Guess values, Priors and Science

Wikipedia defines Guess value as

“In mathematical modeling, a guess value is more commonly called a starting value or initial value. These are necessary for most optimization problems which use search algorithms, because those algorithms are mainly deterministic and iterative, and they need to start somewhere.”

I am not an intuitive mathematician (nor an unintuitive one for that matter), but I have appreciated the importance of good guess value or nominal value developing software that uses computational geometry.

Yesterday I read this excellent long piece- The 60-Year-Old Scientific Screwup That Helped Covid Kill . Personally, I had been convinced by the Aerosol spread theory since I read these two pieces back in the Summer of 2020.

How Coronavirus Infected Some, but Not All, in a Restaurant

How coronavirus spread from one member to 87% of the singers at a Washington choir practice

As a result, I have spent the better part of the previous 14 months wondering “why is the scientific community so slow in accepting potential aerosol dispersion of coronaviruses”. They say science progresses one funeral at a time, but since Jan 2020 we have had far too many funerals that should have sped up the progress. Ever since I read this piece — The “noble lie” on masks probably wasn’t a lie I had been spending limited time I had, browsing through old papers on the spread of respiratory illnesses to disappointing results. The wired piece helped deepen my understanding of how we may have got here. Where the 5-micron boundary for aerosols came from is an extremely fascinating story. Maybe we can expect some Netflix documentary or a long-form book on this issue in the near future.

This entire episode appears exceedingly similar to the Dietary Fat is the villain dogma. Gary Taubes in his books — Good Calories, Bad calories and A Case against Sugar, has described this debate extremely well. As a video suggestion, I would recommend the youtube channel — What I have learned. In retrospect what seems appalling is how the scientific community basically accepted the faulty and weak fundamentals of the Fat theory — as initiated by Ancel Keys and propagated by hundreds after that. Ditto for the demonization of cholesterol and numerous other food items — including milk.

From these two examples at the very least, I would conclude the disproportionate importance of Priors and the outsized impact they have on the journey of academic and peer-reviewed science. Hopefully, the pandemic will correct some bugs in this mechanism, though this is by all accounts a slow and arduous process.

Incidentally, I was chatting with Kushal Mehra about his latest discussion with Shrikant Talageri and Kushal pointed out that he thinks the 1500 BCE dating of Rgveda is also one such Prior which has had an outsized impact on the journey of the Aryan Debate. Prima facie I found this point fair — thought I must say, the 1500BCE date has held up quite well over the centuries (especially since the recent genetic results). But has the date 1500 BCE anchored the research around it, making 1500 BCE appear the best fit for composition of Rgveda and the beginning of consolidation of the Arya into the Indian palimpsest ?

Post script:

I know this place has had a lot of AIT/OIT pieces including this one by me and lot of readers might be fed up by it (even I am to an extent). This isn’t aimed at AIT/OIT but is merely a passing reference to it.

I don’t know how Max Mueller and others came up with the date myself — but it would be an interesting story to research notwithstanding the current research. I would also highly recommend Razib’s podcast with Mallory which delves into the history of the larger indo european question.