This Idea Must Die is a compilation of answers to a question posed by Edge magazine in 2014 to leading scientists and academics: what scientific idea is ready for retirement?
The question is predicated on the often true assumption that new scientific ideas often gain traction at the expense of the old. Judging from the numbers, it seems that scientists in diverse disciplines, such as cosmology, theoretical physics, cognitive science, science studies, psychology, mathematics, biology, the social sciences, philosophy, and many others agree that their respective fields contain much antiquated cruft to be reconsidered.
A broad summary of some controversial scientific ideas that were put forth for retirement include:
- The need for “complete understanding”, leading to pushes for theories of everything, such as string/M-theory
- The idea of the anthropic principle – that the laws of physics are consistent, pre-ordained, or unique, and therefore tuned to support existence
- The idea that certain scientific precepts such as parsimony, falsifiability and replicability should be applied automatically to every scientific venture, especially in a causally complex world
- “Essentialist” analytical lenses that taxonomize based on “innate” criteria, such as race or gender
- Frameworks that rely on “culture” as the primary medium of memetic propagation, without explaining its mechanics (still quite controversial as the alternative is some form of evolutionary psychology)
- Essentializing differences between animals and humans, such as models of animal behavior that characterise them as mindless automata, or that humans are evolutionarily unique because of adaptations like language and tool-making
- Any distinction between “nature” and “nurture”, or the idea that environmental factors can be taken in isolation from genetic ones when analysing traits and behaviors
- The idea that human intuition is consistent across culture, and can therefore be used as a basis for philosophical arguments (i.e. the “intuition pump”)
- Current methodologies of scientific research, such as peer review, current “results-oriented” funding models, large randomised controlled trials, liberal application of arbitrary statistical rules like p-values, regression, standard deviations etc
- Taking statistical results from analysis of datasets as conclusive of a causal relation (such as in Big Data), rather than using those results as a starting point to conduct empirical studies to demonstrate causality
Many of these answers are not conclusive, and often, different answers may contradict each other in small or large ways. The cosmology section has scientists arguing for and against string theory, for example, with some using string theory’s lack of options for empirical falsifiability as a basis for its rejection, with others declaiming the limits falsifiability when applied to this particular case.
The anthology is rich, detailed, and the provocativeness of its question has allowed for a wide range and variety of answers across many fields. This quality makes it an often arresting read. If I had to suggest how to improve this book, however, I would say that it would benefit somewhat from a degree of curation. Not in the sense of content editing, because that would detract from the richness of the material, but in the sense of providing context and structure to the various answers.
The book is already halfway there in that respect. There is a suggestion that the answers are roughly grouped by theme and discipline. Where this book could have done more would be to create explicit categories for certain groups of answers, in order to increase clarity for the reader. And in those categories, it might be beneficial to have a paragraph or two explaining the broad context and schools of thought underpinning those disciplines, in order to give a little bit more context to the layperson.
Despite the above, however, I can see how more curated approach like the one I described might not be feasible from a process or person-oriented standpoint. Answerers may take issue with having their views categorized in ways that they do not like, or having their answers co-opted into a meta framework without their explicit consent. As it is, positioning without commentary may be the best that the editor can come up with to preserve the original texture and richness of his sources.
I give this: 4 out of 5 wrong-but-useful ideas