Does Light ACTUALLY Move?

Ajouté : 2026-06-01

[00:00:00]Have you ever wondered how the speed of light was first measured? Or for that matter, have you ever stopped to think about how strange it is that light has a speed at all? The next time you look outside and see the sun shining and the world around you illuminated, pause for a moment and ask yourself, what is there about that experience that would lead you to think that light is a thing that travels? that it moves through space, that it takes time to get from one place to another, and that this time can actually be measured. Today, I want to tell you one of the greatest stories in the history of science, the story of how the speed of light was first estimated.

[00:00:40]Our story begins with Galileo in the early 17th century. In6009, Galileo heard reports of a strange new Dutch instrument, a tube with lenses that could make distant objects appear near.

[00:00:53]He didn't invent the telescope, but he built his own, improved it, and then did something that changed human history. He pointed it at the sky. And when he turned it towards Jupiter, he saw something astonishing. Jupiter was not alone. Around it were four tiny points of light. And night after night, Galileo watched them shift position. They were orbiting Jupiter. This was the first time anything had ever been observed orbiting another planet. But hidden inside this discovery was another possibility, one that was less philosophical, but enormously practical.

[00:01:30]The moons of Jupiter moved regularly, and that meant they could be turned into a clock. Not a clock made of brass wheels and springs, but a clock hanging in the heavens. And this mattered.

[00:01:44]You see, by the 17th century, ships were crossing oceans and empires were expanding. Maps were becoming instruments of power. But one problem remained stubbornly difficult. And that was the problem of longitude. How far east or west you are relative to some reference point on Earth.

[00:02:03]Latitude, the easiest sibling of longitude, tells you how far north or south you are relative to the equator.

[00:02:09]It's represented as an angle measured from the equator at 0° up to the north pole at 90°. The problem of determining latitude was solved thousands of years ago using the sun. For example, on the day of the equinox, the sun at noon is directly overhead at the equator, meaning a stick placed in the ground casts no shadow. But if you are further north on the same day, say at a latitude of 30°, then a stick placed in the ground at noon will cast a shadow and the line connecting the end of the shadow to the top of the stick will make an angle of 30° with the vertical. And so just by using the sun or even the stars in the sky, you could determine your latitude.

[00:02:52]But longitude was much more difficult because longitude is really a problem of time. To know how far east or west you are, you need to compare your local time with the time somewhere else. The basic idea is simple. As the earth turns, different places face the sun at different times. The line pointing directly towards the sun is the noon line. And when a place rotates onto this line, it is noon in that location. So first imagine one place on earth, let's call it point A. And as A rotates onto the noon line, it's 12:00 noon at A. As the Earth keeps turning, clock A keeps running. After one complete rotation, A returns to the noon line once again. One full day has passed. So 24 hours corresponds to 360°, which means the Earth turns 15° every hour. But now imagine two places, A and B. A could be a shipping port and B could be a ship at C. A reaches noon first. So, clock A reads 12:00 and starts ticking. But B has not reached noon yet. The Earth has to rotate further before B reaches noon. And when it does, the clock at B will read 12:00.

[00:04:09]But clock A is now reading 4:00. So noon at B happens 4 hours after noon at A.

[00:04:17]And since the Earth turns 15° every hour, that 4hour difference corresponds to 60° of longitude. In this case, A is 60° east of B. Okay, so what was the problem? Well, the problem was not finding local noon. A sailor could do that easily simply by noting when the sun reaches its highest point in the sky. The real problem was not knowing what time it was back at the reference point, the port, while you were somewhere else entirely.

[00:04:49]Now, in principle, you could set sail with a reliable clock synchronized to port time. The problem would then be easy. If the ship reaches local noon and the port clock reads 400 p.m., then the port is 60° east of the ship. On the other hand, if when the sun reaches its highest point in the sky, the port clock doesn't read 400 p.m., but instead it reads 8 a.m., then the port is 60° west.

[00:05:18]But at the start of the 17th century, there were no reliable marine clocks capable of keeping accurate reference time through a long sea voyage. That is why longitude became such a valuable problem. So valuable in fact that Spain had already offered a prize for solving it under Philip II and Philip III later renewed the reward for the search for a practical solution.

[00:05:42]And then in 1616, Galileo proposed an extraordinary idea. He suggested using the moons of Jupiter as a way of determining your longitude. Galileo had already noted that the moons appeared to orbit Jupiter with metronomic regularity and in particular the innermost large moon Io was especially useful because it repeatedly passed into Jupiter's shadow with a period of less than two Earth days making it a very useful clock.

[00:06:13]So how can you turn the motion of Io into a way of determining your longitude here on Earth? Well, to answer that question, we first need to understand some of the intricacies of actually observing Io from Earth. The first thing to appreciate is that light from the sun causes a shadow to be cast behind Jupiter. And as Io orbits Jupiter, it periodically passes into this shadow and then emerges from the other side. In other words, Io is eclipsed by Jupiter's shadow on a periodic basis. And this is what Galileo and the astronomers of the early 17th century observed. When Io entered the shadow and vanished, this was called an immersion. And when it emerged again, this was called an immersion. Each disappearance or reappearance was like the tick of a cosmic clock. But things get even more interesting if you remember that both the Earth and Jupiter are orbiting the sun. And so the view of Jupiter and its shadow from Earth changes as the Earth and Jupiter move around the Sun. Broadly speaking, when the Earth is approaching Jupiter before opposition, we see Io enter the shadow, but we don't see it exit because the rest of the shadow is obscured behind Jupiter. On the other hand, when Earth is moving away from Jupiter, we see Io exit Jupiter's shadow, but we don't see it enter the shadow. And so the challenge for astronomers was to see if they could predict when Io would enter or emerge from the shadow of Jupiter at some time in the future.

[00:07:52]And the reason this was important was because if astronomers could predict that an eclipse of Io would happen at say 1000 p.m. in Paris, then if someone were able to observe the same eclipse using their own local time and find that the eclipse happened at say 11:30 p.m., they could compare the predicted Paris time with their own local time. Then they could work out their longitude since the Earth rotates through 15° hour. So an hour and a half corresponds to 22.5°.

[00:08:21]And since their local time was later than Paris time, they must be east of Paris. And so it was realized that the moons of Jupiter could in principle provide the missing reference clock. But to make this method useful, astronomers needed accurate tables predicting the eclipses of Jupiter's moons. And this was exactly the kind of work being done at the newly founded Paris Observatory.

[00:08:46]The observatory had been established under Louis the 14th as part of a larger project to make astronomy useful for navigation, geography, and mapmaking. It was not just about looking at the stars.

[00:08:58]It was about measurement, power, and the ability to place the world accurately on a map.

[00:09:04]And at the center of this work was Giovani Dominico Cassini, an Italian astronomer who had been brought to Paris a few years earlier to help lead the new observatory's scientific program.

[00:09:16]Cassini was already one of Europe's greatest authorities on Jupiter. He had observed the planet in detail, studied the motions of its moons, and published tables predicting their eclipses. But Cassini was not working alone. Another crucial figure at the Paris Observatory was Jean Picar, a French astronomer and a priest with a gift for precise measurement. One of the important sites the French astronomers wanted to connect to Paris was Uranaborg, the old observatory of the 16th century pioneering astronomer Tao Brahe. Tao had made some of the finest astronomical observations ever recorded before the invention of the telescope.

[00:09:57]This video is sponsored by UPDF.

[00:10:00]UPDF is an all-in-one PDF editor that lets you edit documents, convert files, and use AI tools directly inside your PDFs.

[00:10:10]For example, if I open a long research paper, I can use the AI assistant to summarize the document, explain difficult sections, or answer questions about the content. That's especially useful when working through dense technical material. If you deal with PDFs regularly, whether that's research papers, lecture notes, textbooks, or work documents, then this is genuinely useful. What sets UPDF apart for research are the new academic focused features. First, there's paper search.

[00:10:39]It combines traditional keyword-based academic search with an AI mode that answers your research questions directly from real academic sources. You can even search by DOI or title to cut through information overload.

[00:10:54]Then there's the scholar research agent, which digs deep into web sources to gather and assemble accurate information, helping you see the full picture of a topic quickly. Beyond the AI, you can make quick edits to text, images, and links, or convert PDFs into Word, PowerPoint, Excel, and more. The OCR feature supports 38 languages, making it easy to convert scanned paper documents into searchable and editable PDFs. UPDF works across Windows, Mac, mobile, and browser. So, one account covers four devices. If you'd like to try it, there's a free version available, and you can use my link in the description or in the pinned comment for an exclusive discount.

[00:11:36]But to connect these observations with the new work being done in Paris, astronomers first needed to know exactly where Uranaborg was relative to the Paris Observatory.

[00:11:47]And so in 1671, Picar was sent north to Uranborg on the island of Ven. His task was to help determine Uranorg's longitude relative to Paris. And the method depended on coordination. Cassini would observe the eclipse of Jupiter's moons from the Paris Observatory while Picar would observe the same events from Uranaborg in Ven.

[00:12:12]Picar was assisted in Uranaborg by the young Danish astronomer Ole Roma. At this point, Roma was still near the beginning of his career. But these observations of Jupiter's moons would soon place him at the center of one of the most important discoveries in the history of light. But before that discovery could happen, there was the careful practical work of comparing observations of Io's eclipse between Paris and Uranaborg.

[00:12:39]Surviving documents show several paired timings of the eclipses of Io relating to Uranaborg and Paris. We see the date of each eclipse, whether it was an immersion or an immersion that was observed, and the local time of the observation at each location. I've written the times here in modern 24-hour notation, although Picar's original account labeled them simply as morning or evening local times.

[00:13:05]In one such observation on the 4th of January 1672, an immersion of Io was recorded in Paris at 12:42 and 36 seconds. The same event was recorded at Uraniborg at 124 and 45 seconds local time. So the difference between the two clock readings was 42 minutes and 9 seconds.

[00:13:27]By looking at the different times across each pair of observations, Picar was able to determine the time difference of each pair and then calculate an average, finding a value of around 42 minutes and 10 seconds. And then using this mean value together with the fact that the Earth rotates at a steady rate of 15° every 60 minutes, they estimated that Uranorg lay about 10° 32 ark minutes and 30 arc seconds east of Paris. And this result showed why Galileo's idea was so powerful. The eclipses of Jupiter's moons really could be used to measure longitude. But for this method to work beyond a few carefully coordinated observations, astronomers needed tables, predictions of when Io would disappear into Jupiter's shadow and when it would reappear. If those tables were accurate, then an observer elsewhere did not need someone watching the same eclipse from Paris at the same moment.

[00:14:27]Rather they could simply observe an eclipse of Io, compare their local time with the predicted Paris time in the table and determine their longitude. And that is why the work of Cassini in Paris became so important. Cassini had already produced tables of the motions of Jupiter satellites as well as detailed configuration showing what the moons of Jupiter should look like on a given date and time. And this became part of a sustained astronomical program.

[00:14:55]Astronomers would observe the eclipses, compare them with predictions, refine the tables, update the predictions, and then use those improved tables for measurements of longitude. But how exactly do you predict when an eclipse is going to occur? Well, first you need to know how often Io returns to the same eclipse condition. More precisely, you need the average time between equivalent events. For example, one immersion to the next immersion or one immersion to the next immersion. But there is a subtle point here. To see this, imagine first that Jupiter's shadow was fixed in space. Io would enter the shadow, go once around Jupiter, and return to the shadow after one ordinary orbit. But Jupiter's shadow is not fixed in space.

[00:15:45]As Jupiter moves, the shadow turns with it. So after one orbit, Io must travel a little further to reach the shadow.

[00:15:53]Again, for eclipse prediction, the relevant period is therefore the average time from one immersion to the next or from one immersion to the next.

[00:16:04]Importantly, this was not determined from a single orbit. It was found by timing many eclipses over long intervals. And we can quantify this. The average eclipse period can be written as P= deltaT / N where deltat T is the elapsed time between two comparable eclipse events and N is the number of returns of IO to the eclipse condition within the time period considered.

[00:16:32]And this is precisely the kind of averaging idea that was used to determine the period that was ultimately responsible for making predictions of future eclipses in Cassini's tables. For example, if you consider roughly the time taken for Jupiter to complete one orbit around the sun, then in this time, Io makes 2,448 returns to the eclipse condition. And just over 4,332 days have passed here on Earth. And if you then take the average eclipse period equation and sub in these numbers, you find a mean eclipse period of 1 day, 18 hours, 28 minutes, and 36 seconds. And this then represents the basic clock rate of IO that could be used in the prediction tables. Once that rhythm was known, the tables could predict when Io should next disappear into Jupiter's shadow or reappear from it. For example, if an immersion happened at time t0, then after one revolution, the next equivalent immersion would happen at time t1 = t + p. After two revolutions, it would happen at T2 = T + 2 P. And after N revolutions, it should happen at TN= T + NP. So the tables worked by carrying IO's clock forward. Observe one tick, add the period, and predict the next tick. Add the period many times, and you can predict eclipses months into the future. But once astronomers began comparing these predictions with what they actually observed, something strange appeared. If the tables were tracking the eclipses perfectly, then the difference between the predicted time and the observed time should have stayed close to zero. But it didn't.

[00:18:21]Starting near opposition, when the Earth is between the Sun and Jupiter, the predicted and observed eclipse times lined up. But as Earth moved away from Jupiter, the observed immersions drifted later and later than the predictions.

[00:18:34]The differences grew over many cycles of Io, reached a maximum, and then shrank again as Earth came back round towards Jupiter. But what exactly does a point on this right- hand plot represent?

[00:18:48]So let's run the simulation again and pause when the predicted and observed immersion differ by 12 1/2 minutes. On the left is the Earth Jupiter configuration corresponding to that moment. In the middle, the timing residual settles at 12 1/2 minutes. And on the right, the white marker P shows the predicted emergence of Io, while the orange marker O shows what was actually observed. In other words, at this point in Earth's orbit, Io emerged from Jupiter's shadow 12 minutes later than Cassini's tables predicted. And crucially, this delay was not random. It followed the changing geometry of Earth and Jupiter around the sun. Now this is a very strange thing because nothing about Io itself should know whether Earth is moving away from Jupiter or back towards it. Io was still orbiting Jupiter. Jupiter was still casting its shadow. From Jupiter's perspective, Io was simply orbiting with a fixed period passing into and out of the shadow with metronomic precision. And yet from Earth, the apparent timing of that clock seemed to depend on where Earth was in its orbit relative to Jupiter. So what was going on? One possibility considered within the Paris observing program by Cassini and Roma was that the shift in eclipse timings might be explained if light did not act instantaneously, but instead took time to cross the distance from Jupiter to Earth. This was not a completely new idea. Ancient philosophers had argued over whether light was instantaneous or whether it took time to move through space. For example, Impedicles had suggested that light took time to travel, whereas Aristotle rejected the idea that light traveled at all, treating illumination as effectively instantaneous. Later thinkers continued to debate the topic.

[00:20:37]But in the 17th century, the instantaneous view was still powerful and widespread.

[00:20:43]However, Galileo had suggested a way of testing the idea and even potentially measuring the speed of light directly.

[00:20:50]He imagined two observers standing on distant hills with covered lanterns. One uncovers their lantern and the light travels to the other who then uncovers their lantern as soon as they see the first, and the second observer's light then travels back to the first observer who attempts to time the round trip. It was an ingenious thought experiment, but across earthly distances, any delay was far too small to separate from human reaction time. However, in the case of Io and Jupiter, the situation was different. The light was not crossing a valley. It was crossing the solar system. And as Earth moved around the sun, the distance that light had to travel from Jupiter to Earth changed by millions of kilome. So perhaps the delay hidden in Io's eclipses was the first indication that light itself takes time to travel. If light traveled at a finite constant speed, then the time it takes to travel from one place to another is simply proportional to the distance. And if you plot the changing distance between Jupiter and Earth as a function of time, then you see that the distance gradually grows to a maximum and then reduces back down to a minimum. And we see that the shape of this plot bears a striking similarity to the eclipse delay plot that we saw earlier. To see this clearly, we can plot the Earth Jupiter distance plot on the left and the eclipse time delay plot on the right.

[00:22:16]And we see that they have the exact same shape. So perhaps the observed delay in Io's eclipse was actually due to the fact that light travels at a finite speed rather than instantaneously as had been assumed in making the predictions.

[00:22:32]So let's flesh out this idea and see how it works in detail. Suppose an immersion happens at Jupiter at time t. At that moment the light from the event has to cross a distance d from Jupiter to Earth. If light travels at speed C, then the light travel time is approximately d over C. So the time at which the event is observed on Earth is equal to T + D / C. And this is the key equation. Next, imagine two successive immersions of Io that happen at Jupiter at times t1 and t2. The first happens when Jupiter and Io are a distance d1 from Earth, and the second happens when the distance is d2.

[00:23:15]And so following the same logic, the first immersion is observed at T1 + D1 over C and the second immersion is observed at T2 + D2 over C.

[00:23:29]And so the observed interval between the two immersions is equal to their difference. And if we then substitute in the two observed time expressions and simplify, we end up with t2 minus t1 plus d2 minus d1 / c.

[00:23:46]But now think about the interval between two successive immersions. The first immersion happens at Jupiter at time t1 and the next one happens at Jupiter at time t2. So the true period of Io's eclipse cycle is t2 minus t1. And so if we sub this back into our observed interval equation, we find the following expression. So the observed interval is the true eclipse period plus a correction factor that depends on the changing distance between the Earth and Io after one full cycle.

[00:24:21]And so we see that if Earth is moving away from Jupiter, then the second light signal has further to travel than the first. And that means that D2 is greater than D1. and therefore the observed interval is longer than the true eclipse cycle period. In other words, Io's clock appears to run slightly slow and the next immersion is seen a little later than expected. But if Earth is moving towards Jupiter, then the second light signal has less far to travel than the first and that means that D2 is less than D1 and therefore the observed interval is slightly shorter than IO's true eclipse period. IO's clock appears to run slightly fast and the next immersion is seen a little early.

[00:25:06]Now, at this stage, it's important to notice a few things. First, let us write our equation as P + deltat T, where deltat T is D2 minus D1 / C. It's crucial to emphasize that deltat T is not the total lateness of an eclipse after many eclipse cycles. It is the extra delay added to one eclipse interval. It tells us how much longer or shorter the observed gap is between one eclipse and the next. And we can visualize this. We can plot a graph with delta t on the vertical axis which is the cycleto cycle change in light travel time. And on the horizontal axis we have the eclipse number or equivalently time.

[00:25:49]Each point represents the extra delay added between one eclipse of IO and the next. When the points are above zero, the Earth Jupiter distance is increasing. So each observed eclipse interval is stretched by a few extra seconds. When the points are below zero, the Earth Jupiter distance is decreasing. So each observed interval is shortened by a few seconds. The largest positive and negative values occur when the Earth Jupiter distance is changing fastest. And near opposition and again near conjunction, the distance changes only slowly. So delta t falls back towards zero. So this graph is not yet showing the whole accumulated delay.

[00:26:27]It's showing the little extra amount added or subtracted each time Io completes another eclipse cycle. But Cassini and Roma were not just looking at one eclipse interval. They were comparing predictions and observations over many eclipses. And that is where these small differences build up. And we can quantify this accumulation. After one eclipse interval, the accumulated delay is just delta t1. After two intervals, the accumulated delay is delta t1 plus deltat t2. And after three, it is delta t1 plus delta t2 plus delta t3. And so after many eclipses, the total accumulated shift is the sum of all of the small shifts.

[00:27:11]And we can visualize this. If we now plot the accumulated delay next to the cycleto cycle plot, we see a familiar curve emerge. This is precisely the relation we saw earlier. And this is precisely the pattern noticed by Cassini and Roma. Only now we have provided an explanation by assuming that light does not travel instantly but rather at a finite constant speed through space. And that is the key. The finite speed of light idea did not just say vaguely that the tables might be wrong. It predicted the direction of the error. While Earth was moving away from Jupiter, each eclipse interval should be stretched slightly. So the accumulated delay should grow. While Earth was moving back towards Jupiter, each interval should be shortened slightly. So the accumulated delay should shrink.

[00:28:03]So the test was obvious. look at a time when Earth was moving away from Jupiter and see whether Io's immersions really did fall increasingly behind the ordinary predictions. And that is exactly what happened in 1676.

[00:28:19]After Jupiter's opposition that summer, Earth was moving away from Jupiter.

[00:28:24]Immersions of Io were observed in August, and those observations could be used to predict future immersions later in the year. But if light took time to travel, then as Earth moved further away from Jupiter, the light from each later immersion would have further to travel.

[00:28:40]So those immersions should be seen increasingly late compared with the predictions.

[00:28:46]Later that year, the prediction became concrete. Using observations made after Jupiter's opposition, Roma argued that an upcoming immersion of Io in November would be seen later than predicted by about 10 minutes.

[00:29:02]And that is exactly what Roma found. The immersion was observed about 10 minutes later than the ordinary calculation predicted, close to the delay expected if light needed time to cross the increasing distance between Jupiter and Earth. So Roma was not simply reporting that the tables were wrong. He was arguing that they were wrong in exactly the way expected if light needed time to cross space.

[00:29:27]Roma presented his argument to the French Royal Academy of Sciences on the 21st of November. He argued that the inequalities in the immersions of Io were evidence that light does not travel instantaneously. And on the 7th of December, the Janal de Sava printed the famous notice describing the discovery.

[00:29:47]The notice described the delay in Io's observed motion and the extraordinary explanation that light itself takes time to travel. But Roma did not stop there.

[00:29:57]He did not merely argue that light takes time to travel. He also tried to estimate how fast it must be.

[00:30:05]He began by imagining what would happen if light took only 1 second to cross a distance roughly equal to the diameter of the Earth. That already sounds fast, but on the scale of the solar system, he was able to show that it's not fast enough. And his argument was both beautiful and simple. It was already known that the average eclipse period of Io was about 42 1/2 hours. And Roma argued that during this time, the Earth could change its distance from Jupiter by about 210 Earth diameters. And so it follows that if light took 1 second across each Earth diameter, the observed interval between two immersions would be shifted by about 210 seconds, roughly 3 1/2 minutes. Such a large shift over just one period should have been obvious in the observations. But no such delay was seen. So Roma could place a lower bound on the speed of light. It had to travel more than one Earth diameter every second. But this did not mean light was instantaneous. Rather, it meant the effect was too small to reveal itself clearly in just one or two revolutions of IO. And as we've seen, the extra delay added to each eclipse interval was tiny, of the order of seconds, not minutes. But Roma's crucial point was that the small delays could accumulate. A few seconds here, a few seconds there, repeated over many eclipses, could build into a delay large enough to measure. And from that accumulated delay, Roma concluded that light took about 22 minutes to cross the full diameter of Earth's orbit around the sun. In other words, about 11 minutes to cross the radius of Earth's orbit, roughly the distance from the sun to the Earth. That distinction matters.

[00:31:53]Roma did not write down a speed in kilome/s. He did not have the modern value of the astronomical unit, and the meter itself did not yet exist. What he had found was a light time, an estimate of how long light takes to cross the diameter of Earth's orbit. The next step was taken by Christian Huygens in his 1690 treaties on light. Huygens used Roma's light time together with an estimate for the size of Earth's orbit to convert the delay into something closer to what we would now call a speed. A little over 200,000 km/s in modern units. That's well below the modern value of about 300,000 km/s, but still an incredibly fast, almost unimaginable speed to consider in the 17th century. The error in the calculated speed came partly from Roma's estimate for the maximum delay. He used about 22 minutes for light to cross the diameter of the Earth's orbit, whereas the correct value is closer to 16.7 minutes. It also came from the fact that the size of the Earth's orbit was not yet known precisely. But that almost doesn't matter. For the first time, light was no longer beyond measurement.

[00:33:04]It had crossed from philosophy into physics. And that is what makes this story so remarkable. It began with a practical problem. How to find longitude? To solve it, astronomers turned Io into a clock in the sky. But when they watched that clock carefully, it seemed to depend on the changing distance between Earth and Jupiter. The great revelation was not that a measurement error had been made or that the tables needed correcting. The great revelation was that the anomaly could be explained quantitatively by assuming that light itself takes time to travel.

[00:33:41]And once that idea was accepted, light was no longer instantaneous. It had a speed, an astonishing speed, a speed close to 300,000 km/s in a vacuum. and what we now understand as the ultimate speed limit of the universe.

[00:33:58]And that is where we'll end for today.

[00:34:00]So until next time, goodbye.

[00:34:04]And as always, a massive thank you to all my patrons and a special shout out to the following who have been incredibly generous with their support.

[00:34:13]Thank you so much. I couldn't do it without