Facts are facts, they are records or complexes of data. I think it’s our perception of the conclusions drawn from particular facts as true that decays over time. And unlike radioactive isotopes, sometimes particular truths will return or revive as a result of new evidence or reevaluation of the old evidence and data. I hope to read the book; in the meantime I assume from the reviews that this book applies statistical methods to scientific knowledge/facts. So it is subject to one major limitation of its own method: statistics do not apply to an individual event; to use them otherwise perpetuates the fallacy or flaw sometimes called “the gambler’s paradox,” unfortunately too often done in medicine. Even gamblers’ paradox is a misnomer since a paradox is a contradiction that cannot be explained by the accepted science or logic; it’s really the gamblers’ error. It does not matter how many consecutive times the coin came up heads, unless the game is rigged the probability of the next time being tails is still one in two. A radioactive substance may have a half-life of 3 months. An atom of that substance may not have decayed after 1,000 years. Statistics give you probabilities about future behavior of collections or aggregates of objects/events based on and derived from analysis of past outcomes of similar (but necessarily not identical) populations and events. How good is the modeling and similarity?
From The Scientist
By Samuel Arbesman
Current, October 2012
Did you think a “fact” was something proven to be true? Behold its Pluto-like demotion to merely something we think is true—for now. “Facts” live and die, writes Samuel Arbesman in this nifty introduction to “scientometrics,” the ecology of knowledge; facts migrate through social networks, interbreed, evolve, and undergo mass extinctions. Surprisingly, the growth and turnover of knowledge plot the same predictable curves as bacteria in a petri dish or a radioisotope’s decay. But predictable doesn’t mean optimal: errors are perpetuated, effort is duplicated, and pieces of many a lifesaving puzzle lie buried in widely separated studies. Data mining and crowdsourcing are beginning to leverage computers to extract more value from what we don’t know we know.
Arriving just as a seemingly solid fact—random mutation—is shaken by the epigenetics earthquake, this book is well timed. And it’s full of fascinating tidbits: DNA-copying enzymes make the same mistakes as medieval scribes! Unfortunately, the book is marred by clumsy writing and careless editing, overlooking even the odd mathematical howler: included on a chart of technological capacities that double in mere months is “DNA sequencing, dollars per base [pair].” That’s stated backwards—it’s the bang, not the buck, that doubles.
The publisher’s promotional blurb: