Propositional and Procedural innovation

May 24, 2020

Knowledge can be broadly categorized into two types: propositional and procedural.

Propositional knowledge is the kind that answers questions about the why. It is fundamental. It talks about the laws of physics, computer algorithms, mathematical theorems. Procedural knowledge answers questions about the how. It is much more practical. It tells you how to cook a dish, build a car, and shoot a rocket into space. All procedural knowledge is based — at least in part — on propositional knowledge. You need to understand thermodynamics to design a combustion engine.

This distinction can be useful when looking at innovation models. Some organizations focus on propositional innovation, some focus on procedural. But, here’s the catch, very few do both.

This goes against the layman’s view of innovation: you invent something and you build a product out of it. According to my view, you can either invent something or you can build it, but it’s hard to do both.

Some of you would have already realized this looks a whole lot like The Innovator’s Dilemma, the hugely influential book that introduced the concept of disruptive innovation. Let’s take a look at the primary example from the book: the hard-disk drive industry.

This industry was dominated by big players, like IBM, who invested hugely in propositional innovation: fundamental breakthroughs in technology that allowed them to build ever-smaller hard drives — that they never sold. It was always smaller entrant firms who took the technology, build the products, and sold them. They applied procedural innovation.

Other examples of this phenomenon include the Macintosh, ARPANET, GPS, or most tech startups’ offerings. There is some core procedural innovation in all of these, but the fundamental science was figured out by others who often saw no profits from their discoveries.

My favorite application of this distinction is when comparing Google and Amazon.

A way of looking at Google is like an unregulated monopoly whose profits are in part used to fund unorthodox side-projects with little marketability. The company is known for its tough hiring process of engineers with PhDs. Of course, not all of Google’s branches engage in this kind of core research, but it’s hardly deniable at least some part of Google does, including their founders.

On the other hand there’s Amazon, an unregulated monopoly whose profits are used to — increase its monopoly? The company’s motto is about the customer experience, how to bring value to the market. Their most successful products are off-the-shelf technologies distributed in innovative ways. A laudable enterprise no doubt, but hardly propositional. Even Bezos’ most exploratory endeavors aren’t affiliated with Amazon. To me, Amazon looks like the procedural company par excellence.


Addendum: Marianna Mazzucato’s book The Entrepreneurial State explains how public investment often focuses on propositional research (my words, not hers), whereas private institutions invest in procedural investigation: go-to-market strategies. Critics of the book often argue that the latter is as important — if not more — than the former. It’s what drives societies’ progress. They are de-facto defending the prevalence of procedural knowledge over propositional.

Addendum 2: If you think about the differences between both knowledge types you’ll quickly realize the line between them gets blurry in some cases. I think it is a useful distinction nevertheless.

Addendum 3: I know there are plenty of organizations that engage in propositional and procedural R&D, not need to remind me. However, I would argue most private propositional research relies on market distortions that provide huge positive optionality. Big pharma, for example, profits from legally granted monopolies over their discoveries.

Hi there 👋, I’m Gerard, Software Engineer and aspirant polymath. I write this blog, build Factorial, and publish the best I read on Twitter.