Failure?

April 9, 2010 9:15 PM

Nothing succeeds like failure. [H]istory shows that breakthroughs often spring not from carefully laid plans, but from mischance or even sheer, ridiculous accidents. A stovetop spill heralded vulcanized rubber; the potency of uranium was revealed when a rock was left in a drawer among photographic plates. And great research seldom follows an unswerving path. At RCA in Princeton in the 1950s, David Sarnoff exhorted his team to invent a flat television that could hang on a wall. “There were an enormous number of failures,” says Princeton historian of science Michael Gordin — and instead of TVs, the world got the Seiko digital watch in 1973.
posted by caddis (38 comments total)

Some speak of the future; my love she speaks softly. She knows there’s no success like failure, and that failure’s no success at all.
posted by allen.spaulding at 9:17 PM on April 9, 2010 [1 favorite]

Bob always speaks wisdom
posted by caddis at 9:19 PM on April 9, 2010 [1 favorite]

See Also: Born Losers, a cultural history of failure in America, by Scott Sandage.
posted by allen.spaulding at 9:23 PM on April 9, 2010 [2 favorites]

I always heard those "Success by accident" stories growing up, but what I noticed is that most of them were from the 1800s or very early 20th century. Are there any real examples from the past 50 years? It seems like most of the obvious stuff has been discovered already.
posted by delmoi at 9:25 PM on April 9, 2010 [1 favorite]

I always heard those "Success by accident" stories growing up, but what I noticed is that most of them were from the 1800s or very early 20th century. Are there any real examples from the past 50 years?

Sildenafil (Viagra) was originally studied as a drug for hypertension and angina pectoris. It didn't do much for angina but it did produce a certain well-known side effect.

Minoxidil (Rogaine) was also originally a hypertension drug. Its effect on hair growth was originally an unwanted side effect.
posted by jedicus at 9:39 PM on April 9, 2010 [1 favorite]

See also all the innovation that has come out of NASA -- Tang, Velcro, etc. You could say that the mission was to go to the moon, but on the flip side look at all of these products that have either failed to be used as intended or have succeeded in other settings much more than anticipated.
posted by Madamina at 9:40 PM on April 9, 2010 [1 favorite]

Are there any real examples from the past 50 years?
Post-it Notes
posted by eye of newt at 9:42 PM on April 9, 2010 [1 favorite]

A dropped burrito and a pot of frying oil gave us the chimichanga. Top that.
posted by nestor_makhno at 9:58 PM on April 9, 2010 [4 favorites]

Are there any real examples from the past 50 years?

Back in 1964, two physicists were pulling their hair out trying to pin-point the source of a persistent background noise on a new, highly-sensitive radio antenna. Turns out the source was the universe itself, thirteen billion years ago.
posted by dephlogisticated at 10:00 PM on April 9, 2010 [4 favorites]

This makes sense from evolutionary/memetic standpoint. Great strides are made in biology exclusively thru happy accidents. Why would it be different in the world of ideas and culture?

I mean, yeah it's very different because there's the added level of self-awareness and intent, but the laws that govern evolution were here before the laws that govern man.
posted by es_de_bah at 10:10 PM on April 9, 2010 [3 favorites]

Turns out the source was the universe itself, thirteen billion years ago.

Accidentally discovering the universe actually took a fair bit of skill.
posted by five fresh fish at 10:15 PM on April 9, 2010 [3 favorites]

“Whatever you do, do the experiment. Stop thinking how it might come out. Loosen up and let the organism speak to you.” And Gentile warns his summer interns that experiments are always messy: Equipment breaks, fire alarms inconveniently go off, results just won’t come. “Life is not a turnkey experience. Keep going, let’s see where this goes. Find a way around it, over it.”


Truer words never spoken.
posted by lalochezia at 10:43 PM on April 9, 2010

Are there any real examples from the past 50 years?

Aspartame was discovered in 1965 by a researcher licking his finger during an experiment to find an ulcer drug. Similar discovery for grape flavoring while distilling a dye. I wonder how many of those people still taste their experiments though.
posted by Brian B. at 11:09 PM on April 9, 2010 [2 favorites]

Nothing succeeds like failure.

This is a successful example of the failure of confirmation bias. The list of failures that ended in failure has gotta be really long.
posted by three blind mice at 11:10 PM on April 9, 2010 [8 favorites]

Wasn't the discovery of penicillin also an accident of sorts?
posted by marble at 11:17 PM on April 9, 2010

Wasn't the discovery of penicillin also an accident of sorts?
I believe it was accidentally discovered by a Scotsman and deliberately discovered by an Australian.
posted by GeckoDundee at 11:20 PM on April 9, 2010 [3 favorites]

Wasn't the discovery of penicillin also an accident of sorts?

Yes. And so was the discovery of LSD.
posted by Brian B. at 11:22 PM on April 9, 2010

Semiotic information theory was accidentally discovered by Charles Sanders Peirce in his refridgerator, where it was growing on a mushroom biryani he had left there for two weeks.

The city of Truro in Cornwall, England, was discovered by Sir William Herschel, who grew sick of searching the heavens for the planet Neptune and fixed his telescope instead on the second story windows of a nearby building, hoping to see some women undressing.

Melvin De Groote invented process that adheres chocolate to vanilla ice cream quite by accident, after deperately trying to discover something that would make his seven young children shut up for ten goddamn minutes.
posted by the quidnunc kid at 11:54 PM on April 9, 2010 [15 favorites]

The most interesting words in research aren't "Eureka!", but "that's odd...."
posted by DreamerFi at 12:09 AM on April 10, 2010 [6 favorites]

The TV show Connections was full of this stuff. I really wish someone would take another run at the format and some how keep the flavour of the first series as the second and third iterations were much weaker. And it would be cheap.
posted by Mitheral at 12:15 AM on April 10, 2010 [1 favorite]

This is a successful example of the failure of confirmation bias. The list of failures that ended in failure has gotta be really long.

Yeah, and as the article is ostensibly about the problem of grant-giving agencies being risk-averse, it's hard to work out what they are arguing for exactly. They can't seriously be saying that funding should be channelled towards projects likely to fail, even though the article emphasises confirmation-bias stuff like "history shows that breakthroughs often spring not from carefully laid plans, but from mischance or even sheer, ridiculous accidents" and the strapline "Even top scientists sometimes hit a wall – and that’s when breakthroughs happen".

I guess they are saying that funding decisions should be less risk-oriented, and that we should return to "the halcyon days, when you could get funding for just plain interesting studies without promises that anything world-shattering would come out of it." ... well, OK, but some proper research on funding vs. results would be useful here - e.g. "does allocating funding to open-ended exploratory projects work out?" - we're talking about scientists, after all. A list of cherry-picked, amusing anecdotes does not really make a good argument (although it does make a good read).
posted by memebake at 2:18 AM on April 10, 2010

I started to make my own list and accidentally found a list already made:
Serendipity
posted by hypersloth at 4:05 AM on April 10, 2010 [1 favorite]

Scientists act like inventing things is hard work. Nothing can be farther from the truth. Take for example:

"Looking closer, he found to his astonishment that the entire tile had been scrubbed of tritium"

How many great inventions accompany the phrase "to his astonishment?" All of them!

And who do we have inventing things? The least likely people to be astonished: smart people!

So in a flash of inspiration, I realized that in order to maximize the amount of inventions, instead we put in charge of inventing things the people most likely to be astonished: dumb shits!

So fantastic is my idea, that on proofreading I discovered that I have used a total of four exclamation marks!

To my astonishment.
posted by digsrus at 6:11 AM on April 10, 2010 [1 favorite]

See also all the innovation that has come out of NASA -- Tang, Velcro, etc.

Though NASA has had some interesting consumer spinoffs, Velcro is not among them (what with having been invented and patented in Switzerland many years before NASA was founded).
posted by ricochet biscuit at 6:23 AM on April 10, 2010

It seems like most of the obvious stuff has been discovered already.

Sure, then there was iPod, iPhone and iPad!
posted by Brandon Blatcher at 6:31 AM on April 10, 2010 [1 favorite]

Nothing like failing upwards to give you success with humility -- you can't be a magician unless you have the guts to be a fool...
posted by Alexandra Kitty at 7:23 AM on April 10, 2010

The most interesting words in research aren't "Eureka!", but "that's odd...."

Surely the lesson is not that many things come form accidents, but that many things come from investigating accidents and being curious and following through?
posted by The Whelk at 7:51 AM on April 10, 2010 [1 favorite]

Two words: Tollhouse cookies!
posted by pushing paper and bottoming chairs at 7:54 AM on April 10, 2010

"Lets see what happens if we [do/add/change/etc] this.... "
posted by infini at 7:59 AM on April 10, 2010

I'd argue that it depends how you define "failure". An unexpected result is not automatically a failure: if you pursue it, you might go down a blind alley or you might find something interesting. It also depends what you're trying to do - in pure research, even a dead end may corroborate something (e.g., we couldn't find X in specimen Y, consistent with theory Z).

However if you're trying to develop a product, that's when dead ends usually spell failure. Seconding 3BM here - there are zillions of failures that nobody talks about, because they're simply failures and that's boring. In my experience only a tiny tiny fraction of unexpected results lead to something interesting that eventually turns into a product. But serendipity makes for great stories, so we like to tell those.

The best of those stories become part of our collective heritage (see this thread, Connections, and several books and TV shows along this line) but they are statistical anomalies in a great sea of boring failures.
posted by Quietgal at 8:28 AM on April 10, 2010

Accept Defeat: The Neuroscience of Screwing Up

anybody who has done experimental research knows full well that most of the time, the data isn't what you expect (or understand.) the hard part is figuring out what's noise and what's signal.

At the same time, most people seem to think that research consists of confirming preexisting notions. Hence confusing learning something new (and maybe not understanding it) and failure. As if success was defined as confirming the status quo.

It's new information, people!
posted by warbaby at 8:46 AM on April 10, 2010

I wrote a 2007 bestseller called the Myths of Innovation that in large part was about sorting out what really happened in all these legendary stories, and what we can learn from what actually happened.

Three things to keep in mind:

> history shows that breakthroughs often spring not
> from carefully laid plans, but from mischance or even sheer, ridiculous accidents.

1. Whoever wrote this should be careful about 'often'. The large majority of discoveries, breakthroughs and inventions happened for people who dedicated their lives to finding them. As you'd guess it would be. They might not have found the specific one they were looking for, but their pursuit of one thing enabled an accident that led to the discovery of something else. In all cases, they were working very hard trying to do something, not waiting for accidents to happen.

2. People discount how much work follows accidental discoveries. Even if an apple hits you on the head (Newton's story is likely apocryphal btw), there is often years of work necessary to reliably replicate, productize, or prove whatever was discovered.

3. An experiment is, in a way, is 'accident seeking'. A good experiment has some unknowns that are being explored and the word suggests a mindset of having to fail many times to get something right, if it's possible to get it right at all. Many 'accidents' happened to happen in labs, where the people who had the accident had the proper mindset of being interested in the unknown, rather than rejecting it or throwing it away as weird/stupid/strange/odd like most of us would. Much of the chatter about Innovation, and 3M or Google's 20% time is really a recasting of the core ideas of the experimental lab.
posted by Berkun at 11:38 AM on April 10, 2010 [6 favorites]

Sometimes math research feels like an endless branching hall of failure and redirection to a more tractable problem. It goes something like this:

1) Choose a problem.

2) Write a bunch of stuff, trying to get an understanding of it. (ie,prove some shit.)

3) Recite the following mantra: 'Well, I can't see any solution to the problem I actually wanted to solve, but if I restrict to the right special case then the equations will work out kind of nicely.'

4) Return to (1) with the new problem you've posed for yourself. Repeat until you're studying a semi-simple Lie algebra.
posted by kaibutsu at 1:53 PM on April 10, 2010 [1 favorite]

history shows that breakthroughs often spring not from carefully laid plans, but from mischance or even sheer, ridiculous accidents.

How can an incredible discovery every really be called expected? Isn't that why they call them "discoveries"?

Who plans to make breakthroughs?
posted by gottabefunky at 4:25 PM on April 10, 2010

sorta like Columbus went to look for India and accidently discovered america instead?

btw, Berkun, what do you think of taking an exploratory research approach to uncover opportunities for innovation?
posted by infini at 4:28 PM on April 10, 2010

I always heard those "Success by accident" stories growing up, but what I noticed is that most of them were from the 1800s or very early 20th century. Are there any real examples from the past 50 years? It seems like most of the obvious stuff has been discovered already.

Well, the article lists some stuff. Butterfly pigments and red not-platinum -> chemo drugs, for two
posted by rubah at 8:47 PM on April 10, 2010

S. Donald Stookey invented/discovered CorningWare when he took a piece of FotoForm glass and accidentally heated it to 900 degrees Celsius instead of 600 degrees Celsius. He was startled by how hot it was and dropped it - but it didn't break.
posted by Lucinda at 6:20 AM on April 11, 2010

gottabefunky: Plenty of people plan to try to make breakthroughs. They decide they are going to discover the gene structure of cells (e.g. DNA), or cure Polio, or put a man on the moon. There are of course no guarantees, but plenty of people set their goal to be something no one else has quite done before. In some cases a start-up company is doing this, though their work is more derivative, and nearly always involves capitalizing on ideas that have been proven but not been successful delivered in the marketplace (Apple & Microsoft's use of Xerox Parc's work is a great example). Many of the products we, as consumers, call innovations, are actually the final leg of productizing ideas that were proven years earlier (laser is a great example).

infini: There's a whole chapter in the book on 'methods' and my conclusion was that there are many mothers, and many motivations. Necessity is one, greed is another, curiosity a third, and there are more. And similarly when you are trying to pave new ground, there are lots of different approaches, and they all have high rates of failure. There are attitudes that help, but still, as three blind mice mentioned earlier, many failures are just failures.

A good analogy to think about is exploration - like Magellan, DeGama, etc. When you are searching an unknown space for an unknown thing, the more information you can gather on that space (scout ships, previous explorers journals, etc.) the higher your odds. So exploratory research is of course useful, but like all methods, its incomplete. Being systematic helps, but so does intuition and risk taking. Diversity and dedication seem to be the only real advice. As essentially just as there is no way to reliably predict the future (despite futurists and stock market prognosticators who claim otherwise), there is no truly reliable way to break new ground. It has, and will always have, high rates of failure. In many cases even when people are successful with the science, or the invention, the world is rarely as excited about it as the discoverers are (See histories of Galileo, Darwin, Oppenheimer, Doug Englebart (mouse), Tesla (radio), etc.) In some cases they're downright hostile.
posted by Berkun at 9:27 AM on April 11, 2010 [1 favorite]