Today seems to be a good day for empirical philosophy. I wrote this section after lunch, so its freshly converted from LaTeX to HTML, and thus tentative in character. This section is supposed to come after the part on blackboxing, which in turn comes after the prelude to the chapter. Or, for those interested, just read it :D. The philosophical problem is the zeitlichkeit of black boxes, how they invent their own time. The empirical problem is to describe what quantitative social sciences are really doing. Hence, empirical philosophy!
Blackboxing and time
The historicity of a black box is not necessarily so that going back in time means that they are more open. This is sometimes true for machines as they are invented, where you would usually travel back in time to find the origin of of a technology in a research lab. Blackboxing is however ‘anti-genealogy’ because when they break down, when they fail, they are as easily reversed, perhaps to an even more primitive stage as when they were invented. Moreover, due to the implicated actualism, blackboxing as a process never ends. Machines, concepts, methods and procedures have to be constantly used, maintained and upgraded to keep working, even though this is usually invisible to the ‘ordinary user’. Even though we usually never think about it, there are technicians employed around the clock to keep our mobile telephony working, scientists in labs doing everyday research in standard procedures to keep our facts straight (Landström 1998), and teachers in schools repeating the instructions of grammar to pupils five days a week.
On a more profound ontological level, this goes as well for actants, entelechies and hybrids (these concepts will be explained below):
1.2.8 Every entelechy makes a whole world for itself. It locates itself and all the others; it decides which forces it is composed of; it generates its own time; it designates those who will be its principle of reality. It translates all the other forces on its own behalf, and it seeks to make them accept the version of itself that it would like them to translate. (Latour 1988: 166)
It is thus imperative not to study blackboxing in conventional linear time frames. The SOM-institute is actually a good example of this. When going from the 1999 survey, as described above, to the 2010 survey, by looking at the methodological chapter, the box is more open than roughly a decade earlier. The method documentation chapter from the 1999 survey is six pages long (Lithner 2000), the one from the 2010 survey is 33 pages (Nilsson & Wernersdotter 2011). At least in writing, it seems to take more words to account for the same type of survey. Word quantity can however be deceiving, instead the openness of a black box is where to look, to the amount of work needed to make it work:
During recent years it has become more difficult to reach a high response rate, and moreover the SOM-surveys have increased in size. This year’s number [response rate] hence must be regarded as high. In the 1999 survey the response rate was 67 per cent, which is on the same levels as the average fourteen nationwide SOM-surveys made so far along /…/ (Litner 2000: 398)
And for the 2010 survey:
From the 2000:s the level [of response rates] has however gone down. If the average up until 1999 was 68 per cent, the first decade of the 2000:s was 63 per cent. The 2008 survey became the first one to go below 60 per cent, which was also the case in 2009. This year’s survey did however reach 60 per cent again. (Nilsson & Wernersdotter 2011: 557-558)
Then another seven pages are spent analyzing who and why some people are not responding to the questionnaires. What was unproblematic a decade earlier, is debugged, analyzed and progressively made to work again. Black boxes are deceptive in this way, they only withdraw when they function as they were supposed to. Layer by layer (or rather, variable by variable) the missing respondents are localized. One such example is the group of young men:
The lower response rate among the young groups is especially clear among young men. For men in the ages 20-29 years the response rate is 36 per cent, compared to 48 per cent among women in the same age. (Nilsson & Wernersdotter 2011: 562)
The traditional way of thinking scientific discoveries is that phenomena were discovered, then, once they are discovered, they were there all along. The standard way of thinking innovations, is that once someone invented a technology, a method or a formula, it is there for us only to use (if we just can afford it or know how to use it). This, however, only works for studying ready-made science. When analyzing science in action, time is relative to the speed of black boxes. Intercontinental telephony works with the speed of fiber-optic cables, as long as they work. But when they break down, their speed is reduced to the time it takes for technicians to localize and mend the failed components. The quantitative social sciences, I shall argue, are no different in this respect. And as we saw in the prelude of this chapter, they are no less ‘cutting edge’ than the latest cellular networks.
Off topic, kind of: So what then is cutting edge technology? Is it the latest iPhone or the skyline of Dubai? Well, so may be, but it is also your kitchen knife (pun intended), the shoes you wear, even though they might be really old, or the turbofan jet engines on your recent flight, even though its 1960:s tech. The Minicall, the fleece sweatshirt or dialup modems are however not cutting edge. They came after jet engines, still, when nobody uses them, they are no longer in networks, so they retire as obsolete artifacts in museums and deep in your closet.