[Klik hier voor de Nederlandse versie van deze tekst]
To a certain extent, science is the application of technology that helps to arrive at new knowledge. First of all, scientists use instruments to experiment with and to measure. But also, technology organizes the work of scientists. With advances in technology, the character of science will change. But how far can science change without ceasing to be science? This question requires a good understanding of technology, science and their interaction and a critical assessment of new technology’s possibilities and temptations.
Sociologists of technology talk about the affordances of technologies, which means that the physical properties of a device or design allow specific human actions. Affordances say something about how we interact with the devices we use. A technology seems to press certain buttons in our brain on or off so that certain actions are pre-sorted.
When Harrods introduced the first automatic escalator in 1898, the designers were amazed at the people who did not ascend the stairs but stood still, enjoying the escalator’s unexpected affordance. Many of us start scolding other road users as soon as they drive a car. The affordance of the car seems to bring out the worst in us. Or maybe not even the worst: owning a weapon also stimulates gun use, which is even worse.
Affordances play a role in almost all human activities. For example, if I had to write my dissertation on a typewriter, I would never have been able to obtain my PhD. I don’t even think I would have got my master’s degree. A digital word processor allows a different kind of academic work than an analogue typewriter, and my brain simply turns out not to be empowered with the requirements of a typewriter: I have to be able to scroll my texts; the spelling checker must be able to correct all my writing errors; I need to see a thesaurus if I can’t come up with a word. Indeed, I know from experience that the word processor enables a different kind of science.
Technologies are not neutral instruments; they change how we act and interact with each other. Affordances are about the impact of a single device on a person. But we are surrounded by countless devices: we are part of a network of people and things, where things respond to people’s actions, and people respond to the affordances of things: people and things determine each other’s freedom of choice.
This mutual influence of people and things is the starting point of actor-network theory (ANT). In this theory, both people act and form a network of actors together. For many, ANT is the theory you love to hate. Pretentious, vague and, most importantly, it denies the fact that only humans can act intentionally; the actions of a device are not the same; such a device doesn’t know what it is doing.
This is true. But ANT also focuses our attention on several essential issues. Firstly, it concerns how people and things become embedded in an interplay of practices that become so attuned to each other that other practices become virtually impossible. This means that choices from the past determine the choices we make now and in the future. New technologies always build on existing practices, which is why, for example, it is so difficult to introduce successful sustainable technologies − they do not fit with current unsustainable practices.
With ANT, we better understand how practices change with the introduction of new technologies, which usually happens without considering the moral implications. Consider, for example, how everyone looks at their mobile phone all day long. The mobile phone makes us part of a socio-technical network of apps and apps: the device shapes how we interact with others.
ANT shows that our decisions are not only made by our brain, but we have externalized many decisions, such as culture and institutions. Thus, Technologies largely determine our actions and our conscious thinking − perhaps to an even greater extent than our brains do. And that’s a good thing because technologies are often much smarter, stronger and more accurate than ourselves. As I wrote, technology can be seen as part of the intersubjective superorganism. Part of our thinking and doing lies outside of our brain, so we can do much more than our limited physical and cognitive capacities permit.
With new devices, our superbrain is getting even smarter. No wonder our values and meanings are changing. We can do more, see further, and make decisions we could not make before. But we also ‘unlearn’ values and meanings because technology makes them meaningless or because technology sends us in a different direction. Now that tires barely puncture, you no longer have to learn to fix a tire as a child. Now that the casual dress code in the office has become commonplace, you no longer need to be able to tie a tie. Don’t worry; there are always tutorials on YouTube.
What do these insights mean for the work of a scientist? What affordances does she have to deal with, and what network is she a part of? A first example can be found in the availability of knowledge. Until twenty years ago, most of the knowledge was to be found in books and academic journals that were by no means available to everyone. Scientists who worked at a university with sufficient money for a good library or expensive subscriptions had no problem. This inequality was exacerbated because precisely these universities were best regarded and largely determined the reputation of an individual scientist.
In short, the knowledge gathered by science was only accessible to a small group of people (mostly white men with grey hair and glasses). The most renowned universities retained their top position without any effort. They were at the top of the rocks that had to be climbed to be seen as an important scientist. Those who got to the top of the rock determined which knowledge was most valuable.
Thanks to the Internet, scientific knowledge has become accessible to all scientists. Journal subscriptions are still ridiculously expensive, but open access or otherwise illegal download sites such as SciHub ensure that knowledge can be shared universally. Knowledge has actually become something that can be retrieved by all scientists in the world, with which the essence of science can be pursued even more strongly.
But there’s more. What is seen as the most valuable knowledge no longer depends on someone’s reputation. Instead, rankings have been introduced that are more or less objective. Most striking is the H-index, a simple number indicating how many articles a scientist wrote are cited at least as often as that number.
I do not want to say that this index is the measure of things. For example, my H-index is almost twice as high as that of Nobel Prize winner Peter Higgs; you know about the Higgs particle. I have never won a prize other than a bottle of beer at a pub quiz.
But that’s not the point. The criteria of the H-index are the same for everyone; it doesn’t matter which club you belong to. It is a much more democratic measure of the quality of a scientist than the arbitrary criteria before.
The story of the impact factor for scientific journals is comparable. This indicator shows how often an article from that journal is cited. This leads to a clear overview that enables every scientist around the world to estimate the status of a journal. As before, she does not have to spend years in a research field to know which journal is best read. Here, too, this leads to science that is more inclusive, more global and more democratic.
But while a new technology ensures that certain values are realized, and problems are solved, other values and problems come back. With our cell phones on the table, we are always ‘on’, constantly focused on the world that lies beyond our vision; this goes to the extent of real, direct access through eye contact or a good conversation. It’s easy to retype a word on a computer, but a day’s work may have been in vain if you forget to save your work. New errors have replaced the errors that a digital word processor fixes. Scientists and editors have become very creative in artificially boosting their H-index and impact factor, respectively. The perverse incentives that each indicator has to do their job.
As such, it’s no wonder that 15 years after its introduction, the H-index is taken less and less seriously. The NWO, the body that distributes the most research funding in the Netherlands, does not even allow these kinds of quantitative measures in the assessment of the quality of a researcher. This critical approach is justified; with every new technology, we need to question which practices, values and meanings are pursued and which can no longer be pursued. Too often, innovation is blind to these questions question. There is simply a promise that people would like to believe in, such as the promise that a digital word processor makes writing easy or that increased computing power provides more knowledge. These kinds of promises should be treated with scepticism: we should always ask ourselves whether these promises correspond to what we want and need.
This last point is especially important when considering the promises put forward around big data analytics. The presence of an infinite reservoir of information and the unlimited computing power of computers allows the determination of relationships that do not require any human interpretation. All those cookies you accept mindlessly make it possible for the sites you have visited and the links you have clicked to be stored somewhere on a server. Your profile is compiled from this deluge of data.
Whether that profile matches your psychology is irrelevant. The results of the big data operations are correct because they are assumed to be accurate. Because the amount of information is unprecedentedly large, there is no other way – at least, this seems to be the idea − that it must lead to the correct inferences. The human brain is bounded in its rationality, as Herbert Simon stated, and has difficulty processing large amounts of information. Causal relationships drawn are no more than heuristic short-cuts: wouldn’t actual causality presuppose that all information is included?
No, what then remains is a meaningless mash that still needs to be interpreted. Big data suggests that science is little more than inferring relationships from information, as a rule of thumb: the more information, the better. But science isn’t just about collecting data. An interpretive step taken by scientists remains necessary, as this interpretative step adds meaning to the observations. With that meaning, we can understand what is happening (or think we can understand…). Big data adds an extra layer of noise scientists have to extract patterns from.
Adding meaning is indispensable because the reality around us is an amorphous whole. No device could give an accurate picture of that reality. Instead, we use devices to structure that reality. Similarly, there is no such thing as neutral information that can serve as input for big data analytics; this information has already been filtered. As the now well-known examples of discriminatory algorithms show, self-driving cars are more likely to hit a black person than a white person, and friendly chatbots lapse into racist trash talk in no time.
The advantage of devices is not that they show reality as it is but that they show everyone the same image of reality. It helps us rid truth claims of subjective elements. Where in the past, the truth was proclaimed by prophets − you couldn’t get it more subjective than that – we now have come to rely on technologically mediated measurements that do not belong to any person in particular.
I can never be sure that what I see looks the same to someone else. But a telescope always shows the same picture. In effect, this means that science can be seen as applied technology and not − as is so often thought − the other way around: the information that forms the basis of our knowledge must always be mediated by technical instruments. Whether that is a simple yardstick or the large hadron collider from CERN, these instruments make observations independent of the observer, so these observations become comparable. This comparable information enables science to bring order to science by recognizing patterns and drawing meaningful connections.
Ultimately, science is all about the application of objective technologies and the interpretations of subjective scientists. This combination ensures that science is more than ‘just an opinion’. This does mean that technology has a necessary, but not a sufficient, role. New technology will never be able to lead to a science without scientists. Instead, new technology should lead to more accurate tools and more accessible communities of researchers. So let us rejoice in the possibilities of new technology that makes more information available to more people, but let us not be distracted by the affordances of a technology that leaves us with knowledge which has no meaning.