The sad tale of Jacques Benveniste

The following was a comment that was intended to go on this post at Marchway Memoirs. It got a bit big, so I posted it here instead.

Oh yeah, Jacques Benveniste. Here we go. Benveniste was a kook bloke who was a scientist. And, of course, like most scientists, he had his human side...

I have the impression you think scientists are derived from some Vulcan-like race of purely logical beings. We’re human, with human faults.

...and his scientist side. And let’s face it; because scientists are human, they are prone to error. And emotion, unlike Vulcans.

On the whole, scientists are disciplined enough at work to let the cold, hard facts get in the way of petty human emotion. But in rare instances, they do it the other way round,..

Scientists are human. I’ve never met [James] Randi, and have no wish to, because I'm human and one of the things humans do is take a dislike to people.

...and when they do, it can really blind the practitioner to the facts. Particularly if the practitioner is displaying very strong emotions such as love, hunger and greed.

Or dislike.

James Randi is one who certainly inspires very strong emotions from both scientist and non-scientist alike. He's not a scientist. But his knowledge of the scientific method appears to be beyond question.

Until now.

And here's the question: Is it possible that dislike of the man could cause one to temporarily take leave of their senses outright and completely miss the point of a particular story?

This blogger thinks so. But let's be honest here - some of the criticism of Randi makes that very allegation against the man himself...

It's not always purely on logical grounds, although he has done some absolutely disgraceful things in my opinion. One example – do you remember the ‘Memory of Water’ hullabaloo? In the interests of fair play, here’s a page against and a page for the idea.

...and I suspect that this is going to be one of those times.

So it can become a bit of a pissing match, where someone stands accused of losing sight of the facts in their fury over the form of the tactics being employed. On the other hand, we have a contention, possibly with merit, possibly not, that those very tactics are being used in a mean-spirited fashion, and their application can skew the very data where they’re being used.

So what can we do when faced with something like this? I would firstly suggest looking at the claim itself…

Jacques Benveniste, a talented scientist, reported in Nature that water was capable of retaining a ‘memory’ of something that had been dissolved in it, even after dilution to the point where none of the compound could be left.

…to see if there’s anything there. Jacques Benveniste was a chemist who thought that he was on to something when he made the claim that water’s chemical structure somehow altered, allowing it to ‘remember’ substances that had been dissolved in it, even if the substance had been removed through successive dilutions. And this, of course, was why homeopathic remedies worked.

The problem with Benveniste’s claim, was that by explaining how the ‘memory’ of water was supposed to work, he had completely bypassed the bit that hadn’t been established about whether homeopathy actually even worked in the first place.

And if you think that’s bad, Nature published it!

Now to be fair, Nature is one of the world’s most read science journals, and gets tonnes of submissions each year. And as we pointed out before, scientists are human, and prone to error.

Nature is peer-reviewed, but because human error occurs, stuff does slip through. Some celebrated slip-ups include the following:

• Jan Hendrik Schon’s fraudulent piece on superconductivity which was published; and
• Enrico Fermi’s breakthrough piece on the weak interaction theory of beta decay, which was not.

Benveniste’s investigation was underpinned by an assumption that homeopathy worked – which, when you think about it, should have set off alarm bells from here to Timbuktu.

Nature’s alarm bells didn’t go off, unfortunately, until the complaints started rolling in, and they ended up having to send a bunch of investigators round to Benveniste’s lab to investigate.

“And why not?”…

Well, Nature arranged to send a delegation to the good doctor’s laboratory. He agreed to this test because he expected his method to work. It had worked many times before. The idea was simple: they prepared three flasks of water, one of which had been treated as described in the method, the other two were just ordinary water. That’s a ‘blind’ experiment – Benveniste, and indeed most of the 'testers', had no idea which flask contained which sample – and it’s the ideal way to carry out any experiment, wherever possible. If Benveniste’s experiment was right, if he had made no mistakes, if there had been no contamination of his samples, then he should be able to find the treated one.

…said Benveniste.

Benveniste and his team were found to have made some errors that were staggering in their ineptitude. These included the following:

1. Benveniste’s experiments were "statistically ill-controlled", and the lab displayed unfamiliarity with the concept of sampling error. The method of taking control values was not reliable, and "no substantial effort has been made to exclude systematic error, including observer bias"
2. "interpretation has been clouded by the exclusion of measurements in conflict with the claim". In particular, blood that failed to degranulate was "recorded but not included in analyses prepared for publication". In addition, the experiment sometimes completely failed to work for "periods of several months".
3. There was insufficient "avoidance of contamination", and, to a large extent, "the source of blood for the experiments is not controlled".
4. The study had not disclosed that "the salaries of two of Dr Benveniste's coauthors of the published article are paid for under a contract between INSERM 200 and the French company Boiron et Cie."
5. "The phenomenon described is not reproducible". "We believe that experimental data have been uncritically assessed and their imperfections inadequately reported."

(Source: Wikipedia)

It's relevant at this point to discuss the reviewers engaged by Nature to investigate this and report back:

• Nature editor and physicist Sir John Maddox;
• Scientific fraud investigator and chemist Walter Stewart; and
• Skeptic and former magician James Randi.

Of these, Randi appears to suffer the most vitriole from Benveniste supporters. Randi is not a professional scientist...

You’d think that a group of scientists and a representative from Nature would be enough to deal with this, but they took Randi along.

The information on which flask contained which sample was in a sealed envelope. Sealed before Benveniste could have seen it, so he couldn’t possibly cheat. Enough?

...yet he was in a position almost straight away, along with the other investigators, to see that the process was flawed. Although Benveniste was not in a position to see the contents of the sealed envelope containing the details of which flask contained which sample, other people in the lab whom the tester [Benveniste] and the test apparatus came into contact with were.

So they double-blinded the test. Randi is an ex-stage magician. In a combination of flashy showmanship...

No, Randi insisted on taping the envelope to the ceiling. A serious study turned into a circus. Surely it would have been enough to keep the envelope in someone’s pocket? Or did he think scientists are capable of picking his pocket, steaming open the envelope, then resealing it without leaving a trace and putting it back?

...and a wish to ensure that the lab stayed in good spirits, Randi taped the envelope to the ceiling. This also had the effect of ensuring that any attempts to sabotage the study would also be minimised - it is unlikely that Randi would have assumed that he would be the only person in the lab with a practised sleight of hand.

Supporters of Benveniste like to seize upon this stunt of Randi's...

Randi had gone in there not with the premise that Benveniste might have made a mistake, but that he was a deliberate fraud and was likely to tamper with that envelope, given half a chance. Further, he didn’t trust the other scientists, or even the Nature guy, with that envelope either. It was an unnecessary, childish, and grossly insulting act.

...as evidence that Randi had some kind of axe to grind and that he was out to deliberately discredit Benveniste and his team.

Or even that Randi did this without the support of Maddox and Stewart. This is a questionable claim that isn't supported by the account that they subsequently published in Nature.

None of this is relevant.

At the end of the day, all that matters is that Benveniste was brought undone by sheer incompetence and some extraordinarily bad record-keeping.

Benveniste's reputation was ruined by this. But he never gave up on his idea.

Despite being the junior party in the review panel, most supporters of Benveniste see Randi as being the ringleader in Benveniste's "debunking"...

In the event, Benveniste’s experiment failed. So Randi chalked up another ‘debunking’. Benveniste’s career and reputation took a beating (to be fair, he really should have had someone else look at that data before sending in the first paper, but that’s hindsight for you).

...

I don’t like the way he works. It’s more in tune with the stage than the laboratory, and that’s something I really want to stay clear of.

...despite the fact that at the end of the day he was able to contribute towards an important step towards ensuring that yet another erroneous piece didn't contaminate the human knowledge bank.

Benveniste never forgave Randi. Neither did a whole bunch of Randi critics who saw this as yet another excuse to hate the guy.

In his later years, Benveniste lost the plot entirely and was convinced that the "memory" that he was so convinced that water had, could be transmitted over the internet.

He was awarded an Ig Nobel Prize in 1998 for that.