## From my Second A-to-Z: Orthonormal

For early 2016 — dubbed “Leap Day 2016” as that’s when it started — I got a request to explain orthogonal. I went in a different direction, although not completely different. This essay does get a bit more into specifics of how mathematicians use the idea, like, showing some calculations and such. I put in a casual description of vectors here. For book publication I’d want to rewrite that to be clearer that, like, ordered sets of numbers are just one (very common) way to represent vectors.

Jacob Kanev had requested “orthogonal” for this glossary. I’d be happy to oblige. But I used the word in last summer’s Mathematics A To Z. And I admit I’m tempted to just reprint that essay, since it would save some needed time. But I can do something more.

## Orthonormal.

“Orthogonal” is another word for “perpendicular”. Mathematicians use it for reasons I’m not precisely sure of. My belief is that it’s because “perpendicular” sounds like we’re talking about directions. And we want to extend the idea to things that aren’t necessarily directions. As majors, mathematicians learn orthogonality for vectors, things pointing in different directions. Then we extend it to other ideas. To functions, particularly, but we can also define it for spaces and for other stuff.

I was vague, last summer, about how we do that. We do it by creating a function called the “inner product”. That takes in two of whatever things we’re measuring and gives us a real number. If the inner product of two things is zero, then the two things are orthogonal.

The first example mathematics majors learn of this, before they even hear the words “inner product”, are dot products. These are for vectors, ordered sets of numbers. The dot product we find by matching up numbers in the corresponding slots for the two vectors, multiplying them together, and then adding up the products. For example. Give me the vector with values (1, 2, 3), and the other vector with values (-6, 5, -4). The inner product will be 1 times -6 (which is -6) plus 2 times 5 (which is 10) plus 3 times -4 (which is -12). So that’s -6 + 10 – 12 or -8.

So those vectors aren’t orthogonal. But how about the vectors (1, -1, 0) and (0, 0, 1)? Their dot product is 1 times 0 (which is 0) plus -1 times 0 (which is 0) plus 0 times 1 (which is 0). The vectors are perpendicular. And if you tried drawing this you’d see, yeah, they are. The first vector we’d draw as being inside a flat plane, and the second vector as pointing up, through that plane, like a thumbtack.

Well … the inner product can tell us something besides orthogonality. What happens if we take the inner product of a vector with itself? Say, (1, 2, 3) with itself? That’s going to be 1 times 1 (which is 1) plus 2 times 2 (4, according to rumor) plus 3 times 3 (which is 9). That’s 14, a tidy sum, although, so what?

The inner product of (-6, 5, -4) with itself? Oh, that’s some ugly numbers. Let’s skip it. How about the inner product of (1, -1, 0) with itself? That’ll be 1 times 1 (which is 1) plus -1 times -1 (which is positive 1) plus 0 times 0 (which is 0). That adds up to 2. And now, wait a minute. This might be something.

Start from somewhere. Move 1 unit to the east. (Don’t care what the unit is. Inches, kilometers, astronomical units, anything.) Then move -1 units to the north, or like normal people would say, 1 unit o the south. How far are you from the starting point? … Well, you’re the square root of 2 units away.

Now imagine starting from somewhere and moving 1 unit east, and then 2 units north, and then 3 units straight up, because you found a convenient elevator. How far are you from the starting point? This may take a moment of fiddling around with the Pythagorean theorem. But you’re the square root of 14 units away.

And what the heck, (0, 0, 1). The inner product of that with itself is 0 times 0 (which is zero) plus 0 times 0 (still zero) plus 1 times 1 (which is 1). That adds up to 1. And, yeah, if we go one unit straight up, we’re one unit away from where we started.

The inner product of a vector with itself gives us the square of the vector’s length. At least if we aren’t using some freak definition of inner products and lengths and vectors. And this is great! It means we can talk about the length — maybe better to say the size — of things that maybe don’t have obvious sizes.

Some stuff will have convenient sizes. For example, they’ll have size 1. The vector (0, 0, 1) was one such. So is (1, 0, 0). And you can think of another example easily. Yes, it’s $\left(\frac{1}{\sqrt{2}}, -\frac{1}{2}, \frac{1}{2}\right)$. (Go ahead, check!)

So by “orthonormal” we mean a collection of things that are orthogonal to each other, and that themselves are all of size 1. It’s a description of both what things are by themselves and how they relate to one another. A thing can’t be orthonormal by itself, for the same reason a line can’t be perpendicular to nothing in particular. But a pair of things might be orthogonal, and they might be the right length to be orthonormal too.

Why do this? Well, the same reasons we always do this. We can impose something like direction onto a problem. We might be able to break up a problem into simpler problems, one in each direction. We might at least be able to simplify the ways different directions are entangled. We might be able to write a problem’s solution as the sum of solutions to a standard set of representative simple problems. This one turns up all the time. And an orthogonal set of something is often a really good choice of a standard set of representative problems.

This sort of thing turns up a lot when solving differential equations. And those often turn up when we want to describe things that happen in the real world. So a good number of mathematicians develop a habit of looking for orthonormal sets.

## My 2019 Mathematics A To Z: Norm

Today’s A To Z term is another free choice. So I’m picking a term from the world of … mathematics. There are a lot of norms out there. Many are specialized to particular roles, such as looking at complex-valued numbers, or vectors, or matrices, or polynomials.

Still they share things in common, and that’s what this essay is for. And I’ve brushed up against the topic before.

The norm, also, has nothing particular to do with “normal”. “Normal” is an adjective which attaches to every noun in mathematics. This is security for me as while these A-To-Z sequences may run out of X and Y and W letters, I will never be short of N’s.

# Norm.

A “norm” is the size of whatever kind of thing you’re working with. You can see where this is something we look for. It’s easy to look at two things and wonder which is the smaller.

There are many norms, even for one set of things. Some seem compelling. For the real numbers, we usually let the absolute value do this work. By “usually” I mean “I don’t remember ever seeing a different one except from someone introducing the idea of other norms”. For a complex-valued number, it’s usually the square root of the sum of the square of the real part and the square of the imaginary coefficient. For a vector, it’s usually the square root of the vector dot-product with itself. (Dot product is this binary operation that is like multiplication, if you squint, for vectors.) Again, these, the “usually” means “always except when someone’s trying to make a point”.

Which is why we have the convention that there is a “the norm” for a kind of operation. The norm dignified as “the” is usually the one that looks as much as possible like the way we find distances between two points on a plane. I assume this is because we bring our intuition about everyday geometry to mathematical structures. You know how it is. Given an infinity of possible choices we take the one that seems least difficult.

Every sort of thing which can have a norm, that I can think of, is a vector space. This might be my failing imagination. It may also be that it’s quite easy to have a vector space. A vector space is a collection of things with some rules. Those rules are about adding the things inside the vector space, and multiplying the things in the vector space by scalars. These rules are not difficult requirements to meet. So a lot of mathematical structures are vector spaces, and the things inside them are vectors.

A norm is a function that has these vectors as its domain, and the non-negative real numbers as its range. And there are three rules that it has to meet. So. Give me a vector ‘u’ and a vector ‘v’. I’ll also need a scalar, ‘a. Then the function f is a norm when:

1. $f(u + v) \le f(u) + f(v)$. This is a famous rule, called the triangle inequality. You know how in a triangle, the sum of the lengths of any two legs is greater than the length of the third leg? That’s the rule at work here.
2. $f(a\cdot u) = |a| \cdot f(u)$. This doesn’t have so snappy a name. Sorry. It’s something about being homogeneous, at least.
3. If $f(u) = 0$ then u has to be the additive identity, the vector that works like zero does.

Norms take on many shapes. They depend on the kind of thing we measure, and what we find interesting about those things. Some are familiar. Look at a Euclidean space, with Cartesian coordinates, so that we might write something like (3, 4) to describe a point. The “the norm” for this, called the Euclidean norm or the L2 norm, is the square root of the sum of the squares of the coordinates. So, 5. But there are other norms. The L1 norm is the sum of the absolute values of all the coefficients; here, 7. The L norm is the largest single absolute value of any coefficient; here, 4.

A polynomial, meanwhile? Write it out as $a_0 + a_1 x + a_2 x^2 + a_3 x^3 + \cdots + a_n x^n$. Take the absolute value of each of these $a_k$ terms. Then … you have choices. You could take those absolute values and add them up. That’s the L1 polynomial norm. Take those absolute values and square them, then add those squares, and take the square root of that sum. That’s the L2 norm. Take the largest absolute value of any of these coefficients. That’s the L norm.

These don’t look so different, even though points in space and polynomials seem to be different things. We designed the tool. We want it not to be weirder than it has to be. When we try to put a norm on a new kind of thing, we look for a norm that resembles the old kind of thing. For example, when we want to define the norm of a matrix, we’ll typically rely on a norm we’ve already found for a vector. At least to set up the matrix norm; in practice, we might do a calculation that doesn’t explicitly use a vector’s norm, but gives us the same answer.

If we have a norm for some vector space, then we have an idea of distance. We can say how far apart two vectors are. It’s the norm of the difference between the vectors. This is called defining a metric on the vector space. A metric is that sense of how far apart two things are. What keeps a norm and a metric from being the same thing is that it’s possible to come up with a metric that doesn’t match any sensible norm.

It’s always possible to use a norm to define a metric, though. Doing that promotes our normed vector space to the dignified status of a “metric space”. Many of the spaces we find interesting enough to work in are such metric spaces. It’s hard to think of doing without some idea of size.

I’ve made it through one more week without missing deadline! This and all the other Fall 2019 A To Z posts should be at this link. I remain open for subjects for the letters Q through T, and would appreciate nominations at this link. Thank you for reading and I’ll fill out the rest of this week with reminders of old A-to-Z essays.

## A Leap Day 2016 Mathematics A To Z: Orthonormal

Jacob Kanev had requested “orthogonal” for this glossary. I’d be happy to oblige. But I used the word in last summer’s Mathematics A To Z. And I admit I’m tempted to just reprint that essay, since it would save some needed time. But I can do something more.

## Orthonormal.

“Orthogonal” is another word for “perpendicular”. Mathematicians use it for reasons I’m not precisely sure of. My belief is that it’s because “perpendicular” sounds like we’re talking about directions. And we want to extend the idea to things that aren’t necessarily directions. As majors, mathematicians learn orthogonality for vectors, things pointing in different directions. Then we extend it to other ideas. To functions, particularly, but we can also define it for spaces and for other stuff.

I was vague, last summer, about how we do that. We do it by creating a function called the “inner product”. That takes in two of whatever things we’re measuring and gives us a real number. If the inner product of two things is zero, then the two things are orthogonal.

The first example mathematics majors learn of this, before they even hear the words “inner product”, are dot products. These are for vectors, ordered sets of numbers. The dot product we find by matching up numbers in the corresponding slots for the two vectors, multiplying them together, and then adding up the products. For example. Give me the vector with values (1, 2, 3), and the other vector with values (-6, 5, -4). The inner product will be 1 times -6 (which is -6) plus 2 times 5 (which is 10) plus 3 times -4 (which is -12). So that’s -6 + 10 – 12 or -8.

So those vectors aren’t orthogonal. But how about the vectors (1, -1, 0) and (0, 0, 1)? Their dot product is 1 times 0 (which is 0) plus -1 times 0 (which is 0) plus 0 times 1 (which is 0). The vectors are perpendicular. And if you tried drawing this you’d see, yeah, they are. The first vector we’d draw as being inside a flat plane, and the second vector as pointing up, through that plane, like a thumbtack.

Well … the inner product can tell us something besides orthogonality. What happens if we take the inner product of a vector with itself? Say, (1, 2, 3) with itself? That’s going to be 1 times 1 (which is 1) plus 2 times 2 (4, according to rumor) plus 3 times 3 (which is 9). That’s 14, a tidy sum, although, so what?

The inner product of (-6, 5, -4) with itself? Oh, that’s some ugly numbers. Let’s skip it. How about the inner product of (1, -1, 0) with itself? That’ll be 1 times 1 (which is 1) plus -1 times -1 (which is positive 1) plus 0 times 0 (which is 0). That adds up to 2. And now, wait a minute. This might be something.

Start from somewhere. Move 1 unit to the east. (Don’t care what the unit is. Inches, kilometers, astronomical units, anything.) Then move -1 units to the north, or like normal people would say, 1 unit o the south. How far are you from the starting point? … Well, you’re the square root of 2 units away.

Now imagine starting from somewhere and moving 1 unit east, and then 2 units north, and then 3 units straight up, because you found a convenient elevator. How far are you from the starting point? This may take a moment of fiddling around with the Pythagorean theorem. But you’re the square root of 14 units away.

And what the heck, (0, 0, 1). The inner product of that with itself is 0 times 0 (which is zero) plus 0 times 0 (still zero) plus 1 times 1 (which is 1). That adds up to 1. And, yeah, if we go one unit straight up, we’re one unit away from where we started.

The inner product of a vector with itself gives us the square of the vector’s length. At least if we aren’t using some freak definition of inner products and lengths and vectors. And this is great! It means we can talk about the length — maybe better to say the size — of things that maybe don’t have obvious sizes.

Some stuff will have convenient sizes. For example, they’ll have size 1. The vector (0, 0, 1) was one such. So is (1, 0, 0). And you can think of another example easily. Yes, it’s $\left(\frac{1}{\sqrt{2}}, -\frac{1}{2}, \frac{1}{2}\right)$. (Go ahead, check!)

So by “orthonormal” we mean a collection of things that are orthogonal to each other, and that themselves are all of size 1. It’s a description of both what things are by themselves and how they relate to one another. A thing can’t be orthonormal by itself, for the same reason a line can’t be perpendicular to nothing in particular. But a pair of things might be orthogonal, and they might be the right length to be orthonormal too.

Why do this? Well, the same reasons we always do this. We can impose something like direction onto a problem. We might be able to break up a problem into simpler problems, one in each direction. We might at least be able to simplify the ways different directions are entangled. We might be able to write a problem’s solution as the sum of solutions to a standard set of representative simple problems. This one turns up all the time. And an orthogonal set of something is often a really good choice of a standard set of representative problems.

This sort of thing turns up a lot when solving differential equations. And those often turn up when we want to describe things that happen in the real world. So a good number of mathematicians develop a habit of looking for orthonormal sets.