"Believers in scientism flaunt the virtue of science as being a “self-correcting process.” They portray the scientific method as approaching an asymptotic limit, where each correction brings the scientist, while unable to claim absolute truth, at least closer and closer to the truth.
Q: But what if science in practice is more like Brownian motion?
Q: And to what extent do human frailties factor into the view of self-correction?
Let’s look at some recent science news to flesh out these questions.
Proof of biased behavior of Normalized Mutual Information (Mahmoudi and Jemielniak, Nature Scientific Reports, 19 April 2024). For those willing to wade through pages of abstruse derivations, this paper claims to have proved that a widely-used method for assessing community networks is inherently biased and misleading. This is not a small matter:
"Community detection (CD) within social networks has emerged as a pivotal area of research, given its potential to unravel intricate patterns of interaction and group dynamics. It is used in many
disciplines, including biology, criminology, economics, and urban planning, to mention just a few examples. In particular, this topic has also emerged as a critical field in the battle against disinformation. Social networks are often the primary conduits for the spread of disinformation, with communities within these networks playing a significant role in the dissemination and amplification of misleading content. By identifying and understanding these communities, we can gain valuable insights into the dynamics of disinformation spread, enabling more effective interventions. It is essential, as the rapid spread of medical, political, social, as well as scientific misinformation and disinformation are among the greatest civilization challenges of our times."
This means that one of the primary sources of scientific data leading to policies of censorship in social media has come from a flawed method of detecting mis/disinformation!
Consider the irony: NMI (normalized mutual information) has been used to evaluate the source and spread of disinformation – but the method itself produces disinformation!
Q: Have scientists been inadvertently lying about who is lying?"
CEH