will automatically compare the cost (or effort) to keep something, to the potential loss if we actually need it in the future. With cheap and easy digital memory the cost is minute, and so it would be irrational to throw something away, since it would just be for the sake of it. Even if setting the deletion date in advance re-balances the effort between keeping and destroying, having the option at all will always result in such an cost-benefit analysis, and there is only one rational solution : keep keep keep.

Furthermore, one could easily re-write the John and Jane story in various ways in which it is beneficial that she finds an old email. Maybe she was conned by John in the past, and being reminded of it saves her from falling for him again. Or, she might have broken off contact with John due to a misunderstanding, maybe she had been confused or angry for other reasons, and in re-reading the emails in a new light could see her mistake, and now make things right. This I think captures a point that Mayer-Schoenberger misses, that given variations in human reasoning at any time, the influence of mood and situation, change of context can also be a good thing. The problem is one can argue with plenty of examples either way, however my own personal opinion is that if one has to decide, then surely it is better to leave open the option for change in the future, rather than close down all opportunity. To do otherwise is to assume a kind of supertemporal ominpotence, that what we know and think now will always be valid, and I for one am not self-confident (or arrogant) enough to believe that. Mayer-Schoenberger ironically uses this point in his argument against perservation, people can and do change, and deletion might allow a second chance with the slate wiped clean, but old facts in new contexts can help as well as hinder this.

Another interesting idea of his would be to implement what he calls 'digital rusting' , a mechanism where by information and data randomly degrades, in a process mirroring our own memory. But again a cost-benefit analysis has to be applied if we were to allow such a thing, and again I would be of the opinion that better some unwanted facts are retained then some wanted ones are lost. Why I feel this way, and I think it is normal to do so, probably has something to do with the natural difference in how we morally judge actions versus omissions, and doing wrong versus failing to do good (in both cases the former being judged more harshly). Allowing data to decay would be an action, since we choose the mechanism, and it is I think consistent that we would feel worse having caused harm through destroying data, then having failed to ensure some good by keeping it.

But perhaps the biggest obstacle preventing such solutions from succeeding is we wouldn't want them to. Mayer-Schoenberger himself claims that one reason for our constant drive for better and better memory is the notion that it somehow brings a kind of immortality. I think I agree with this, especially if it is put the other way, that to destroy our past is to kill off a part of ourselves for ever. Despite that fact that ultimately as Woody Allen put it, we can't really achieve immortality through our works, only through 'not dying', emotionally there is something in the idea of living on through what we did, or created. We can picture these things outlasting us, and since we identify them with our selves, we thus can view, or more importantly feel, it as a kind of survival. And it is the attempted satisfaction of this desire that I think , at least in my case, would be hardest to overcome.

But even if this kind of endurance beyond our lifespan is an illusion, there is another way in which our past selves can live on, in our current and future lives. Mayer-Schoenberger remarks rather dismissively on how odd it is to read an old diary entry, implying that since it can feel like written by someone else, it is somehow now longer relevant. My opinion would rather be that in precisely this way it can bring an old forgetten self back to life.
For example just recently when searching my email I stumbled  upon some old correspondence and found a passage I had written and which, on re-reading now, I found both insightful and relevant, indeed probably more so than the original recipient had. Encountering the viewpoint of a previous 'me'  can be like bumping into an old friend, re-awakening memories for enjoyment, as well as reminding me of points I once knew but had now forgotten. This for me is the great thing about digital memory, that in its scope and durability our actions and thoughts can resonate across time, and far from being only relevant to their original contexts, they can often transcend them. And this is something wonderful, since to me always one of the saddest descriptions of a person or an idea is that they or it were  'before their time'. There is something poignant about the wasted chances this implies, but external memory helps to overcome this. It is only through recording and later discovery can ignored and missed ideas or thoughts perhaps sometime find life, and the current expansion in the types of digital memory extends the domains in which this can happen.

The point is that destruction destroys more the thing itself, it also wipes out the potential, the opportunity, that leads from it; the possibility that it might some day be relevant. In other areas of life physical restraints (basement size and tolerance of our wives!) prevent unlimited hoarding, in digital technology we are creating a space without such restrictioons. So let's make use of it.  Before people asked God to save them, now with a click of a button, we can do it ourselves.