Bringing technological determinism back in

Information, in theory and practice

Technological determinism gets a bad rap, probably because it’s obviously false. Technology allows us to do many things that we couldn’t otherwise do, but it doesn’t determine which of those things we actually choose to do.

But the critical move to deny technological determinism has a problem. The internet is the most important communication technology in the history of humanity.

This is a matter of faith, at this point. And I admit the base rate (ie what is the probability that I am in fact living in this critical juncture versus what is the probability that I’m just an academic who thinks his object of study is, like, really important) is low.

But I see the internet as a major actor in the politics of the past two decades. Its impact on economics and especially culture are undeniable. We haven’t reached a steady state, and I don’t think we will for decades. This belief is at the base of much of my work.

The way a given technology is used, or how it operates in the world, is determined by existing power structures or by the ideologies of the users. If our goal is to explain “what will this technology do?”, it does make sense to think about who has the power to use that technology, and for what.

But the internet has already upended power structures, toppled regimes and destroyed industries. The mutually constructed internet is impossibly complex; no one understands how to map causes and effects. The most honest way to think about the internet is as a living being, an alien or a god.


Another branch of social science (arguably among the most powerful) is completely technologically determinist. Rational choice-style formal theorists modelling communication see in new communication technology a shift in the capacity of information to be spread more quickly, more broadly, or more (or perhaps less) credibly. Consider this assumption in Andrew Little’s excellent 2016 article in the JOP:

When interpreting the model, I assume that one effect of better ICT is increasing the tactical precision of the signal, that is, giving the citizen more information about how others will protest.

The credibility point is the most important for many of these theorists. The perfect example came in Turkey’s 2016 attempted coup. President Erdogan used Facetime to credibly and publicly signal to the nation that he had not been captured, potentially turning the tide of the coup. In an environment of uncertainty, the technological capacity to broadcast a single credible signal served to co-ordinate the pro-regime camp.

Of course, the battery life on Erdogan’s smartphone if he were holed up in a basement in Ankara shouldn’t change the coup dynamic at all. The capacity to broadcast a national signal used to be far more correlated with success on the ground; this is why coups tend go straight for the television or radio broadcasting station. If everyone has a smartphone, the capacity to broadcast a national signal contains almost 0 information about regime strength. My interpretation (and I really don’t know much about this case, please correct me if I’m wrong in some key detail) is that this was so effective because of heterogeneity in the understanding of information technology of the population.

I think that we are currently experiencing an epochal disequilibrium in this key parameter. Modern information technology has unfathomable effects on the flow of information in the world.

The formal rational choice approach involves stripping away the irrelevant epiphenomena to examine the central political dynamics. I love this approach in the context of the internet because it takes the technology seriously.

Like empirical research, the task of formal theory becomes more complicated if there exists significant heterogeneity in the actors in the system. Several theorists have also recently devoted attention to this problem, and some have developed much more rigorous models of the dynamic I describe below.

We are exiting the era of mass media, an era in which the intuitions about information of many currently living people (and most of the most powerful people) were shaped. I believe that digital literacy is a useful first approximation, but also that the heterogeneity is deeper than we have yet been able to measure.

Humans are constantly learning about the world, but it’s much harder to learn how to learn. Knowing how to learn makes a parameter out of our responses to the stimuli we receive from the world. It makes sense that this process would slow down as we age.

Think about how you learned about gravity. Not 9.8 whatever, but the intuitive calculations your brain makes when it sees something falling. As a baby, you watched a bunch of things and determined how that process works: things fall. If you saw a balloon floating, you develop a more sophisticated things-fall theory; airplanes, another wrinkle. But if you saw, at age 2, a real-life UFO, one which could accelerate, slow down and then hit hypersonic speeds, this would merely be evidence for the theory of how things move. As adults, with fully formed theories of how things move, seeing this UFO would force us to either conclude that a) aliens are real or b) there was an error in our perception. From a Bayesian perspective, the latter is dramatically higher likelihood.

The problem is that the internet is aliens.


What many older Americans see and post on Facebook is literally nonsense. There are threads of failed communicative acts; people don’t understand the structure of the platform, why certain words appear in front of them or what happens when they press certain buttons. Anecdotes abound of older users creating a new account whenever they get logged out, or of posting private messages on public pages.

Perhaps this problem can be ameliorated with better platform designs, ones more interested in improving the well-being of their users than in making money. The anti-technological determinists are clearly right about this, and from a harm reduction standpoint, there’s plenty that could and should be done.

But the internet has shifted one of the deepest parameters in how humans understand information. I can’t understand, let alone describe, the extent of this change. But I think the formal literature is on the right path: the internet has already shifted some deep parameter about the amount of information contained in a signal.

In the broadcast era, the existence of a signal (here, a video) of Erdogan saying that he still had power necessarily contained the information that he was at least in control of the broadcasting apparatus and thus that the coup had not been entirely successful. In the modern era, this signal merely contains the information that Erdogan had power in the sense of battery life of his iPhone.

No, quitting iPhone apps doesn't help save battery, says Apple | iPhone |  The Guardian

The situation is similar for a broad range of digital signals. In 1950, if you received a signal consisting of text from a Nigerian prince who wanted help smuggling his fortune to America, you might well take it seriously. The time and materials needed to transmit such a signal made it costly, and thus necessarily containing the information that someone paid a price to transmit that signal. In 2010, on the other hand, the existence of ctrl + c, ctrl + v, and email lists means that this signal in fact contains zero information.

If all of the actors were uniform, the receiver of the signal would be aware that the signal was costless and discount it accordingly. The pace of the development of ICT is currently very fast, however, fast enough to intersect the human lifespan. This means that there are many older people in positions of relative or absolute power in Western society who have no chance of “grokking” this development.

Young people’s minds are simply more pliable, and many of the younger generations can remember only the internet. This doesn’t mean they’re “digital natives” and thus computer whizzes; many younger people are dramatically better at typing on a smartphone than on a laptop, for example, and are if anything less likely to understand the architecture of an operating system than were people who could only operate machines without slick UIs.

But they (we) learned how to learn in age of digital reproduction. Older people, regardless of how skilled they are, are unlikely to be able to update this base parameter. I thus anticipate a few more decades of disequilibrium, one in which the older generations still running most countries will be continuously more confused by the new informational content in digital media objects.

The recent Boog/Antifa conflicts show this dynamic isn’t restricted to Turkey. Regardless of whether your greatest fear is left- or right-wing street violence, think about how much video footage of it happening in the US used to exist. There’s now countless hours of it.

The absolute amount of political street violence has increased, but not by much—not by nearly as much as the supply of videos of it has. It’s possible (but not trivial) for me to discount the information in these videos by thinking through the fact they exist because of a technological shift in the capacity to produce video. But for older people who have a job other than “political media scholar,” the plausible reaction is to think that threat posed by their most feared enemies is far greater than it currently is.