Time and Clocks (1960)

I've made it look as if I were going to write a definition of time. But the fact is I can't. Nor can anyone else. If you look up time in a dictionary, you'll find that the definition contains words like succession, event, before and after. Words that already contain the notion of time. I can define a watch. There's no problem there. But I can't define time. And this reflects the subtlety of this notion that we all take for granted. Still, there's no use pretending that we don't know what time is. For instance, 1 second is a reasonable time. Uh your heart is beating perhaps once a second. And on the long side, 100 years is certainly an extreme. You perhaps can't live that long. Now, what about shorter intervals down in this region? Well, this movie that you're seeing is made up of many individual still pictures. And my gestures seem perfectly smooth. But if we slow up the camera so that there are fewer frames per second, your eye and brain now begin to distinguish individual frames, and the motion of my hand seems jerky. Here you see the usual still picture at about a 12th of a second instead of the usual 1/24. Now, let's look at it for about a sixth of a second. Here we are showing only six different still pictures a second, and it has become definitely uneven. So, you see that our ability to distinguish short time intervals falls somewhere between 0.01 and 0.1 seconds. There are some experiments that you can do yourself that check this sort of thing that are described in the film notes. Now, what about all of this these shorter times here that we can't appreciate directly? There are a variety of ways that we can use to expand our time scale and find out what happens. Here's a picture of drop splashing into a cup of milk. Now, let's take one of these and examine the pictures the camera took one by one. This is some idea of what we see, but it doesn't tell us very much. Now, let us imagine that here are some of the pictures that we just saw. They are separated in time by about a 24th of a second and about this much on the film itself. Suppose that we had a high-speed motion picture camera that took many thousands of pictures during a 24th of a second. Then, the film would be longer and by looking at these pictures, we would be able to see just what happened in between the pictures that we've have already looked at. It seems quite unbelievable in a way, and yet the milk drop does get into the cup with some kind of splash. It gets from the state in which we first saw it to the final state by a series of continuous processes. Now, we're never going to be able to observe directly what happens in these very short time intervals. And yet we feel that this technique we have of expanding our time scale and looking is perfectly legitimate. Yet we can't prove it. And this is another example of the subtlety of time. Now just for fun, I can take a something that takes place in a long time and squeeze it down, telescope it so it takes place in a short time. Here is a rose and it's growing, but nothing seems to be happening. You can watch it for 5 minutes and then you get bored. But if we take a picture once an hour and show them rapidly like this, you can see the flower unfolding in an almost explosive way. Now I'd like you to notice something about these two sets of pictures we've just looked at. They each contained a clock. In the case of the unfolding rose, the clock was just an ordinary clock. And in the case of the splashing drop, the clock was just the shaft on the on a little motor which was turning at 1800 rpm and which had a mark on it. Now this is the essence of all time measurements. One compares the start and finish of an event with the position of a clock's hands. Now let us talk some more about clocks. A clock is a device that makes regular events. For instance, it could ring a bell or flash light or the hands can move by in front of a face. Now there's no trouble with those events, but the difficulty is to to know what perfect regularity or what even ordinary regularity mean. Now what do we mean by regular intervals of time? That's kind of hard to test. We have to compare one clock with other clocks and keep doing it. It's fortunate that we have a tremendous body of evidence that tells us our clocks aren't off much. We don't notice any serious discrepancies, a minute a year perhaps or less. But when we get down into higher precision, we need better standards. Now, a well-isolated standard that has been used for a long time is the solar system. Here's a sketch of the solar system for our purposes. It has the sun in the middle, the earth right here, and the moon going around the earth, and planets are moving around here. Now, for I mean, for centuries we've observed night and day as the earth turned around its axis, and the moon the phases of the moon have been observed from month to month, and this is the seasons themselves. And these all serve as approximate clocks. Nowadays, we use much the same motions, but we observe in a much more refined way with telescopes. And then we still find that the motions are not very regular because they're all influencing each other. And in fact, the second, which is defined as a certain big fraction of a particular year, uh is not really very conveniently reproduced by this sort of standard. So, we either need a quicker standard of time or convenient ways of subdividing time. Now, one way of doing this that's crude and convenient is to use one's pulse. And I can tell you an interesting story about Galileo, who was watching a lamp swinging in a church and timed it using his pulse. Here's a bowling ball suspended from the ceiling of the studio, which is about 25 ft high. Let's imagine that it's a lamp such as Galileo might have observed in a church. And he found that whether it was swinging in a big arc like this or a relatively small arc such as this, that it took the same time as he found by timing it with his pulse. Galileo also found that the period of a pendulum depended on its length. Here is the pendulum and it's taking a certain time to swing. If I shorten it, it swings faster. If I shorten it still further, it swings even faster. Not much else affects the time of a pendulum. Now, here are three lamps or rather bowling balls, uh a little one over there and a big one here held up by strings and another little one here also held up by a string and much further up. And when I burn the threads, they will all be released nearly at once and they should swing together. Now, let's try this experiment. You will see that Galileo's statement is certainly right. At least uh if he was just timing it with his pulse. And in fact, the pendulum makes a remarkably good regulator for clocks. And how would you check it? Well, you'd compare several of them together just as we are doing now. So, the pendulum became the standard device for dividing up long intervals of time. Now, here's a pendulum clock whose pendulum is arranged to beat seconds. In this way, we can easily compare the time told by a clock like this with astronomical time. These clocks have been refined to a high degree, but they're still not good enough. They're not isolated well enough. If the weather gets warmer, the pendulum gets longer and the clock gets off. If one takes it up on top of a mountain, the force of gravity is different and the clock will also not tell the right time. We can make clocks with atoms now and we have learned to do this since the Second World War. Here is an atomic clock, one of a couple of dozen in the world. It uses the vibration of cesium atoms to control its output, which is down here. The front half of this apparatus consists entirely of electronic equipment. This kind of thing. In the back is the apparatus that contains the cesium atomic beam. Here is the part of the atomic clock that has the cesium in it. The cesium atoms start down here and travel up through this little tube, vibrating as they go, and eventually strikes a detector. As you can see, you don't see very much. Now, what sort of evidence do we have that cesium clocks keep good time? You will remember that when we wanted to tell whether a clock was regular, we compared two of them. And this has been done with cesium clocks. Here's some data that show how they perform. This is a curve of the difference in frequency of two cesium clocks over a three-day period. Here's one day, two days, three days. And let's see, the height of each of these boxes is two parts in 10 to the 11. So, you'll see that over a 3-day period, these two cesium clocks didn't differ in frequency by more than about a few parts in 10 to the 11th at most. Now, how much is a part in 10 to the 11th? Well, there are 30 million seconds in a year. So, that in 3,000 years, we would have about 10 to the 11 seconds. So, that one part in 10 to the 11th is 1 second out of 3,000 years. And that's the sort of discrepancy that would arise between two cesium clocks had they been running for 3,000 years. Now, it would seem that cesium clocks agreed with each other. And in fact, they work so well that they may become our standard of time instead of using astronomy the way we do now. It's a good question to ask whether cesium clocks would keep the same time as astronomy does. And experiments are now being done to find this out. We don't yet know what the answer is. But if there is a discrepancy, it might throw some additional light on the fundamental nature of time. For many purposes, we don't need the precision that is available from an atomic clock. But we do want to observe events that happen in short time intervals. And one way of doing this is to use high-speed motion picture photography. Professor Harold Edgerton has developed this technique to a high degree. And he will now tell you something about how it works and the way he uses it. You are now in the stroboscopic light laboratory at MIT. Here we study events which happen in a very short interval of time. To do this, we use electronic flashlights and various sorts of cameras that are associated with these lights. At the right here is one of these lamps that you will shortly see in action. In an ordinary motion picture camera, the film uh goes through at 24 pictures per second. In this camera, it goes through at a rate of many thousands. Then when the pictures are shown on the screen at the normal rate, you see the motion uh greatly reduced in speed. I'm now going to show you the uh stroboscopic light. Each flash of light lasts approximately 1 millionth of a second. I will now run it several times a second so that you can see it glow. To your eye, it looks as though it were continuous light when it is running at these very high speeds. I now wish to uh show you an example of how high-speed photography can slow down very rapid motions. Consider as an example this balloon. You are going to see this balloon broken with uh a camera running at normal speed. Everything happened instantaneously before your eye. Uh the balloon in one picture was complete, in the next frame, it was collapsed. With the high-speed pictures, you will be able to see what happens when the knife breaks the balloon. I will now show you some examples of the usefulness of such pictures. This picture shows smoke flowing through the blades of an electric fan. The slow motions permit us to study these currents and to understand how they are formed. Pictures like these help us study how a bird flies. Uh actually, we seldom use high-speed motion picture techniques in studying rapid motions. Instead, we use simpler techniques involving single pictures or even double pictures taken at exact times. Uh this one example shows a bullet in flight, and here you see two pictures of the same bullet which are taken 100 microseconds apart. In other words, the light that took this picture took one flash 1 minute into the second, and then the second flash occurred 1/10,000 of a second later. The examples that you will have seen were taken with exposures down to approximately 1/20 of a microsecond. Beyond this speed, it is almost impossible to get pictures because of the dimness of the light or the insensitivity of the film. Frequently, we don't need the detail of a photograph. And we all we have is an electrical signal. Now, microphones put out an electrical signal from sound. And photoelectric cells put out an electrical signal when light strikes them. And there's a whole host of devices of this sort. So, how do we study electrical signals? Well, we use a cathode ray oscilloscope. Here are two of them. Now, this is a very useful one that you could build yourself from a kit. And here is a considerably more complicated and elaborate instrument. These devices are so useful that it's worth taking some time to talk about their operation a little. This tube here is very much like the tube in the television set. A bright spot travels back and forth. Here it is. And when I apply voltages to the scope like this, I can make the spot go up or down at will. I can also make it go sideways like this. And in fact, the scope itself makes the spot go sideways. It It'll move from the left to the right at a slow rate and then come rapidly back. Watch. Now, if if while it's doing this, I apply a voltage, an electrical signal, you will notice that the thing will go up, stay up, and come down when I stop. Or alternatively, it'll do that and so on. In this way, I can see what happened during the time that the spot is traveling across the scope face. Now, you'll notice that it's taking a couple of seconds for the spot to get across the screen here. I can speed this up a lot. Here it's going 10 times faster, and here it's going 100 times faster. When it's going at that rate, it's logical to look at a common phenomenon, namely the AC power line. Let me connect the AC power line to this scope like this and plug it in here. Now, we see what the power line looks like on a cathode ray oscilloscope. If I count the little boxes from here to here and read all these dials, I can figure out that it takes a 60th of a second for the spot to go from here to here, which is just what it should be. Now, let me show you something. Here is a pendulum which is equipped with a brush that writes on this piece of paper. Now, if I set it swinging like this and then pull the paper down at as steady a rate as I can manage, I get a trace of the motion of the pendulum plotted against time. If this distance from here to here represents the period of the pendulum, so is this the period of the pendulum from from here to here, from any corresponding points. And this is just the same sort of curve that we have on our cathode ray oscilloscope right now. Many phenomena look like this, many vibrations. Now, if I wanted to continue having this plot, I'd have to keep on pulling out paper. The cathode ray oscilloscope gets around this by using the same piece of paper over and over again, so to speak. After it's got this far, it comes back and retraces the same curve. Now, instead of using our cathode ray oscilloscope to look at something that repeats all the time like the power lines, let's put in some signals that occur just once, and for instance a flash bulb, and see how long a time it takes for a flash bulb to to flash. To do this, I connect up a photoelectric cell of this sort. You can buy one of these for uh $2 or so at your local radio supply house, and here it's connected up to a battery and so on. Now, let me connect it to the scope. This here. And now you see if I make some adjustments here so that uh I can pick up the lights from the studio. There's a little opening here. I've covered it with uh tape so as not to put too much light in. And here's a little opening and you can see as I move it around I'm picking up the studio lights here. And when I put my hand over it, the spot goes down. So, the more light I let in, the higher the spot is. And if I put my fingers across here in in rapid succession, you can even see four little wiggles. Now, let's use this with our flash with a flash gun. I'll put it here. And uh I've predetermined roughly where this should be set here. And um I'll also speed the sweep up a little to be about right. And you it usually takes quite a few tries before you get the scope adjusted so it does just what you want. And now that I think would be about right. And now here's a flash gun. And I'll put that about here. And I'll connect up two wires. A remarkably unstable arrangement. And now set it off. There, I think you can see this pulse that occurred here. Let's do it again. Now, let's stop the film and look at this image. You see that it occupies about two boxes here. And since there's um by reading these dials, I know it's 50 milliseconds per box, it takes about a tenth of a second for this flash to be over. Now, that's uh not very fast. And it would be interesting for you, using a much simpler scope, to compare that with a strobe, which um gives a much shorter flash. Now, with cathode ray oscilloscopes, we can measure times as short as 10 to the minus eighth seconds, and with very special ones, one can get down to 10 to the minus ninth seconds. Now, one can go further. Uh one can use coincidence techniques to measure intervals that are short as 10 to the minus tenth seconds or thereabouts. Uh suppose two pulses occur in time like this. One can delay the first one until it sits on top of the second. And by measuring the delay time, which one can do in terms of the length of a cable, one can get down to the vicinity of 10 minus tenth. And that's about as far as we can go with electronic techniques. Now, the still further, still shorter times can be measured and can be observed by allowing particles to make their own tracks in film, this kind of picture. Like this. Such particles often travel with very high velocities, such as the pretty close to the velocity of light, which is this. So, if one can, using one's microscope, one can look at the film, one can measure distances that are small as uh say three times 10 to the minus fifth centimeters. And since the velocity is this of the particle, the time that it took the particle to traverse the little track is of the order of about 10 to the minus fifteenth seconds. You can see that this is not as direct a measurement as either of these sorts. One is inferring this from looking at the track. Now, one one can imagine shorter time still. After all, light can cross an atom. Uh and how long does that take? Well, an atom has a diameter of about 3 * 10 ^ - 8 cm. And again, light is has that velocity. So, that the time it takes light to cross an atom is about 10 ^ - 18 seconds. And now if we consider a nucleus, we get to even shorter times. Uh nucleus has a diameter of around 10 ^ - 13 cm. And now the time for light to cross a nucleus would be 10 ^ - 23rd seconds. Now, you can't observe those times. You can only imagine them. And of course, one can talk about them. And we're not sure just what they mean. And in fact, uh nobody knows. And it may be that time of these short intervals cannot be thought of the way we've been thinking of it in when when it was in measured in longer intervals. It's just rather like matter. Uh if you have say some material such as gold and you divide it and keep on dividing, you eventually reach one atom. Uh you can imagine this anyway. And now you can't divide any further and still have gold. Perhaps time is like that. And one just can't go on to arbitrarily small intervals and have time as we think we understand it. Now, where are we? Well, first we started talking about time and uh we were a little bit uh vague about what time really meant because we couldn't help it. And we talked about those intervals that we could appreciate directly. And how we measured them, how we set up clocks and standards. Then we went on to shorter intervals. Time intervals where we either had to use photography or cathode ray oscilloscopes to see what was happening. And then we got less and less definite and went to even shorter times and finally we got to a point where we can't observe anything and where we just talk. Now in the next movie we're going to talk about longer time intervals and see what happens as one goes towards billions of years.