The Knowledge Machine

How Irrationality Created Modern Science

By: Michael Strevens

Narrated by: Julian Elfer

Length: 8 hrs and 16 mins

Publisher's Summary

A paradigm-shifting work that revolutionizes our understanding of the origins and structure of science.

Captivatingly written, interwoven with historical vignettes ranging from Newton's alchemy to quantum mechanics to the storm surge of Hurricane Sandy, Michael Strevens' wholly original investigation of science asks two fundamental questions: Why is science so powerful? And why did it take so long, 2,000 years after the invention of philosophy and mathematics, for the human race to start using science to learn the secrets of nature? The Knowledge Machine's radical answer is that science calls on its practitioners to do something irrational: By willfully ignoring religion, theoretical beauty, and, especially, philosophy - essentially stripping away all previous knowledge - scientists embrace an unnaturally narrow method of inquiry, channeling unprecedented energy into observation and experimentation.

Like Yuval Harari's Sapiens or Thomas Kuhn's 1962 classic, The Structure of Scientific Revolutions, The Knowledge Machine overturns much of what we thought we knew about the origins of the modern world.

©2020 Michael Strevens (P)2020 HighBridge, a division of Recorded Books

Almost there. Scholarly review.

I wanted the book to succeed, but it fails. Science is distinct from mere engineering precisely because more than prediction and control matter; theory does too. The Iron Rule (i) marginalizes or altogether ignores the role of the creation of new forms of mathematics (e.g. statistics, calculus, information theory, causality, topology etc.),(ii) the development of regimented reference to real objects which expand language & linguistic inferential capacities, and (iii) the generation of "styles of reasoning" (cf. Hacking) which expand what can be talked about and connect experimental phenomenon to theory. All three components are part of scientific communication, not just informal conversation and thing, and any can decisively decide debates. These new kinds of logics, empirically motivated or forced (e.g. Fourier analysis) allow for kinds of debate distinct from before the Scientific Revolution(SR), but which don't suffer the pathologies of "natural philosophy": endless cycles of debate and the generation of distinct schools of thought.
Mathematics can alone constrains the antics of the world and saves us from countless unnecessary experiments, trials and controls. Science is not a pile of data or an Iron Rule "to look", but also a developing theoretical framework increasingly free of Bacon's language idol and replaces old words with an ever-growing empirically rich language that unifies and can falsify in practice as well as experiment.. The generation and effort into regimenting a language with references to actual, causally relevant objects, and the languages role in gluing together phenomenon and providing inferences similar to mathematical ones, capable of deciding debate is ignored.
This shortcoming produces another; how is it that a scientific operationalization (reduction of a hypothesis to an experiment) counts as good or relevant? Pragmatists have an answer: it comes down to generating a desired power over nature, but pandemonium is in the details, and Strevens ignores this mystery. (He almost gets it with his rhyming thought experiment). As a former practicing scientist the creative act of testing a hypothesis often included a concomitant persuasive act of convincing others the experiment is relevant, and this requires a "Style of Reasoning", which is not quite a paradigm, but which requires education and practice to develop. The Iron Rule presumes operationalization is easy, when it is not.

Summary: I wish the inter-workings of the knowledge machine were fully exposed, but much remains a black box.
(1) How does a new mathematics form and become authoritative?
(2) How do new ways of talking generate right material inferences?
(3) What sorts of "Bridge Law" consensuses connect a proposed operationalization from experiment to verbal/mathematical theory?


These are severe shortcomings. The knowledge machine is held together by theory, not merely facts or tricks or hacks (like machine translation or GPT-3, so much AI work, engineering formulae or Sui Dynasty Chinese canal building). Scientific data connect like lego blocks into a structure of stable theories, with a shape that allows for mathematical and linguistic inference and the appropriate placement of new blocks of observations.

Sociologically, the works failure to regard the role of replicability and technology in science. Accumulation of data doesn't just happen, because people re-test general relativity, for example, but because subsequent experiments require scientists to reproduce previous experiments to move forward. In the biological sciences, re-using a strain or protein for further inquiry critically augments knowledge. And, in areas where science translates to commercial applications, successful commercial technologies can settle debates (Heavier than air controlled flight happens, contrary to theorists who denied the possibility.).
As for the historical scholarship, there is nothing novel. But, too little credit is given to Boyle who set forth the rules of communication, and too much to Newton whose remarks were ignored on the Continent, while Boyle was not. Newton broke his metaphysical "shallowness" when it came to light and "absolute space".

A worthy attempt, part of the story for sure, but lots missing. We still don't know how the knowledge machine works.

Beautifully written, fascinating thesis

Strevens must have studied the flowers of rhetoric, because this book is so beautifully written. The thesis is that science is successful and has so much continuity because it demands an "irrational" separation between cold, empirical reporting and the aesthetic judgements of individual scientists (who are nonetheless permitted to express their views informally). Strevens argues that this kind of separation is unnatural and counterintuitive, and that it was only in the peculiar, Newtonian, post 30 Years War milieu of 17th century Europe that such a separation could arise.

Overall, I find the argument compelling, but I'm not expert enough to pronounce judgment on its correctness. I was hoping to hear the author connect his theory with Bayesian models of inquiry, but I think I know how he might do that.

If you think science is a simple matter of falsification, this book will help set you straight. Either way, this was a really fun listen.

The best book I have consumed in a very long time.

Hard to overstate the breadth of so short a work, The Knowledge Machine attempts to explain the rationality, irrationality, objective and subjective pieces that come together to form science and it's demarcation from natural philosophy. A must read for any science or science adjacent person.

Don't buy audiobook

There are many illustrations that are referred to in this book and the audio book loses the richness of the story because of this

Too little material for a whole book

This becomes a bit repetitive and gloomy for a topic that one could cover with some pointed fingers and a lot of respect.

Could be better

Overall, this book is a sweeping account of why science is dogmatic - this point is not up for discussion. I agree with the author and have come to this conclusion a number of years ago, so while the book is interesting, it didn't really enlighten me very much. It's a good overview of the inherent contradictions and problems within the scientific method, however, I'm not sure if it really manages to get beyond the problems posed by Hume, Popper or Kuhn - although it is a far easier read! I did find it frustrating that Schrödinger - who saw the separation of humanities and science as a great backward step (in Nature and the Greeks and Science and Humanism) was not mentioned at all.