Thursday, September 29, 2005

Hedgehogs, Foxes and Feynman

Freeman Dyson reviews the recently published collection of Feynman's letters, collected by his daughter Michelle. The essay is long, but worth reading in full. I feel Dyson underrates Feynman a bit, and his classification of thinkers as hedgehogs (Einstein) or foxes (Feynman) is too black and white. Our scientific activities are shaped by the questions of our day. Einstein had the opportunity to solve foundational puzzles (the problems and data were at hand) -- did Feynman? (Also true of Dirac, who Dyson rates as a "greater genius" than Feynman.) Certainly, Feynman had broad interests, but it is also clear that he kept returning to deeper issues like turbulence (strangely not classified as a deep issue at the time), interpretation of quantum mechanics (EPR, probability, etc.) and quantum gravity. I might add that 100 years from now his early interest in quantum computing could be remembered as particularly significant. Although he didn't develop any key algorithms (or even, if I recall correctly, explore the notion of universal quantum computers built from simple gates), he was the first to note the exponential power of quantum computers and the qualitative difference between quantum and classical computing.
Why should we care about Feynman? What was so special about him? Why did he become a public icon, standing with Albert Einstein and Stephen Hawking as the Holy Trinity of twentieth-century physics? The public has demonstrated remarkably good taste in choosing its icons. All three of them are genuinely great scientists, with flashes of true genius as well as solid accomplishments to their credit. But to become an icon, it is not enough to be a great scientist. There are many other scientists, not so great as Einstein but greater than Hawking and Feynman, who did not become icons. Paul Dirac is a good example of a scientist greater than Feynman. Feynman always said, whenever the opportunity arose, that the "space-time approach" that led him to his new way of doing particle physics was directly borrowed from a paper of Dirac's.[6] That was true. Dirac had the original idea and Feynman made it into a useful practical tool. Dirac was the greater genius. But Dirac did not become an icon because he had no wish to be an icon and no talent for entertaining the public. Scientists who become icons must not only be geniuses but also performers, playing to the crowd and enjoying public acclaim. Einstein and Feynman both grumbled about the newspaper and radio reporters who invaded their privacy, but both gave the reporters what the public wanted, sharp and witty remarks that would make good headlines. Hawking in his unique way also enjoys the public adulation that his triumph over physical obstacles has earned for him. I will never forget the joyful morning in Tokyo when Hawking went on a tour of the streets in his wheelchair and the Japanese crowds streamed after him, stretching out their hands to touch his chair. Einstein, Hawking, and Feynman shared an ability to break through the barriers that separated them from ordinary people. The public responded to them because they were regular guys, jokers as well as geniuses. The third quality that is needed for a scientist to become a public icon is wisdom. Besides being a famous joker and a famous genius, Feynman was also a wise human being whose answers to serious questions made sense. To me and to hundreds of other students who came to him for advice, he spoke truth. Like Einstein and Hawking, he had come through times of great suffering, nursing Arline through her illness and watching her die, and emerged stronger. Behind his enormous zest and enjoyment of life was an awareness of tragedy, a knowledge that our time on earth is short and precarious. The public made him into an icon because he was not only a great scientist and a great clown but also a great human being and a guide in time of trouble. Other Feynman books have portrayed him as a scientific wizard and as a storyteller. This collection of letters shows us for the first time the son caring for his father and mother, the father caring for his wife and children, the teacher caring for his students, the writer replying to people throughout the world who wrote to him about their problems and received his full and undivided attention.


Calculated Risk said...

Professor, I'm looking forward to reading this book. As Freeman Dyson points out, Feynman was an inspiration for everyone (not just scientists / physicists).

Off topic: I hope you saw DeLong's post today. When I read the last part, I thought of you ...

And, while we're at it: never get involved in a land war in Asia; do not read My Pet Goat when death is on the line; never play poker with a man named 'Doc'; never accept a battle of wits where iocane powder is a factor; never blithely download and install a file from Microsoft without carefully, carefully researching what it will do beforehand; never get involved in an argument over Noam Chomsky; and never post about human genetics on you weblog.

Best Regards!

Anonymous said...

I believe it's obvious that Murray Gellman was a greater genius than Feynman. It really bugs me that Feynman got all the glory and all the chicks.


Anonymous said...

Just for the record (and fun!)

From "Simulation Physics with Computers", R.P. Feynman 1982:

"But I don't know whether the general theory of this intersimuilation of quantum systems which are really intersimulatable-which are equivalent-as has been done in the case of classical computers. It has been found that there is a kind of universal computer that can do anything, and it doesn't make much difference specifically how it's designed. The same way we should try to find out what kinds of quantum mechanical systems are mutually intersimulatable, and try to find a specific class, or a character of that class which will simulate everything. What, in other words, is the universal quantum simulator (assuming this discretization of space and time). If you had discrete quantum systems, what other discrete quantum system are exact imitators of it, and is there a class against which everything can be matched? I believe it's rather simple to answer that queastion and find the class, but I just haven't done it."


"I know, almost certainly, that we could do that for any quantum mechanical system which involves Bose particles. I'm not sure whether Fermi particles could be described by such a system."

Feynman was, I believe, the first to ask for a universal quantum computer! Well, actually, his insistance of an "exact" simulator is one which we nowdays remove: we require only that a universal quantum computer could approximate a general quantum computation. And we know how to deal with fermions now (one wonders if Feynman knew the Jordan-Wigner transform.)

Anonymous said...

From the article:
Feynman always said, whenever the opportunity arose, that the "space-time approach" that led him to his new way of doing particle physics was directly borrowed from a paper of Dirac's.[6] That was true. Dirac had the original idea and Feynman made it into a useful practical tool. Dirac was the greater genius. But Dirac did not become an icon because he had no wish to be an icon and no talent for entertaining the public.

Um, huh? Reading Feynamn's Nobel lecture gives a little different story:

The minor point is that Dirac wrote that the infinitesimal propogator was "analogous" to the exponential of i times a Lagrangian. But as is clear from the lecture, it was Feynman who realized that "analogous" should be "equal" (propotional to, I guess). And, even more importantly, it was Feynman who took this idea and ran with it to create the path integral formulation of quantum theory. To say that the idea for the path integral was "borrowed" from Dirac is like saying that Shakespeare borrowed from the Normans who invaded in 1066. OK, maybe not that extreme, but still...the insight to extend this to a sum over paths is the _real_ insight and one for which Feynman deserves the credit.

It's strange the Dyson would write this.

Steve Hsu said...

CR: thanks for pointing that out! :-)

"...never post about human genetics on you weblog." Especially if you don't know what you are talking about!

DaveB: thanks for looking that up - I guess Feynman knew very well about the concept of universal computation... But I guess I am used to seeing others claim credit for really working out the details.

This issue of the path integral is quite interesting. Is the Nobel lecture where Feynman reports asking Dirac about whether he knew the relation was an equality, and Dirac answers "Is it?" (Showing that Dirac hadn't known before Feynman.)

Anonymous said...

Indeed, the record on what exactly Feynman knew about universal quantum computers is sketchy. I think many would point to David Deutsch's 1985 article as the "inventor" of the universal quantum computer.

I've also read the story ("Is it?") somewhere, but it's not in the Nobel Prize lecture.

Blog Archive