Sunday, October 07, 2007

Welcome to the Memory Hole

The Memory HoleIn George Orwell's novel 1984, the memory hole is a mechanism for removing embarrassing documents, previous crimes and inconvenient bungles. Old documents are revised, and the original copies are consigned to the memory hole where "not even the ash remains". It is a mechanism for "smoothing over" the actions of leadership. It is a mechanism of censorship. It is about collective amnesia.

History helps us to understand what has happened so that we can control what happens again. Even if we cannot control it, we will know not to be surprised.

In the words of historian Daniel Boorstin
"Our past is only a little less uncertain than our future, and, like the future, it is always changing, always revealing and concealing. We might better think of Prophecy as History in reverse."

I have been working hard to gather dated records of corruption and poor leadership in science and academia in general, editorial misconduct, the forgotten lessons we should not forget, and a few things that just happen to interest me. They tell a story of meaningless attempts to define and redefine what we mean by scientific misconduct, while ignoring many actual problems and the meaning of science itself. They tell a story of cover-up, deception and the bullying of those who have raised concerns. There are stories of shoddy science, some mistakes, fraud, and a few very evil acts.

Forgetting history is why some wax lyrical about the wonders of a new clinical trials register while ignoring what has already happened. It is why we ignore the fact that we already had a clinical trials register - it was called the FDA (and the MHRA). Those charged with maintaining that previous register colluded (through legislation and practice) to help commercial companies hide inconvenient clinical trials. It is why the wrong questions are asked.

"If they can get you asking the wrong questions, they don't have to worry about the answers."
(Pynchon T, 1995 Gravity's Rainbow. ISBN 140188592)

How many know that Pfizer was involved in human military research in an attempt to produce incapacitating agents that damaged humans and produced (ironically) retrograde amnesia. Does that make us think differently about their more recent actions? Should it? How many recall the radiation experiments, scientific racism, the pseudoscience of eugenics and the illegal experiments carried out on poor, African or mentally disturbed children? How will our actions of 2007 be viewed in 2037?

I did a statistical calculation to work out how many events one would need to collect in order to make sure that the collection included at least one event per day. Assuming that bad events are randomly distributed in the calendar, I calculate that, given 3000 events, the chance that any one day will have no events is around 0.02664% (perhaps a Procter and Gamble statistician would check that out for me). Anyone who has events from the past to add to the database, please send along. So welcome - starting today.

Earlier|Later|Main Page


Anonymous said...

While I don't work for P&G, it is correct that your 0.02664% [=(364/365)^3000] applies to any particular date - in other words, it gives the probability of no events on a single, given, date - say, June 20. The chance that one of the 365 days in a year has zero events, given a sample of 3,000 randomly dated events. This is much bigger. VERY roughly speaking,you can think of it as 365 chances, each of 0.02664%, so that the overall probability of at least 1 day in the year where there are no events is something aound 10%. That's ignoring leap years...Gary

Aubrey Blumsohn said...

That is indeed correct

The relationship is incredibly steep


2000 events there is a 78% chance at least one day will be unoccupied
3000 events there is a 9.3% chance at least one day will be unoccupied
3500 events there is a 2.5% chance at least one day will be unoccupied
4000 events there is a 0.63% chance at least one day will be unoccupied
5000 events there is a 0.0403% chance at least one day will be unoccupied