Steven Beckwith (PhD '78), Astronomer and Explorer of the Origin of Life

The question of how life emerged from the primordial soup of elements is one of the most profound and far reaching that humans can ask. Astronomers will undoubtedly play a key role in ultimately finding the answer. After a long and productive career in astrophysics, Steve Beckwith is now investigating what he calls "the leap" from a pre-biotic Earth to life as we know it today, with the obvious implication for what this might mean about the development of life elsewhere in the universe. At the same time, he is working on an exciting search for black holes and an investigation on the phenomenon of gravitational lensing. Together, the lesson is that in retirement, great scientists can become graduate students all over again, posing ever deeper questions.

In the discussion below, Beckwith recounts his undergraduate education at Cornell, his deep involvement with the development of infrared astronomy at Caltech, and a research career that brought him back to Cornell, then to the Max Planck Institute of Radio Astronomy, then to a joint appointment at the Space Telescope Science Institute and Johns Hopkins, and then to the University of California. Through it all, Beckwith has toggled between fundamental research and administrative leadership and a commitment to sustaining and improving the best of science culture. That was an ideal in which he was fully immersed during his years at Caltech, and they have shaped his views and his research ever since.

DAVID ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It's Monday, April 22, 2024. It's my great pleasure to be here with Professor Steven Beckwith. Steve, it's wonderful to be with you. Thank you for joining me.

STEVEN BECKWITH: It's a pleasure to be here.

ZIERLER: To start, would you please tell me your title and institutional affiliation?

BECKWITH: Professor of Astronomy Emeritus, Professor of the Graduate School. I retired formally a year and a half ago, January of 2023.

ZIERLER: I know there's a distinction between retirement and emeritus. Where are you still active, both administratively and in the research?

BECKWITH: I'm not active administratively, and I don't have any service duties, nor do I have any teaching duties. But I have been fairly active in research. My interests changed over the years. Most recently, I was learning biochemistry because I'm interested in the origins of life. Then COVID came, and because of COVID, it became very difficult for me to maintain the collaborations I had at the University of California, San Francisco, which provided me with avenue to work with real experts. It's a one mile walk from my condo in San Francisco to the new UCSF campus in Mission Bay. I was attending all the faculty seminars, I had a collaborator, Professor Adam Frost, and COVID disrupted my ability to go to meetings and work with Adam.

I have also kept my hand in astronomy, especially in the space business. I have been quite interested in what one could do with small satellites counter to the usual trends in which astronomers tend to concentrate on building bigger telescopes even in space. Eventually, the sought-after facilities get too big, and that has happened now with the biggest facilities on the ground or in space.

But what I find remarkable, learned during my time at the Space Telescope Science Institute, is what amazing things are possible once you get telescopes above the atmosphere. I've been working in a collaboration under Principal Investigator Professor Jessica Lu to create a constellation of several hundred small satellites with small telescopes. The telescopes would have 15-centimeter apertures, very small relative to space facilities such as the Hubble Space Telescope. But they would be incredibly powerful, nevertheless. With a constellation of these telescopes, you could observe almost the whole sky all the time, 24/7. You would be monitoring everything that went "bump in the night"–transient light sources. This is very exciting.

I've had a long trajectory in science over the course of my career, and I had gotten far away from one of the things I really enjoyed doing when I started, which is building instruments. The new project lets me get back in the instrument-building business. These satellites will not work well for the kind of photometric observations we want unless they're pointed very stably. And you cannot go out on the market and buy commercial control systems, called ADCS (Attitude Determination and Control System) for a small satellite with the accuracy we need. For a 12U CubeSat, we want to stabilize the pointing to better than one second of arc to allow accurate measurements of brightness, and they are not available for small satellites. That didn't make sense to me, so I've been doing experiments here in my condo in San Francisco to show that we can deliver pointing information to very high accuracy with a simple setup.

I'll show you the small satellite demo I built to test my ideas. A 60mm telescope (really a finding telescope) is mounted it on a tripod with a modern CMOS camera and a crude stabilization system. Our condo is very long and thin as is common in old buildings converted to lofts. The setup has 20-meter distance to a group of four LEDs that serve as surrogate stars. Two of them are mounted on a microscope stage to adjust their positions, and two of them are fixed in position. I observe these sources and do experiments to see how well the mock-satellite can track them and stabilize pointing jitter. And this setup can easily determine the pointing of the satellite to better than 100 milliseconds of arc. The commercial star trackers you can buy now can't get anywhere near that accuracy.

I've been doing the experiments at home. I built almost all the hardware myself. I have a machinist two blocks away who's got his own shop for the more demanding parts, and bought the telescope, camera, reaction wheel motor, and the on-board computer from vendors who sell mostly to amateur astronomers. The work required me to learn all sorts of new skills. I had to learn to program in C for the camera readout and motor control. I had to learn how to control motors, and I had to relearn some of my lost skills in fabricating metal parts and equipment. A lot of this stuff, I just built by hand.

ZIERLER: You're a graduate student all over again.

BECKWITH: I'm a graduate student. I'm probably the most highly compensated graduate student in the system right now. And I started doing some of this work even before I retired. It has been great.

ZIERLER: Is this project self-funded? Are you on a grant?

BECKWITH: Sort of. When I left the Office of the President, they generously gave me money so I could restart my research. I've been drawing that down very slowly. And I'll draw the rest of it down this year, mostly for this project. Our project actually has a launch for a demonstration space mission. We are partnering with a young guy in Toronto who founded his own company after getting a Master's Degree at UT. He wants to build star trackers that provide the very high accuracy we need, and we may want to buy their star trackers once they are available. They got funding for a small satellite and a launch from the Canadian Space Agency, and they offered us payload space if we supplied our own payload, so that's what we're doing. The launch is about two years out, and we will be able to test out some of our ideas for small space telescopes. My home efforts have been to verify experimentally that we can do what we think we can do on paper.

It's going well. If we can make these lab experiments work, it will be support for the rest of the project. We're working with a group at Livermore National Labs to get one of their solid optics telescopes (their invention), which would be ideal for main instrument. Jessica and her graduate student, Hannah Gulick, have been working on the cameras. The new CMOS detectors are almost perfect in their performance compared to the limits of nature. They have quantum efficiency of almost one and almost no noise. These detectors have changed the nature of what we can do in designing instruments for astronomy. We have a small and very able team. We put in one proposal with NASA that was turned down, but it was a tough round this last year, and we will succeed with perseverance. NASA has less money for new projects than we would like. My role is primarily to support Jessica. The primary work is the home experiments, and I'm just having a blast.

ZIERLER: These small satellites, are they an array? Are they all combined?

BECKWITH: They'd all be free-floating satellites. They will be small, about the size of a bread box, perhaps 20 by 20 by 30 centimeters or 12U. Even these small satellites can be very powerful because of the quality of the new detectors, and if we can stabilize their pointing, they should be provide accurate photometry of the stars and other objects we observe.

We have several science projects which would be well-suited for these small telescopes. And I always like the discovery potential that comes with opening up a new window for observations that can lead to new discoveries. I hope there will be some discoveries as well as the planned science projects.

New Black Hole Searches

ZIERLER: On that basis, what are the science objectives of this new project? What do you hope to discover?

BECKWITH: The primary objective is to look for black holes that would pass in front of distant stars and increase their brightness through gravitational lensing. And you can't predict when a lensing event will happen. You have to monitor a lot of stars, and every now and then, a black hole will pass in front of one of them. With enough events, the statistics will tell us about the number and nature of free-floating black holes. This is not a new idea. It's been done for years on the ground. But in space, especially if you have enough telescopes, the number of stars that you can monitor is much greater than what you can do on the ground. And the accuracy of photometry free of the Earth's atmosphere greatly expands the information we can assemble on these free-floating black holes.

The primary project is to study black holes, but there are lots of other things that may create brief flashes of light that we would like to discover when they brighten. Gravitational wave detection has changed our view of the possibilities for transient sources, and we know there are a lot of candidates at any one time. When a gravitational wave source is detected, the position on the sky can be determined very crudely by astronomical standards. The error circle is typically 30-degrees, a huge swath of the sky.

If there are instantaneous or nearly instantaneous signatures associated with events like those that are picked up by the new gravitational wave observatories, we will not see them unless we are looking everywhere at all times. You can't do it any other way. That's another element of the science that could be enabled by a constellation of telescopes. We know of other transient sources such as fast radio bursts, and of course, gamma ray bursts where it would be interesting to look for visual signatures. Caltech's Professor Shri Kulkarni, a close friend and colleague of mine, has done a lot of work in this area. It's called transient astronomy or time-domain astronomy, and it is a field that has recently matured. Our multi-telescope project could contribute very much to this field.

This had initially been sort of a side project for me. I was planning to spend maybe a third to half my time on it and the other half on biochemistry. But I haven't actually been able to do any biochemistry for about a year and a half, so I am concentrating on astronomy.

My wife has a health problem, which makes it dangerous for us to be exposed to viruses like Covid. It's safer for me to spend most of my time at home away from other people, so I rarely go into a lab. My work with Adam and his group was lab-based, and I have given that up for the foreseeable future. Remarkably, I can do quite a bit of good experimental work right here where I live. There's a bit of friction over the lab clutter that we have worked out, but it's otherwise a great way to remain active in research. When I say "active," I'm spending time, I'm spending money, and I'm having a lot of fun. It's a good mix for me. I love to work on the hardware, and I also love figuring things out. I love the fact that you can use mathematics to learn things about the world, and I've been approaching the experimental work by calculating the ultimate limits imposed by fundamental physics and comparing them with the experimental results. The results I am getting correspond well to the limits I derived, and that's always satisfying. It's very much getting back to the things that drew me into astrophysics in graduate school.

ZIERLER: The work on detecting black holes and thinking about gravitational waves, is this complementary to LIGO, LIGO II, and Virgo? Are you part of that effort at all?

BECKWITH: We're not part of the LIGO effort. Our work would be complementary, certainly. Jessica's main interest is in black holes–she's a black hole person. I'm not a black hole person by history or trade, but I like learning new fields. The goal is to get a census of how many black holes are out there, because we really don't know. Black holes don't radiate. It's the common problem for astronomers that if it doesn't radiate, you may not know it's there except indirectly just like dark matter. The signatures are subtle. We're confident that black holes are the normal end state of matter, what matter would like to become if not prevented by things like nuclear fusion or degenerate fermion pressure. A black hole is the highest entropy state for matter, so it's the natural end state of gravitational collapse. It should be important to know exactly how many black holes we've got, what their distribution and population is. But it's very difficult to really know because they don't radiate light, and you need to find a way to count them through these accidents of alignment with distant stars.

ZIERLER: These very foundational questions about the nature of black holes, where they are, how many of them there are, do you think understanding those questions gets us closer to understanding dark matter? Do you see a connection?

BECKWITH: I don't know. An experimental scientist should be open to lots of possibilities, so I really don't know. The smart money would not bet on black holes making a big contribution to dark matter, because there are many peripheral constraints from other observations on the extent to which they can contribute. I am a bit pessimistic that we'll make an impact on the dark matter problem through this kind of a project. On the other hand, you never know. I know particle physicists have these great ideas of candidates for dark matter, but nothing has worked out so far in pushing the Standard Model. And we do know that there is a big problem in relating the very small to the very large, that quantum mechanics and General Relativity don't work very well together, and yet they both provide exquisitely accurate descriptions of nature at the different scales they operate. There's some problem in physics that we haven't figured out. Because of that, I think that when you're doing new experiments, you need to be open to the idea that it will introduce some insight into the way that matter or the universe is constructed that you just didn't have the imagination to foresee. We already know the main physical models are incomplete.

ZIERLER: A similar question, thinking about black holes the way you are, where does that leave us with quantum gravity? Could gravity be quantized? Should it be quantized?

BECKWITH: I don't know. I'm a physicist by training, and I make my living as an astronomer, so I love to follow that stuff. There has been a big hope that you'd get quantum gravity through string theory, but from everything I've read string theory is actually in trouble because there are just so many different string theories that all could do the same thing, and you can't easily break that degeneracy. Lee Smolin wrote a very interesting book "The Trouble with Physics" where he argues that it's time to start moving away from string theory and consider other ways of looking at physics to resolve the tension between General Relativity and quantum mechanics. I thought it was a very persuasive book, but I don't have enough expertise, especially in General Relativity, to have a good sense of what the future will bring. It's beyond what I feel good at.

ZIERLER: You've emphasized space-based observational projects and astronomy. What about the world of land-based astronomy? Where are you involved there?

BECKWITH: I'm not involved in that at all, and it used to be my bread and butter. I spent years going to ground-based observatories. I loved observing, and I loved going to the mountain and taking data. It was hard work, but then coming back and looking at the new observations was just great. But for many reasons, I don't do that anymore. The travel was tough, especially on my family. It's not easy to accommodate the tension between the demands of observing and raising children. And in any case, the nature of ground-based observing has changed with computers controlling telescopes and the increased sophistication of instrumentation that makes it more efficient to schedule observations automatically with little intervention by the observers. When I started out, telescopes didn't even have guiding systems. I'd sit in the cage of the 200-inch telescope with my eyes frozen to the eyepiece in the winter just guiding the telescope to keep it on track. Everything required manual control. We built all of our own equipment and had to nurse it through the observing runs. That's where I developed the skills that I use in the laboratory now. The observer was a very critical part of the whole operation.

And slowly, we've been replaced by machines and computers. This is a good thing, because the telescope time is used more efficiently. I know my colleagues still make decisions during observing runs, but they typically don't go to the mountain. At Berkeley, if you want to observe at Keck, you go to a control room in Campbell Hall. You don't have to go to Hawaii anymore. It does remove a lot of the romance for me, although it is, I think, a more efficient way to work. But the other thing that's happened is that as we've done many of easy things with relatively simple instruments–people are not constructing instruments the way we used to. When I was a student in the 70s, you could have an idea and go into the lab, and honestly, within a week or so, working in the machine shop, soldering new electronic boards, we could create something which was a significant augmentation of the instrumentation we were working on, and go to the telescope and use it. It was a rapid turnaround cycle.

Now, that easy stuff has been done, which is a normal trajectory for new technology. And the quality of the detectors has improved to an extent that we couldn't even dream about in those days. We used to use individual detectors of essentially one pixel to measure light intensity. We'd image light through a small hole, measure how much light was going through. Now, the detectors are large-format two-dimensional arrays that produce high resolution pictures of everything. And they are almost perfect detectors capturing nearly 100% of the incident light. The instruments and telescopes have gotten very big, too, making it impractical to try out new technology without a lot of investment. Once the telescopes get to a size like–well, Keck or the VLT and certainly the next generation, the very large telescopes, EELT and the TMT–they are so big that the development cycle for new instruments is more like that for space experiments than what we used to do on the ground. Everything has to be well planned out in advance. The instruments cost hundreds of millions of dollars, actually more than the Keck Telescopes themselves. They'll require huge collaborations of scientists, engineers, and manufacturers, and the work will be very finely divided.

The management of those projects is almost as important as the scientific or technical prowess of the team members to make them work well. This is a normal maturation. But it changes the nature of the work. In my view, if it's all going to be automatically and remotely controlled anyway, I'd love to have something up above the atmosphere. Furthermore, I think these very large ground-based telescopes like the EELT are designed to be photon buckets. They plan to make progress with adaptive optics to get very high resolution images, and I'm sure they'll make great strides, but it's just not like being above the atmosphere. The projects, I think, will be targeted very much toward spectroscopic observations of very faint things, the fuzzy things at the edge of the universe. And there are plenty of people who can do that science much better than I can. I'm not going to get involved in it.

Whereas I feel I can make real contributions to the telescope constellation project I'm a part of, especially to the instrumentation. It's an appropriate task for me to take on. It is fun, of course, but I don't want to be spending my time on projects to which I cannot make meaningful contributions. Small satellites are a pretty good niche for me. As for ground-based telescopes, I am not even sure when my last observing was. I took a trip to see Andrea Ghez observe at the Keck in probably 2009 or 2010 at her invitation, because she's a good friend of mine, and we were in graduate school together. It was after I started working at the UC Office of the President. It was fun to be in the control room in Waimea, but I played no useful role. I don't think I've been to a ground-based observatory since then.

ZIERLER: As a UC professor, are you following developments with the TMT, or lack thereof?

BECKWITH: Well, yes, but at a far remove. I was the Vice President for Research for the UC system, and when I came to UC the TMT project was being featured as the next big thing for UC and Caltech astronomy. Henry Yang, chancellor at Santa Barbara, decided to take the lead for UC in promoting that project–and he has a great set of skills to help them at the early stages–and it meant that I could step back and let him do his work on UC's behalf. I was also frankly a little skeptical of the viability their initial plans, which made it difficult for me to be an unabashed salesman. They were (and still are) badly undercapitalized. I did not think they had a good understanding of the cost. The project team was absolutely superb, and the technical work was outstanding, but there were two groups in the US working on the next generation of telescopes who did not cooperate–in fact, they competed–and it seemed to me that anything built at scale would have to be a single national project with the financial backing of sovereign states. I had a long history watching how large scientific projects developed, and it looked like the two projects were traveling down well-worn paths that did not lend confidence to their pronouncements. It's not just in science. Large industrial and government projects like the Oakland Bay Bridge typically overran their costs and require deep pockets to bring them to completion.

I got into high level management roles in science before having any real training for that kind of work. I tried to make up for my early lack of experience by doing a lot of reading. I subscribed to the Harvard Business Review among other things, and I've got a whole shelf full of management books that sparked my interest in the challenges of leadership. One thing that is widely written about is why it's difficult to understand the cost and time of very large new projects at the outset when the projects reach a certain scale. It's difficult for anybody, not just scientists but people in industry who are building new chip plants or factories or public works projects, to know in advance what the biggest challenges to the cost will be. Most of the psychological problems identified in these articles that make accurate prediction hard, I could see in the TMT, and it worried me. So it seemed appropriate to let Henry be the UC lead.

TMT is a beautifully designed project, and if they can get it built, it will be a great addition to astronomy. For a variety of reasons, the odds in its favor have not been very good for several years, so we'll need to watch and see.

Astronomy and the Origin of Life

ZIERLER: Your interest in origins of life, that's obviously something that interests all scientists to one degree or another. When did you get involved in that, and what was the spark for you?

BECKWITH: All of us are interested in life's origins at least a little bit. In astronomy, the typical interest is in alien life. Carl Sagan was my colleague during my 13 years on the faculty at Cornell, and while we were not close, I knew him well, and I followed his work. He tapped into this visceral need we all have to know whether or not we're alone in the universe. He evolved as a thinker over the course of his life, and it was fun to watch. In some ways, it motivated my research in the 1980s, especially with Anneila Sargent, who's one of Caltech's great professors, and we worked on a subject which was a precursor to figuring out if there is abundant extra-terrestrial life: are there a lot of planets?

We got into the field very early by working on circumstellar disks around young stars. That work was pioneered by Mel Dyck and Ben Zuckerman, with whom I was an early collaborator, and blossomed into the work with Anneila culminating in the mid-1990's. So, I was always interested in extra-terrestrial life. But it wasn't clear to me exactly how someone would really figure out life's origins even if aliens are discovered. The jump from the physics and chemistry to biology requires a complete change in orientation of basic assumptions about what drives the creation of the patterns in each of the fields. This looks like a similar challenge to resolving the discrepancies between General Relativity and quantum mechanics.

One of the great things about the job of Vice President for Research is that you have to try to understand the particulars of research across an enormous spectrum of intellectual thought, even in the humanities. In our office, we had to worry about things like computers for the digital humanities–it turns out, the humanists wanted more computing power than most of the scientists wanted or were using. It was interesting to watch the relative developments of different research areas. And I got quite interested in biology because, of course, it's a huge fraction of the research enterprise and ascendent in the progress of science. The size of the NIH budget relative to the NSF gives you an example of its dominance.

It was also interesting because I began to see, as I learned more and more about it, why it was ascendent. Science fields can tick along for a long time, making incremental progress, and suddenly, a big insight or a new technique that allows us to see the patterns in a different way suddenly accelerates our understanding and opens up all sorts of new avenues for research. That has clearly been happening in biology for quite a while. And it's ongoing. Our response to the COVID crisis is a good example. Understanding of RNA vaccine design had been developing for many years before COVID arose. Once there was a nucleotide sequence for the virus, scientists designed a vaccine on a computer in a few hours. It's literally that fast. Then, the biochemists were good enough to synthesize the vaccine based on a computer design almost as quickly. That kind of rapid development wasn't possible 15 or 20 years ago. And it was my privilege to know some of the people who did the critical work.

I know Jennifer Doudna just well enough–she has an office not far from far from my office in Berkeley–that I closely followed the development of CRISPR when it became widely publicized ten years ago. I had educated myself well enough that by the time CRISPR was announced I could really appreciate what that meant for gene editing and what it meant for editing of nucleotide sequences and large molecules. It seemed to me that all of the tools and the developments in biochemistry were at a point where we might start to think about experimental methods for discovering the transitions that had to occur to enable life to happen. And there are a series of needed steps that you might not think about but are, nevertheless, critical to the complicated set of mutually dependent chemical reactions needed to sustain the most rudimentary life.

My friend and collaborator Adam Frost is an expert in protein synthesis. The whole idea of protein synthesis, that you read out a gene with messenger RNA that subsequently serves as the template to create new proteins is a remarkable thing. What's more remarkable is that almost all of the biochemistry that goes on to enable life involves these very specialized proteins, including those necessary to make all other proteins. Large, complex proteins are required in almost all the critical reactions needed to sustain life. And yet, how do those reactions come about before you have the proteins and the apparatus to make proteins? The transition that life made to rely on the mechanism of making proteins, is one of the many mysteries of how we get to the very complex state we're in. I thought it might be possible that the modern tools developed to study cells and reactions have become good enough to give us a stab at answering some of those questions. And there are a number of research groups out there who are doing it, and I think they're making interesting progress.

Adam and I were talking about the origins of protein synthesis because he's an expert in the field. Sadly, our work has been put on hold. My interest has not waned in any way. My ability to work with him in the lab has disappeared for a while to attend to my wife's health. But the remarkable developments in laboratory biology were, for me, the key that said, "Maybe it's time to jump in."

Plus, as you get older, it's easy to stop doing hard intellectual things, and I did not want to stop. I discovered quickly I couldn't begin to fathom the origins literature unless I knew organic chemistry, so I taught myself the rudiments of organic chemistry. I studied textbooks, and I went through and did all the problems in the same way I would have as an undergraduate. I didn't have the lab experience, but that was okay, I could see where it was going. There are courses online now that allow you to teach yourself these subjects. There are some beautiful lectures that are available.

There is a brilliant professor at Yale, J. Michael McBride, who posted his freshman introductory course on organic chemistry online. I went through all of his lectures for the first-year students. I'd even binge watch lectures on airplanes. I could watch four or five organic chemistry lectures in a row on a cross-country trip. It was total immersion. And the same with biology. It was engrossing to dive into a whole new intellectual activity in my late 60's. That provided a completely new intellectual challenge for me, and I reverted back to the level of an undergraduate. I was aware that I wasn't as good as most organic chemistry majors at Berkeley or Caltech, but I still picked up enough to be able to read the literature and appreciate the questions and advances.

I can read articles on organic chemistry in Science Magazine and figure out why they are interesting. It has been invigorating. I sat in on a biology course at Berkeley for a while, but eventually, especially after COVID, it was necessary to do everything at home. I got deeply into biochemistry, developed many of the needed tools, and I started working with Adam in his lab, but then I had to put it on hold. Fortunately, I still have astronomy, which I love.

ZIERLER: Between astronomy and your interest in the origin of life, I'm curious if exoplanets is sort of a natural connecting point for you.

BECKWITH: Well, it sort of is, so I follow the developments, but I don't do any research on exoplanets. The research I mentioned earlier with Anneila and Ben in the 80s and early 90s concentrated on how disks are created and subsequently condense into planets. We were working on those questions before extra-solar planets were discovered. In the mid-90s, I thought we had largely exhausted avenues for big advances on disks. It would require a new facility like ALMA to make more progress, and ALMA was more than a decade off. It was time for me to stop working on the disks.

But about the same time, the first exoplanets were discovered, and I thought, "Okay, the disk business is interesting, but it's not nearly as interesting as the exoplanet business." Because in the end, what's mostly interesting about the disks is that they might create planets. Research in exoplanet work was undergoing a shift, and I didn't bring any special skills to the new research that made me want to jump in. As I've watched over the years, I'm less optimistic that we're going to be able to discover many things which will be very relevant to the hard questions about the origins of life from astronomical research. I don't think that exoplanet research will give me clues to the answers to the questions that I'm most interested in.

ZIERLER: I wonder if more fruitful, would be things like Mars Sample Return or exploration of the icy worlds in our solar system.

BECKWITH: It might be. Certainly, missions to the moons of Jupiter and Saturn, moons like Europa, where there's clear evidence for possible ice-covered oceans with plumes that could be sampled with a fly by satellite to look for organic molecules have the potential to tell us a lot. In the early research of Uri and Miller, where they were creating amino acids and building blocks of proteins in the laboratory from inorganic matter, it was fashionable to think that the methods would tell us something about how life came to be. The truth is, though, it's such a leap from an amino acid to even a reasonable-sized protein that'll do anything for you, it's hard to see how producing the monomers gives you much insight into making the polymers. It isn't going to happen by statistical fluctuation in a sea of building blocks.

If you have a whole sea of amino acids, they might create some larger molecules, but that's just so far away from the size of the molecules that are used in life today, that it isn't obvious to me that their presence tells us a whole lot about how the transition took place. Even, for example, if you go to Europa or one of these moons, and you fly around, and you find a lot of amino acids, unless you're finding complex proteins, which you probably can't capture at the speeds you're going around these planets without destroying them, it's not clear to me that you're going to get the answers. I don't want to be discouraging, it's just not something that I would necessarily bet on myself.

Now, it is true that within life, there is a unique handedness to a lot of these molecules, which probably set in early and was reinforced through reproduction so that many of the molecules have clear chirality, a clear handedness, that makes the distribution incredibly asymmetric. And one of my colleagues, Rich Mathies, believes that if they could identify that kind of asymmetry of chirality in molecules from the moons around these distant planets–presumably, you could do this from Mars Sample Return as well–that would be a smoking gun for the presence of life, because that's the only way we know of creating such an asymmetry. That's plausible. I'm not sure if it still tells us what we need to know to understand how to make the shift to biology, but it would be an important contribution to knowing if the shift was easy or hard.

ZIERLER: You mentioned it's great for you to be in such close proximity generally to UCSF. Is University of California at San Francisco a center for origin of life research? Is there really interesting stuff happening there?

BECKWITH: No, I don't think so. These are professional biologists. As far as they're concerned, they've been given biology, and it's interesting. There are so many interesting problems to work on, they don't need to worry too much about how it came to be. In some ways, it's the way that physicists work, saying, "Okay, the Big Bang came about. It's interesting to find out why, but honestly, look what it gave us," and they just go and work on other things. To my knowledge, there's no real work going on there on origins of life.

ZIERLER: Let's go back and establish some personal history. Tell me about where you grew up in and your family background.

BECKWITH: I grew up in Milwaukee, Wisconsin, in a suburb just north of Milwaukee. We were comfortable and I was privileged in that sense. I went to Shorewood High School, which was a good public high school. I didn't take it seriously enough.

ZIERLER: What were your parents' professions?

BECKWITH: My father was a trial lawyer, a litigator. My mother was a homemaker. I have two brothers, John and David. David lives in Menlo Park, and the two of us ride bikes together almost every weekend while we're out here. We had developed a very close relationship through our shared activities, biking, hiking and skiing. I was reasonably good at sports and enjoyed them in high school. I loved mathematics but I found high-school level math too easy to pay much attention to. I think a lot of us who find that easy also probably don't pay as much attention in high school as we should have. In any case, I was so fascinated by physics and mathematics that I studied it on my own.

When I was in middle school maybe or early high school, I retrieved a physics textbook from a trash can somebody had thrown away. And I read it just because I wanted to see how things worked. I thought it was really interesting. I was good at math, so I'd study math books. I think One, Two, Three... Infinity was the one that got me into calculus, so I studied calculus when I was quite young, and it seemed pretty sensible, didn't seem to be that hard. I did a lot of studying on my own. I just loved science, but especially I loved physics and mathematics. And evidently, I had some talent for it. My grades were mediocre, but I was good at taking tests, and I did well enough on the SAT tests that I was able to get into a very good university, the College of Engineering at Cornell. It was a great opportunity for me.

When I got to Cornell, I was quite conscious of the fact that it was an expensive school and that I was lucky enough to have parents who could pay for it all–not all of my peers were so fortunate. So I dedicated myself entirely to my studies to take advantage of the opportunity. I immersed myself completely into what a university like Cornell could offer. And it was a fabulous experience. Cornell is a sea of opportunity for intellectual stimulation and learning. In high school, I was on a swimming team, and I was a springboard diver. I was good enough to go to the state meet, and I did okay. I wasn't number one, but I was up in the top six, I think. When I went to Cornell, the diving coach tried to get me on the team, and I worked out with him for a while. And it was fun, I enjoyed it, and I was a pretty good athlete, but I decided that if I was going to do physics, I didn't think I could do that and pursue a sport. It's not easy to mix, at that level of dedication, sports and college. I quit the swim team and dedicated myself to my classes. I chose engineering physics, which was one of the best majors at Cornell. It was demanding. A lot of very good people have come out of that program.

Physics From Cornell to Caltech

ZIERLER: I've often heard it said of engineering physics at Cornell that it's as much physics as it is engineering. I wonder if that was your experience.

BECKWITH: Oh, absolutely. Yeah, it was probably more physics than engineering. The engineering was sort of–we had to learn a few engineering things but mostly we learned basic science. When we were studying electrodynamics, we had to learn how to do certain microwave calculations in case we went into the microwave industry. But really, it was just physics. And at a very, very rigorous level. I loved it. I worked very hard. I took more than 20 credits a semester to accelerate my graduation. I graduated in three years. I think I was the first person to graduate from that program in three years, but only technically because there was a Chinese guy in the same year whose name started with a C, whereas mine started with a B, who also graduated early, and I think he was a little better than I was. But I did well enough, and I loved it. I did well enough in school to get into Caltech among other schools and decided to go there. That's how I have the affiliation that brings you to this Zoom call. That was my early history.

ZIERLER: Going to college in the early 1970s at Cornell, were there the remnants of the 60s at all on campus? Was there unrest and all of that?

BECKWITH: Some. And the Vietnam War was still going on. I was in the lottery for the draft, and I had a low lottery number. I think it was in the 70s. My lottery number was low enough that I would've been drafted, but I was in the last year that was given a student deferment, and it was in part because I was studying engineering physics, which was considered to be of national importance. I was able to study without being drafted. Then, by the time I graduated in '73, the War was winding down to the point where, even after graduation, I wasn't going to get drafted. Although, it's possible that going to Caltech would've prevented being drafted as well. I do remember worrying about that, though. It wasn't a small thing.

ZIERLER: Did you have any exposure to astronomy as an undergrad? Did you ever get to take a class with Carl Sagan?

BECKWITH: I never did take any astronomy; I was totally into physics. Not because I didn't like astronomy. I did. When I was a kid growing up, I had a small telescope that I'd use in the backyard to observe, and I attached a single lens reflex camera to it to take pictures of bright objects like the moon and the planets. I had a dark room and developed the film for those pictures. It was fun. But I didn't do much more than that, just sort of dabbled around and learned a little bit about it. I didn't really discover astronomy in a true way until I got to Caltech.

ZIERLER: When you were applying for graduate schools, you were focused on physics.

BECKWITH: Completely, and that's what I was admitted to at Caltech.

ZIERLER: What kind of physics?

BECKWITH: Well, I wasn't sure. There were a lot of things I liked as an undergraduate. I was doing some research with professors at Cornell, but it was almost all in plasma physics. And I didn't want to do plasma physics. I could see already it was a bit like trying to predict weather. Every time they figured out how to solve some plasma instability, five more would pop up. They were mainly trying to confine these plasmas for fusion reactions. I didn't think that was in my future, so I wasn't quite sure what I would do in graduate school. But I was sure that when I got to Caltech, just like as an undergraduate, there would be so much that was interesting that I'd be able to pick something. And I did.

The admission letter from Caltech said, "You will get full support as a research assistant." I showed up on the first day and went into the physics secretary's office. I think her name was Nancy Durham, if I recall correctly. And I said, "I have this letter that says I'll be supported through this research assistantship. Where will I work? What's my assignment?" She said, "Oh, you have to find someone to work for." I said, "Oh, what if I don't do that?" She said, "Well, then you don't get paid." I thought, "This does seem a little bit different than what I thought this letter actually said." But it was okay. That's the way the system worked, and I understood it. I looked at all the research groups. They each had little blurbs on what they did. And the two that struck me as the most interesting were radio astronomy and infrared astronomy.

I talked to Marshall Cohen about radio astronomy, and he talked about what they did, and it was very interesting. And I talked to Gerry Neugebauer about infrared astronomy. In the end, I thought the infrared astronomy looked like it had a bit more potential for new discoveries, because it was an immature field. Radio astronomy had been around for a while. All the traditional fields had been around. But the infrared work was very, very new. The technology came from military surveillance work developed during the Korean and Vietnam Wars just as radio astronomy had come from the work on radar for World War II. And Gerry Neugebauer, who would become my advisor, had been drafted and sent to JPL to work on infrared technology, because that's what they needed for surveillance.

He learned all of the most advanced techniques for infrared detection, and when he got out of the army, he starting working with Bob Leighton, and they wanted to apply the new technology to astronomy. Of course, he was given the cold shoulder by the traditional astronomers who thought it was completely peripheral to what they were interested in. But I saw, when he talked, what he saw, that nobody had looked at the universe this way, and there had to be tremendous potential to discover new things. In hindsight, it was exactly right. The rate of progress, even in those days, was so great that new things were happening all the time. I asked him if I could have a job, and he said yes. And that was how I chose that direction.

ZIERLER: I want to back up. Thinking about first starting with physics, where else did you apply? Why did you ultimately choose Caltech on the assumption that you would pursue physics?

BECKWITH: I applied to Cornell, and I got in. I was one of their good students. A few people said privately, "You're probably better off going away. Otherwise, you become too ingrown." I applied to Colorado. I love to ski. I thought, "That sounds like a pretty good place to be." I can't remember where else, maybe Wisconsin and Caltech. I thought Caltech might be a bit of a reach, but not too much, given that I was one of the top students in a very tough program at Cornell. And I believe there's a good Cornell-Caltech connection. A lot of people have gone from Cornell to Caltech, and some have gone the other way. Part of that was probably the aftermath of World War II and the fact that Hans Bethe, Ed Salpeter, Richard Feynman, and Kurt Gottlieb, went Cornell after Los Alamos.

Anyway, I took the decision seriously. My father said, "You ought to go visit all these places before you go make up your mind." He was generous and offered pay for my trip. My trip to LA was in March, because I had to make up my mind by April. And I left Ithaca on a day when it was about 33 degrees and precipitating in some mixture of solid and liquid water. I don't know if you've spent much time there, but the weather can be pretty unpleasant, especially in the spring. I changed planes in Pittsburgh, and the weather was even worse there. It was sleeting down. It was falling ice.

This kind of spring weather was normal. I grew up in Milwaukee and lived in upstate New York, and I really did not know any other climate. I landed in LA when a Santa Ana wind was blowing. I got out of the plane and couldn't believe it. It was 85 degrees at the airport and dry. I could see the San Gabriel Mountains from the airport. The air was so clear. There were palm trees waving around and all these people walking around half-dressed. And I thought, "Wow." I'd never been to LA before. It was pretty attractive in early March. I had a couple days' visit, I talked to several professors when I was in Pasadena and came away completely impressed with the atmosphere and, of course, the climate.

ZIERLER: In the early 1970s, your world is kind of constrained as an undergraduate. Did you know what an exciting time this was in particle physics, that the Standard Model was in the process of getting built up?

BECKWITH: I did, but I also saw that it involved enormous collaborations, where scientists were small cogs in a big machine. It's not my personality to work that way. I eventually learned to be a team player on large teams, especially when I got into administration., but my preference is to work individually or in small teams. I like doing things on my own. And it didn't look like high energy physics would be an easy field to do that. A lot of my fellow students felt the same way. Yes, it was very interesting, but the question is, "How much of a role are you going to be able to play in advancing this field?" I just didn't know. Plus, I think that particle physics required a level of mathematical sophistication that I did not have at the time–well, maybe not for building of the accelerators, but certainly the theory.

Switch to Infrared

ZIERLER: How long did it take once you got settled in at Caltech before you made the serious interest and switch into astronomy?

BECKWITH: A few weeks. I asked to work in the infrared group, and they said yes. Once I got in to that group, I dedicated myself to learn everything I could about the field, and do what I could to contribute, learn about the equipment, and the scientific problems they were working on. I didn't have the background that the astronomy students had, so I had to learn a lot of astronomy. But it wasn't that hard, and it was lots of fun. Later when I started teaching courses at Cornell, I had to teach a lot of material I had never taken, but I'm an autodidact, and I really enjoy learning astrophysics as well as traditional astronomy. Stellar structure and nucleosynthesis were very interesting topics. I learned a subject in a way that I could teach it to undergraduates. Back to your question, once I decided I was going to go in that direction, I just pushed very strongly to come up to speed.

ZIERLER: Do you have a clear memory of meeting Neugebauer for the first time?

BECKWITH: Yeah, I have a pretty clear memory. He was in his office with Eric Becklin, and they were chatting. I went in and told them where I'd come from. Turns out, he was an undergraduate at Cornell also, in regular physics, not in engineering physics. I gave my spiel, and they were interested, polite, a little quizzical, and listened well. Gerry said, "Okay, fine. Go away, and we'll let you know." And the next day, he said, "Sure, yeah, we'll hire you if you want to work here." Subsequently, of course, working in the labs of the third floor generated lots of great stories about how the group functioned.

ZIERLER: What was Neugebauer's focus at that point? What was he working on?

BECKWITH: They were working on star formation. The broad questions they were looking at was how stars are born from clouds of gas and dust. It was clear that stars were created in the cocoons of these large clouds, and maybe even in cocoons they made themselves. Those cocoons could be penetrated at infrared wavelengths, especially if you got out to the thermal infrared, but you couldn't do much at visual wavelengths until they had emerged from their cocoons. He tackled the questions his technology could uniquely address: "What scientific problems would be best addressed by using that technology?" And star formation was one of them.

It wasn't the only one. Because of the technology focus, the group would work on anything that was interesting to observe in the infrared. We looked at novae, various kinds of stellar occultations, a range of planetary science. The group's scientific interests were wide-ranging. There was also an ethic that we were good at making measurements that we could completely defend, that people could trust, and that was the most important standard to maintain. You wanted to make sure you got the observations right. Exactly how you interpreted the observations, it wasn't so critical. I know Eric Becklin was very vocal about the value of that approach. He often said, "The interpretation's going to change in five years anyway, so it doesn't matter too much what you think is it now." They always made a good effort to get the interpretation right, but the theoretical side was of passing interest to them. I was a little different in the sense that I was also very interested in the theoretical side.

Even when I did my thesis, I'd write things, and they'd say, "Do you really want to write this? Because after all, it's the observations that matter, and this interpretation might change." In retrospect, I think they were wise and probably right overall. But I was a young kid, full of myself, so I did what I wanted and worried about what was going on. Nevertheless, the notion that, "When we publish, we want to be clear about what we did, and we want people to be absolutely able to trust the measurements we put out because that will be critical to any progress in science" is the most important lesson I took from my mentors, and I agree with it. Gerry insisted that we had to understand everything about our techniques and the uncertainties in what we published.

Chas Beichman told me a very Gerry-like story of the time when he was working on IRAS. They were putting out a catalog of infrared sources, and there was something in the catalog that was just a little bit off. Not much, but a few percent difference between what they saw and what they expected.

Chas said, "Of course there are going to be small discrepancies." But Gerry absolutely would not let that catalog be published until they figured out how to resolve that discrepancy. And it turned out to be an important discrepancy to resolve for the final catalog; it made a difference to the interpretation of one of their topcs. That focus on quality was the most important part of the education I had in graduate school, the attentiveness to the scientific method and the almost German thoroughness about making sure that you understood every element of what you did. Gerry was particularly well suited to be an experimental scientist, and he passed that trait on to his students and postdocs.

A lot of the education you get in science involves learning technical skills including mathematical techniques and how to build things. In many ways, it's the way that you think about science and what you're willing to do to make sure that what you're doing is correct and really does contribute to pushing knowledge forward that's more important than any particular skill. Emphasis on quality was an extremely strong part of the group culture. And to be honest, it was extremely strong overall at Caltech. That's one of the great things that Caltech does in its education of young scholars. The professors don't suffer fools gladly, and they don't want you to just gloss over small problems for convenience. If you're cutting corners, you get weeded out. That's incredibly valuable. Certainly, it's one of the strongest things I took away from my experience there.

Mandatory Instrument Building

ZIERLER: Tell me about the ethos of instrument building in the Neugebauer group. Was the idea that there was an expectation everybody should build their own stuff? Was there anything that was coming off the shelves?

BECKWITH: No, we all had to build something. Some things, you took off the shelf. For example, there was a point at which I was building an instrument, and I needed to find a way to monitor a micrometer setting that I was changing manually in the dark. It was an Ebert-Fastie spectrometer, and you changed the angle of the diffraction grating to change the wavelength across a slit feeding a single detector; we had primitive detectors. I needed a way to make that simple change that didn't require me to look at a micrometer with a flashlight every time. I didn't know much about electronics, so I had to learn. Keith Matthews suggested that I read the Texas Instruments manual to learn about integrated circuit chips. It's really well-written. So I read the manual and designed an electronic system to automatically control the grating angle. remotely. But did I put together everything with transistors? No, I bought integrated circuits. I learned how to buy an integrated circuit and wire it up with feedback and create a working system from off the shelf parts.

There were certainly many things that we did use and didn't build per se. The Ebert-Fastie spectrometer was an existing optical instrument that was lying around unused, but with a different diffraction grating and an infrared detector, we could make it into an infrared instrument. I repurposed that for what I was doing. But there was an expectation that we all had to build something.

My first project was to build an articulated secondary mirror for the 60-Inch Telescope at Mount Wilson so we could make thermal infrared measurements. By an accident of fate, when I was in middle school, I took a class in drafting because I guess a lot of guys from Milwaukee were going to become mechanics. I became pretty good at drafting and design.

So I thought up a design with a lot of advice from Keith Matthews our resident technical guru, then I drew up plans for all the parts in a semi-professional set of drawings, submitted them to the machine shop, and they were impressed because science students didn't usually come to graduate school will skills in drafting and design. We had it built, assembled it on the telescope, and it worked. It was a great feeling for me. We then used it for a number of different projects, and we made a few discoveries. Neil Evans who was a postdoc then and now a professor at UT Austin was my mentor. We were good friends and collaborators, and he showed me how to carry out a scientific project from start to finish, which I didn't know when I arrived. My responsibility was to make the instrument work. It was exhilarating.

When it came time for a thesis, Gerry said, "I know you built one instrument, but we expect our students to build at least two things to show the first one wasn't a fluke." I thought it might be another bait and switch, but it led to my work on the spectrometer, which was richly rewarding and led to more interesting science than the first project.

Yes, everybody was expected to build something in that group. And I think, at least for me, it was very good.

ZIERLER: Do you think having the engineering physics degree from Cornell, not just a physics degree, was specifically useful? Did you draw on that aspect of your education?

BECKWITH: I did. If I had to do it again, I'd do the same thing. I'd get an EP degree at Cornell, and I would not do the regular physics curriculum. Because first of all, we got all of the same physics training that those guys got. We did all the heavy mathematical work with the same textbooks like Jackson's electrodynamics. We did all of the same demanding problems, but we also had some practical problems, where we had to apply the concepts to real world situations, essentially engineering problems. That is not only useful, but also very gratifying because one of the main reasons I was interested in physics is that it explains real world phenomena.

In graduate school, it was important to feel comfortable about calculating anything having to do with instrument performance. I'm using those same skills right now. For my satellite pointing experiment, I worked out the ultimate performance permitted by quantum mechanics to compare with my observations, the so-called shot noise limit. My early results tracked the theoretical performance really well down to a rather low level, then they bottomed out and exposed some systematic floor that I had to understand. Eventually, I figured out where the floor came from and was able to remove it an improve the results of the experiments.

It seemed likely that somebody must've solved this problem before, and I started looking at the literature after I had worked it out on my own. Sure enough, my derivations were not new. Fortunately, the result I derived is what other people worked out, too, so it added to my confidence. But that is part of what anyone should do when working on an instrument: work out from first principles what you should be able to do, and if it wasn't getting close to that limit, figure out how to correct it. That is practical physics.

ZIERLER: Were you contemporaries with Andrea Ghez?

BECKWITH: She is a couple years younger than I am. I think my last year of graduate school was maybe her first year of graduate school, and then I left. Andrea and I got to know one another pretty well in those days.

ZIERLER: Were there any other women in Neugebauer's group?

BECKWITH: There were. Anneila worked in the group for a while, but she didn't really want to build instruments. That's how we got to know one another. I said, "Maybe I can help with some of the engineering details, and you can help with new science from millimeter wave astronomy."

I think it was a tough place for most women in those days, owing to a whole combination of things. Gerry was an absolute gentleman and a leader in trying to create an equitable atmosphere for everyone. But many people worked in those labs with a variety of poorly developed social skills, and the group had a tough-love approach to people surviving, and it was a very male-dominated culture. There was not a whole lot of empathy or handholding. That approach served some people well and other people not so well. It served me well because my personality was such that when given a challenge, I said, "Great, fine. Leave me alone, I'm going to do this." I didn't want handholding, although I was not shy about asking for advice. But there are other people, and I've certainly learned this throughout my career, who would have done better with more encouragement and help. Those people didn't do as well, I think, in that lab.

Andrea's tough and very bright. She did not take any nonsense, but she also had a great personality that smoothed over any friction. She was one of the students who really did well in the infrared lab. Anneila was more senior, her husband was a very distinguished professor at Caltech, and she had a vast social net to draw on, so like Andrea she could hold her own easily. Others had a harder time especially if they were shy or reserved as is typical of scientists.

It has been interesting to watch the evolution of our approach to graduate education. It seems we have gone from one extreme to another. I don't think there's necessarily one true way to provide optimal education for aspiring scholars; different people respond well to different approaches. I was lucky to be in a group where my own personality benefitted from the approach they used. It helped to quell my arrogance and forced me to improve myself, so I'm grateful for that. More coaching, encouragement, and empathy would have served some of my fellow students much better.

ZIERLER: Was anybody in Neugebauer's group or more broadly around Caltech talking about black holes in a rigorous manner?

BECKWITH: Sure, Kip Thorne's group was all about that. Saul Teukolsky and Bill Press were there. These guys were a little older than I was. They were finishing just as I was starting. Dick Bond was there. Dick is a good guy, I really liked him. I think he worked in Kip's group. Anyway, yes, black holes were very much part of the theoretical effort. Gordon Garmire was working on black holes. He was doing X-ray astronomy. The X-ray astronomers, although it was early days, could see from working with the theorists that black holes were one of the few ways you were going to generate enough X-rays to detect with the early technology. Also, that was not long after the discoveries of quasars that prevailing wisdom ascribed to enormous black holes at the centers of galaxies. Maarten Schmidt and Jesse Greenstein, mostly Maarten Schmidt, discovered the quasars in the mid-60s. I arrived in '73, so the discoveries were about 10 years old. There was a huge effort to understand quasars. I recall many seminars during which almost all the theorists said, "The only way you're going to get that kind of continual energy output is through a gravitational field created by a huge black hole." There was skepticism on the part of some of the observers because this was quite an exotic interpretation, and nobody knew how you might get these enormous concentrations of mass. But there was no doubt that it was interesting.

I don't think Andrea was thinking about that at the time, she was working on T Tauri stars, just like many of us in the infrared group, because that's kind of the thing you could do with the early infrared detectors. I don't think she got into the black hole business until she had already graduated and was at UCLA. But you're going back half a century, so I'm lucky I can even remember everyone I worked with.

ZIERLER: Did you get to know Neugebauer on a personal level at all?

BECKWITH: I did.

ZIERLER: What was he like?

BECKWITH: He was a very intense guy. He had very high standards, and he also cared a lot about people. It wasn't always obvious from the way he interacted with them, but it was clear when you got to know him, he really, really did care. There were times when I had some problems, and he was very empathetic to me and helped me overcome my own difficulties. I thought he was a great guy. He did have some habits that struck people as odd. He had corner office on one side of Downs Lab, and Eric Becklin, his closest collaborator, had a corner office on the other side of the building on the same floor. When he wanted to speak with Eric, he didn't pick up the phone or walk around the hall. His door was open, and he just screamed, "Eric," at a volume that would boom around the hallway. Eric would eventually saunter over, although not too quickly.

There was a little interplay between them. I would be sitting in Eric's office, and when Gerry yelled for him, he'd look at me and continue talking. Eric wouldn't respond to Gerry right away. He needed to poke him just a little bit for this. Gerry would yell for me in the same way, and I'd have to get up and come around. There was a lot of behavior that today would be considered dysfunctional and no doubt have HR stage an intervention, but those of us who knew him personally understood this behavior as a sign of affection. He was ultimately a very generous man. I have good memories of Gerry. He was a great role model at Caltech because of his strong moral and ethical compass. Of course, when I got into management in my own career, I had to actively unlearn a lot of the things that I was taught in that group–a lot of bad behavior that I modeled is not acceptable, but it wasn't because of any lack of attempts by the group's leaders to help us develop our scientific careers. They just didn't worry too much about appearances.

ZIERLER: Beyond Palomar and Mount Wilson, was Neugebauer involved anywhere else?

BECKWITH: They went down to Chile to make observations. We did go to Hawaii at least once before I left. Hawaii was still rather new as a desirable site. It had just the 88-Inch Telescope. None of the new large telescopes had been built. But we spent more of our time at Palomar and Mount Wilson. And we were very lucky. We had so much telescope time, we had an abundance of opportunities to try things out that astronomers at other institutions could not easily do. The surfeit of time allowed us to make mistakes and figure them out later to correct them and go back to the telescope having lost only a bit of what we now consider a precious resource. Some of our success came about from the same strategy to win at the lottery. If you buy more tickets, your chances of winning are higher. We had an abundance of time on those telescopes.

The Question of Molecular Hydrogen

ZIERLER: How did you develop your thesis topic?

BECKWITH: I had been working on star formation and related topics, which most of our group did. Right around the time I had to choose a thesis topic, Gary Grasdalen, Steve Strom, and their collaborators made an observation at Kitt Peak with one of their new infrared spectrometers, where they saw emission lines from molecular hydrogen in the Orion Nebula. And their observations attracted a lot of attention, because the infrared vibration-rotation lines they detected had never been seen before in astronomy. Molecular hydrogen had been detected at ultraviolet wavelengths from a rocket-launched telescope, but the infrared lines were thought to be too weak to see, and H2 doesn't have a radio signature. However, it had to be the most abundant molecule in the universe, because it combines two atoms of the most abundant element. No one understood the Grasdalen observation. The data looked kind of crappy. In Neugebauer's group, the Grasdelen group was not considered to be very good at making quality observations, an arrogant attitude for sure, but with an element of truth.

So there was skepticism about the actual discovery. I looked at it and said, "This is really kind of interesting," because all of a sudden, you've got a new way of looking at these clouds and a molecule that is considered to be the most abundant but never before seen. We had technology that was well-suited to making these observations, and I thought that if we're really better than these guys, as everybody in Gerry's group was saying, we ought to be able to do better than they did. I told Gerry, " I want to work on molecular hydrogen." There wasn't a deep scientific question that I was going to address by observing the molecular hydrogen. It just seemed to me that the subject had tremendous potential to tell us some new things about these molecular clouds, and we were in a good position to make better observations.

And he smiled and said, "Oh, okay, that's interesting." And then, Eric Becklin, who was in many ways a mother to all of us, who wanted us to succeed, said, "Steve, I don't think this is a good idea. This is just one observation. This is a random result. You should work on something that you know you'll get a predictable result." My own reaction was, "Screw that. I want to take a chance. Because why else am I doing science? I want to discover new things. If it doesn't work out, I will be okay, I'll just do something else." I talked to Gerry about why I wanted to pursue that topic, and he said, "If you think it's more interesting, that's what you should do." He gave me money to build a better instrument than we had at the time, and that's what I did. We had an abundance of telescope time to try out new projects, and it worked out incredibly well in the end.

One of the first observations we made after I got the instrument working–I think the paper still gets a few citations from time to time–turned out to be a great expansion of the discovery Grasdelen and Strom had made. Eric Persson was a young astronomer at the Carnegie Institute who wanted to work with me on this topic, and he helped with all the initial observations and provided tremendous mentorship to me for our foray into the new area. He was my partner when we started making molecular hydrogen discoveries.

And Eric Becklin was one of the first to get involved when it became obvious that we could detect H2 emission from many different sources. The first night we looked at Orion to verify the Grasdelen results we discovered that the Grasdalen map was too small; it had missed a lot of the emission. [Ed. note: All observations required at least two observers. All of these observations were made with help from various people in the infrared group: Eric Persson, Ian Gatley, and Jay Elias were the most frequent companions, and my recollection was that either Eric Persson or Jay Elias was present for the first Orion work.] We started pointing the telescope at different spots to make a map of the extent of the emission, and I called down the mountain the next morning to either Gerry or Eric Beckin and told them that we had discovered molecular hydrogen emission all over the Orion Nebula. What happened next is that Eric Becklin drove up to Mount Wilson and said, "Okay, we're going to do this together." And Eric was great because Eric was one of the most gifted observers I've ever known. He could make decisions in real time at the telescope that were just astonishing–he was a genius at it.

We did have technical advantages that allowed us to do better than the group at Kitt Peak. As a result, we started discovering molecular hydrogen all over the place. First of all, we found the Orion observation wasn't necessarily wrong, but it was so insensitive that it didn't show the extent and bipolar symmetry of the H2 emission region. Our new discoveries led to some interesting papers. I was on the mountain at the 100-Inch Telescope one night with Ian Gatley. Ian was down in the data room, and I was guiding the telescope. Guiding the 100-Inch Telescope required looking at stars through an eyepiece while sitting on what we called the diving board, a wooden platform about two feet wide extended fifteen feet from the side of the dome on a hydraulic lift over a concrete floor 30 feet below. We would probably not be allowed to use it these days, and you didn't want to fall off the platform while guiding the telecope in the dark. I smelled smoke and then the telescope stopped moving. When we turned on the lights to see what happened, it turned out that the electric motor that controlled the declination of the telescope had caught fire. These were very old direct current motors dating back decades to the early life of the telescope, and they had to be lubricated with kerosene. For some reason, the lubricated motor caught fire that night. Our night assistant put the fire out quickly, and it was not a hazard.

But we could no longer move the telescope in declination except with the fine guidance motor restricting the angles we could observe to a narrow band of a few degrees in declination. The maintenance crew could not fix it until the next day. Ian and I looked at our list of potential targets, and Ian said, "Let's see what's in the catalog close to the 34-degree angle we were stuck with." We went through the catalog, we found one of the famous stars, T Tauri, was at the right declination. So we decided to look at the famous T Tau instead of quitting and going to bed. In the first few minutes we were observing T Tau, we discovered molecular hydrogen emission even though no one would have predicted it to be a molecular emission source.

It turned out to be interesting and gained a lot of attention, because the H2 emission was one of the indicators of mass outflow from the young star hitting a static cloud surrounding it that did not obscure the star itself, something that that none of the pundits would have predicted. We certainly wouldn't have predicted it. But because there wasn't much else we could look at we took a chance and made an important discovery. There are many more stories like that. Much of my thesis was exploring unexpected signatures of the universe's most abundant molecule, and it was great fun. We'd be constrained by some equipment failure, and it would take us down a whole new avenue.

ZIERLER: How much of that is about intuition when you talk about Eric's talents, just a gut feeling for how to do it?

BECKWITH: The truth is, you can call it intuition, but I think it was thousands of hours on the mountain making various kinds of observations. I think his intuition developed from an almost unrivaled amount of practice. He was incredibly good at observing, and he had done so much that those telescopes were like extensions of his body. You can call it intuition, but I think that in Eric's case, it really came from an enormous amount of effort to hone his native skills. I remember how we made that first map of the molecular hydrogen emission in Orion. The group had a green Chevrolet truck to drive everything to and from the observatories. I had put the numbers we got for H2 emission at different telescope pointings on a grid. Eric and I drew the contour map by hand on the hood of that truck, and it was that hand-drawn contour map that ultimately got published. And that's the paper in 1976 that still gets cited. Eric was another of my mentors, and he had talents that none of us could touch, but I did think that it was not just a God-given gift, it came from a lot of hard work.

Mount Wilson and Low Turbulence

ZIERLER: Something I'm very curious about, doing research up on Mount Wilson in the 1970s. When you think about Hale, and Carnegie, and the origins of all this in the San Gabriel Valley in the 1910s, there's no light pollution, there's no smog. Are these things that you're dealing with? How do you get around them?

BECKWITH: It's a good question. First of all, it wasn't a great site for optical astronomy anymore because the Los Angeles was visually very bright. We were working in the infrared, so it didn't make very much difference for our observations.

ZIERLER: What is it about infrared that mitigates these issues to some degree?

BECKWITH: The light pollution is from visible light. It's the wavelengths we see, and it scatters off the atmosphere, clouds, and any particulates in the air. Optical observations are at a wavelength of about 5,000 angstroms, or half a micron, where we see light. It's blue-green light. The artificial lights in the city are at visible wavelengths by design, and the visual astronomy observations are affected by the scattered light and extra background from the city. Whereas at longer infrared wavelengths, two microns to 10 microns, the city does not produce much interfering light. You do have to worry about water vapor in the atmosphere that emits thermal radiation at those wavelengths, but that is a problem at any ground-based observatory.

Mount Wilson and Palomar Mountain are at relatively low elevations, so there is a lot of atmosphere above the telescopes. The background radiation and absorption from water vapor in the sky also affects the observations. That's why Hawaii at an elevation above 13,000 feet is a better site. But southern California has a dry desert climate, and Mount Wilson turns out to be a reasonably good site to do infrared astronomy. And the atmospheric turbulence is very low. What we call the "seeing" was very good. The excellent seeing compensated for a lot of other problems. Those were also fine telescopes, and we did well with them. I didn't really appreciate the importance of all of the work carried out by Hubble, and Hale, and all the great astronomers of that era until I was a little older and began teaching astronomy to university students. I did get to know Allan Sandage pretty well because I got dark time on the 60-Inch Telescope and none of the optical astronomers wanted to use the 60-Inch Telescope, so I was often on the mountain when he was observing with the 100-inch in dark periods making observations of galaxies.

Alan was a very odd duck. But he impressed upon me the history of the place and how much it meant to human civilization. When I got to be a little older, and I started giving a lot of public talks, I would rank the greatest telescopes in history. I was usually talking about the Hubble Space Telescope, which I ranked as the third most important telescope in history. The first was always Galileo's first telescope, and the second the 100-Inch Telescope at Mount Wilson. I thought it was without a doubt the second most important telescope in the history of science.

ZIERLER: This is the one that Hubble used.

BECKWITH: This is the one that Hubble used, first of all, to establish the scale of the universe with the Cepheid variables, right? That was, in a way, what the telescope was built to do. But then, he established that the universe was expanding–which it turns out wasn't really an original observation. It was first done by Slipher in between 1912 and 1917. In 1917, Slipher measured the radial velocities of 25 galaxies and saw that all but three had lines that were red-shifted and therefore moving away from us. Twelve years later, Hubble established this behavior as a law that today we call Hubble's law. His first dataset was pretty crummy, too, but he did understand what it meant. That discovery totally changed our view of how the universe is put together. Even Einstein was surprised. So yes, I began to appreciate, when I was still working on the 100-Inch Telescope, what an important role it had played in the history of science. Later, when I needed to explain to the public why they might want to invest some money in the otherwise impractical science of astronomy, I would use that as one of my prime examples.

ZIERLER: Did you have a sense of the relationship institutionally and interpersonally between Caltech and the Carnegie Institute?

BECKWITH: I knew all about enmity, and I knew about the institutional divorce and the fights. Gerry and Eric, but mostly Gerry, would complain endlessly in the observing rooms, using many expletives, about what they thought of their fellow astronomers–although all of us were actually physicists–especially those at Carnegie. And I'm pretty sure his colleagues felt the same way about him. There was a lot of tension, and I stayed away from it as best I could.

I prefer to not get involved in gossip if I can avoid it, even to this day. I'm the last person to know what the latest rumors are, because I try to avoid talking about them. We couldn't avoid it during observing runs, because we'd spend many nights on the mountain isolated together, and we all talk, so I did hear a lot about the latest drama from my mentors. The fights and disagreements arise from enormous egos trying to work through the issues of shared access to common facilities on which their success depends. And those ego issues continue to this day. They have greatly affected the TMT and GMT projects to the extent that the US may not build either of these telescopes because the players involved are never going to find a way to make the compromises needed to work together.

ZIERLER: What was Neugebauer's style as a mentor? Did you work closely with him? Did you bring him problems, and he helped you work through them?

BECKWITH: No, not really. I think his role as a mentor–first of all, I think there was this general sense then that we had to prove ourselves without too much help. "You're on your own, kid," as the Taylor Swift song goes. You needed to demonstrate that you could carry out research independently. I loved that approach because it was said, "If you're going to be a pioneer, you have to show that you're comfortable going off by yourself with a rifle and a bag of gunpowder and dealing with whatever comes up." It was pretty well-suited for my personality. It wasn't for everybody, and it's not necessarily the way great science is done, but it has its place and was promoted in the infrared group. Now, it could be that some of the other students got much more mentoring that I didn't see. But my recollection was that it was every man or every woman for themselves in the sense of demonstrating what they could do. There was mentoring in the sense of correcting errors and misbehavior. And just speaking for myself, I did plenty of things that needed correction. I broke things, I'd write things up which were wrong, and every now and then, I'd make a bad judgment on where I should be and what I should be doing. I'd get called out on those mistakes, and that was very good for me. I would consider that very much part of the mentoring process, that the group didn't overlook missteps.

ZIERLER: This is going to sound like a long time ago, but did you use computers at all for your thesis research?

BECKWITH: We had access to an IBM 360, a mainframe computer run by a staff that would accept jobs submitted on punched card decks and run programs for you. A lot of our data was written down in notebooks. I've still got those notebooks from my graduate school days. The electronic signals were converted to digital form as a stream of numbers at regular intervals. This analogue to digital converter would then print the numbers on a paper roll, and I'd write them down in my notebook for redundancy, so I had reams of printed numbers. Gerry and Eric would take their paper rolls and submit them to the Caltech computer center, where they had typists who entered the data on punch cards. Normally, every paper roll was transcribed by two independent typists so that they could catch errors in the cases of entries that did not match. Very rigorous.

The punched cards could be analyzed by Fortran programs to do statistics: averages, standard deviations , and things like that. But for my own data, I often used an HP-67 hand calculator to perform the same analyses on my own data reading the numbers from my notebooks. I was lucky in that my father was willing to buy me that calculator–I think it was 100 bucks at the time, not cheap for a student–and it became a valuable tool to allow me to do all my analysis on my own. I analyzed all the data for my thesis in my living room in Pasadena, punching in the numbers from my notebook into the calculator to do all the statistics rather than going to the computer center and having that done. In the end, it was more work for me but gave me the results more quickly than waiting for the data center to do it. By the time you get organized to hand them a coherent set of numbers to enter and write the programs to analyze it, you can just sit down and get it done in a few hours, so that's how I did it. And it turns out often to be very valuable, because reducing data by hand alerts you to patterns in the data, sometimes from systematic noise, that would escape notice in a computer calculation.

ZIERLER: When did you have enough data where you felt like you were ready to defend? Or your sense of when the project felt complete.

BECKWITH: I thought it was never really going to be complete, because we just kept discovering new things that were different and not obviously answering a single question. I would have an interpretation for some of the observations, which was less important in our group than the fact that the data were robust. We just started publishing papers, and eventually everyone thought it was enough for a thesis. Eric Persson was the biggest mentor I had and most influential role model. He contributed enormously to the work and the observing and is a coauthor on every paper. There came a point where we had published maybe four or five papers on the subject, and he thought it would be enough for a thesis. Gerry agreed. Since I've been a professor myself, I've mentored graduate students and can see that my thesis wasn't done in the traditional way testing a specific hypothesis. And in some ways, I regret that lack of pedagogical rigor, because our ideal for research is to apply that scientific method of testing hypotheses. On the other hand, most of those papers were very highly cited, and I had unusually good job opportunities right out of graduate school, so I was quite lucky. The different approach seems to have been accepted by people who appreciated the skills needed to make those discoveries. I was fortunate.

ZIERLER: Were there any other professors at Caltech besides Neugebauer that you were close with or that you considered a mentor?

BECKWITH: I paid a lot of attention to a few. I listened to Wal Sargent because his wife Anneila and I worked together closely. He was a great astronomer. But he was a very different kind of researcher than me. I had dinner at their house a lot, and I listened to him and tried to learn from his approach to choosing scientific problems. Peter Goldreich had a big influence on me. Peter and I were not close, but it was obvious how brilliant he was, and I spent some time with him on road trips to local conferences. There were a few things he said to me during my time at Caltech that had a profound impact on me. I'll just give you a trivial one. I used to smoke cigarettes when I went there, and we were riding in a car together once, and Peter said something to the effect of, "That's just stupid. What would be the purpose of that?" And I quit. I just quit smoking. That was it. Never smoked another cigarette.

I asked him once about how he selected problems to work on. He had published a paper on the pumping of OH masers around giant stars that stimulated a whole line of research I worked on with Neil Evans my earliest mentor and collaborator. Neil and I did some good work on OH masers, and Peter provided the theory of how these OH masers worked and how we could test the theory. But he was sometimes criticized by his colleagues who asked, "Peter, why do you work on this minor topic as opposed to something important such as the origin of the Big Bang?" He told me that there are some patterns seen in the universe that are so regular, they must have a simple explanation, and those are good problems to attack. These patterns can't be random. The universe just didn't fluctuate into a highly ordered state. In contrast, there are some patterns that are so amorphous that there are probably lots of ways to explain them.

Molecular clouds are an example of amorphous patterns. How does a molecular cloud come to look like it does? With gravity and loosely bound matter, it is not surprising that it accumulates in a large, amorphous blob. But the details of the structure? Those are probably too random to be explained by a single idea.

Peter used the example of Saturn's rings as a good pattern problem. He said, "You look at Saturn's rings with their sharp edges, very thin vertical size compared to the horizontal size, and they're highly organized. One overarching cause has to be driving that pattern. That's something I should be able to figure out." A similar insight came to him when the first observations of the light fluctuations in the solar surface showed regular structure reminiscent of a cardigan sweater. These were patterns driven by some kind of regular turbulence. He worked on that problem, and I remember going into his office and seeing the very complicated equations he was working through to explain these patterns. He was a brilliant mathematician. Very involved. And I talked to him about his work to get a sense of how a great thinker approached problems.

Peter had been an engineering physics major at Cornell 10 or more years before I got there. We had the same training. What I took away from Peter was that first of all, a lot of what we do in science is looking for patterns in the universe and explaining these patterns in simplified ways. That's largely what astronomy is all about, and it may be all of science, in fact. A star is a pattern, an object, that needs explanation for its regular properties. A galaxy is a strikingly beautiful pattern, even the amorphous elliptical galaxies; they are examples of observed structures with regular patterns.

This kind of thinking got me interested in the dynamical patterns we call life. Biologists explain the patterns they see as different life forms, species, using Darwin's idea of evolution by descent with modification. That key insight solves the mystery of why you have all these highly organized species and why these species can develop from simple structures like cells in a never-ending competition to adapt to a changing environment. It's a powerfully simple explanation for an enormous variety of things we see in biology.

We now use the word adaptation. Before Darwin, it wasn't a concept that received much attention. Going down in scale by a few orders of magnitude to see what makes life possible, you can see incredible patterns in molecular reactions that also seem to beg simple explanations. Protein synthesis is a good example. It is an elaborate series of complex reactions with very large molecules choreographed to produce more very large complex molecules with extraordinary fidelity. This mechanism was designed by German engineers. It's intricate, it's beautiful. and it is difficult to believe that this complex set of reactions came about through random trial and error with little positive feedback until it was nearly complete. Nature didn't just fluctuate suddenly into protein synthesis. It may have fluctuated among certain paths that led to the final state, but there had to be something that drove it down those pathways and something that selected the right directions along the way. In astronomy, the driving force for most patterns is gravity. I think you can make the argument that every important pattern we astronomers work on, is ultimately driven by gravity. After the Big Bang, we believe gravity separated out from the other forces, and as matter arose, it started to collapse under its own self-gravity. The big patterns we call objects–galaxies, stars, nebulae, etc.–are just frustrated states, often transient, on the way to the ultimate object, a black hole. Even nucleosynthesis is just a reaction to the pressure of gravity in a collapsed object we call a star.

Collapsing matter wants to end up as a black hole. But sometimes it can't, because the nuclear reactions it creates as it gets smaller produce heat that counters the pull of gravity, and it self-supports. Biologists take metabolism and the reproductive mechanisms as given (sort of like our Big Bang) and apply adaptive evolution to the products of reproduction to explain the different life forms. Life's patterns arise automatically through natural selection in an existing ecosystem. But that ecosystem had to arise from inorganic circumstances where there was no metabolism or reproductive mechanism to allow for adaptation. The likelihood that a working system arose just from fluctuations in mixtures of very small molecules seems to me to be very small. It would be preferable to discover a force like gravity with an abundance of free energy to drive pattern formation as frustrated local equilibriums in a gradient of change from one mix to another. Perhaps there are statistical principles or a certain level of feedback along certain patterned directions which make it more likely that you're going to do X instead of Y, and we haven't discovered those yet.

And tying it back to Peter, I remember very clearly Peter's advice about investigating the most highly organized things. If you see a pattern that hasn't been figured out, and it's regular enough that it's likely to have a simple explanation, and if you have the tools, maybe experimental, maybe theoretical, to look for the simplifying drivers, that's how science is going to move forward. My choice of trying to learn enough about biology and biochemistry now to think about this big problem was motivated in part by the highly organized nature of life at the molecular level coupled with the new tools in biochemistry that might make some headway possible in understanding the origin of the molecular organization. To me, the organization of life is more mysterious than organization of galaxies, because we have a pretty good idea of broad principles governing galaxy formation. It's not such a great mystery that when a large collection of matter collapses under its self-gravity that it will make structures like the galaxies we see. Whereas, it is still a great mystery how you get protein synthesis before there were any proteins around to make it viable. That's not so obvious.

I favor the development of an energy engine like metabolism as the first to arise and the more exotic elements like information storage and retrieval coming later. Many biologists have the opposite view favoring the use of information and its preservation in nucleic acids like RNA to be the precursors to all other developments. It seems to me that a source of energy like metabolism to drive flows along gradients as discussed more than 60 years ago by Harold Morowitz is needed to produce more involved organizations that encode themselves as information to store and use. So, I think the information storage and readout came later in contrast to most biologists I know. Anyway, this seems like a great question to pursue in my old age with no other job responsibilities.

When I started thinking about this topic a few years ago, I wasn't sophisticated enough to know if we did have the tools to make inroads. My work as the Vice President for Research at UC taught me that the progress in molecular biology was astonishingly rapid, and I thought the new tools might provide a good hook to make progress on some of the molecular mechanisms. I wanted to be knowledgeable enough to either make a contribution or at least understand the papers when they started being written on how these patterns arose.

Remaining Mysteries of Star Formation

ZIERLER: We're coming up on the second hour now, and we haven't even gotten to the beginning of your professional career, so what I suggest is that we schedule a round two and continue the story. But let's end wrapping up your time at Caltech. Given the fact that your research was making waves, you were getting highly cited, what was the significance of your discoveries, would you say?

BECKWITH: It became clear when we saw the molecular hydrogen emission that it had to be stimulated by some fairly energetic phenomenon. There was a lot of argument about where that energy came from, but there was enough energy that it may have been important for the evolution of the clouds and the formation of new stars. One of the ideas was that the molecules were heated in shock waves driven by some kind of explosion or violent outflow from the young stars in the centers of these clouds, and that was my view as well. The existence of the shockwaves meant that there were things going on in the interiors of these clouds or in the surroundings of these young stars, like the T Tauri stars, which people were just beginning to understand as important to the whole process of the evolution of the clouds and the objects themselves. In the T Tauri stars, the disks of gas and dust yet to be discovered at the time were probably the source of the molecular hydrogen emission.

And I think this is still our understanding, although I must admit, I haven't read any papers on the subject in a couple decades. One of the big mysteries about star formation at the time of this work was that an a priori calculation of the collapse times for clouds implied the rate at which stars should be forming was at least 10 times higher than what was actually observed. Something was preventing very vigorous star formation from occurring. There were a lot of ideas, but no one was sure. One idea was that as stars are assembled, they spew out matter to shed angular momentum and allow compact forms. The mass ejection might be like what is seen from old giant stars. Maybe there's some mechanism to limit the size of any star by shedding unwanted matter. And it's not very different from what happens to prevent a star from becoming a black hole when the interior gets hot enough to create heat through thermonuclear reactions, which then generate pressure to halt the flow.

Perhaps, stars do the same thing to the clouds, and the clouds disperse because of the internal pressure of mass loss. The shock waves heating the molecular hydrogen are a signature of that process. Now, to establish the links showing this is how clouds evolve requires a lot of painstakingly detailed work that I did not have the interest in after we had gotten the first big returns. Some people do. I don't. And it was not clear if all of the things preventing cloud collapse, say, were necessarily going to be observable with the techniques we had, so it may have been an incomplete study at that.

Eventually, people discovered beautifully ordered outflows from young stars in the shape of jets. The Hubble Space Telescope has produced incredible images of young stars in Orion as well as T Tauri stars that demonstrate these young stars are pretty dynamic. They're not just collapsing, they are also ejecting quite a bit of mass and energy. In retrospect, the importance of the H2 emission was to provide a flashlight or beacon on something that showed there were other forces at play in these clouds of broader importance to the creation of young stars and the life cycle of their surrounding cocoons. But even that statement is probably oversimplified, and I am happy that I stopped working on my thesis topic. There's a limit to what you're going to get out of any technique without great technological advances, and we had gone about as far as we could go at the time.

ZIERLER: Last question for today. If I understand correctly, your degree is in physics, not in astronomy. Is there a backstory there?

BECKWITH: I always thought of myself as a physicist. I applied to work in physics at Caltech and discovered the enormous variety of areas where physics training could make an impact. There were people who effectively worked in astronomy, but they weren't thought of as astronomers because they mostly built equipment to do observations using their experimental physics backgrounds. That was true for X-ray and infrared astronomy. Radio astronomy was originally that way, but it matured, and the radioastronomy group was in the astronomy department.

I take a broad view of the skills needed to be a scientist. I don't want to be too siloed. Astronomy and astrophysics are a subset of physics to me. And astronomy puts some of the greatest demands of applied physics on the investigator. You really have to know more applied physics to do astronomy well than any other field, in my view. You have to know about nuclear, atomic, and molecular physics. You have to know something about solid state physics and plasma physics. You need to be comfortable with large scale dynamical phenomenon as well as gravity and turbulence. You have to know a lot of stuff. The breadth of astrophysics is very attractive to me.

Do I consider myself an astronomer or physicist? Honestly, I consider myself a scientist. I love physics, and I love astronomy, and now I love biochemistry and biology as well. I have been very lucky in life to be in the position to work on many interesting things. I don't even know what my current identity would be; whenever I am asked, however, I say that I am an astronomer.

It's been a lot of fun to return to my roots as a builder of instruments. I got a chance to go back to Mount Wilson a few years ago, when I was the director of the Space Sciences Lab at Berkeley. I toured the dome of 60-Inch Telescope, a place where I spent months of time as a graduate student. The chopping secondary I built in 1973 is still up there. It's completely rusted, the mirror is in the state of whatever happens to aluminum after 50 years of sitting out in the smog and almost black. But there it was, and I knew that I built it when I was half a century younger than I am today.

ZIERLER: Well, we'll begin next time when you begin your professor's life at Cornell, and we'll go from there.

[End of Recording]

[Begin Recording]

ZIERLER: This is David Zierler, Director of the Caltech Heritage Project. It is Tuesday, May 28, 2024. It is wonderful to be back with Professor Steven Beckwith. Steve, is great to be with you again. Thank you so much.

BECKWITH: Thank you.

Back to Cornell

ZIERLER: We're going to pick up in 1978. You are named an Assistant Professor of Astronomy at Cornell. We talked about this a little last time. You just leapfrogged the postdoc. You went straight from Caltech to Cornell.

BECKWITH: Yes.

ZIERLER: What does that tell us about the job market at that time and how quickly some of the departments wanted to snap you up?

BECKWITH: I had two offers. I had offers from Cornell and the University of Texas. It was great fortune for me, and it was lucky that two departments were searching for instrumentalists at the time. I wouldn't say it was all to the good because there were still things that I could've learned as a postdoc, especially if I'd had a senior mentor, that would have helped me make the transition to faculty work easier. I had a chance to work with Charlie Townes, for example, that I turned down. I am sure studying under Charlie would have been a great education.

One of my early experiences as a young faculty member would be considered inappropriate today but turned out to be good for me. I approached one of the senior professors, a good friend and probably the person most responsible for having me hired at Cornell, after about six months on the job and asked him for advice. I said, "I've been here for about six months. I've been working hard at research and teaching. Can you give me any pointers, or advice, or feedback on what you think?" He said, "Steve, let me tell you what we think. At Cornell, if you have to ask what it takes to be a professor, you shouldn't be one." Now, these days that would be considered an unacceptably harsh reply. It would probably require trigger warnings. But for me, this was an exceptionally good response. Because it taught me the rules of the game. You're a professor here, and you're expected to be a leader. You're not expected to be a follower or a trainee. Show us what you got, and don't complain about the opportunity.

I was young, and because of my luck, I was too self-assured. And that response was exactly the sort of thing that I needed put me back in my place and tell me to figure out what I had to do to get to the next level. Not everyone would respond well to what I was told, and today's system formal requirements for mentoring and feedback and being attentive to the feelings of young scholars is probably preferable to help most people succeed. But there was an advantage to giving someone like me a wakeup call, and I appreciated it.

It was also clear that if I had been a postdoc first, I would've been better positioned to jump right in and take up my faculty duties more easily. The first five years I was an assistant professor, I worked punishing hours, medical intern levels of commitment. I had a lot of learning to make up. I was teaching astronomy without ever having taken a formal course in astronomy. I was trying to get funding for my research for the first time. But I was also quite lucky to be in the position I was in, and I my training at Caltech was good enough to help me learn what I needed to pretty quickly.

I guess I was attractive to these schools because I had the skills to build instrumentation and was also skilled enough in theoretical work to be able to understand and use the kind of theory that we teach our students. I excelled at some of my class work at Caltech. Cornell's EP program was as good a preparation for those challenges as any I could imagine. And, of course, much of that early success was just good luck. I was lucky to have picked a risky thesis topic that actually succeeded. It was work that got a lot of attention at the time. And departments wanted instrumentalists. It was still a time in astronomy where building instruments was seen as a way to unearth novel discoveries. That thinking was true at Cornell, and it was true at Austin. For the most part, my peers were typically quite good at the technical things required to make instruments, or they were gifted doing mathematics and physical theory, desirable traits for a professor, but not always both. I think I had that combination at a time when there was a market for it.

But the biggest factor was that I got lucky at the right time. New discoveries don't happen all the time. There's a tendency for those of us who are successful to look back and attribute that success to our unique and wonderful qualities. And everybody has some unique and wonderful qualities. In many cases, what distinguishes successful people is that they had the right timing for the work they decided to do, and that was true for me.

I loved Cornell. It is a great place to live and work. It is a great intellectual university. It has a different character than Caltech. Caltech seems to be uniquely focused on pushing forward certain boundaries and making sure their scientists are winning the top honors in their fields. Cornell is more eclectic. They really like to have very broad thinkers and tolerate non-conformity: Carl Sagan or Tommy Gold or Martin Harwit. Cornell hired people in the humanities who took a similarly eclectic approach to their scholarship. I loved that part of Cornell. That was my start. I worked very hard learning the subjects that I was going to teach and getting the benefits of being embedded in an intellectual culture.

The first courses I taught were the junior-senior astrophysics courses for students concentrating on astrophysics. We didn't have an astronomy undergraduate major, but there were people who concentrated in the subject. I used Martin Harwit's book for some of these courses, and Frank Shu's book for others. I learned a tremendous amount. I often had to start learning the subjects from scratch. But I found out, as most professors do, that the best way to learn a subject is to teach it. Except for a few embarrassing moments where I was in front of a large audience, and I said something that turned out to be inaccurate–I misunderstood some element, and somebody called me out on it. Except for a few of those, it was a great experience, and taught me a lot about astronomy, and helped me to put many of the research projects in better perspective.

ZIERLER: As you got situated in the astronomy department, you learned from your colleagues. What was Cornell astronomy known for? What were its areas of strength?

BECKWITH: We had four main concentrations. One was infrared astronomy. Martin Harwit and Jim Houck were the two other professors in that area when I came in. That was one pillar. Which was interesting because that was not common. Caltech had it, too. A second was planetary science. We had very strong planetary science faculty: Joe Ververka, Peter Gierasch, and Carl Sagan. Later Steve Squires, and I may be forgetting someone. Cornell had a library of all the data from missions like the Voyagers. And of course, Carl had a very high public profile, so that brought attention to Cornell. A third was theoretical astrophysics, and it was an extraordinarily strong department. Tommy Gold was a bit of a dilettante but an original thinker. Peter Goldreich was Tommy's student. We had Ed Salpeter, who was one of the reigning theorists of his generation.

Hans Bethe, who was also a reigning theorist. He was in the physics department, but he was the one who figured out nuclear fusion, how that all worked. Bethe had brought a bunch of people from Los Alamos when he came to Cornell. Ed Salpeter was one of them. Richard Feynman was another at the time. Stu Shapiro and Saul Teukolsky, they were brilliant theorists who worked on gravity problems and gravitational waves. It set the standard for a lot of what we see now in the gravitational wave detectors. Saul was Kip Thorne's student at Caltech when I was a student, so I knew him there. Just a brilliant mathematician. Ira Wasserman and Dave Chernoff were hired after I came. They were both good friends of mine.

The fourth area was radio astronomy. Cornell managed the Arecibo observatory, and there were a number of faculty who used Arecibo for their work, particularly my friend Jim Cordes, but also Frank Drake who was very well known. Because we ran the Arecibo Observatory, we usually had one or two specialists in radar astronomy, who bounced radar waves off the moon and nearby planets. It was quite an interesting business. Those were the four pillars. And we had our own building that they eventually expanded as the faculty grew.

ZIERLER: I know you were focused on molecular hydrogen emission work during these years. What were some of the big questions? What was your work in this area?

BECKWITH: The original question was why you see this stuff in the first place. And we knew it was there before it was observed. Molecular hydrogen had to be the most abundant molecule in the universe. The problem with it is, it's homopolar. It's a diatomic molecule, but the two atoms are identical, hydrogen in this case. And that means it doesn't have a dipole moment. Even though molecules emit radiation according to quantum mechanical principles, the same ideas from classical electricity and magnetism apply, which means dipoles produce the strongest radiation. The next term in the expansion of radiation moments is quadrupole. Quadrapole radiation is substantially weaker than dipole radiation, but it's the dominant form when a radiator has no dipole moment. Molecular hydrogen is a quadrapole radiator, and it is not very intense even when excited. This is true for the emission caused by vibration or rotation of the molecule, but not for the electronic bands that depend on electron transitions in the molecular orbitals. Those were in the ultraviolet and not so easy to observe. You needed a space telescope.

The primary vibration-rotation bands are in the near infrared. The radiation strengths are weak. What we saw in our observations was pretty bright. The question was, why is it so bright? Because to excite those energy levels in a hot gas required temperatures of a few thousand degrees. We were detecting line emission from regions where the temperatures were thought to be very cold. There were two ideas. One was that ultraviolet light from very hot stars stimulated the molecules into higher electronic levels, and the subsequent cascade transferred energy into the infrared lines. It's called ultraviolet pumping.

The other was that the molecules were heated to temperatures of several thousand degrees, and the molecular levels were excited by the high temperatures. The heating had to come from shockwaves that heated the gas for a short time as they passed through the molecular clouds. No one had detected large reservoirs of hot gas except through the molecular hydrogen lines, so hot gas had to be a very small fraction of the cloud at any time. The next question is where did the shock waves come from? What's that all about? We saw molecular hydrogen emission all over the inside of a dense molecular cloud, which was thought to be cold. Everybody began to realize that during the course of star formation, the stars don't just collapse to form new stars within quiescent environments, but they violently spew out enormous amounts of energy and matter as well. Many of the outflows are narrowly channeled into bipolar jets. This pattern of mass loss was even more difficult to explain than a more uniform spherical expansion.

The molecular hydrogen emission pointed to the introduction of a lot of energy into these clouds from the stars themselves. This was also true of the T Tauri stars. T Tauri stars were a little bit farther along in their early evolution, but they continued to spew out a lot of matter. That was interesting. It had a number of potential consequences. One is that if you're putting a lot of energy into the cloud, it probably affects the way the cloud creates other stars. One of the big puzzles at that time was that the rate of collapse based on the free fall time, an easy calculation, implied a rate of star formation about 100 times bigger than what was actually observed. Something was preventing new stars from being made as easily as they should have been.

The prevailing wisdom became that these episodes of mass loss during the collapse of a young stars kept further collapse at bay and stopped them from forming new stars at a high rate. It presumably plays a critical role governing the rate of formation in the larger clouds. A second question is, why is the matter not coming out at all angles as from an explosion. It's typically channeled into two narrow opposing jets that emanate along an axis that probably coincides with the net angular momentum axis of the star. That's not an intuitive result to me. I don't think anybody would've predicted that result ahead of time until it was actually observed. In the scramble to figure out what's going on, most theorists decided that magnetic fields had to be responsible for the channeling of the outflows.

It is interesting that this phenomenon of narrow collimation of mass outflow spans many orders of magnitude in celestial objects. The same narrow jets were seen in distant quasars with the VLBI techniques looking at radio waves from hot plasmas. The quasars, which were presumably million- to billion-solar mass black holes, also got rid of their energy and mass through narrow jets. Those jets are orders of magnitude bigger, longer, and more energetic than the ones from young stars. It suggests, following Peter Goldreich's principle, that there's some simple universal mechanism driving mass loss across six orders of magnitude or more of size and energy scale. That is an interesting discovery. Figuring out those mechanisms is what we do in science.

The same thing is true when you look at life. I don't think the universe just wandered into life. I think it drove life into existence. We don't understand what the primary driving force is, but it's probably some kind of natural energy flow. Back in the days when jets were first discovered in the massive black holes that are responsible for quasars and also in young stars, they indicated a general pattern in astrophysics requiring some deep explanation for something that's common across many scales.

I don't know if we've ever quite agreed on the explanation. There has been a lot of good work on this subject, especially by Roger Blandford who was at Caltech for many years before moving to Stanford. His group did much of the original work on one of the prevailing theories. There were other groups such as the one headed by Richard Lovelace, a professor in Applied and Engineering Physics at Cornell, with competing theories that could also explain these patterns. A lot of the work involves plasma physics, and Cornell has traditionally been strong in plasma physics.

Ralph Pudritz at McMaster University had yet another mechanism to explain the collimation that differs from these other theories. The competing theories all use magnetic fields to channel the energy and mass of these outflows. All these ideas are slightly unsatisfactory in the sense that they are not simple. All create a clear, simple pattern from an amalgam of complex circumstances. And sure, that can happen. We have to figure out what nature's actually doing. But it's a bit counterintuitive. When you see a very simple pattern, you hope there is a simple explanation, as I learned from Peter Goldreich.

The Utility of Speckle Astronomy

ZIERLER: A technical question. Tell me about speckle interferometry. What is that?

BECKWITH: One of the things you always want for an instrument is to reach the fundamental limits imposed by physics for any technique. For a telescope, the best angular resolution should be limited only by the diffraction of the light from the entrance pupil of the telescope, the primary mirror or primary lens. The diffraction limit for a large telescope, such as the 100-Inch telescope on Mount Wilson, is about 50 milliseconds of arc. That's a very, very small angle. A ground-based telescope looking up through the atmosphere does not get close to that resolution. Under periods of extraordinarily low turbulence or good astronomical seeing, the images might be slightly smaller than one second of arc, 20 times larger than the diffraction limit. The seeing is often much worse than that, depending how much turbulence is in the upper atmosphere at a particular site.

The air has an index of refraction that changes with its density, and there are density fluctuations in the atmosphere because of temperature fluctuations. Think of them as frozen bubbles in air that move with the air as the wind blows. Light from sources above the atmosphere is refracted by these turbulent bubbles before it reaches the telescope, and that refraction breaks up the wave front and distorts the images. The images in the focal plane of the telescope are large blobs of light with constantly changing structure. I spent hours of my youth looking at these blobs while guiding large telescopes. A snapshot of the image with a very high-resolution camera, perhaps a 10-millisecond exposure, shows that the image is broken up into a collection of substructures called speckles, and each speckle is about 50 milliseconds of arc across, the same size as the diffraction limit of the telescope.

The effect of the atmospheric distortion is to preserve the best resolution of the telescope in a zillion little images called speckles. These speckles move around rapidly inside the image as the wind blows. People studying these images–especially specialists in military surveillance–realized that even if you can't correct the distortion, you can recover most of the resolution if you record images fast enough to freeze the speckles and use the speckles to synthesize an image with a resolution limited only by the telescope's diffraction. The speckles are randomly dispersed within the image, so you need some technique to either line them up and stack them, or what we usually do is work with spatial Fourier transforms of the images and recover the amplitude, not the phase, to reconstruct higher resolution images.

If you take a lot of snapshots, analyze their Fourier transforms, and use them to find the unchanging features, you can, in principle, reconstruct images of the thing you're looking at, at the diffraction limit of the telescope. That's a big gain, right? Even on a night of good seeing, you could gain a factor of 20 in angular resolution, which is a huge jump. That technique is called speckle interferometry, because you use the speckles with the Fourier analysis techniques reminiscent of what is done with interferometers to reconstruct your image. It has a lot in common with what radio astronomers do with distributed radio telescopes.

I got interested in the technique because my experience as a student at Caltech was that large advances in observing power inevitably led to discovering all sorts of great stuff. Sort of like with molecular hydrogen. I was introduced to infrared speckle interferometry by Mel Dyke at the University of Hawaii and Ben Zuckerman who was at the University of Maryland at the time, I think. Bob Howell was working with Mel and developed a lot of the software needed to do the work. Mel had his own technique of observing infrared speckles. It was crude because we didn't have infrared cameras only single detectors, so Mel would focus an image on a narrow slit in the focal plane of the telescope in front of the infrared detector and sweep the image back and forth across the slit to record the speckles as brief flashes of light in the time series of output from the detector. One dimensional Fourier transforms of those time series allowed him reconstruct the image in one dimension. By rotating the orientation of the slit and the image motion, it was possible to get diffraction-limited information about the two-dimensional image. I loved this approach.

I wasn't even most interested in the initial science Mel and Ben were doing–I think there was an emphasis on observing close binary stars for which speckle interferometry is ideally suited–but I saw tremendous potential to explore new kinds of observations because of the big gain in angular resolution. The gain at infrared wavelengths was not as great as in the optical–instead of a factor of 20, we got a factor of 5 or 10–but that is still a substantial gain. The instrumental challenges also matched the skills I had. It was fairly easy to modify existing instruments to record the speckle patterns. The main new requirement was computer programming. For my own first instrument, I bought one of the first IBM PCs–I think it was in 1981–with 64k on the motherboard. Imagine that. The Apple watch I am wearing has something like half a terabyte. It's unbelievable to think what we were working with then.

One of the first tasks was to see if we could calculate the Fourier transforms in real time. I looked at technical specs for the CPU in that first IBM PC and calculated the time it would take to do a fast Fourier transform of our data using the standard algorithm. I coded it in Basic, which was the main language of the PC, and it was about 10 times slower than what the CPU should allow. So I taught myself the assembler language for the Intel 8086 and 8087 processors and wrote the Cooley-Tukey fast Fourier transform butterfly algorithm in assembler for the IBM PC. And that code performed at the calculated speed and allowed us to get real-time results from our simple speckle observations at the telescope. I still have the HP-16C calculator that I used work through all the binary and hexadecimal code needed to write the assembly program, and I am using it now to do some programming on the project for Jessica Lu and her team.

Mel, Ben, Bob, and I made a number of discoveries with this new technique. The first big discoveries were in the young stars HL Tau and R Mon. At very high resolution, they looked like they had little disks of dust, something that had not been seen in young stars, although Martin Cohen had suggested that HL Tau had a disk existed based on lower resolution data. A disk is exactly the sort of structure that was thought necessary to create planets. And we were the first ones to see one of these disks around a young star. We were a good team to make this kind of discovery. Bob Howell was a gifted programmer, Mel and I were pretty good at building instruments and making them work, and Ben had a deep knowledge of many interesting astronomical questions and results on young stars, so it was a good group to work with. That work started my interest in planet-forming disks around stars, which turned out to be the best work I ever did.

Circumstellar Disks Are Real

ZIERLER: This is circumstellar disks you're referring to.

BECKWITH: That's right. We were the first ones to really demonstrate they existed. I realized at some point that if we were looking at dust disks, we ought to be able to see their thermal emission at much longer wavelengths–the speckle observations were of scattered light at 2 microns–and the thermal emission might even be seen in millimeter or submillimeter observations. Anneila Sargent was one of my closest friends in graduate school. She was part of the Caltech Millimeter Observatory, so I got in touch with her and suggested we look at some of these stars with millimeter waves. The interferometer produced super high resolution. We could actually resolve the disks." She was up for it. We were good friends. Half of science is having fun, it seems to me. We said, "Hey, this is going to be fun. Let's do that."

She then took the initiative to use some of her own time on the interferometer; she didn't have to go through any committees to try out our ideas. A new element was to look for the gas in the disks using the CO signatures. We made the observations, and sure enough, we started detecting these disks around the young stars in gas and in dust. HL Tau was the first one. We could see the disk, but it was still a bit fuzzy leaving some room for doubt. Today, you can get pictures with ALMA of these disks, and there's no doubt about the structure. Our first observations showed structures that were sort of blobby, and we had to use the velocity patterns from the CO emission to argue that the gas was in Keplerian rotation that could only come from a disk of gas orbiting the star. Even our collaborators, Ben and Nick Scoville, were skeptical of our interpretation, since it did have profound implications for planet formation. Anneila and I had to pressure them to get our interpretation in the first paper. But subsequently, we did a lot more work with far better tools to understand the physics of what was going on, and our original ideas were confirmed.

We also realized that if we wanted to look for disks in a large sample, we needed a lot of collecting area and a way to observe stars quickly. Interferometers were hard to use for this task. We wanted a big millimeter-wave dish with a broad-band detector for the thermal emission. The biggest one at the time was in Europe. It was run by the Max Planck Institute for Radio Astronomy, and I think they worked jointly with the French and Spanish researchers. IRAM was its name. We asked if we could get some time to carry out a survey with their millimeter-wave bolometer, and they gave us some time with the stipulation that one of their researchers actually carry out the observations. We ultimately wound up looking at–I think we did our survey of 100 stars with this telescope and detected emission that we attributed to disks from about half of them.

The large fraction of stars with disks was a big deal. First of all, we argued that the dust we were seeing in these disks was the kind of matter that it took to create planets. It looked very much like what theorists thought the early solar system had to look like. The amount of dust in these disks was about the amount thought to be in the early solar system disk, too, so the quantity of matter matched well with ideas about planet formation. But nobody really knew if these disks were common or rare before our survey. And our observations showed that it was so common, we made the bold statement that if these things evolve into planets, then most stars have planetary systems. That was quite a discovery. We published our results in 1990.

ZIERLER: This would've really been the intellectual foundation of exoplanet research.

BECKWITH: Well, sort of. Let's give the exoplanet people their own credit. We know there is a solar system, right? And you can calculate what it would take to detect our solar system around a nearby star. Independent of any disks or any of that, going out and looking for solar systems with the right instrument was a very good and interesting scientific project. The problem was, it was very hard to do because the signatures of the planets on the motion of the star–the reaction of the star to the orbital motion of the planets–are pretty small. Astronomers looking for those signatures had to make huge improvements in instrumentation and use a lot of observing time to have a hope of detecting these signatures. It was so interesting, though, that I think exoplanet research would have proceeded whether or not we made our discoveries, and I believe the early work in that area started before we published our results on disks, although the results on exoplanets did not come out until five years after our disk work.

But I have to believe that when we came out with our results in the early 90s, it gave those people a lot of encouragement to not give up. Because their early techniques were pretty limited, and they might have easily worried that the solar system was unique. In science, you need patience. And while I don't think our work was necessarily the foundation for exoplanet research, I do think it must've given everyone a lot of hope, and I do think we established early on that planets were probably common as opposed to being rare. And it turned out to be true. Now that you can actually detect the planets, people have demonstrated that most of these stars do have planets.

ZIERLER: Just a terminology question, circumstellar disks and protoplanetary disks, what's the difference?

BECKWITH: In principle, a circumstellar disk can be around an old star. Or any star. When you're doing star-formation research, there is a very subtle difference. If you look deep in a molecular cloud like the Orion Nebula, and you find sources that you only see at infrared wavelengths, optical light doesn't escape, and some of these objects have the characteristics that they just collapsed to that state, you could call those protoplanetary disks because you believe they will create planets after some time. You would call those stars protostellar. The same is true for the young T Tauri stars. Whereas a much older star might have a disk that was born much later and they would not be likely to create planets. Those would still be called circumstellar disks. But the terminology is loose, and in most cases it does not matter much which term you use.

The Move to Max Planck

ZIERLER: You mentioned the Max Planck Institute. I wonder what the initial connection point there was that made you think that there'd be a big move to Germany in the cards for you.

BECKWITH: I didn't think that there would be a big move to Germany in the cards. That was a complete accident, another piece of luck actually. I first made the connection to the Max Planck Institute for Radio Astronomy–which was not the Institute I eventually went to–because they were the ones running the millimeter facility we needed. That's what we needed for the science, so that's who we contacted, and it worked out. The people at the MPIfR did not regard me as a potential candidate for an MPI Directorship. Moving to Germany was a completely different thing.

My wife, Susan, is a linguist by training. She got her PhD in linguistics at Cornell. Her specialty is northern Germanic languages. She can read old Icelandic eddas, old English, Latin and Greek, and I am really in awe of her abilities with language. Her main specialty was German, and she is fluent in the language. She was working for a company in Middletown, New York, and that company also had a branch office in Germany. I was up for a sabbatical leave, and we were married. And I thought, "Well, I could take a sabbatical leave in Germany the would give her the ability to work for her company in Germany while we were abroad, so it wouldn't disrupt her work. Now, Germany would not normally have been my first choice. I went to the Max Planck Institute for Astronomy, because I thought it was close to her company's office. The MPIA was a sleepy backwater place not making a big impact in astronomy at the time. But I was open to new experiences, and I wanted to support her career, too, so we went to Germany and spent six months there.

And that visit was a long saga because our daughter, our first child, was born there in June just two days before we were scheduled to return to Ithaca. She was born prematurely, and she turned out to be the youngest baby they had ever been able to save. Of course, we delayed our return. We were supposed to come back to the US on July 2, as I recall. Our daughter was born on June 30 at a low birth weight, and we stayed for four more months while she was getting well enough to travel with us. That was a tough time of our lives. I didn't speak German with any fluency. Susan was fluent, of course, but that put a lot of pressure on her because especially in the intensive care unit, whenever I had a question, she had to translate, and she was a young mother who did not know from day to day if her daughter would survive. The doctors all spoke English, but most of the nurses didn't, so she'd translate, and it was all very stressful. We did get through it, our daughter survived without any long-term medical problems, and that became my main connection to Germany and the Max Planck Institute for Astronomy. And I must say, the Germans in the hospital and in the Institute were tremendously supportive and helped us in many, many ways for which we were very grateful.

I'd made friends at the Max Planck Institute for Astronomy, and I got to appreciate the system, and how it worked. A couple years later, they were searching for another director, and evidently my name was put in the hopper because I had that connection, and I had a high enough profile as a professor to be a credible candidate. I had never planned to leave the United States or to leave Cornell. I loved Cornell. Susan didn't really want to move, but she was willing to do it on my behalf, and the fact that she was fluent in German, and had taught there and been a student there meant that she thought she could accommodate my ambition.

Ultimately, I got the call. And when I got the call, I couldn't believe the opportunity it presented. I could not believe the resources that are available to a Max Planck Institute director and the security to try out new things with minimal interference. It was like nothing I'd ever heard of, and I was at a stage in my work where I needed the resources available at the Max Planck to remain competitive for building instruments. The world was no longer using single detectors for infrared instruments but used new semiconductor chips with thousands of detectors to make full images possible. These detectors cost half a million dollars, and it cost another half a million for the electronics to read them out. Nowadays, you can buy detectors for amateur telescopes for a few hundred dollars that are orders of magnitude more powerful than what was available back then. I've got one here. I think I spent maybe $500 on this camera. Anyway, it was becoming a big-money business, and it was clear to me that if I wanted to remain competitive in building instrumentation on my own, I needed a source of funding much greater than I could expect from the NSF.

Having the resources to be able to build instruments with modern detectors required jumping to a higher level of support, and the Max Planck directorship offered exactly those resources, so I took the opportunity, and it was a great experience. It was the best job I've ever had, without a doubt. I loved the job. I learned German and really tried to understand the European astropolitics to influence the course of European astronomy. I loved all the new challenges. But it was hard on Susan. Even with her fluency and knowledge of the culture, she was what they call an Ausländer, and she had to deal with labor laws that discriminated against her.

She had good jobs working for IBM and then SAP, but every year the government labor department, the Arbeitsampt, required those companies to advertise her job to see if a German citizen was available to take it. That was not a sustainable situation for us. The difference in our status in German society was enormous. At one point, I took the family to Euro Disney when our kids were old enough to appreciate it. We booked a trip to Euro Disney and planned to take the train. It wasn't that expensive, and I was pretty well-paid by European standards. But the big challenge was for me to buy the train tickets. We went down to the station about half an hour before the train was scheduled to leave, and as I stood in line to buy the tickets, and I realized that the time required just to get to the ticket window meant we might miss the train.

I pushed aggressively and wound up getting tickets at the last minute and had to run to the platform to meet Susan and the kids. I was sweating. We got on the train just before the doors closed, and they gave me quite a bit of grief because it turns out I had never before had to buy train tickets for myself. Someone at the Institute always took care of it. If I wanted to travel, someone else would deal with the details. I had no idea what it was like to even buy a train ticket. And there were plenty of my fellow Max-Planck directors who had their own drivers. It's a pretty comfortable life. But it did highlight the tension in my family, because Susan was dealing with this kind of hassle every day, and I was shielded from it. She also had chores dealing with the kids and the schools that are difficult in any culture but especially difficult in a foreign one. Our lives were very asymmetric, and it was unsustainable, so there came a time when I needed to look for another job in the US.

ZIERLER: What were your most important priorities directing the Max Planck Institute for Astronomy?

BECKWITH: Number one, I wanted to reform the culture so that the Institute itself was a great place for scientists to come, and work, and produce. I believed that the best science is done primarily because you get the best people. You could see that at Caltech. That's the Caltech philosophy, and they're absolutely right about it. A small institution like Caltech has been world-class because they pay attention to attracting top talent in the world. And I thought, given the resources we had and the situation–Heidelberg's a very pretty city, it's a nice place to live, really–attracting top talent was the number one priority. And it wasn't attractive to outsiders when I arrived, because my co-director ran it in an old-fashioned hierarchical manner, where he largely dictated what people had to do. The local people were used to that system, but it was completely stultified and unattractive to foreigners and even other Germans not from Heidelberg. No top intellectual wanted to come and work in that environment.

That goal meant a culture change. Culture change in established institutions is exceedingly difficult, if not impossible, although I did not know that when I started. I learned that especially when I came back to America and tried unsuccessfully to make some changes in other institutions. But it worked out well at the MPIA probably because there was an unspoken hunger for it. And part of the reason it did work out is because I had so much power and authority as a Max Planck director that people were willing to give it a shot. The MPIA is vibrant now. It serves as a magnet for young students and young scientists to spend a few years. They've attracted top talent in new directors, top talent in scientists, and their output is great; they're doing interesting things. Reforming the culture did not help my personal science very much, but it seemed to me that a great privilege had been given to me by German society, and I wanted to produce something with lasting value more than just promoting my own scientific research.

The expectation of a Max Planck director is to use the resources to win a Nobel Prize, and most people try to do that often successfully. But I felt that I had been given a larger responsibility–this is maybe an American attitude–to maximize the use of their resources for German science in general, and it should go beyond me. I wanted to create an institution that constantly fosters great science. And I wanted to either identify, find, stimulate, or dictate one or two good science projects and instrumentation projects, which would distinguish the Institute from other places in the world, so that we had unique ways of doing our research, and we had unique scientific problems we were working on. I worked hard to do that, and I think with some success.

And a third goal was to help raise European astronomy in general. Europe had lagged the United States in astronomy for over a century since the telescopes on Mount Wilson and Palomar had been built. The US had the largest optical telescopes, built some of the largest radio telescopes, and started a space program that spawned x-ray and gamma-ray astronomy. They had also taken the lead on the Hubble Space Telescope and got the bulk of the time for US astronomers. Europe had been through two devasting wars, and investment in astronomy was not something they put much effort into until late in the 20th century. The European Space Agency was relatively young, whereas NASA was mature and well-funded and committed to doing research in astronomy among other fields.

And, of course, the very diverse cultures, languages, and governments made it difficult for Europeans to cooperate on large scale research projects. In Alsace, you could drive across the border from Germany to France, and the restaurants would claim they spoke no German. This was nonsense. Of course they spoke German. They wouldn't speak German with us. There's still quite a bit of enmity across borders, and you could see all the jockeying among the French, and the Germans, and the Italians at advisory council meetings trying to shape European-wide investments in astronomy. A lot of the cultural stereotypes are accurate. But the European scientists realized that they were only going to succeed if they worked together to get resources vastly larger than what any one country could provide. They had to become a kind of the United States of Europe, and they were beginning to do that. It was really interesting for me to participate in that progress. And they have succeeded beyond what most people would have envisioned. Honestly, the VLT, the Very Large Telescope, is a more powerful observatory now than the Keck Observatory or the Gemini observatory is. Europe has claimed that lead. The next-generation large telescope, the European Extremely Large Telescope, EELT, is fully funded, it is under construction, and it will be bigger than whatever the United States has, whether the US builds the Thirty-Meter Telescope or the Carnegie one.

Even if we wind up building the TMT–and that seems increasingly unlikely right now–Europe will still have a more powerful facility. And they are doing some beautiful science with new space instruments, too. It was exciting for me to be part of that renaissance. It was not easy. I barely spoke German, although I speak English well, and English is the common language of European astro-politics. But being able to speak French or Italian would've been an enormous advantage in these meetings. The most sophisticated of my colleagues switched seamlessly from French to German, to Italian, to English. They could carry out politics in any of these languages, and that was very important to get people to cooperate.

ESO hired Riccardo Giacconi to be the Director General the European Southern Observatory when they were starting to build the VLT. Riccardo had a successful career managing very large projects as well as doing Nobel-prize-winning scientific research. Riccardo was a bit aggressive for European tastes; they were often at odds with his approach, but he was an incredibly good manager and largely responsible for the success of ESO during the period I was there. I was the head of ESO's Science and Technical Committee for a few years. I loved the role and my peers on the committee and did my best to promote European science. I was also on the European Space Agency council for a while and worked hard on that. I found the roles interesting and although none of them helped my personal science, I found it rewarding to promote science overall, but especially European science. They had tremendous potential. In these respects, I was probably not a traditional Max Planck director.

ZIERLER: What are you most proud of when you look at your legacy at Max Planck?

BECKWITH: The culture.

ZIERLER: Which, as you mentioned, persists to this day.

BECKWITH: It wasn't just me. I helped hire Hans-Walter Rix, the successor of my co-director Hans Elsasser. Same thing. He was German, but he was working in Arizona, and I helped attract him back to Germany. He was young, 33 years old when he was hired. It was a risk to hire someone that young as a director. But he continued the cultural changes I brought about and, in some ways, accelerated them. Then Thomas Henning was hired to replace me, and the two of them continued to build on what I had started and took the Institute much farther than I ever did.

I have no doubt that starting down that direction is what created the successful culture at the MPIA, which will ultimately give rise to much more, much better science than I was likely to do on my own. And I am proud of that. I used to go back to Heidelberg every year, just to soak in the ambiance and reconnect with old friends. I was distinguished older visitor, although otherwise useless. But the Max Planck Society would pay for me my travel, and the trips were always fun. I could speak a little German, walk the forest paths, visit my old haunts, and sit at cafes with colleagues I liked.

That early experience taught me a lot, too. In subsequent leadership jobs, I felt that fostering the right culture was more important than most other things. But it's also very difficult to do with an established culture, especially in the United States where scientific directors are often viewed skeptically. It is so difficult that, in fact, I'm not sure I really succeeded anywhere else. But those changes in Heidelberg are very much what I'm proud of. On a personal level, I enjoyed embracing and learning another culture and another language. It gave me a much broader perspective on the world in general, and a lot broader perspective on the United States. Not a lot of Americans do that.

Joint Appointment at Hopkins and STSci

ZIERLER: Were you specifically looking for opportunities back in the United States? And then, once it became available, how did the Space Telescope Science Institute directorship come onto your desk?

BECKWITH: Someone called me and said, "What do you think about applying for this job?" I thought it was a good opportunity. It required some of skills that I had acquired by being a director in Germany. Not all people in science have that kind of leadership experience. I applied for the job at the time when I had promised Susan to bring the family back to the US where she would be treated more equally, and the timing was just good luck. It was good luck that I got the job, and it was also good luck that it came at the right time for us. I didn't see many other things that were going to work out as well.

ZIERLER: And did the professorship at Hopkins come as part of a package deal, or did it happen sequentially?

BECKWITH: It was a package deal. It was important to me, actually. I value my identity as a professor. Would I have taken the job without that appointment? Maybe. I thought the potential to help science by being head of the Space Telescope Science Institute was as great as any other job. Even though I had far less power and probably even a lot less influence than I did in Germany, the Hubble Space Telescope was the most important astronomical facility in the world. As I said earlier, it was the third most important telescope in history after Galileo's first telescope and the 100-Inch Telescope at Mount Wilson.

Of course, I was a bit naive about what I was actually going to be able to do as STScI Director, but seemed to me that the potential to produce great science for the world by running that operation well was greater than anything else I could do. Certainly greater than going back to being a university professor. I had already had a taste of how good it felt to help other people succeed in a way that would outlive me, which is the potential reward of leadership. And I came to understand how important good management and administration are to the academic enterprise, despite the normal disdain of the faculty–and some of these jobs are brutally hard. Imagine the challenges of the president of Harvard right now. Not for me.

But I understand why these jobs can be so enticing to people. Our universities and research organizations are incredibly important for modern society to flourish, and they depend on having top leaders to keep them healthy. Most professors don't see what goes on at the high levels of administration; they don't understand the tradeoffs faced every day by the heads of institutions, because the administrators are the interface between the real world and the university world. The university world is a nice world, but it's not the real world, and there are many forces from outside the university working to derail its mission. I sensed that the Space Telescope Science Institute had similar challenges that would need to be managed well to keep the world's most important astronomical telescope healthy and available to astronomers to carry out their research. The jobs are politically fraught, of course. They're fragile. They don't guarantee tenure, although most have secure jobs to fall back to when it becomes impossible to remain in them. But if you do those jobs well, you have the potential to have a huge impact on science. And that turned out by an unfortunate accident, to be true for me at the Space Telescope Science Institute in ways that I not only couldn't have seen but would not have wanted to see before I went there.

ZIERLER: Did you come to Space Telescope Science with the idea of also a research culture change? Or were things good in that regard?

BECKWITH: No, the research culture at STScI was really good. In fact, I didn't have to make any major change at all. It had been set up well by my predecessors Riccardo Giacconi and Bob Williams. The one thing that I was asked to do when hired was to raise the standards for promotion and tenure at STScI. The governing bodies thought the promotion criteria had gotten too lax, and they were tenuring people who might not have been promoted anywhere else. My instructions were to apply the same criteria that would be applied at a place like Caltech or Cornell for promotion and tenure. And I did that. I don't think there were any exceptions to my new high standards.

But enforcing a higher bar for promotion was not popular with the staff. It created a lot of controversy, because I turned people down for tenure who were popular. They may have had the support of 80% of the faculty. But my policy was that at a place like Cornell or Caltech, 20% who doubted a candidate's desirability for tenure is too much doubt. I expected the successful candidates to have 90% or more support from their peers to be promoted. There is often enough personal friction that generates some opposition to tenure that must be allowed for, but it should be minimal. Somebody at Cornell once told me, "The way we see it is, if there's any doubt, that's too much doubt." That's a judgment call that I enforced, but it wasn't a great way to be a popular director. At first, the advisory councils liked the high standards, but eventually they had to worry about whether my approach allowed me to be an effective leader of the staff. It wasn't an easy question.

But I thought that an important part of my job was to foster a culture of intellectual excellence over many decades. To the extent that there was any culture change at STScI, I enforced a higher standard for tenure. Otherwise, the Institute was functioning very well. We brought about a lot of interesting changes in the way we approached proposal review that were important for selecting the best science to do with the space telescope. Many of these did not originate with me but came from the staff. Meg Urry, who is now a professor at Yale, was one of the people who proposed major improvements in our proposal selection process that continue to this day.

Meg looked closely at the past selection cycles and discovered several practices that appeared to give unfair advantages to the reviewers themselves. She proposed changes that eliminated those problems. She was also worried about the ability of groups to do big projects that required a lot of telescope time and yet fared poorly in proposal reviews in which committees tended to reward small proposals over larger ones preventing us from supporting some important and unique science. I recall that Duccio Macchetto and Robert Brown were also leaders helping us make big changes in the way we allocated time.

In response to Meg's analysis, we changed the entire structure of the proposal review committees and also the way we competed different sized projects against one another for peer review. Many of the Institute's staff helped out with these policies. And they had a profound impact on the science projects we selected. Only a few of the ideas originated with me, but they all spoke to my desire to get the best scientific programs to maintain support for the mission and to be sure we were not discriminating in favor of a few people because of our selection processes. For example, we instituted a policy as part of my director's purview to dedicate 25% of the telescope time to large proposals, projects requiring more than 100 orbits of time. All the large proposals competed against one another instead of competing with the small projects. That approach balanced the scales and removed the social barriers to getting large allocations of telescope time. That change alone was successful at bringing in some of our highest impact science projects during the time I was there.

These are some of the accomplishments that I found gratifying and outlived my time at STScI. That was the way I used my administrative role. And then, of course, there was the Ultra-Deep Field. Once again, a deep field wasn't my original idea, because Bob Williams had already shown the power of doing it with the telescope's first instruments. We calculated the likely impact of doing a new deep field with one of the new instruments planned for the final servicing mission, the Advanced Camera for Surveys and found ACS could improve our capability to do deep, wide-field imaging by about a factor of 10 compared to Bob's original work. That was significant. There was a question of whether the project should be competed or simply carried out at the Director's discretion. We had found out that the CCD detectors used in the camera slowly degraded from radiation damages when they got into orbit, and I felt that if we were going to do another deep field and get the maximum benefit, we should do it when the detectors were fresh immediately after the new instrument was installed after the servicing mission. If we had put the project into peer review, we believed there was a high risk of not being ready in time to take full advantage of the fresh detectors, The actual work needed to plan and carry out the imaging was straightforward to handle at STScI but would have required a lot of additional effort by any group without access to our internal resources and staff. The timing was important, so with the blessing of our oversight committees, I dedicated a large fraction of director's discretionary time to the Ultra-Deep-Field.

And then, of course, it was important that it become a community resource and not an advantage for me or our scientific staff. The data had to become public immediately. And the people who got the most scientific results from the Ultra-Deep Field were not at STScI, although a few of our staff did do a lot of good work on the data. We supported a lot of scientists in the community with money, too, to help them support their teams. That was consistent with Bob Williams' original intention of using a public facility for the wide benefit of the astronomical community.

Washington and Science Policy

ZIERLER: Being at Space Telescope, did that give you a close view, a front-row seat, as it were, to NASA, Congress, and the federal aspects of astronomy policy?

BECKWITH: It's half the job. Maybe more than half. In the end, it was the whole job, right? When Sean O'Keefe decided to cancel the final servicing mission to the Hubble Space Telescope, it was going to kill the telescope and probably require laying off a large fraction of the staff at STScI. My sense at the time was that there was almost nobody else well positioned to prevent that from happening besides the Director of the Space Telescope Science Institute. So I worked actively and openly to reverse O'Keefe's decision. That was all politics.

I knew Senator Barbara Mikulski well. She was the head of the Senate Appropriations Committee, one of the most powerful positions in Congress. I helped raise money for her campaign. Bruce Margon, our deputy director for science, is a gifted communicator, and Bruce and I would walk the halls of Congress, talking to different senators and representatives about the need to support Hubble and NASA to carry out a final servicing mission. And we worked with the National Academy to get a recommendation to reverse the decision from the nation's most prestigious scientific body.

This occurred during the George W. Bush Administration. Sean O'Keefe was appointed by Bush as I recall, and we weren't going to get traction with the President to have O'Keefe change his mind. We did have some good luck that encouraged the administration to have O'Keefe reconsider by working with journalists to make our case to the public. This is a beautiful story. There was a guy in the White House named Andrew Card, who I think was the president's chief of staff. He was a very senior, very powerful guy. Bruce and I cultivated our relationships with the newspapers, especially the science writers in the Washington Post and the New York Times. And we got to know a young science writer at the Washington Post named Guy Gugliotta [sp?]. We invited him to the Institute, and we talked to him about all the science that the telescope could do with another servicing mission. Guy soaked it all up and seemed to be really eager to write an article in support of another servicing mission. Then, we saw nothing in the few weeks after his visit. No story came out in the Washington Post. We thought we had wasted our time. Because everybody in Washington reads The Post. Then, I believe it was the first week of May a few months after Guy's visit, around my wedding anniversary, one day, Guy published an article on the front page of The Washington Post with the story: "School Children Donate Lunch Money for the Hubble Space Telescope." Guy had gone to Ohio, and he visited a classroom of elementary school kids who were so enamored with the stories coming out of the Hubble Space Telescope about astronomy that they all decided to pool their lunch money and send it to NASA to support another servicing mission.

It was about 20 bucks. A local doctor said, "Well, this isn't much," so he chipped in another 100 dollars to make it credible. The Hubble Space Telescope cost $10 a second for observations, so in 10 seconds, 100 bucks is gone. Anyway, they sent $120 to Washington, and it became a big human interest story that appeared on the front page of The Washington Post. And we heard that morning the story was published, Andrew Card picked up the phone, called Sean O'Keefe and said, "Sean you've got to fix this problem." That was a big part of our political education and ultimate success.

ZIERLER: It must've worked.

BECKWITH: It was part of a campaign. There were other parts where we had to work with Senator Mikulski to stop NASA from reassigning employees in the Hubble program to other missions to prevent the team from falling apart. It was a complex business. I worked a lot with John Bahcall. He was instrumental in doing some of the things I could not do. He'd do anything to save the telescope. I'd get material to John from NASA so that he could use it in the National Academy and Congress. Susan and I were at home one winter morning sitting in our living room, and there was a knock on our door. One of the Goddard guys had driven up to hand me a stack of papers with orders for internal staffing changes, and he said, "If these are implemented, we're dead."

I got them to the right political people so that they stopped those changes from being implemented. It was an interesting time in the sense of the Chinese curse. In the end, it was like the battle of Troy. All the kings and princes died. Sean O'Keefe and I both left our jobs in a big shakeup, and others succeeded us. But the Hubble Space Telescope was saved. Bush appointed a new NASA director who wanted to service the telescope. And that telescope's still up there as we speak, and it's still producing good science. This is 2024. The events I related were in 2004, 20 years ago. If I were asked what my legacy was at STScI, it's getting a servicing mission to keep the observatory active. The achievement wasn't mine alone, but I don't think it would have succeeded without my efforts.

ZIERLER: Was the professorship at Hopkins sort of your scientific refuge? Is that where you went to get the science done?

BECKWITH: When I stepped down from the directorship, I went across the road and took up my professor's post for three years. I was intending to stay for a long while until the job at UC came up. I liked Baltimore. Susan liked it, too, and it was where we raised our kids. I was in my 50s and liked being a professor, so it was a nice landing spot. I planned to teach and write a book. Johns Hopkins is a great university. It was a good place for me and my family. I never did teach because they gave me a few years before having formal duties, and by the time I was ready to teach my first class, UC offered me a new job, and I left Hopkins. But I did teach when I came to UC.

Administrative Leadership at the University of California

ZIERLER: How did the UC job come available? Why was it attractive to you?

BECKWITH: I got a call from someone who said, "There's this job available, and you might be suitable for it." It did look interesting. There were two elements of it that I liked. One element was, and this was true for my career since the Max Planck, and that it had responsibility for the allocation of a lot of resources. The Vice president for Research at the Office of the President doesn't spend the resources for his own research, and much of the funding is earmarked for specific projects. But it was a big flow of money, about $150 million a year, if I recall correctly. And the Vice President's office had a big staff. With that level of resources, I thought, I would find something really interesting to do to promote science.

It seemed to me that anybody who's talented in one of these jobs can find ways to channel resources to things which should advance knowledge. That was a big part of the attraction. But the other part of the attraction is that while I do love astronomy, I think of myself as a scientist first with wide interests. For a long time, I thought of myself as a physicist, but then even that was too narrow. I love figuring out how things work in the world, and if you're the vice president for research, you've got a responsibility to understand how all the different fields of research work. Maybe an impossible task, but at least it's a good reason to learn new areas particularly the ascendant science of molecular biology.

There are a lot of interesting research problems in the humanities, too. It turns out that the humanities had one of the biggest demands for computing power of any research area in the UC: digital dance. For reasons I still don't understand, dance groups from different campuses wanted to coordinate their performances at a distance using digital video transmission. The computing power needed to do what they wanted to do was enormous. The challenge was to figure out how to help these people get what they need. Anyway, for me, at that time of my career, I felt the breadth of the intellectual challenges was tremendously attractive.

I had been in leadership positions long enough that I began to appreciate the joy of having breadth as opposed to depth. Academics are rewarded for deep new insights into specific topics and not breadth of interest. So taking on roles that required breadth rather than depth meant giving up the opportunity for the usual academic rewards most people seek especially when they are young. But the intellectual rewards in understanding many new areas were the most important to me; it was just so much fun exploring new territory that I was very lucky to have a career that would give me those choices.

ZIERLER: To clarify, being vice president for research in graduate studies, it's located physically in Berkeley, but it's for the entire UC system?

BECKWITH: Located in Oakland. There is a difference. I worked in downtown Oakland. I also got an appointment as a professor at the campus of my choice. I was really flattered that any of the campuses were happy to have me as a professor. Berkeley was the natural one. Santa Cruz would also have been a great choice, but it was farther away from where we lived. My wife and I wanted to live in the city. We had lived in a German village, we had a big country place, a rural place, in Ithaca, and we lived in the suburbs of Baltimore, where we raised our kids. We knew that San Francisco is one of the great cities in the world and thought "Why don't we just live in San Francisco and let me commute across the Bay?" So we did. We bought a condo here in San Francisco and live right in the heart of the city south of Market Street. But proximity to campus makes a big difference, and Berkeley was the closest one.

With Berkeley, I could hop on the BART and ride to either UCOP in Oakland or Berkeley in just a few minutes. Or I could drive my car over. It's very easy. Santa Cruz and Davis are much farther to commute to, and every other campus would have required me to move when I left the Office of the President. I chose Berkeley, and I really love the choice. It's one of the great universities of the world. I'm very proud to be associated with it. But my work as vice president was in the Office of the President in Oakland. And if you walked into that office, it has no quality of a university that you can detect except for logos on the walls. That's it. Otherwise, it looks like an insurance agency. I'm not knocking insurance agencies, but they're not universities.

ZIERLER: How did you get a sense of your priorities? Did this come from the president? How much autonomy did you have in figuring out where to place your efforts?

BECKWITH: I had a lot of autonomy. I was given broad instructions and told to use my judgment. Some things were particularly important to Mark Yudof, the president who succeeded Bob Dynes shortly after Dynes hired me, and I carried them out as best I could. There was a sense among the regents that the Office of the President had grown too big and powerful, and as one of the regents said, bloated. One of my jobs was to reduce bloat, and I did that. There was a lot of it in my office. I had 225 people on my staff when I came, and it looked like we could do the core work we needed with less than half that many. Downsizing is not a popular thing to do as a leader, but the whole Office of the President was trying to shrink, and we systematically started reducing staff.

And that's painful to do. It's painful for the people affected, and it's painful for the person making the decisions. We were trying to find ways to move people to the campuses without losing their work. Some people just had to be let go. It was hard work of the sort that is not, at least for me, very pleasant. But it was also the right thing to do because honestly, we got down to 100 people before Mark Yudof said, "Okay, Steve, I think that's enough," and we were able to function quite well with that smaller number.

I also had a lot of other priorities to improve the way we were supporting research. We provided funding from the Office of the President to the campuses for what were called multicampus research initiatives. And this funding had become a sinecure for the people who had gotten it many years ago. It was obvious that some of the groups we funded were underperforming at a level that was an embarrassment. And yet, the system was stuck, and it was politically difficult to use that money to fund fresh initiatives. One of the reasons the staff was so large was that each campus needed to have some of their own people in my office to preserve the money flowing to their campus research groups. I believe we spent $23 million dollars a year on these groups in 2008 when I arrived.

I made the decision that we should recompete all of the funding in a round of new proposals. Everything was going to be up for competition: existing programs and any new programs the campuses wanted to propose, and we solicited proposals widely across the UC system. This was not popular with the people who were already funded, and the people who were going to get new funding did not know who they were, so the political pressure was to keep the status quo without changes. But it was the right thing to do to remove the uncompetitive programs and reward the good programs and provide a way for new efforts to get support.

We recompeted almost everything and made the selection with a system of peer review panels using the best practices I learned from working at STScI and the very effective review system that evolved over more than a decade of experience. It was interesting that 55% of the programs we were already funding did well under peer review and kept their support. They put in excellent proposals and demonstrated that they were worth funding. But fully 45% of what we were funding did not do well in competition. A fair fraction of them were in the bottom quartile of their competed areas. We had been funding some very poor efforts. We also got a whole lot of new programs that were often the most interesting in their areas. The very top proposals were new and were better than anything we had been funding previously. The opportunity cost of the existing system was enormous. One program led by Dr. Laura Esserman, a renowned doctor and medical researcher at UCSF, was to combine the data from all five hospitals in the UC system to improve breast cancer research. The UC system has a lot of patients, and the combined data would bring a qualitative improvement to the statistics of clinical trials. Before Laura made her proposal, we could not coordinate research between any of them because they all have their separate IRBs, institutional review boards, and they all do things differently.

That proposal was such an obvious multicampus effort that UC could do that no other institution in the world could do that it showed why we needed to find new ways to use the university's money if we wanted UC to remain a premier place. Laura's was the top proposal in her area. And there were other new ideas that became the best in their areas but were unknown before we opened up the opportunity. It produced a lot of good for UC research.

Now, my political mistake was that when we removed the internal political pressure to maintain specific funding to groups on the campuses by putting them up for competition, it also meant that the senior administrators who needed money could easily cut the budget for this funding without worrying about the political pain. A budget cut just meant we had a smaller pool for the next competition and not specific cuts to individuals who could then lobby the president and the chancellors to keep their support. The president wound up cutting that program by quite a bit, and it was a shame because we supported some really interesting research in those multicampus units which was unique and could not be done in any other system. But those are the things that you learn when you're in upper management; we were always jockeying for resources, and those were lean times. I did not come into that job with a great skill in competing for resources in a standing bureaucracy, but I did come to appreciate how large institutions work and how difficult it is to change their directions even in the face of overwhelming evidence that they need to change.

ZIERLER: Of course, the backdrop to all this is the financial crash of 2008. How did you ride that out, and when did things start to feel like they were getting better?

BECKWITH: That was part of the reason for cutting down money and staff, because we didn't have enough money to do everything. But it was necessary. We rode it out successfully because our presidents made it our top priority. I think most of the credit goes to Mark Yudof. Mark was sophisticated in working with the regents to affect change. He didn't always do what they wanted, but he did get them on his side using a combination of persuasion and pressure. And he worked well with the state government. He had to work closely with Jerry Brown at the end of his term, which wasn't so easy, but he managed.

I thought he navigated that period pretty well. Yes, we were taking cuts, and yes, we had to reduce staff, but at least in my area, those cuts did not greatly reduce the effectiveness of the overall operation. When an institution needs money, one of the first things that's cut is research.

There are a lot of interesting problems at the level of vice president for research that most faculty don't see and are only aware of in their particular areas. It was interesting to see all of these to understand how institutions, particularly UC, had made decisions about what it was going to do and still preserve its academic character.

The Native American Grave Protection and Repatriation Act is something that few people have heard of. Whenever there is digging for construction, it's common to find human bones, and someone has to decide whether they belong to local native tribes or are super ancient and can't be attributed to any modern group. In either case, they can useful for scientific research. That was a problem that affected all the paleontology museums at campuses like Berkeley and also arose at some university construction sites. Several controversies about the universities museum collections and bones unearthed at construction sites came to my office and needed mediation between our faculty and native American tribes. Another policy that affected our ability to carry out research freely was export controls. The federal government shifted the oversight of export control policy in the Bush Administration from the Department of Commerce to the State Department. And as a result, because State was worried more about security than commerce or research, suddenly many technologies which were commonly used in research became restricted because they were considered to be defense related. That shift restricted a lot of important equipment especially for international collaborations. It was something we had to constantly navigate.

We had a lot of work on the ownership of patents and decisions about investing money to patent inventions by the faculty. Most of the regents were very successful businesspeople, and when the money started getting tight, many wanted us to monetize all the great discoveries made at the campuses to make up some of the shortfall, so there was a big push to make my office concentrate more on patents and licenses rather than research oversight. I believe one of my successors was hired specifically for her expertise in commercializing technology. Unfortunately, it was never going to yield enough money to make a meaningful contribution to the university's budget, but that wasn't something you could tell the regents.

We had the database of all the patents and licenses that had ever been granted to UC going back half a century. I made a spreadsheet and did some analysis of our history and found that even the most successful patents were never going to make up more than a tiny amount of income for the University. Supporting faculty patents is still a good thing to do, but it cannot address the budget issues caused by a lack of state support and pressure to keep tuition low, so it should have not taken priority over other issues. In any of these jobs, there's a certain fraction of your time you spend on things that are probably useless, but you have to do them anyway.

Small Satellites for the Next Generation of Astronomy

ZIERLER: We'll bring the conversation right to the present. Tell me the most important things you're currently working on and how they relate to some of the questions you've always been pursuing.

BECKWITH: I'm part of a consortium headed by Professor Jessica Lu, where we want to create a constellation of small satellites with telescopes to do astronomy from space. The main departure from traditional advances in astronomy instrumentation is to think small. Astronomers always want to build a bigger version of whatever telescopes they have at any time. What do we want after building the 10-m Keck telescope? We want a Thirty-Meter Telescope. We never do something smaller than what we had before. But if you understand what can be done with any telescope in space, even a telescope that is 15 centimeters in diameter is remarkably powerful for a lot of science, then if you can build them sufficiently inexpensively enough, you can envision a constellation of hundreds of such satellites looking independently at different parts of the sky to cover almost the entire sky at all times, and that would be a big advance for astronomy without developing larger instruments than we have now.

There are a lot of the interesting transient phenomena, for example, the gamma ray bursts and gravitational waves that are bright but fleeting. You don't need a powerful telescope like JWST or Hubble to see them. But you do need to be pointed at them when they go off, if you want to observe the changes in brightness. Most of the time, a single telescope is looking somewhere else and would miss the brightening even if it were sensitive enough to see it. The idea is to create satellites which are small, about the size of a bread box, with small telescopes, self-contained, and make them cheap enough using the advances in CubeSat technology, the off-the-shelf technology you buy and slap together, so you put several hundred of these satellites in orbit. We're thinking maybe 300 satellites.

And with 300 satellites and small telescopes with wide fields of view and modern detectors, you could look at most of the sky all of the time. If something goes bump in the night, so to speak, one of your telescopes is going to be pointed at it and will be able to observe the way it brightens and dims. That observation opens up a number of interesting possibilities for science. Jessica's science is primarily on black holes, so she's interested in looking for those rare times when a black hole passes across the face of a distant star and brightens the star through gravitational lensing. With a complete observation of the brightness variations, you can calculate all the properties you need to about the black hole. By counting a lot of these events, you can figure out how many black holes are out there and how they are distributed over different masses, which is not known. And it is a very fundamental new area to understand because all matter wants to collapse into a black hole, the maximum state of entropy for matter.

Jessica is interested in how many free-floating black holes are out there and how they vary in mass. The constellation of telescopes is a perfect way to find out. Any time a black hole passes across the face of a star, we would see it. Even these small telescopes are powerful enough to observe very faint stars, allowing an enormous statistical sample. And there are other transient sources that it would be interesting to observe. One of my colleagues from Cornell, Dave Chernoff, pointed out to us that if there are cosmic strings, and there are plausible reasons there might be, you'd observe the brightening caused by cosmic strings passing in front of stars. These events would be very rare, 1 out of 100,000,000 stars might be affected each a year, so you'd need something that could look at a huge number of stars simultaneously to see even one event. They leave a very big signature, so if cosmic strings are out there, we'd see them. Gravitational wave sources are going off all the time. The facilities are incredibly sensitive, but they can't localize the sources to better than a 30-degree angle in the sky. Not all of them put out light that can be seen at visual wavelengths, but some of them do, and some of those are particularly interesting. When two neutron stars come together and coalesce, they produce visible light signatures. There's a whole series of sources that could be interesting.

One of the reasons I like this project for Berkeley is that it could be a new anchor facility for Berkeley. Berkeley, like Caltech, has had in the past facilities that were its own that gave it an advantage for certain types of science compared to other astronomy departments in the world. Whether it's the Owens Valley radio observatory for Caltech, Hat Creek observatory for Berkeley, or the Mount Wilson, Palomar, and Keck observatories, these unique facilities attract the very best scientists to our institutions. That is what keeps our institutions at the forefront of research.

The other thing I like about Jessica's project is that if it works and I can contribute to it, it leaves Berkeley with a legacy of a brand-new scientific facility, an anchor facility, which would create, I think, many new science opportunities for the department and continue to attract top talent. It combines science with the ability to maintain the science culture that keeps a place like ours at the top of the game.

ZIERLER: It really sounds like a capstone for your career, enabling the science and building the instruments.

BECKWITH: It's what I got interested in doing fifty years ago when I started out in astronomy.

ZIERLER: I want to thank you so much for spending this time with me. It's been a wonderful set of conversations. Your perspective is just great to have captured for history. I want to thank you so much.

[END]