It took me fourteen years to read this book. Old bus tickets used as bookmarks tell me how far I got each time, but this time I finished. It was distinctly underwhelming.
There are lots of interesting insights in the book, but it suffers in the end from trying to explain everything. In the middle it suffers from having been written in the 1950s, so there are page-long paragraphs referring to scientific debates that have been long forgotten and the contemporary literature cited, with its casual references to lobotomies, torture of animals, and racism, is a good reminder that the 1950s were not the ethical highpoint of the 20th century. These historical vignettes are interesting, in a sense, and even serve to underscore some of his points, but the discussion often assumes a context that no longer exists and the logic rests on points that have long since evaporated so the reader has to fill in the gaps best they can.
It suffers throughout from being written in a style that belabours every single point and, towards the end, becomes increasingly repetitive. It's a shame, because when you work out what he means - which sometimes requires reading the same sentence five or six times (over a period of fourteen years) - you realise it could have been said much more simply and to greater effect. When he does unclench and lets the words flow, the effect is there and he can be almost memorable, but the moments are few and far between and the high points are never that high.
The argument, to paraphrase, goes something like this: induction is logically impossible, but we know stuff. Actually, I, Polanyi, really do know stuff, I'm sure of it. I swear on my head that I do. Mumblemumble. Therefore, god exists. The last bit, reached by an extremely dated critique of evolution, was an obvious overreach even at the time and he knew it. He fudges the whole thing by not saying it directly but it's clear what he means. The early part of the argument is stronger, but the whole thing is interesting.
All of which is a shame because a lot of what he says (before he goes off the rails in the final section of the book) will likely still resonate with the working scientist. He gets deep into what it means to *be* a scientist and places the scientist himself (it was the 50s, so it's always him) at the centre of science which sounds right to me (not the pronoun thing). He doesn't quite finish the job off though. He talks a lot about scientists individually, but less about scientists as a community. He also dodges other sources of interesting messiness. Having neatly pierced the idea that science can be purely objective, the messiness starts to ooze out and he recoils somewhat from the subjective chaos he's unleashed. He still talks a lot about the beauty of science and discovery and his scientists sound like warrior saints, aiming at a higher ideal and chasing the grail. To be fair, they are only aiming, but what happens when they miss, or stop aiming so high?
There are also passages on freedom, society, science, authoritarianism, and totalitarianism whose time has come round again and feel vital and fresh. He was writing in the aftermath of the Second World War with an active totalitarian state to help make his points. It was critically and obviously important then, and continued to be so even if we forgot. There's particular pertinence in his discussion of how science conceived purely for humanitarian ends (and not as an end in itself) can be hijacked by ideology. He was talking about Marxism, but you can substitute the world view of the past 40 years that human flourishing is a consequence of a strong and growing economy and therefore places economic pursuits above all others. In this view (as in the Marxist view) science is forced in a more technological direction by the insistence on immediate "impact" and "pull through" while also being criticised for anything that does not conform to the ideological view. In the final stages - which the US seems to be entering right now - scientific criteria of right and wrong are wholly replaced by ideological criteria. The US might already be there, but everyone else is cheerfully following along.
He also talks about thinking machines a lot. Technically, it's rather naive stuff - the machines weren't doing much thinking in the 50s - but through it you can trace the beginnings of an argument against, say, AI and its use in science, that's more sophisticated even in its nascent state than much of what passes for thoughtful discussion on the matter today.
In short, a difficult but occasionally worthwhile read.