by Jason Crawford · October 13, 2020 · 3 min read
Progress is messy. On the whole, over the long run, the advance of technology and industry has improved life along almost every dimension. But when you zoom in to look at each step, you find that progress is full of complications.
Intensive agriculture achieves high crop density (which is good because it improves land and labor productivity), but this takes fertility out of the soil faster and makes fields more susceptible to pests. To solve these problems, we then need things like artificial fertilizer, pesticides, and improved crop varieties.
Burning lots of coal provided us with warmth in our homes, with industrial processes such as iron smelting, and with motive power from steam engines. But it also caused air pollution, blackened our skies and deposited soot on everything—including our lungs. London in 1659 and Pittsburgh in 1861 were both likened to hell on earth because of the oppressive clouds of black smoke. Improving air quality has been a long process that included moving coal-burning away from human habitation, switching to cleaner-burning fuels such as gasoline and natural gas, and the introduction of electricity.
City life provided people with many opportunities for work, commerce, and socialization; but crowding people together in filthy conditions, before sewage and sanitation systems, meant an increase in contagious disease and more frequent epidemics. In the 1800s, mortality was distinctly higher in urban areas than rural ones; this persisted until the advent of improved water and sewage systems in the late 1800s and early 1900s.
Automated manufacturing in the factory system was far more productive than the previous system of home production or “cottage industry”. In that system, a weaver, for instance, would perform his craft at home, using his own loom; keep his own hours; and be paid by the piece. The factory system created a need to commute, and resulted in a loss of autonomy for workers, as they could no longer set their own hours or direct their own work. This has mostly been a permanent change, although recent decades have seen a slight reversal, as the Internet enables flexible “gig” work, lets some employees work remotely, and makes it easier to start small businesses.
Nor can we, in every instance, fall back on “revealed preferences” to argue that people actually want the new thing, since they chose it: sometimes industrial shifts take away old options, as when weavers could not compete against the power loom; or technology runs ahead of governance, as when coal began to pollute common skies.
So technological changes can be an improvement along some dimensions while hurting others. To evaluate a technology, then, we must evaluate its overall effect, both the costs and the benefits, and compare it to the alternatives. (One reason it’s important to know history is that the best alternative to any technology, at the time it was introduced, is typically the thing it replaced: cars vs. horses, transistors vs. vacuum tubes.) We must also evaluate not only the immediate effects, but the long-term situation, after people have had a chance to adjust to the new technology and its ramifications: mitigating its downsides, working out governance issues.
Conversely, a common error consists of pointing to problems caused by a technology and concluding from that alone that the technology is harmful—without asking: What did we gain? Was the tradeoff worth it? And can we solve the new problems that have been created?
This is well-understood in some domains, such as medicine. Chemotherapy can treat cancer, but it can also give you nausea. The unpleasant side effects are acceptable given the life-saving benefits of the treatment. And there are ways to mitigate the side effects, such as anti-nausea medication. Nausea might be a reason to avoid chemotherapy in a specific case (especially since there are alternative cancer treatments), but it’s not a good argument against chemotherapy in general, which is a valuable technique in the doctor’s arsenal. Nor is it a sufficient argument even in a specific case, without evaluating the alternatives.
Other domains don’t always receive the same rigorous logic. The argument “pesticides aren’t necessary—they’re just a response to the problems caused by monocropping!” is analogous to “anti-nausea pills aren’t necessary—they’re just a response to the problems caused by chemotherapy!” Perhaps—but what problem is being solved, and what are the alternatives? There are alternatives to monocropping, just as there are to chemotherapy—but just because alternatives exist doesn’t mean they are viable in every (or any) situation. A case must be made in the full context. (Understanding the context is part of industrial literacy.)
That’s not to say that we can’t identify the drawbacks of pesticides, or monocropping, or chemotherapy, or coal, or factories. We can and should, and we should seek better solutions. No technology is sacred. Indeed, progress consists of obsoleting itself, of continually moving on to improved techniques.
But if you want to criticize a technology, show that there is a viable alternative, and that it doesn’t sacrifice important properties such as cost, speed, productivity, scalability, or reliability; or that if it loses on some dimensions, it makes up for it on others.
« “The growing gap between our power and our wisdom” Video: Manjari Narayan (@NeuroStats) on statistics past and present »