The Drake Equation Turns 50: An interview with Frank Drake

Ordinarily, science journalists do not report on the work of their close relatives. But on the occasion of certain special anniversaries, an exception may be allowed. In this case, the occasion is the 50th anniversary of the Drake equation, the formula for predicting how many detectable civilizations exist in the Milky Way galaxy. That equation was first written on a blackboard on November 1, 1961, at a conference organized by Frank Drake, who is today chairman emeritus at the SETI Institute, the organization whose mission is the search for extraterrestrial intelligence. He is the father of Science News astronomy writer Nadia Drake, who recently interviewed him about the origin of his equation and its relevance to the search for intelligent life in faraway stellar systems today.

Why did you write the Drake equation?

I was motivated by a desire to understand what governs how many civilizations there are to detect in our galaxy, and to see if I could quantify that number. It became the agenda for a meeting.

(Interview continues below the box.)

What was the meeting?

Well, Project Ozma [the first SETI search] happened in the spring of 1960. And it got a lot of press attention, much more than it deserved, and this caused the National Academy of Sciences to say, “Well, gee. This is a surprise, maybe we should have some kind of study to see how real this is, and what the chances are of actually succeeding.” So they asked me to convene a meeting of experts in the field to discuss the subject and arrive at what it’s going to take to succeed.

And that was in 1961.

I organized the meeting in the summer of 1961, almost a year after the original experiment. It ended up being a three-day meeting held at Green Bank [West Virginia], just after Halloween.

I was the whole organizing committee — the local organizing committee, the scientific organizing committee, the entertainment organizer.

Around Halloween, I realized we couldn’t all just get together in a room and start talking. So I thought, “Well, we should have a subject for each session. We ought to talk about key aspects of this whole life in the universe thing, what we knew about it, and what needed to be done.”

I’d been thinking about it for months, and all that thinking was based on the history of the solar system and life on Earth. It had to do with planetary formation, how many planetary systems there are, how often life arrives, how inevitable is the evolution of intelligence, how often do you develop a detectable technology, and how long that technology lasts.

I knew all these things were crucial to the end result, which was the total number of detectable civilizations.

So I decided to make each one of them the topic for a session. And I realized then that if you multiplied these factors together, you got the thing you were after, which is the number of detectable civilizations.

Then?

Then I just wrote that thing down and started the meeting on November 1 by putting the equation on the blackboard. There is now a plaque on the wall saying, “This is where the equation was first written.” I didn’t jump up and down yell “Eureka!” or wave my arms or anything. It was just something that had been in my mind for months. It wasn’t “a-ha,” it was just — to me — obvious.

Who was at the meeting?

I invited everybody in the world who I knew was interested in this subject, and that amounted to 12 people. Among them were Carl Sagan, Otto Struve — the great astrophysicist of the 20th century, who was then the director of Green Bank — John Lilly, who’d dealt with the intelligence of dolphins and was making the case that intelligence was a very widespread phenomenon. And Melvin Calvin, who was very big on the fundamentals of biology, and the origins of life and all that. He’s the one who identified the chemical processes that take place during photosynthesis [Calvin’s Nobel Prize was announced during the meeting]. And then there were some other people from the electronics industry, and people who worked on the theories of the origin and abundance of planetary systems. Barney Oliver was one of them.

What was their reaction to the equation?

They thought it was a wonderful way of describing, with one equation, the history of intelligent life in our solar system. They thought it was a succinct, clear way to tell what was important and what we needed to quantify. As a thinking tool, it’s very good.

What was the reaction of the broader astronomical community? Were they critical, accepting?

They accepted it, didn’t consider it some great breakthrough or something. They noted that it was an interesting observation, if you will.

An observation?

An observation of the way to quantify the history of life, and intelligent life, on a place like the Earth. And the equation slowly took hold because people started realizing it was really helpful in organizing their thinking. The main product or benefit that’s come from it is that it organizes people’s thinking, gets them to thinking about what’s important.

The nature of the equation is that everything is equally important — it’s all there in the first power. Nothing is squared, there are no logarithms, no trigonometry. It’s all very simple. It shows that everything is equally important.

Now it appears in most of the astronomy textbooks. There’s one book out about history’s most important equations, and there are about three equations in it, and it’s one of them. That boggles my mind. It’s right there next to E=mc².

A couple of weeks ago, I was wearing my Drake equation T-shirt, and people would stop and say, “Oh, I know what that is!” Does it surprise you that it’s so popular, even outside of astronomy?

It used to surprise me, but it doesn’t surprise me now. People do use it often in teaching and in science magazines. It’s a great teaching tool, and it’s been repeated over and over in articles about ET life. A lot of people know about it.

Sometimes people use it to make jokes, or use the form of the equation to find something else. There’s one on the Web where the person does a calculation of the probability of finding a desirable date in Chicago. And the “Flake Equation” calculates the number of people who claim they’ve been abducted by UFOs. Its comes out to about the right answer….

Is there anything you would change or modify, if you were writing the equation today?

I wouldn’t change it. I wouldn’t change the factors, but I would discuss each one, and what you mean by them. Because what you mean depends on what answer you’re after — whether it’s the number of civilizations, or the number of detectable ones, or something else.

What are some of the criticisms you’ve heard, and what is your response?

Mostly the criticisms are actually suggesting the adding or modifying of factors in the equation. My usual response is, “what you’re saying is correct but what you’re wanting to add is something that’s in the existing factors.”

For example, when it comes to the fraction of creatures that have developed intelligence, you may have to exclude creatures like dolphins, which are intelligent, but could never develop a technology. Same thing with developing a detectable technology. You can’t think every intelligent creature could do that. On a water planet, for example, it seems unlikely that creatures could build giant telescopes and radio transmitters.

Nobody ever says the whole thing is wrong or needs replacing. It has survived for 50 years without really needing any additions or changes.

But you do have to be very precise in just what the factors mean. For example, the equation can be used in several ways. L could be the lifetime a civilization exists, in general. Then the answer is just the number of civilizations, not detectable civilizations. It’s the lifetime, period, and has nothing to do with whether they ever have detectable technology or not. The number of detectable civilizations is going to be much less than the total number of intelligent civilizations.

Another problem is the number of detectable planets – does a thing like Europa count? It’s an ocean world, there’s no land mass, there’s no way to process ores, there’s no fire. Does that count or not? If you’re just trying to get the number of habitable planets, then you include the Europas. But if you’re trying to arrive at the number of technology-using planets, then you might exclude the Europas. You have to be precise about what you’re talking about. Do you really just mean habitable, or do you mean the potential for having detectable technologies?

And of course, Europa is actually a moon.

That’s right. When we talk about the number of habitable planets, what we really mean is the number of habitable bodies. There may be systems where there are more habitable moons than planets.

So if we’re talking about n-sub-e, and that actually includes habitable moons, that would change the estimate.

That’s right. Originally the term did not include moons. It was strictly planets. But now, n-sub-e should probably include any body that can support technology. It should include the moons if they could possibly contribute. And we should always keep in mind that life is extremely opportunistic and adaptable.

How does something like NASA’s Kepler planet-searching mission change what we know about these factors?

The Kepler search has already shown that planetary systems exist in huge numbers. It’s showing that the majority of sunlike stars have planetary systems. It is also showing that in these systems, there are more low-mass planets than large-mass planets. Unfortunately, the range of observed masses doesn’t extend to Earth-mass yet, but the distribution says there are many, many Earth-mass planets, and of course that makes n-sub-e larger than we used to imagine. In the beginning, we thought only half the stars have planetary systems because half the stars are double stars. But now we know that double stars have planets, too.

But if you include moons around planets that are bigger than Earth – and there are many planets that are more massive than Earth – then that would also affect n-sub-e, if you’re considering all bodies instead of just planets.

That’s right. And just because a planet is bigger than Earth doesn’t mean it isn’t good for life. Bigger planets might be richer in life than Earth-mass planets. We always seem to think that the solar system is optimum, and that it’s the very best place for life, and that can’t be absolutely correct. We can’t be absolutely optimum — that’s so improbable.

It used to be thought that beyond the orbit of Mars, you could have no life. That’s a big misconception. What we’ve learned is that the surface temperature isn’t just dependent on how far a planet is from the star, but how much insulation is being provided by the atmosphere, or an ocean. Venus is incredibly hot, with a deep atmosphere. There’s a liquid ocean on Europa because of insulation from a layer of ice. The giant planets — people don’t realize their atmospheres are so deep that temperatures deep down are the same as on Earth. And, every plausible gas in planetary atmospheres creates a greenhouse effect. It is an interesting fact that there’s no such thing as a negative greenhouse. You can be very far from your star, and still be warm enough for life. You can be subterranean because the temperature goes up as you go down. The habitable zone may go all the way to infinity.

There are lots of ways to make a planet warm enough for life. 

For life as we know it.

Well, yeah. Every time we’re talking about life, we’re talking about carbon-based life with DNA and all that…

What were some of the more difficult terms to quantify, when you wrote the equation?

At that time, we were really very much in the dark. There were many things we did not know — which we still do not know today. And since everything in the equation has equal weight, I knew the answer was going to be only as good as the thing we knew the least about.

What were some of these things you did not know?

Well, one is the number of habitable planets in a planetary system [F-sub-l]. There were simple models that people put too much faith in, which predicted that the number of habitable planets was either very small, or zero. Of course, zero couldn’t be the right answer because we knew of at least one.

Another one was the fraction of biota that would evolve an intelligence [F-sub-i]. There are still a lot of misconceptions in this subject to this day.

But the thing that was most troublesome was the longevity of civilizations in a detectable state [L] — which to this day is a very vexed subject. A great difficulty here is that the answer depends a great deal on what is possible to detect.

Can you elaborate?

If you could detect the signals from people’s cell phones, then you could detect every civilization in existence, practically. If you could only detect very powerful transmissions like those from Arecibo, then the number is very small. The answer to that one is very dependent on your assumption of the capabilities of your detection systems.

Is that a question that you think we will ever be able to answer?

I think we will.

All of these things will be answered as soon as we detect a few other civilizations. It will just open the floodgates, and these things that now have error bars of a factor of a million will become very well known. But until we detect other civilizations, we’re just speculating about what they’re using that makes them detectable. It’s almost in the realm of science fiction.

What kinds of technology do you think they’re using?

In my optimism, with no basis for it, I am predicting that the use of stars as gravitational lenses will be very widespread because it is such a powerful technique for studying the universe. I think that every civilization that has the technical capability will do it, big-time. This greatly affects not only the detection capability but the power of the signals you can send to other civilizations. It works both ways. You can use the lens for detection, but also to amplify outgoing messages.

If the use of stars as gravitational lenses is extremely widespread, then the number of detectable civilizations is huge.

A little ways back, you said, “I knew the answer to the equation was going to be only as good as the thing we knew the least about.” Is “L,” the length of time a civilization is detectable, the thing we know the least about?

L is definitely the thing we know the least about. The only source of information we have about L is our own history. We have to use ourselves as a model, and assume that we’re “typical,” although we may not be. But we have only been easily detectable since the second world war, when very powerful radio transmitters were invented. So we have been easily detectable for 70 years.

The equation collapses to N [the number of detectable civilizations in the galaxy] equals L. So if we have a minimum value of L being 70 years, based on us, does that accurately reflect what N is, at this point?

If 70 is actually the rule, there will only be 70 detectable civilizations in the Milky Way at the present time.

At the present time.

Yeah. That’s good news and bad news. It’s good news that there’s more than one. But it’s bad news because it means the number of stars you have to search is just enormous. It is way more than a million. That’s based on our only being able to detect very powerful radio transmitters, which is our situation right now.

Do you have your own favorite number for L?

Yep.

And?

First, it is a curious and intriguing fact that the difficulty of detecting other intelligent life depends as much, or maybe more, on how widespread altruism is in intelligent civilizations, as it does on the qualities of search technology. 

L is the average length of time a civilization is detectable, and this means L is the average length of time a civilization is using some detectable technology. Now if it’s only using the technology for its own purposes, L could be pretty short. That’s what we do at the present time. We’ve gone 70 years and now we see that our visibility is actually going to start decreasing pretty soon. We’ve invented superior ways of delivering television to people’s homes — we do that by cable, which doesn’t release any indication of our existence to space, or the direct-to-home transmission of TV from satellites.

None of that scenario requires any special freaky situation with respect to our civilization or our planet. It’s a scenario that you could expect to play out very similarly anywhere, which says that the lifetime of detectable radio transmission might be on the order of hundreds of years.

There are people who argue that it’ll be longer than that because we need to use radar to detect dangerous asteroids and such, so they would say thousands of years. There are also those who think that civilizations — the creatures of other civilizations have much of the same thinking and philosophies as humans do. This gets interesting!

One of the key parts of being human is that we are altruistic. The big question is whether intelligent beings in the universe are, on the whole, also altruistic like we are. Because if they are, then they’re likely to do something of an altruistic nature, understanding that there are other civilizations that would like to know about them. They will intentionally transmit very powerful signals for the benefit of other civilizations for a very long time. But that’s an act of altruism because it won’t benefit you. You just think it’s the right thing to do.

If there’s altruism, and even only occasionally, it changes everything. You get into the math — the value of L is the average L for civilizations, the arithmetical average. So if 100 percent are only visible for 70 years, then L is 70 years. But if 1 percent has an L of, say, a billion years because they’re altruistic and continually send detectable signals, then the value of L is 0.99 times 70 plus 0.01 times a billion. Which is 10 million.

All it takes is 1 percent to be altruistic and transmit a detectable signal essentially indefinitely. That’s a very interesting result, because it’s strictly correct. It’s not crazy. And so, assuming only one in 100 civilizations does this thing, it’s not far-fetched.

What this says is that L could be much larger than our intuition would suggest. And therefore, the value of N is much larger.

So do you think N could actually be something like 10 million?

It could. But when people really press me on this…the assumption that 1 percent transmit for a billion years totally dominates the result, which is intuitively offensive. The answer depends on this one outlier data point, which has no observational statistical basis. An escape from this is to take the geometric mean instead of the arithmetic mean, in which case the outliers don’t so greatly dominate. You come up with an answer of 10,000 years. That’s the number I usually throw out as being plausible.

That makes the whole enterprise more promising because there’s 10,000 places we could detect. But at the same time, it means the fraction of stars with transmitting civilizations is 10,000 over maybe 100 billion. So it ends up being one in 10 million. That’s the challenge to anybody designing a search. If they’re going to be realistic about it, they need to design a search for 10 million stars. That’s the bottom line. 

What do future SETI searches need to be like? Big science, small science…citizen science?

It has to be a big science approach, with very high sensitivity. You have very little chance with backyard dishes or amateur telescopes because you need a lot of sensitivity to detect even very powerful transmitters like Arecibo. The way I describe that is, you can build a very beautiful small airplane, but it’ll never get you to the moon.

The future of SETI is to build an antenna that will look in all directions at all times. Actually, you need two of them, at opposite sides of the Earth so you can watch everything all the time, because every once in a while, somebody sees a signal which looks like the real thing. They’ll see it for a minute or so, and then it goes away. They’ll go back and look at the star and see nothing.

There’s a famous signal, called the “Wow!” signal, detected by a big SETI project at the Ohio State University in 1977. They were using the big antenna that then existed in Ohio. This antenna used the rotation of the Earth to look at various places in the sky. If there was a signal coming from the star, the signal would appear and get strong and decrease because of the response pattern of the antenna. They saw a very strong signal that produced exactly the pattern it would if it was coming from a source that was fixed in the sky. That’s called the “Wow!” signal, because when the person saw it — in those days there weren’t computers, they were recording all this on graph paper — he wrote “Wow!” alongside it.

We know of ways that these things can be a product of fortuitous events in our radio equipment, but they’re improbable, but not ridiculously improbable.

These signals that look real suggest that real signals might not be on all the time. Other civilizations might be using a strategy where they are sending intentional signals, but can’t send intentional signals to all the stars all the time. You make a very powerful beam, and turn it from one star to the next, over and over and over. Every star may only see it once a year or so. This makes the search much more difficult because not only do you have to look at 10 million stars, you have to look at them over and over, or watch the whole sky continuously.

We know how to do that, we can build an antenna that would do that. But it’s very expensive. You create a multitude of beams, they cover the whole visible sky, and the amount of computing that’s required — I happen to know — is of the order of 10²⁰ flops. In the year 2000, when this was proposed, it would take a computer facility the size of a football field to make this thing work. It’s getting better all the time, but it’s still beyond our reach and available finances.

Have there been others like the “Wow!” signal?

Yeah, another 35, at least. The whole sky was searched by Harvard, in the same way as was done at Ohio State. They saw more than 35 candidate signals. This was done by Paul Horowitz, who’s still very active at Harvard.

The Harvard project saw about 35 “Wow!” signals. The well-known SETI@home program saw more than 100. Both Horowitz and the people at Berkeley have gone to places like Arecibo and actually searched for signals at the appropriate locations and frequencies and none of the signals has ever been seen a second time. So here are all these detections that do all the right things, but there are no confirmations. And there hasn’t been observing time to just sit and watch one of these places for a year. But that’s what we need to do.

So do you think it’s possible that we’ve already seen evidence for another civilization but we can’t confirm it?

Well, some of those candidate signals could be real, but I think they’re all artifacts. This opinion is based on our experience at the SETI Institute where we have actually identified many, many, candidate signals as artifacts, with no residue of unexplained detections.

What else they could be? You said they’re hard to produce by accident.

About the only other thing it could be is an airplane that just by accident gets radiation into your beam. Radio telescopes have a main beam which is capturing the bulk of what they’re seeing. The telescope captures radiation from a small area, called the beam. But all of them have some additional sensitivity in all directions. If there’s a very powerful signal within the view of the telescope, you could see it and it will create a signal in the system that looks just like a real one, but is phony.

Most people say that after all this searching — hundreds, millions of channels, hours and hours and hours — it’s not surprising that occasionally the improbable happens. At the present time the view is that we have to, as serious scientists, assume that all of these candidate detections are artifacts of our equipment.

When do you think we might have an answer for the equation?

That’s a very standard question. The answer is, when we’ve got enough money. It’s all money-driven. We have the technology, we know how to build radio and optical telescopes of any size that have just about as much sensitivity as is permitted by the laws of physics. But the search has to be so comprehensive in where we look and what wavelengths we look at, and how long we look. It’s got to be a search which can cover many, many, many possibilities. Our search capability is totally limited by available funds. We depend still on donations from private individuals. It’s not a lack of technical know-how, it’s a lack of resources to construct and carry out the technology we know how to build.

If we could build an ideal telescope, how long do you think it would take before we found a signal?

Depends on what L is. If we take the 10,000-year value, which can’t be all that far wrong, we’re talking 30 to 100 years. With a big, big comprehensive telescope system.

What do you think the impact on the world would be, if we found another civilization?

Enormous.

OK.

Well, the first key fact is that almost any civilization we can detect will be much older than we are. They will know much more than we do, have much more experience in dealing with all the things we have to deal with, like environmental destruction. We will be students, we will be the new kids on the block and it will be possible to learn a great deal from them. That’s the main thing.

The second thing is that we may not be able to learn much right away. We will just detect evidence of a civilization — a signal, clearly of intelligent origin, but with not enough sensitivity to extract any information from it. This will require a big crash project to build big telescopes. No signal ever dies. The signal always exists — if you build a big enough telescope, you can capture it. We’ll have to build big telescopes so that we can detect signals with enough collecting area, or sensitivity, that we can learn things and capture information from the civilization.

In the best of all universes, they would be sending us their encyclopedia once a year or in some reasonable time interval, and telling us all about their history, what works, what doesn’t work, what’s good, what’s bad….

We’re not sending our encyclopedia.

No, but we’re sending a tremendous amount of information about ourselves through our television. Capturing television is really the best avenue for gaining formation about a civilization. You don’t have to ask a question and wait 2,000 years for an answer.

In the future, how do you hope people will look back on the Drake equation, or be using it?

They should think about it and what it says, what its limitations are, particularly the fact that L could be much larger than your intuition suggests. Which makes you optimistic. Almost all the things that would seem peculiar lead to more detectable civilizations. And you should use it as a way of guiding yourself and planning searches. How do you optimize them? We do that now in SETI. When we’re building a radio receiver, do we use more channels, or narrower bandwidths? We trade things off, make calculations about how this allows you to explore the search space more efficiently. The equation will continue to be very useful in planning the search for extraterrestrial intelligent life.