Proof: Adam Kucharski: May 2025. The book boasts two subtitles – British version has “The Uncertain Science of Certainty”, while the American version has “The Art and Science of Certainty”.
Basic Books Hachtte Book Group. New York.
306 pp: with Notes (& citations) and an index
This is an interesting (I might even call it fascinating) book about reasoning and various tactics that could be used to allow us to approach statements of truth. Several of the author’s anecdotes relate to his work on CoVid in Great Britain and might not be easy for the general audience to follow. Early on, he looks at various methods of demonstrating that a statement is potentially true, including: by contradiction, by contraposition, by construction, probabilistic and “beyond a reasonable doubt”. He points out that proof by assertion and by intimidation or potentially by HIPPO[i] may be dangerous.
Kucharski points out the motto of the Royal Society, “Nullis in Verba”[ii], at least twice. However, he never proposes a definitive concept of what truth actually is, or how we go about “proving” truth.
In Chapter 4 (“Tasting Statistical Tea and Brewing Scientific Beer”), he outlines several of the “stories” about the progression of reasoning and mathematics in a way that is engaging and even entertaining. These stories include the story of William Gossett, who went by the pseudonym of “Student”, for anonymity, when confirming optimal methods of brewing beer in larger quantities for the Guiness family in 1900. He also relates the legend of Ronald Fisher and “The Lady Tasting Tea” (1920s), which gave rise to the concept of randomized, blinded trials, as opposed to observational trials. Kucharski also tells how Fisher in 1925 used Gossett’s table of the of the null hypothesis in fact being true only 5% of the time to show that a difference is probably real. He also tells us how Janet Lane-Caypon used several ways of testing (and confirming) a hypothesis, thus demonstrating that there is not only ONE “right” way to come up with answers to a scientific question
He has an interesting perception of legal definitions of truth AND the concept of how (?socially) important a conclusion from information might be. He talks about what is called “the Blackstone ratio” that “better 10 guilty persons escape than one innocent suffer”, as an example of which type of being right may be more “correct”. In addition, he summarizes the US Supreme Court “Daubert Standard” of five criteria for “validity” in answering a question: Can it be tested? What is the error Rate? Have the observations been subject to “peer review” and published in a scientific Journal? And are the conclusions “generally accepted”?
Chapter 5 (Paradigm Rifts) is the longest and, for me, the hardest chapter in the book to read. Kucharski tries to put several concepts under a single rubric. His premise is that to allow us to make a more correct prediction, we may, like Janet Caypon, need use multiple techniques simultaneously, or in sequence, to evaluate available data so that we may subsequently be able to make a prediction. In order to use a mathematical (statistical) concept, we must first define what question the concept is trying to address. The questions may be about interpersonal relationships, engineering problems, or basic science questions about biology, chemistry or physics among others. Once the question is defined, then one should use an investigative strategy/tactic that might be best suited to answering it. He discusses several of the tools for thinking that are commonly used, including: some classical types of reasoning from philosophy, the ladder of causality, the “wisdom of crowds”, and Bayesian logic. He introduces a concept of triangulation, where one may use several differing tools and then make a “best estimate”[iii] of their intersection, very much as classical navigation uses triangulation to help “fix” the position of a craft on a chart. That several constructs may not always agree in the details means that we may, in the end, have to make an “educated presumption” (hence the term “rift” in the chapter title).
He approaches mathematical methods on viewing statements and ponders whether computers, including Augmented Intelligence[iv] will enhance our ability to determine truth. Unfortunately, Kucharski doesn’t describe his own concept of “truth”. In reality, each of us, the readers, has our own definition of what “truth” is. In the last chapter, he quotes Austin Bradford Hill’s statement, from 1965, that “All scientific work is liable to be upset or modified by advancing knowledge. That does not confer on us a freedom to ignore the knowledge we already have or to postpone the action it appears to demand at a given time”. He does not say that we can prove that science or scientists are always right, only that they should keep an open mind, try to avoid allowing biases to obstruct our view and that they should evaluate evidence, information and opinion as they form their own, personal viewpoint.
The theme of this book does not suggest that science is wrong in general, but that we must continue to be skeptical. There are examples throughout where misunderstanding data and attributing information as fact can lead a thinker astray.
The reader might also be interested in some thoughts on what “science” is: https://winslowmedical.com/archives/226
[i] Highest Paid Person’s Opinion (p127)
[ii] Latin for “Take nobody’s word for it”
[iii] Many people might use the term “best guess”. However, when giving an opinion that may be reviewed in the non-scientific press, use of the word “guess” will potentially be subject to ridicule
[iv] Also called “Artificial Intelligence”
