WEBVTT
00:00:00.600 --> 00:00:06.800
This is a video Iโve been excited to make for a while now.
00:00:07.360 --> 00:00:13.360
The story here braids together prime numbers, complex numbers, and ๐ in a very pleasing trio.
00:00:14.040 --> 00:00:21.680
Quite often in modern math, especially that which flirts with the Riemann zeta function, these three seemingly unrelated objects show up in unison.
00:00:22.280 --> 00:00:28.800
And I wanna give you a little peek at one instance where this happens, one of the few that doesnโt require too heavy a technical background.
00:00:29.560 --> 00:00:31.000
Thatโs not to say that this is easy.
00:00:31.320 --> 00:00:35.960
In fact, this is probably one of the most intricate videos Iโve ever done, but the culmination is worth it.
00:00:35.960 --> 00:00:41.600
What weโll end up with is a formula for ๐, a certain alternating infinite sum.
00:00:42.600 --> 00:00:47.000
This formula is actually written on the mug that Iโm drinking coffee from right now as I write this.
00:00:47.440 --> 00:00:56.200
And a fun, but almost certainly apocryphal story is that the beauty of this formula is what inspired Leibniz to quit being a lawyer and instead pursue math.
00:00:57.240 --> 00:01:03.720
Now, whenever you see ๐ show up in math, thereโs always gonna be a circle hiding somewhere, sometimes very sneakily.
00:01:04.360 --> 00:01:10.160
So the goal here is not just to discover this sum, but to really understand the circle hiding behind it.
00:01:11.200 --> 00:01:19.040
You see, there is another way that you can prove the same result that you and I are gonna spend some meaningful time building up to, but with just a few lines of calculus.
00:01:19.760 --> 00:01:29.120
And this is one of those proofs that leaves you thinking, โokay, I suppose thatโs true,โ but not really getting a sense for why or for where the hidden circle is.
00:01:29.920 --> 00:01:42.760
On the path that you and I will take though, what youโll see is that the fundamental truth behind this sum and the circle that it hides is a certain regularity in the way that prime numbers behave when you put them inside the complex numbers.
00:01:43.680 --> 00:01:50.640
To start the story, imagine yourself with nothing more than a pencil, some paper, and a desire to find a formula for computing ๐.
00:01:51.560 --> 00:01:53.840
There are countless ways that you could approach this.
00:01:54.160 --> 00:02:00.920
But as a broad outline for the plotline here, youโll start by asking how many lattice points of the plane sit inside a big circle.
00:02:01.760 --> 00:02:11.520
And then that question is gonna lead to asking about how to express numbers as the sum of two squares, which in turn is gonna lead us to factoring integers inside the complex plane.
00:02:12.360 --> 00:02:23.960
From there, weโll bring in the special function named chi, which is gonna give us a formula for ๐ that at first seems to involve a crazy complicated pattern dependent on the distribution of primes.
00:02:24.480 --> 00:02:29.600
But a slight shift in perspective is gonna simplify it dramatically and expose the ultimate gold nugget.
00:02:30.400 --> 00:02:32.680
Itโs a lot, but good math takes time.
00:02:33.040 --> 00:02:33.720
And weโll take it step by step.
00:02:33.720 --> 00:02:44.160
When I say lattice point, what I mean is a point ๐, ๐ on the plane, where ๐ and ๐ are both integers, a spot where the grid lines here cross.
00:02:45.080 --> 00:02:52.080
If you draw a circle centered at the origin, letโs say with radius 10, how many lattice points would you guess are inside that circle?
00:02:52.080 --> 00:02:56.680
Well, thereโs one lattice point for each unit of area.
00:02:57.200 --> 00:03:06.160
So the answer should be approximately equal to the area of the circle, ๐๐ squared, which in this case is ๐ times 10 squared.
00:03:06.160 --> 00:03:18.160
And if it was a really big circle, like radius 1000000, you would expect this to be a much more accurate estimate, in the sense that the percent error between the estimate ๐๐ squared and the actual count of lattice points should get smaller.
00:03:19.160 --> 00:03:25.240
What weโre gonna try to do is find a second way to answer the same question: how many lattice points are inside the circle.
00:03:25.880 --> 00:03:33.760
Because that can lead to another way to express the area of a circle and hence another way to express ๐.
00:03:33.760 --> 00:03:35.440
And so, you play and you wonder.
00:03:35.920 --> 00:03:43.480
And maybe, especially if you just watched a certain calculus video, you might try looking through every possible ring that a lattice point could sit on.
00:03:44.760 --> 00:03:52.000
Now if you think about it, for each one of these lattice points, ๐, ๐, its distance from the origin is the square root of ๐ squared plus ๐ squared.
00:03:52.680 --> 00:03:57.240
And since ๐ and ๐ are both integers, ๐ squared plus ๐ squared is also some integer.
00:03:57.840 --> 00:04:03.200
So you only have to consider rings whose radii are the square roots of some whole number.
00:04:04.200 --> 00:04:07.280
A radius of zero just gives you that single origin point.
00:04:08.200 --> 00:04:11.680
If you look at the radius one, that hits four different lattice points.
00:04:12.280 --> 00:04:16.000
Radius square root of two, well that also hits four lattice points.
00:04:16.640 --> 00:04:19.160
A radius square root of three doesnโt actually hit anything.
00:04:19.800 --> 00:04:22.240
Square root of four, again, hits four lattice points.
00:04:22.800 --> 00:04:28.240
A radius square root of five actually hits eight lattice points.
00:04:28.240 --> 00:04:36.520
And what we want is a systematic way to count how many lattice points are on a given one of these rings, a given distance from the origin, and then to tally them all up.
00:04:37.680 --> 00:04:45.160
And if you pause and try this for a moment, what youโll find is that the pattern seems really chaotic, just very hard to find order under here.
00:04:45.680 --> 00:04:49.280
And thatโs a good sign that some very interesting math is about to come into play.
00:04:50.120 --> 00:04:56.680
In fact, as youโll see, this pattern is rooted in the distribution of primes.
00:04:56.680 --> 00:05:00.080
As an example, letโs look at the ring with radius square root of 25.
00:05:00.880 --> 00:05:05.280
It hits the point five, zero since five squared plus zero squared is 25.
00:05:06.000 --> 00:05:10.640
It also hits four, three since four squared plus three squared gives 25.
00:05:10.640 --> 00:05:16.440
And likewise, it hits three, four and also zero, five.
00:05:16.440 --> 00:05:27.600
And whatโs really happening here is that youโre counting how many pairs of integers ๐, ๐ have the property that ๐ squared plus ๐ squared equals 25.
00:05:28.440 --> 00:05:31.600
And looking at the circle, it looks like thereโs a total of 12 of them.
00:05:32.760 --> 00:05:35.960
As another example, take a look at the ring with radius square root 11.
00:05:36.600 --> 00:05:38.240
It doesnโt hit any lattice points.
00:05:38.680 --> 00:05:44.560
And that corresponds to the fact that you cannot find two integers whose squares add up to 11.
00:05:45.080 --> 00:05:45.360
Try it!
00:05:48.520 --> 00:05:59.080
Now, many times in math, when you see a question that has to do with the 2D plane, it can be surprisingly fruitful to just ask what it looks like when you think of this plane as the set of all complex numbers.
00:05:59.280 --> 00:06:09.360
So instead of thinking of this lattice point here as the pair of integer coordinates three, four, instead think of it as the single complex number three plus four ๐.
00:06:10.600 --> 00:06:21.360
That way, another way to think about the sum of the squares of its coordinates, three squared plus four squared, is to multiply this number by three minus four ๐.
00:06:21.360 --> 00:06:22.400
This is called its complex conjugate.
00:06:22.680 --> 00:06:28.520
Itโs what you get by reflecting over the real axis, replacing ๐ with negative ๐.
00:06:28.520 --> 00:06:32.560
And this might seem like a strange step if you donโt have much of a history with complex numbers.
00:06:33.200 --> 00:06:37.040
But describing this distance as a product can be unexpectedly useful.
00:06:37.760 --> 00:06:43.800
It turns our question into a factoring problem, which is ultimately why patterns among prime numbers are gonna come into play.
00:06:44.840 --> 00:06:48.400
Algebraically, this relation is straightforward enough to verify.
00:06:48.920 --> 00:06:55.160
You get a three squared and then the three times minus four ๐ cancels out with the four ๐ times three.
00:06:55.880 --> 00:07:02.840
And then, you have negative four ๐ squared, which because ๐ squared is negative one becomes plus four squared.
00:07:04.160 --> 00:07:06.040
This is also quite nice to see geometrically.
00:07:06.560 --> 00:07:15.360
And if youโre a little rusty with how complex multiplication works, I do have another video that goes more into detail about why complex multiplication looks the way that it does.
00:07:15.920 --> 00:07:23.680
The way that you might think about a case like this is that the number three plus four ๐ has a magnitude of five and some angle off of the horizontal.
00:07:24.560 --> 00:07:40.280
And what it means to multiply it by three minus four ๐ is to rotate by that same angle in the opposite direction, putting it on the positive real axis, and then to stretch out by a factor of five, which in this case lands you on the output 25, the square of the magnitude.
00:07:43.480 --> 00:07:49.720
The collection of all of these lattice points ๐ plus ๐๐, where ๐ and ๐ are integers, has a special name.
00:07:50.320 --> 00:07:53.800
Theyโre called the Gaussian integers, named after Martin Sheen.
00:07:54.440 --> 00:07:56.760
Geometrically, youโll still be asking the same question.
00:07:57.440 --> 00:08:04.160
How many of these lattice points, Gaussian integers, are a given distance away from the origin, like square root of 25?
00:08:04.920 --> 00:08:07.720
But weโll be phrasing it in a slightly more algebraic way.
00:08:08.360 --> 00:08:14.960
How many Gaussian integers have the property that multiplying by their complex conjugate gives you 25?
00:08:16.520 --> 00:08:18.400
This might seem needlessly complex.
00:08:18.760 --> 00:08:25.560
But itโs the key to understanding the seemingly random pattern for how many lattice points are a given distance from the origin.
00:08:26.560 --> 00:08:33.920
To see why, we first need to understand how numbers factor inside the Gaussian integers.
00:08:33.920 --> 00:08:40.720
As a refresher, among ordinary integers, every number can be factored as some unique collection of prime numbers.
00:08:41.480 --> 00:08:48.120
For example, 2250 can be factored as two times three squared times five cubed.
00:08:48.680 --> 00:09:02.760
And there is no other collection of prime numbers that also multiplies to make 2250, unless you let negative numbers into the picture, in which case you could just make some of the primes in this factorization negative.
00:09:03.560 --> 00:09:08.840
So really, within the integers, factorization is not perfectly unique.
00:09:09.120 --> 00:09:15.280
Itโs almost unique with the exception that you can get a different-looking product by multiplying some of the factors by negative one.
00:09:18.080 --> 00:09:22.640
The reason I bring that up is that factoring works very similarly inside the Gaussian integers.
00:09:22.640 --> 00:09:33.480
Some numbers, like five, can be factored into smaller Gaussian integers, which in this case is two plus ๐ times two minus ๐.
00:09:33.480 --> 00:09:40.000
This Gaussian integer here, two plus ๐, cannot be factored into anything smaller, so we call it a Gaussian prime.
00:09:41.080 --> 00:09:44.320
Again, this factorization is almost unique.
00:09:44.600 --> 00:09:51.080
But this time, not only can you multiply each one of those factors by negative one to get a factorization that looks different.
00:09:53.800 --> 00:10:03.080
You can also be extra sneaky and multiply one of these factors by ๐ and then the other one by negative ๐.
00:10:03.080 --> 00:10:07.480
This will give you a different way to factor five into two distinct Gaussian primes.
00:10:08.440 --> 00:10:18.480
But other than the things that you can get by multiplying some of these factors by negative one or ๐ or negative ๐, factorization within the Gaussian integers is unique.
00:10:20.040 --> 00:10:30.480
And if you can figure out how ordinary prime numbers factor inside the Gaussian integers, that will be enough to tell us how any other natural number factors inside these Gaussian integers.
00:10:31.240 --> 00:10:35.040
And so here, we pull in a crucial and pretty surprising fact.
00:10:35.920 --> 00:10:47.600
Prime numbers that are one above a multiple of four, like five or 13 or 17, these guys can always be factored into exactly two distinct Gaussian primes.
00:10:48.720 --> 00:10:56.000
This corresponds with the fact that rings with a radius equal to the square root of one of these prime numbers always hit some lattice points.
00:10:56.560 --> 00:11:00.200
In fact, they always hit exactly eight lattice points, as youโll see in just a moment.
00:11:00.200 --> 00:11:13.800
On the other hand, prime numbers that are three above a multiple of four, like three or seven or 11, these guys cannot be factored further inside the Gaussian integers.
00:11:14.520 --> 00:11:21.320
Not only are they primes in the normal numbers, but they are also Gaussian primes, unsplittable, even when ๐ is in the picture.
00:11:22.200 --> 00:11:28.480
And this corresponds with the fact that a ring whose radius is the square root of one of those primes will never hit any lattice points.
00:11:28.480 --> 00:11:38.800
And this pattern right here is the regularity within prime numbers that weโre gonna ultimately explain.
00:11:39.640 --> 00:11:42.600
And in a later video, I might explain why on earth this is true.
00:11:43.160 --> 00:11:55.120
Why a prime numberโs remainder when divided by four has anything to do with whether or not it factors inside the Gaussian integers, or, said differently, whether or not it can be expressed as the sum of two squares?
00:11:55.960 --> 00:11:58.520
But here and now, weโll just have to take it as a given.
00:11:59.680 --> 00:12:03.840
The prime number two, by the way, is a little special because it does factor.
00:12:04.040 --> 00:12:07.320
You can write it as one plus ๐ times one minus ๐.
00:12:08.080 --> 00:12:11.880
But these two Gaussian primes are a 90-degree rotation away from each other.
00:12:12.480 --> 00:12:15.680
So you can multiply one of them by ๐ to get the other.
00:12:16.480 --> 00:12:22.440
And that fact is gonna make us wanna treat the prime number two a little bit differently for where all of this stuff is going.
00:12:22.680 --> 00:12:27.080
So just keep that in the back of your mind.
00:12:27.080 --> 00:12:31.760
Remember, our goal here is to count how many lattice points are a given distance away from the origin.
00:12:32.400 --> 00:12:37.840
And doing this systematically for all distances, square root of ๐, can lead us to a formula for ๐.
00:12:38.800 --> 00:12:52.840
And counting the number of lattice points with a given magnitude, like square root of 25, is the same as asking how many Gaussian integers have the special property that multiplying them by their complex conjugate gives you 25.
00:12:53.960 --> 00:12:57.960
So hereโs the recipe for finding all Gaussian integers that have this property.
00:12:58.840 --> 00:13:04.360
Step one, factor 25, which inside the ordinary integers looks like five squared.
00:13:04.880 --> 00:13:12.640
But since five factors even further as two plus ๐ times two minus ๐, 25 breaks down as these four Gaussian primes.
00:13:13.440 --> 00:13:19.040
Step two, organize these into two different columns, with conjugate pairs sitting right next to each other.
00:13:20.200 --> 00:13:23.280
Then, once you do that, multiply whatโs in each column.
00:13:23.680 --> 00:13:26.400
And youโll come out with two different Gaussian integers on the bottom.
00:13:27.320 --> 00:13:36.640
And because everything on the right is a conjugate with everything on the left, what comes out is gonna be a complex conjugate pair, which multiplies to 25.
00:13:37.960 --> 00:13:43.120
Picking an arbitrary standard, letโs say that the product from that left column is the output of our recipe.
00:13:44.600 --> 00:13:50.200
Now, notice there are three choices for how you can divvy up the primes that can affect this output.
00:13:51.240 --> 00:13:55.400
Pictured right here, both copies of two plus ๐ are in the left column.
00:13:55.400 --> 00:13:57.520
And that gives us the product three plus four ๐.
00:13:58.440 --> 00:14:05.000
You could also have chosen to have only one copy of two plus ๐ in this left column, in which case the product would be five.
00:14:05.760 --> 00:14:16.640
Or, you could have both copies of two plus ๐ in that right column, in which case the output of our recipe wouldโve been three minus four ๐.
00:14:16.640 --> 00:14:22.480
And those three possible outputs are all different lattice points on a circle with radius square root of 25.
00:14:24.640 --> 00:14:29.200
But why does this recipe not yet capture all 12 of the lattice points?
00:14:30.120 --> 00:14:38.080
Remember how I mentioned that a factorization into Gaussian primes can look different if you multiply some of them by ๐ or negative one, negative ๐.
00:14:38.920 --> 00:14:48.040
In this case, you could write the factorization of 25 differently, maybe splitting up one of those fives as negative one plus two ๐ times negative one minus two ๐.
00:14:49.000 --> 00:14:52.480
And if you do that, running through the same recipe, it can affect the result.
00:14:52.680 --> 00:14:54.920
Youโll get a different product out of that left column.
00:14:55.960 --> 00:15:03.200
But the only effect that this is gonna have is to multiply that total output by ๐ or negative one or negative ๐.
00:15:03.880 --> 00:15:08.200
So as a final step for our recipe, letโs say you have to make one of four choices.
00:15:08.800 --> 00:15:19.040
Take that product from the left column and choose to multiply it by one, ๐, negative one, or negative ๐, corresponding to rotations that are some multiple of 90 degrees.
00:15:21.720 --> 00:15:28.800
That will account for all 12 different ways of constructing a Gaussian integer whose product with its own conjugate is 25.
00:15:28.800 --> 00:15:32.160
This process is a little complicated.
00:15:32.160 --> 00:15:35.680
So I think the best way to get a feel for it is to just try it out with more examples.
00:15:35.880 --> 00:15:40.400
Letโs say, instead, we were looking at 125, which is five cubed.
00:15:41.200 --> 00:15:47.920
In that case, we would have four different choices for how to divvy up the prime conjugate pairs into these two columns.
00:15:48.400 --> 00:15:58.800
You can either have zero copies of two plus ๐ in the left column, one copy in there, two copies in there, or all three of them in that left column.
00:16:00.000 --> 00:16:16.840
Those four choices multiplied by the final four choices of multiplying the product from the left column by one or by ๐ or negative one or negative ๐ would suggest that there are a total of 16 lattice points at distance square root of 125 away from the origin.
00:16:18.960 --> 00:16:27.480
And indeed, if you draw that circle out and count, what youโll find is that it hits exactly 16 lattice points.
00:16:27.480 --> 00:16:33.600
But what if you introduce a factor like three, which doesnโt break down as the product of two conjugate Gaussian primes?
00:16:34.400 --> 00:16:36.320
Well, that really mucks up the whole system.
00:16:36.960 --> 00:16:42.120
When youโre divvying up the primes between the two columns, thereโs no way that you can split up this three.
00:16:42.560 --> 00:16:45.720
No matter where you put it, it leaves the columns imbalanced.
00:16:46.280 --> 00:16:53.080
And what that means is that when you take the product of all of the numbers in each column, youโre not gonna end up with the conjugate pair.
00:16:54.000 --> 00:17:01.200
So for a number like this, three times five cubed, which is 375, thereโs actually no lattice point that youโll hit.
00:17:01.720 --> 00:17:06.960
No Gaussian integer whose product with its own conjugate gives you 375.
00:17:08.080 --> 00:17:12.320
However, if you introduce a second factor of three, then you have an option.
00:17:12.920 --> 00:17:17.040
You can throw one three in the left column and the other three in the right column.
00:17:17.880 --> 00:17:29.800
Since three is its own complex conjugate, this leaves things balanced, in the sense that the products of the left and right columns will indeed be a complex conjugate pair.
00:17:29.800 --> 00:17:31.640
But it doesnโt add any new options.
00:17:31.920 --> 00:17:42.760
Thereโs still gonna be a total of four choices for how to divvy up those factors of five, multiplied by the final four choices of multiplying by one, ๐, negative one, or negative ๐.
00:17:42.760 --> 00:17:51.400
So just like the square root of 125 circle, this guy is also gonna end up hitting exactly 16 lattice points.
00:17:51.400 --> 00:17:52.520
Letโs just sum up where we are.
00:17:52.960 --> 00:17:59.560
When youโre counting up how many lattice points lie on a circle with a radius square root of ๐, the first step is to factor ๐.
00:18:01.000 --> 00:18:14.720
And for prime numbers like five or 13 or 17 which factor further into a complex conjugate pair of Gaussian primes, the number of choices they give you will always be one more than the exponent that shows up with that factor.
00:18:17.200 --> 00:18:28.720
On the other hand, for prime factors like three or seven or 11, which are already Gaussian primes and cannot be split, if they show up with an even power, you have one and only one choice with what to do with them.
00:18:29.600 --> 00:18:31.440
But if itโs an odd exponent, youโre screwed.
00:18:31.440 --> 00:18:33.320
And you just have zero choices.
00:18:34.280 --> 00:18:37.600
And always, no matter what, you have those final four choices at the end.
00:18:40.240 --> 00:18:44.680
By the way, I do think that this process right here is the most complicated part of the video.
00:18:45.320 --> 00:18:49.960
It took me a couple times to think through that, โYes! this is a valid way to count lattice points.โ
00:18:49.960 --> 00:18:53.920
So donโt be shy if you wanna pause and scribble things down to get a feel for it.
00:18:54.920 --> 00:19:00.000
The one last thing to mention about this recipe is how factors of two affect the count.
00:19:01.120 --> 00:19:07.440
If your number is even, then that factor of two breaks down as one plus ๐ times one minus ๐.
00:19:08.200 --> 00:19:11.880
So you can divvy up that complex conjugate pair between the two columns.
00:19:12.720 --> 00:19:20.600
And at first, it might look like this doubles your options, depending on how you choose to place those two Gaussian primes between the columns.
00:19:21.360 --> 00:19:36.520
However, since multiplying one of these guys by ๐ gives you the other one, when you swap them between the columns, the effect that that has on the output from the left column is to just multiply it by ๐ or by negative ๐.
00:19:36.520 --> 00:19:45.560
So thatโs actually redundant with the final step, where we take the product of this left column and choose to multiply it either by one, ๐, negative one, or negative ๐.
00:19:46.600 --> 00:19:53.200
What this means is that a factor of two, or any power of two, doesnโt actually change the count at all.
00:19:53.720 --> 00:19:55.600
It doesnโt hurt, but it doesnโt help.
00:19:56.400 --> 00:20:00.880
For example, a circle with radius square root of five hits eight lattice points.
00:20:01.240 --> 00:20:05.680
And if you grow this radius to square root of 10, then you also hit eight lattice points.
00:20:06.080 --> 00:20:10.160
And square root of 20 also hits eight lattice points, as does square root of 40.
00:20:10.960 --> 00:20:12.880
Factors of two just donโt make a difference.
00:20:12.880 --> 00:20:18.400
Now whatโs about to happen is number theory at its best.
00:20:18.960 --> 00:20:26.160
We have this complicated recipe telling us how many lattice points sit on a circle with radius square root of ๐.
00:20:26.160 --> 00:20:28.680
And it depends on the prime factorization of ๐.
00:20:28.680 --> 00:20:43.440
To turn this into something simpler, something we can actually deal with, weโre gonna exploit the regularity of primes that those which are one above a multiple of four split into distinct Gaussian prime factors, while those that are three above a multiple of four cannot be split.
00:20:44.280 --> 00:20:48.840
To do this, letโs introduce a simple function, one which Iโll label with the Greek letter ๐.
00:20:49.720 --> 00:20:54.720
For inputs that are one above a multiple of four, the output of ๐ is just one.
00:20:55.840 --> 00:21:02.400
If it takes in an input three above a multiple of four, then the output of ๐ is negative one.
00:21:02.400 --> 00:21:04.600
And then on all even numbers, it gives zero.
00:21:04.600 --> 00:21:18.560
So if you evaluate ๐ on the natural numbers, it gives this very nice cyclic pattern: one, zero, negative one, zero, and then repeat indefinitely.
00:21:18.800 --> 00:21:24.080
And this cyclic function ๐ has a very special property.
00:21:24.560 --> 00:21:26.520
Itโs whatโs called a multiplicative function.
00:21:27.520 --> 00:21:39.680
If you evaluate it on two different numbers and multiply the results, like ๐ of three times ๐ of five, itโs the same as if you evaluate ๐ on the product of those two numbers, in this case ๐ of 15.
00:21:40.840 --> 00:21:46.440
Likewise, ๐ of five times ๐ of five is equal to ๐ of 25.
00:21:46.440 --> 00:21:50.160
And no matter what two natural numbers you put in there, this property will hold.
00:21:50.640 --> 00:21:51.760
Go ahead; try it if you want.
00:21:53.240 --> 00:22:06.160
So for our central question of counting lattice points in this way that involves factoring a number, what Iโm gonna do is write down the number of choices we have, but using ๐ in what at first seems like a much more complicated way.
00:22:06.640 --> 00:22:09.280
But this has the benefit of treating all prime factors equally.
00:22:09.280 --> 00:22:19.080
For each prime power, like five cubed, what you write down is ๐ of one plus ๐ of five plus ๐ of five squared plus ๐ of five cubed.
00:22:19.640 --> 00:22:26.120
You add up the value of ๐ on all the powers of this prime up to the one that shows up inside the factorization.
00:22:27.280 --> 00:22:32.200
In this case, since five is one above a multiple of four, all of these are just one.
00:22:32.800 --> 00:22:42.600
So this sum comes out to be four, which reflects the fact that a factor of five cubed gives you four options for how to divvy up the two Gaussian prime factors between the columns.
00:22:42.600 --> 00:22:54.000
For a factor like three to the fourth, what you write down looks totally similar: ๐ of one plus ๐ of three, on and on up to ๐ of three to the fourth.
00:22:55.040 --> 00:22:59.880
But in this case, since ๐ of three is negative one, this sum oscillates.
00:22:59.880 --> 00:23:03.600
It goes one minus one plus one minus one plus one.
00:23:04.480 --> 00:23:15.440
And if itโs an even power, like four in this case, the total sum comes out to be one, which encapsulates the fact that there is only one choice for what to do with those unsplittable threes.
00:23:16.160 --> 00:23:20.680
But if itโs an odd power, that sum comes out to zero, indicating that youโre screwed.
00:23:20.680 --> 00:23:22.720
You canโt place that unsplittable three.
00:23:22.720 --> 00:23:33.400
When you do this for a power of two, what it looks like is one plus zero plus zero plus zero, on and on, since ๐ is always zero on even numbers.
00:23:34.200 --> 00:23:38.040
And this reflects the fact that a factor of two doesnโt help and it doesnโt hurt.
00:23:38.280 --> 00:23:40.600
You always have just one option for what to do with it.
00:23:41.880 --> 00:23:47.840
And as always, we keep a four in front to indicate that final choice of multiplying by one, ๐, negative one, or negative ๐.
00:23:49.080 --> 00:23:50.560
Weโre getting close to the culmination now.
00:23:51.000 --> 00:23:52.520
Things are starting to look organized.
00:23:52.520 --> 00:23:54.160
So take a moment, pause and ponder.
00:23:54.360 --> 00:23:56.320
Make sure everything feels good up to this point.
00:23:57.520 --> 00:23:59.440
Take the number 45 as an example.
00:24:00.120 --> 00:24:02.560
This guy factors as three squared times five.
00:24:03.120 --> 00:24:12.400
So the expression for the total number of lattice points is four times ๐ of one plus ๐ of three plus ๐ of three squared times ๐ of one plus ๐ of five.
00:24:13.120 --> 00:24:21.240
You can think about this as four times the one choice for what to do with the threes times two choices for how to divvy up the Gaussian prime factors of five.
00:24:22.560 --> 00:24:30.760
It might seem like expanding out this sum is really complicated, because it involves all possible combinations of these prime factors, and it kind of is.
00:24:30.760 --> 00:24:39.000
However, because ๐ is multiplicative, each one of those combinations corresponds to a divisor of 45.
00:24:39.840 --> 00:24:52.800
I mean, in this case, what we get is four times ๐ of one plus ๐ of three plus ๐ of five plus ๐ of nine plus ๐ of 15 plus ๐ of 45.
00:24:52.800 --> 00:24:58.160
And what youโll notice is that this covers every number that divides evenly into 45, once and only once.
00:24:58.960 --> 00:25:00.440
And it works like this for any number.
00:25:00.640 --> 00:25:02.520
Thereโs nothing special about 45.
00:25:03.680 --> 00:25:06.760
And that to me is pretty interesting, and I think wholly unexpected.
00:25:07.400 --> 00:25:18.320
This question of counting the number of lattice points at distance square root of ๐ away from the origin involves adding up the value of this relatively simple function over all the divisors of ๐.
00:25:20.280 --> 00:25:22.720
To bring it altogether, remember why weโre doing this.
00:25:23.160 --> 00:25:30.200
The total number of lattice points inside a big circle with radius ๐
should be about ๐ times ๐
squared.
00:25:30.200 --> 00:25:40.360
But on the other hand, we can count those same lattice points by looking through all of the numbers ๐ between zero and ๐
squared and counting how many lattice points are at distance square root of ๐ from the origin.
00:25:41.240 --> 00:25:44.480
Letโs go ahead and just ignore that origin dot with radius zero.
00:25:44.800 --> 00:25:46.880
It doesnโt really follow the pattern of the rest.
00:25:46.880 --> 00:25:50.880
And one little dot isnโt gonna make a difference as we let ๐
grow towards infinity.
00:25:50.880 --> 00:26:06.080
Now, from all of this Gaussian integer and factoring and ๐ function stuff that weโve been doing, the answer for each ๐ looks like adding up the value of ๐ on every divisor of ๐ and then multiplying by four.
00:26:07.160 --> 00:26:11.560
And for now, letโs just take that four and put it in the corner and remember to bring it back later.
00:26:12.720 --> 00:26:18.320
At first, adding up the values for each one of these rows seems super random, right?
00:26:18.800 --> 00:26:25.880
I mean numbers with a lot of factors have a lot of divisors, whereas prime numbers will always only have two divisors.
00:26:26.520 --> 00:26:33.080
So it initially seems like you would have to have perfect knowledge of the distribution of primes to get anything useful out of this.
00:26:34.160 --> 00:26:39.040
But if, instead, you organize these into columns, the puzzle starts to fit together.
00:26:40.040 --> 00:26:43.800
How many numbers between one and ๐
squared have one as a divisor?
00:26:44.520 --> 00:26:45.240
Well, all of them.
00:26:45.840 --> 00:26:49.120
So our sum should include ๐
squared times ๐ of one.
00:26:50.320 --> 00:26:52.120
How many of them have two as a divisor?
00:26:52.120 --> 00:26:58.400
Well, about half of them, so that would account for about ๐
squared over two times ๐ of two.
00:26:59.040 --> 00:27:01.920
About a third of these rows have ๐ of three.
00:27:01.920 --> 00:27:06.240
So we can put in ๐
squared divided by three times ๐ of three.
00:27:06.240 --> 00:27:10.560
And keep in mind weโre being approximate, since ๐
squared might not perfectly divide two or three.
00:27:10.560 --> 00:27:14.440
But as ๐
grows towards infinity, this approximation will get better.
00:27:15.360 --> 00:27:23.680
And when you keep going like this, you get a pretty organized expression for the total number of lattice points.
00:27:23.680 --> 00:27:37.920
And if you factor out that ๐
squared and then bring back the four that needs to be multiplied in, what it means is that the total number of lattice points inside this big circle is approximately four times ๐
squared times this sum.
00:27:38.640 --> 00:27:52.120
And because ๐ is zero on every even number and it oscillates between one and negative one for odd numbers, this sum looks like one minus one-third plus a fifth minus one-seventh, and so on.
00:27:52.960 --> 00:27:54.400
And this is exactly what we wanted!
00:27:54.720 --> 00:28:02.960
What we have here is an alternate expression for the total number of lattice points inside a big circle, which we know should be around ๐ times ๐
squared.
00:28:02.960 --> 00:28:08.040
And the bigger ๐
is, the more accurate both of these estimates are.
00:28:08.520 --> 00:28:13.280
So the percent error between the left-hand side and the right-hand side can get arbitrarily small.
00:28:14.240 --> 00:28:20.240
So divide out by that ๐
squared, and this gives us an infinite sum that should converge to ๐.
00:28:21.200 --> 00:28:23.120
And keep in mind, I just think this is really cool.
00:28:23.360 --> 00:28:35.360
The reason that this sum came out to be so simple, requiring relatively low information to describe, ultimately stems from the regular pattern and how prime numbers factor inside the Gaussian integers.
00:28:37.440 --> 00:28:44.280
If youโre curious, there are two main branches of number theory: algebraic number theory and analytic number theory.
00:28:44.280 --> 00:28:51.760
Very loosely speaking, the former deals with new number systems, things like these Gaussian integers that you and I looked at and a lot more.
00:28:52.360 --> 00:29:02.160
And the latter deals with things like the Riemann zeta function or its cousins, called ๐ฟ functions, which involve multiplicative functions like this central character ๐ from our story.
00:29:03.160 --> 00:29:07.440
And the path that we just walked is a little glimpse at where those two fields intersect.
00:29:07.920 --> 00:29:12.440
And both of these are pretty heavy-duty fields with a lot of active research and unsolved problems.
00:29:12.960 --> 00:29:22.600
So if all this feels like something that takes time to mentally digest, like thereโs more patterns to be uncovered and understood, itโs because it is and there are!