In defence of ideas in a world of data

One result of our current economic and technological preoccupations is that we are blind to all possible alternative approaches to a problem. Our approach is not flexible, it is not supple enough to cope with all the contours of complexity which are innate to the world. Thus generally there are two approaches, whereby if we encounter a difficulty we will attempt to address it either via (i) speed or (ii) quantity. The first is where things such as computers and automation come in. Nominally they are here to make the execution of a task quicker, and they are almost always successful at this. The second is a corollary of the first, in that being able to do something faster (i.e., compute) allows us to do more of an action. So the rise in computing power leads in some sense to considering all problems in terms of the raw inputs for computing – data.

What is necessary, of course, is to ask whether we should actually want this. When we focus on data (and consequently information, in that specific sense) we become blinded by the possibilities beyond data, and unable to consider the contradictions which become inherent to an attitude which regards all the world as a flux of data. Cory Doctorow’s “Metacrap”, or the seven problems with metadata leads us in the direction of the problems within this mindset, the faith that any difficulty can simply be overcome with more data. Make the problem about access to more data, rather than anything specific to the emergent qualities that may be found with a given problem itself. Doctorow has one line in this piece which is what I wish to take as my point of departure. Of point 4 (‘Mission impossible: know thyself’) he writes the following:

Entire religions are formed with the goal of helping people understand themselves better; therapists rake in billions working for this very end.

The injunction to know thyself is still with us as it was in the time of the Delphic oracle, but the point of this in the context of what Doctorow is saying is that the fault, dear reader does not lie with people, but with data itself.

Take what Doctorow says at face value, with no pinch of salt. In a world of data and input/output, what happens in a religion or in therapy? Are we neutral recorder of sense data, waiting for a priest or analyst to come along, and effect a phase transition pushing us into a state of coherence? What has been our status up to this point? Were we but crude meat machines, emptily crunching through the brute sum of daily experience until we went through this transition, and cohered this mass of lived life into some intelligible – that is to say processable – form? I can’t think of anyone who would sincerely profess this opinion, when asked, but this is the implication of the crude data-centric worldview.

Consider, then, religion and psychotherapy from another angle. This perspective is that of human beings as relying upon coherence, just as the world of data does, but for a very specific, and very different reason. The coherence in this realm is the coherence that comes from narrative, from ideas (Paul Ricoeur, the French philosopher of hermeneutics, approached both religion and Freud in just this manner). There is very little that computation has in common with ideas and thinking as performed by humans, beyond the sphere of analogy. The coherence of human thought is not the same as the coherence of the search algorithm. Our intentionality as thinking creatures introduces what we could call a qualitative difference, whereas the data-centric attitude can only conceive of quantitative difference. The data-miner has the restricted conceptual greed of one of Tolkein’s dwarves, simply wanting more. The alternative I am suggesting is that we accept more will never be enough.

Coherence does not necessarily follow from data and information (though these are more than likely prerequisites in thought as they are in computation),  but these must coincide with the presence of a structure. This structure can be covered by many terms: shape, schema, form, model, hierarchy, architecture, program, metric, ontology (in the computer science sense). One thing all of these have in common? Structure ≠ data. The words in the above list are simply not covered by data and information, and this is something which the data-centric outlook doesn’t face up to. A quick summary: data and information are necessary but insufficient. Accordingly, they are not ‘wrong’ but rather they are not the whole story. This has some interesting implications, which allow us to retrospectively reconceive the data-centric in this light. First, this data-centric attitude is not data. This may seem a strangely redundant thing to say, but it is worth making explicit. The data-centric attitude is a way of (coherently) presenting the role played by data and information in our social reality. So, it is an idea of data.

But what happens when we reach a limit within the realm of data, the data-centric attitude? Well this is when, for example, Web 1.0 became Web 2.0, and we are reassured by proselytes of data that there is an arithmetical inevitability in combining these two to derive Web 3.0. The problem with this is that such a linear progression has excluded something, namely that the idea of considering the world as data brought us far, indeed to the limits of the possibilities of data. This exhausted the idea of data, the coherence of this attitude and approach. Accordingly, we have begun to look towards metadata (which Doctorow dismisses as metacrap) as a means of overcoming this limit, and here we reach the crux of the problem, the conflict between ideas and data.

An idea is meant to draw coherence from chaos. It is a singular response to a set of concrete circumstances. As such it is particular. What happens with metadata is that the idea of data ceases to be an idea, because it has been universalized. All that is happened is that we have changed the name from data to metadata, and some of the processes which connect with the processing of data (this is the structuring of data, the idea of data). The form of thought, the intention behind the idea was not interrogated at the limit which it met, and this is the failure of metadata. It accepts as valid, for now and all time, the approach and method that the idea of data implies, that quicker and more are sufficient.

This is not to say that metadata “won’t work”. Metadata is the foundation of functioning business models and information tools, and clearly for now it is enough for the needs of many. The point is that this will soon become insufficient. We need a significant confrontation with our ideology of information, to begin interrogating so much of what we take for granted. Indeed, with the evolutionary  algorithms we are doing just this, ditto with advances in parallel processing. Yet some of the greatest theorists of information go unread because their focus has been on realms which we do not deign to regard as respectable, serious, or rigorous media for the collection and dissemination of data and information as was appropriate at the time. Computers cannot manipulate ideas, ergo ideas are irrelevant, ergo I can safely ignore entire fields of human thought. I believe, in contrast to this, that figure as diverse as Hans-Georg Gadamer, Paul Ricoeur, Ernst Robert Curtius, Erich Auerbach, Harold Innis, et al. are giants in the realm of which the philosophy of information and theories of data and communication are but a part. These figures are important because they have considered the sum ecology involved in human thought, as overlapping networks involving data, ideas, information, social structures, economic flows.

Computing, data, and the information infrastructure need to be related to the above areas, and to be given a history and a context, to be made coherent. Until we begin to do this, to see that a tool is not simply a tool, to move beyond the ideology of technology (which is party to the nonsense of “guns don’t kill people, people kill people”), we will awkwardly be forcing ourselves to contort to fit idealized notions of a world of data and information, rather than using data and information in the service of thought and living themselves. It may be that there is an information ecology, and that technology evolves, but there is no reason why this need be in opposition to our own human development, and the concerns specific to us as biological, thinking beings. The idea of data, the data-centric ideology needs to grow up, and realise that the ideal of abstraction, of pure data and accurate information is a fantasy, and a dangerous one at that.

[This is a basically abstract examination of a data- and info-centric worldview that neglects ideas and thought as the context which they actually inhabit. I hope to return to some of these ideas in a somewhat more specific, applied context at some stage, but the substantive argument is the one presented above.]

Enhanced by Zemanta
Advertisements

6 thoughts on “In defence of ideas in a world of data

  1. I think this is an important subject, and it’s especially interesting to note the fractal break-down of the technical thought process on this point. Experts in machine learning and data mining readily acknowledge that a broken model can never develop into a useful hypothesis, no matter how much data is poured into it. Nonetheless, when we step outside the realm of techne and ask broad questions of direction and purpose about useful endeavor (e.g. “Why should we mine all of the text for references to agriculture?” Or: “Is it really sane to value such pedestrian information so highly?”) this awareness of underlying models and motives seem to go away, so that the problem once again reduces to more techne and more data. A great many people who turn up their noses at the model in a philosophical treatise will uncritically accept the model presented by a high-degree polynomial.

    Perhaps it’s a naive reading, but you’ve made me wonder if this pathological obsession with data is what Baudrillard was referencing (or foreseeing) in his discussions of symbolic exchange and death.

  2. This is a great essay, Andrew. A few weeks ago you inspired me to dig out my old dusty copy of Paul Ricoeur’s Fallible Man- which I had never read. Unfortunately, my 2 year old daughter got a hold of it and it became a coloring book, so Ricoeur will have to wait.

    I have a post planned for this week that may interest you, or that would almost definitely benefit from your criticism, but your post has inspired a question now: how do the philosophers you mention deal with the gap between our ideas and the thing in itself. It has struck me lately that this is one of the greatest sources of error- this delusion that our ideas are reality in itself and not just mere maps that describe a mere portion of it?

    I am thinking here about the Wall Street “Quants” it wasn’t so much their mathematical models that got them into trouble, but the fact that they had this delusional faith that there models were near perfect reflection of reality.

  3. I don’t think it’s a naive reading, and even though my familiarity with Baudrillard is via secondary sources, I have an interest in Bataille which he definitely shared. Even at that, it’s reassuring that perhaps there’s a principle here which is common to the two perspectives, though I tend to try stay in the realm of a formal outline for as long as I can. The notion of general economy definitely feeds into the above perspective, given that it too was an early form of knowledge or information ecology. I would be interested to hear what your perspective is on this, since it’s not an area of any expertise for me!

    Rick, I say definitely give Ricoeur some time (after your daughter is finished with him, of course) as his approach towards mythopoeisis may be something with which you would have sympathy. Regarding the gap between thought and things-in-themselves, I have been thinking through this recently (as with my posts about Graham Harman, and slowly slogging through Kant). I am coming down more in favour of the distance between mind and world being infinitesimal rather than some kind of infinite, unbridgeable gap such that the essence of things are forever lost to us (which is more Harman’s view). I think there is equally a delusion (as you rightly call it) that thought in itself is sufficient, and this is of course what I am in part aiming at with this post. The first thing I want to address, however, is that even the notion of thought as is found in the data-centric approach is dangerously impoverished. It’s an abstraction to the power of an abstraction, which leads to diminishing cognitive returns. As a consequence these diminishing returns will have a social and political impact.

    I have been looking forward to your piece on Quants after the last one you did on the financial crisis and Milton. Nassim Nicholas Taleb calls the attitude which misled the Quants the “ludic fallacy” (and he himself used to be one!), the notion that international finance is a closed system and that accordingly it can be codified along the lines of game theory (which is sadly not a theory in the scientific sense, but closer to a “wouldn’t it be nice” thought experiment). Anyway, I think you should put up some of your daughter’s art as an accompaniment to a post some day, Ricoeur would probably have loved that!

    • I should probably set up a Kickstarter for myself, cheers Eoin! Until the benefaction comes, I would settle for a position as philosopher in residence, or *gasp* even get paid to research in an academic environment! No, surely the latter is too much of a fantasy in 21st century Ireland.

  4. Pingback: Why wisdom is redundant nonsense in the light of information theory « Wetwiring

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s