Comments on “Klueless Massimo”s Account (Massimo Pigliucci) of the Mary Beard (Twitter) debate

Lesson: some people are in the business of producing high noise/signal ratios in discussions and are really caught.

We need to worry about people who write discourses without having a clue about the subject nor the discussion at hand. Of course it will be read by people who, not having seen the original material, also don’t know what it is all about and will get a completely deformed view of what the true problem is.

***

A philosopher of science, Massimo Pigliucci tried to jump in on the Mary Beard debate solely from twitter comments, not getting what the argument was about, then wrote the highest ratio of BS commentary over original text (only a few tweets). Had he understood anything about statistics he would have worried about the high noise signal ratio he produced.

My point is here. His …

I got angry with him because a philosopher should not engage in strawman arguments. It would be bullshitting [clearly in the Franfurt sense of the word]. Yet he did it and thought he could do it with impunity. My issue with Beard was representativeness: you do don’t mess with people’s perception of the past. He went off about scientism, something I spent my life fighting using probability limits.

But it seems that, to him, statistics and probability is something you get out of STAT101 with standard hypothesis testing. Here is my response commenting on his text. I glanced randomly at his article, finding some substantial error every time.

Core of the Debate

Klueless Massimo wrote all this crap not realizing what the core of my argument was about.

Time Series Exist

I spent my life working with historical data.Time series (longitudinal) data points don’t repeat themselves, by definition — cross sectional does. Quants work with time series. So there is a branch of probability and statistics that Klueless Massimo knows nothing about.

Quantitative Historiography

Klueless Massimo didn’t realize you can use a Brownian Bridge methods: conditional on X finishing here, where could it have been in past? Then think rigorously about it.

Repeating my Black Swan Argument

I wrote a whole book The Black Swan, around the N=1 asymmetry. He doesn’t realize that probability tells you something about what probability should not tell you; probability theory is more rigorous that STAT101 students. See Silent Risk (my technical work backing up The Black Swan)

Genetics

Of course we can use genetics as information. Not the whole story but some points are probabilistically potent: absence of SOME haplogroups/subclades. Why is it we have some subclades (like mine) as Roman army genes in the UK and not others? He may have started as a geneticist but doesn’t know probabilistic methods exist to extract some information with maximal rigor.

Indeed it is a mathematical problem.

Absence of genes along the paternal or maternal lines is largely a probabilistic problem. Klueless Massimo committed the carpenter fallacy.

The fact that we find no Norman genes should lead to the uncomfortable conclusion that there a high probability they were not numerous, as we are discovering about invaders in other places. That was the question Cavalli-Sforza had about the unexpected rarity of Turkish genes in Turkey.

You can ask: given t0 and t1, what could have happened, or what story can possibly explain the deviation, or how far fetched is it? Or how much is it that we can’t explain?

Repeats Beard’s Misrepresentation

Kluless Massimo read the exchange and didn’t see that Mary Beard kept asking me for credentials and discounting my works.

Other

Conclusion

Klueless Massimo needs to understand the debate.