Of course I can’t just take a break for the sake of having a break. I feel like I have to do something of interest. So why not make better use of my several thousand past entries and repost one? I’d just reblog it except WordPress’s system for that is kind of rubbish. So here’s what I wrote, when I was first doing A-to-Z’s, back in summer of 2015. Somehow I was able to post three of these a week. I don’t know how.
I had remembered this essay as mostly describing the boring part of tensors, that we usually represent them as grids of numbers and then symbols with subscripts and superscripts. I’m glad to rediscover that I got at why we do such things to numbers and subscripts and superscripts.
Tensor.
The true but unenlightening answer first: a tensor is a regular, rectangular grid of numbers. The most common kind is a two-dimensional grid, so that it looks like a matrix, or like the times tables. It might be square, with as many rows as columns, or it might be rectangular.
It can also be one-dimensional, looking like a row or a column of numbers. Or it could be three-dimensional, rows and columns and whole levels of numbers. We don’t try to visualize that. It can be what we call zero-dimensional, in which case it just looks like a solitary number. It might be four- or more-dimensional, although I confess I’ve never heard of anyone who actually writes out such a thing. It’s just so hard to visualize.
You can add and subtract tensors if they’re of compatible sizes. You can also do something like multiplication. And this does mean that tensors of compatible sizes will form a ring. Of course, that doesn’t say why they’re interesting.
Tensors are useful because they can describe spatial relationships efficiently. The word comes from the same Latin root as “tension”, a hint about how we can imagine it. A common use of tensors is in describing the stress in an object. Applying stress in different directions to an object often produces different effects. The classic example there is a newspaper. Rip it in one direction and you get a smooth, clean tear. Rip it perpendicularly and you get a raggedy mess. The stress tensor represents this: it gives some idea of how a force put on the paper will create a tear.
Tensors show up a lot in physics, and so in mathematical physics. Technically they show up everywhere, since vectors and even plain old numbers (scalars, in the lingo) are kinds of tensors, but that’s not what I mean. Tensors can describe efficiently things whose magnitude and direction changes based on where something is and where it’s looking. So they are a great tool to use if one wants to represent stress, or how well magnetic fields pass through objects, or how electrical fields are distorted by the objects they move in. And they describe space, as well: general relativity is built on tensors. The mathematics of a tensor allow one to describe how space is shaped, based on how to measure the distance between two points in space.
My own mathematical education happened to be pretty tensor-light. I never happened to have courses that forced me to get good with them, and I confess to feeling intimidated when a mathematical argument gets deep into tensor mathematics. Joseph C Kolecki, with NASA’s Glenn (Lewis) Research Center, published in 2002 a nice little booklet “An Introduction to Tensors for Students of Physics and Engineering”. This I think nicely bridges some of the gap between mathematical structures like vectors and matrices, that mathematics and physics majors know well, and the kinds of tensors that get called tensors and that can be intimidating.
I never heard of today’s entry topic three months ago. Indeed, three weeks ago I was still making guesses about just what Gaurish, author of For the love of Mathematics,, was asking about. It turns out to be maybe the grand union of everything that’s ever been in one of my A To Z sequences. I overstate, but barely.
The specific thing that a Young Tableau is is beautiful in its simplicity. It could almost be a recreational mathematics puzzle, except that it isn’t challenging enough.
Start with a couple of boxes laid in a row. As many or as few as you like.
Now set another row of boxes. You can have as many as the first row did, or fewer. You just can’t have more. Set the second row of boxes — well, your choice. Either below the first row, or else above. I’m going to assume you’re going below the first row, and will write my directions accordingly. If you do things the other way you’re following a common enough convention. I’m leaving it on you to figure out what the directions should be, though.
Now add in a third row of boxes, if you like. Again, as many or as few boxes as you like. There can’t be more than there are in the second row. Set it below the second row.
And a fourth row, if you want four rows. Again, no more boxes in it than the third row had. Keep this up until you’ve got tired of adding rows of boxes.
How many boxes do you have? I don’t know. But take the numbers 1, 2, 3, 4, 5, and so on, up to whatever the count of your boxes is. Can you fill in one number for each box? So that the numbers are always increasing as you go left to right in a single row? And as you go top to bottom in a single column? Yes, of course. Go in order: ‘1’ for the first box you laid down, then ‘2’, then ‘3’, and so on, increasing up to the last box in the last row.
Can you do it in another way? Any other order?
Except for the simplest of arrangements, like a single row of four boxes or three rows of one box atop another, the answer is yes. There can be many of them, turns out. Seven boxes, arranged three in the first row, two in the second, one in the third, and one in the fourth, have 35 possible arrangements. It doesn’t take a very big diagram to get an enormous number of possibilities. Could be fun drawing an arbitrary stack of boxes and working out how many arrangements there are, if you have some time in a dull meeting to pass.
Let me step away from filling boxes. In one of its later, disappointing, seasons Futurama finally did a body-swap episode. The gimmick: two bodies could only swap the brains within them one time. So would it be possible to put Bender’s brain back in his original body, if he and Amy (or whoever) had already swapped once? The episode drew minor amusement in mathematics circles, and a lot of amazement in pop-culture circles. The writer, a mathematics major, found a proof that showed it was indeed always possible, even after many pairs of people had swapped bodies. The idea that a theorem was created for a TV show impressed many people who think theorems are rarer and harder to create than they necessarily are.
It was a legitimate theorem, and in a well-developed field of mathematics. It’s about permutation groups. These are the study of the ways you can swap pairs of things. I grant this doesn’t sound like much of a field. There is a surprising lot of interesting things to learn just from studying how stuff can be swapped, though. It’s even of real-world relevance. Most subatomic particles of a kind — electrons, top quarks, gluons, whatever — are identical to every other particle of the same kind. Physics wouldn’t work if they weren’t. What would happen if we swap the electron on the left for the electron on the right, and vice-versa? How would that change our physics?
A chunk of quantum mechanics studies what kinds of swaps of particles would produce an observable change, and what kind of swaps wouldn’t. When the swap doesn’t make a change we can describe this as a symmetric operation. When the swap does make a change, that’s an antisymmetric operation. And — the Young Tableau that’s a single row of two boxes? That matches up well with this symmetric operation. The Young Tableau that’s two rows of a single box each? That matches up with the antisymmetric operation.
How many ways could you set up three boxes, according to the rules of the game? A single row of three boxes, sure. One row of two boxes and a row of one box. Three rows of one box each. How many ways are there to assign the numbers 1, 2, and 3 to those boxes, and satisfy the rules? One way to do the single row of three boxes. Also one way to do the three rows of a single box. There’s two ways to do the one-row-of-two-boxes, one-row-of-one-box case.
What if we have three particles? How could they interact? Well, all three could be symmetric with each other. This matches the first case, the single row of three boxes. All three could be antisymmetric with each other. This matches the three rows of one box. Or you could have two particles that are symmetric with each other and antisymmetric with the third particle. Or two particles that are antisymmetric with each other but symmetric with the third particle. Two ways to do that. Two ways to fill in the one-row-of-two-boxes, one-row-of-one-box case.
This isn’t merely a neat, aesthetically interesting coincidence. I wouldn’t spend so much time on it if it were. There’s a matching here that’s built on something meaningful. The different ways to arrange numbers in a set of boxes like this pair up with a select, interesting set of matrices whose elements are complex-valued numbers. You might wonder who introduced complex-valued numbers, let alone matrices of them, into evidence. Well, who cares? We’ve got them. They do a lot of work for us. So much work they have a common name, the “symmetric group over the complex numbers”. As my leading example suggests, they’re all over the place in quantum mechanics. They’re good to have around in regular physics too, at least in the right neighborhoods.
These Young Tableaus turn up over and over in group theory. They match up with polynomials, because yeah, everything is polynomials. But they turn out to describe polynomial representations of some of the superstar groups out there. Groups with names like the General Linear Group (square matrices), or the Special Linear Group (square matrices with determinant equal to 1), or the Special Unitary Group (that thing where quantum mechanics says there have to be particles whose names are obscure Greek letters with superscripts of up to five + marks). If you’d care for more, here’s a chapter by Dr Frank Porter describing, in part, how you get from Young Tableaus to the obscure baryons.
Porter’s chapter also lets me tie this back to tensors. Tensors have varied ranks, the number of different indicies you can have on the things. What happens when you swap pairs of indices in a tensor? How many ways can you swap them, and what does that do to what the tensor describes? Please tell me you already suspect this is going to match something in Young Tableaus. They do this by way of the symmetries and permutations mentioned above. But they are there.
As I say, three months ago I had no idea these things existed. If I ever ran across them it was from seeing the name at MathWorld’s list of terms that start with ‘Y’. The article shows some nice examples (with each rows a atop the previous one) but doesn’t make clear how much stuff this subject runs through. I can’t fit everything in to a reasonable essay. (For example: the number of ways to arrange, say, 20 boxes into rows meeting these rules is itself a partition problem. Partition problems are probability and statistical mechanics. Statistical mechanics is the flow of heat, and the movement of the stars in a galaxy, and the chemistry of life.) I am delighted by what does fit.
I’ve been reading Elke Stangl’s Elkemental Force blog for years now. Sometimes I even feel social-media-caught-up enough to comment, or at least to like posts. This is relevant today as I discuss one of the Stangl’s suggestions for my letter-V topic.
So sometime in pre-algebra, or early in (high school) algebra, you start drawing equations. It’s a simple trick. Lay down a coordinate system, some set of axes for ‘x’ and ‘y’ and maybe ‘z’ or whatever letters are important. Look to the equation, made up of x’s and y’s and maybe z’s and so. Highlight all the points with coordinates whose values make the equation true. This is the logical basis for saying (eg) that the straight line “is” .
A short while later, you learn about polar coordinates. Instead of using ‘x’ and ‘y’, you have ‘r’ and ‘θ’. ‘r’ is the distance from the center of the universe. ‘θ’ is the angle made with respect to some reference axis. It’s as legitimate a way of describing points in space. Some classrooms even have a part of the blackboard (whiteboard, whatever) with a polar-coordinates “grid” on it. This looks like the lines of a dartboard. And you learn that some shapes are easy to describe in polar coordinates. A circle, centered on the origin, is ‘r = 2’ or something like that. A line through the origin is ‘θ = 1’ or whatever. The line that we’d called before? … That’s … some mess. And now … that’s not even a line. That’s some kind of spiral. Two spirals, really. Kind of wild.
And something to bother you a while. is an equation that looks the same as . You’ve changed the names of the variables, but not how they relate to each other. But one is a straight line and the other a spiral thing. How can that be?
The answer, ultimately, is that the letters in the equations aren’t these content-neutral labels. They carry meaning. ‘x’ and ‘y’ imply looking at space a particular way. ‘r’ and ‘θ’ imply looking at space a different way. A shape has different representations in different coordinate systems. Fair enough. That seems to settle the question.
But if you get to calculus the question comes back. You can integrate over a region of space that’s defined by Cartesian coordinates, x’s and y’s. Or you can integrate over a region that’s defined by polar coordinates, r’s and θ’s. The first time you try this, you find … well, that any region easy to describe in Cartesian coordinates is painful in polar coordinates. And vice-versa. Way too hard. But if you struggle through all that symbol manipulation, you get … different answers. Eventually the calculus teacher has mercy and explains. If you’re integrating in Cartesian coordinates you need to use “dx dy”. If you’re integrating in polar coordinates you need to use “r dr dθ”. If you’ve never taken calculus, never mind what this means. What is important is that “r dr dθ” looks like three things multiplied together, while “dx dy” is two.
We get this explained as a “change of variables”. If we want to go from one set of coordinates to a different one, we have to do something fiddly. The extra ‘r’ in “r dr dθ” is what we get going from Cartesian to polar coordinates. And we get formulas to describe what we should do if we need other kinds of coordinates. It’s some work that introduces us to the Jacobian, which looks like the most tedious possible calculation ever at that time. (In Intro to Differential Equations we learn we were wrong, and the Wronskian is the most tedious possible calculation ever. This is also wrong, but it might as well be true.) We typically move on after this and count ourselves lucky it got no worse than that.
None of this is wrong, even from the perspective of more advanced mathematics. It’s not even misleading, which is a refreshing change. But we can look a little deeper, and get something good from doing so.
The deeper perspective looks at “differential forms”. These are about how to encode information about how your coordinate system represents space. They’re tensors. I don’t blame you for wondering if they would be. A differential form uses interactions between some of the directions in a space. A volume form is a differential form that uses all the directions in a space. And satisfies some other rules too. I’m skipping those because some of the symbols involved I don’t even know how to look up, much less make WordPress present.
What’s important is the volume form carries information compactly. As symbols it tells us that this represents a chunk of space that’s constant no matter what the coordinates look like. This makes it possible to do analysis on how functions work. It also tells us what we would need to do to calculate specific kinds of problem. This makes it possible to describe, for example, how something moving in space would change.
The volume form, and the tools to do anything useful with it, demand a lot of supporting work. You can dodge having to explicitly work with tensors. But you’ll need a lot of tensor-related materials, like wedge products and exterior derivatives and stuff like that. If you’ve never taken freshman calculus don’t worry: the people who have taken freshman calculus never heard of those things either. So what makes this worthwhile?
Yes, person who called out “polynomials”. Good instinct. Polynomials are usually a reason for any mathematics thing. This is one of maybe four exceptions. I have to appeal to my other standard answer: “group theory”. These volume forms match up naturally with groups. There’s not only information about how coordinates describe a space to consider. There’s ways to set up coordinates that tell us things.
That isn’t all. These volume forms can give us new invariants. Invariants are what mathematicians say instead of “conservation laws”. They’re properties whose value for a given problem is constant. This can make it easier to work out how one variable depends on another, or to work out specific values of variables.
For example, classical physics problems like how a bunch of planets orbit a sun often have a “symplectic manifold” that matches the problem. This is a description of how the positions and momentums of all the things in the problem relate. The symplectic manifold has a volume form. That volume is going to be constant as time progresses. That is, there’s this way of representing the positions and speeds of all the planets that does not change, no matter what. It’s much like the conservation of energy or the conservation of angular momentum. And this has practical value. It’s the subject that brought my and Elke Stangl’s blogs into contact, years ago. It also has broader applicability.
There’s no way to provide an exact answer for the movement of, like, the sun and nine-ish planets and a couple major moons and all that. So there’s no known way to answer the question of whether the Earth’s orbit is stable. All the planets are always tugging one another, changing their orbits a little. Could this converge in a weird way suddenly, on geologic timescales? Might the planet might go flying off out of the solar system? It doesn’t seem like the solar system could be all that unstable, or it would have already. But we can’t rule out that some freaky alignment of Jupiter, Saturn, and Halley’s Comet might not tweak the Earth’s orbit just far enough for catastrophe to unfold. Granted there’s nothing we could do about the Earth flying out of the solar system, but it would be nice to know if we face it, we tell ourselves.
But we can answer this numerically. We can set a computer to simulate the movement of the solar system. But there will always be numerical errors. For example, we can’t use the exact value of π in a numerical computation. 3.141592 (and more digits) might be good enough for projecting stuff out a day, a week, a thousand years. But if we’re looking at millions of years? The difference can add up. We can imagine compensating for not having the value of π exactly right. But what about compensating for something we don’t know precisely, like, where Jupiter will be in 16 million years and two months?
Symplectic forms can help us. The volume form represented by this space has to be conserved. So we can rewrite our simulation so that these forms are conserved, by design. This does not mean we avoid making errors. But it means we avoid making certain kinds of errors. We’re more likely to make what we call “phase” errors. We predict Jupiter’s location in 16 million years and two months. Our simulation puts it thirty degrees farther in its circular orbit than it actually would be. This is a less serious mistake to make than putting Jupiter, say, eight-tenths as far from the Sun as it would really be.
Volume forms seem, at first, a lot of mechanism for a small problem. And, unfortunately for students, they are. They’re more trouble than they’re worth for changing Cartesian to polar coordinates, or similar problems. You know, ones that the student already has some feel for. They pay off on more abstract problems. Tracking the movement of a dozen interacting things, say, or describing a space that’s very strangely shaped. Those make the effort to learn about forms worthwhile.
Today’s is technically a request from Elke Stangl, author of the Elkemental Force blog. I think it’s also me setting out my own petard for self-hoisting, as my recollection is that I tossed off a mention of “defining the Ricci Tensor” as the sort of thing that’s got a deep beauty that’s hard to share with people. And that set off the search for where I had written about the Ricci Tensor. I hadn’t, and now look what trouble I’m in. Well, here goes.
You’re not doing that right, by the way. I expect what you’re thinking of is a universe that’s a big block of space that doesn’t happen to have any things clogging it up. Maybe you have a natural sense of volume in it, so that you know something is there. Maybe you even imagine something with grid lines or reticules or some reference points. What I imagine after a command like that is a sort of great rectangular expanse, dark and faintly purple-tinged, with small dots to mark its expanse. That’s fine. This is what I really want. But it’s not really imagining nothing existing. There’s space. There’s some sense of where things would be, if they happened to be in there. We’d have to get rid of the space to have “nothing” exist. And even then we have logical problems that sound like word games. (How can nothing have a property like “existing”? Or a property like “not existing”?) This is dangerous territory. Let’s not step there.
So take the empty space that’s what mathematics and physics people mean by “nothing”. What do we know about it? Unless we’re being difficult, it’s got some extent. There are points in it. There’s some idea of distance between these points. There’s probably more than one dimension of space. There’s probably some sense of time, too. At least we’re used to the expectation that things would change if we watched. It’s a tricky sense to have, though. It’s hard to say exactly what time is. We usually fall back on the idea that we know time has passed if we see something change. But if there isn’t anything to see change? How do we know there’s still time passing?
You maybe already answered. We know time is passing because we can see space changing. One of the legs of Modern Physics is geometry, how space is shaped and how its shape changes. This tells us how gravity works, and how electricity and magnetism propagate. If there were no matter, no energy, no things in the universe there would still be some kind of physics. And interesting physics, since the mathematics describing this stuff is even subtler and more challenging to the intuition than even normal Euclidean space. If you’re going to read a pop mathematics blog like this, you’re very used to this idea.
Probably haven’t looked very hard at the idea, though. How do you tell whether space is changing if there’s nothing in it? It’s all right to imagine a coordinate system put on empty space. Coordinates are our concept. They don’t affect the space any more than the names we give the squirrels in the yard affect their behavior. But how to make the coordinates move with the space? It seems question-begging at least.
We have a mathematical gimmick to resolve this. Of course we do. We call it a name like a “test mass” or a “test charge” or maybe just “test particle”. Imagine that we drop into space a thing. But it’s only barely a thing. It’s tiny in extent. It’s tiny in mass. It’s tiny in charge. It’s tiny in energy. It’s so slight in every possible trait that it can’t sully our nothingness. All it does is let us detect it. It’s a good question how. We have good eyes. But now, we could see the particle moving as the space it’s in moves.
But again we can ask how. Just one point doesn’t seem to tell us much. We need a bunch of test particles, a whole cloud of them. They don’t interact. They don’t carry energy or mass or anything. They just carry the sense of place. This is how we would perceive space changing in time. We can ask questions meaningfully.
Here’s an obvious question: how much volume does our cloud take up? If we’re going to be difficult about this, none at all, since it’s a finite number of particles that all have no extent. But you know what we mean. Draw a ball, or at least an ellipsoid, around the test particles. How big is that? Wait a while. Draw another ball around the now-moved test particles. How big is that now?
Here’s another question: has the cloud rotated any? The test particles, by definition, don’t have mass or anything. So they don’t have angular momentum. They aren’t pulling one another to the side any. If they rotate it’s because space has rotated, and that’s interesting to consider. And another question: might they swap positions? Could a pair of particles that go left-to-right swap so they go right-to-left? That I ask admits that I want to allow the possibility.
These are questions about coordinates. They’re about how one direction shifts to other directions. How it stretches or shrinks. That is to say, these are questions of tensors. Tensors are tools for many things, most of them about how things transmit through different directions. In this context, time is another direction.
All our questions about how space moves we can describe as curvature. How do directions fall away from being perpendicular to one another? From being parallel to themselves? How do their directions change in time? If we have three dimensions in space and one in time — a four-dimensional “manifold” — then there’s 20 different “directions” each with maybe their own curvature to consider. This may seem a lot. Every point on this manifold has this set of twenty numbers describing the curvature of space around it. There’s not much to do but accept that, though. If we could do with fewer numbers we would, but trying cheats us out of physics.
Ten of the numbers in that set are themselves a tensor. It’s known as the Weyl Tensor. It describes gravity’s equivalent to light waves. It’s about how the shape of our cloud will change as it moves. The other ten numbers form another tensor. That is, a thousand words into the essay, the Ricci Tensor. The Ricci Tensor describes how the volume of our cloud will change as the test particles move along. It may seem odd to need ten numbers for this, but that’s what we need. For three-dimensional space and one-dimensional time, anyway. We need fewer for two-dimensional space; more, for more dimensions of space.
The Ricci Tensor is a geometric construct. Most of us come to it, if we do, by way of physics. It’s a useful piece of general relativity. It has uses outside this, though. It appears in the study of Ricci Flows. Here space moves in ways akin to how heat flows. And the Ricci Tensor appears in projective geometry, in the study of what properties of shapes don’t depend on how we present them.
It’s still tricky stuff to get a feeling for. I’m not sure I have a good feel for it myself. There’s a long trail of mathematical symbols leading up to these tensors. The geometry of them becomes more compelling in four or more dimensions, which taxes the imagination. Yann Ollivier here has a paper that attempts to provide visual explanations for many of the curvatures and tensors that are part of the field. It might help.
Gaurish gives me another topic for today. I’m now no longer sure whether Gaurish hopes me to become a topology blogger or a category theory blogger. I have the last laugh, though. I’ve wanted to get better-versed in both fields and there’s nothing like explaining something to learn about it.
So, category theory. It’s a foundational field. It talks about stuff that’s terribly abstract. This means it’s powerful, but it can be hard to think of interesting examples. I’ll try, though.
It starts with categories. These have three parts. The first part is a set of things. (There always is.) The second part is a collection of matches between pairs of things in the set. They’re called morphisms. The third part is a rule that lets us combine two morphisms into a new, third one. That is. Suppose ‘a’, ‘b’, and ‘c’ are things in the set. Then there’s a morphism that matches , and a morphism that matches . And we can combine them into another morphism that matches . So we have a set of things, and a set of things we can do with those things. And the set of things we can do is itself a group.
This describes a lot of stuff. Group theory fits seamlessly into this description. Most of what we do with numbers is a kind of group theory. Vector spaces do too. Most of what we do with analysis has vector spaces underneath it. Topology does too. Most of what we do with geometry is an expression of topology. So you see why category theory is so foundational.
Functors enter our picture when we have two categories. Or more. They’re about the ways we can match up categories. But let’s start with two categories. One of them I’ll name ‘C’, and the other, ‘D’. A functor has to match everything that’s in the set of ‘C’ to something that’s in the set of ‘D’.
And it does more. It has to match every morphism between things in ‘C’ to some other morphism, between corresponding things in ‘D’. It’s got to do it in a way that satisfies that combining, too. That is, suppose that ‘f’ and ‘g’ are morphisms for ‘C’. And that ‘f’ and ‘g’ combine to make ‘h’. Then, the functor has to match ‘f’ and ‘g’ and ‘h’ to some morphisms for ‘D’. The combination of whatever ‘f’ matches to and whatever ‘g’ matches to has to be whatever ‘h’ matches to.
This might sound to you like a homomorphism. If it does, I admire your memory or mathematical prowess. Functors are about matching one thing to another in a way that preserves structure. Structure is the way that sets of things can interact. We naturally look for stuff made up of different things that have the same structure. Yes, functors are themselves a category. That is, you can make a brand-new category whose set of things are the functors between two other categories. This is a good spot to pause while the dizziness passes.
There are two kingdoms of functor. You tell them apart by what they do with the morphisms. Here again I’m going to need my categories ‘C’ and ‘D’. I need a morphism for ‘C’. I’ll call that ‘f’. ‘f’ has to match something in the set of ‘C’ to something in the set of ‘C’. Let me call the first something ‘a’, and the second something ‘b’. That’s all right so far? Thank you.
Let me call my functor ‘F’. ‘F’ matches all the elements in ‘C’ to elements in ‘D’. And it matches all the morphisms on the elements in ‘C’ to morphisms on the elmenets in ‘D’. So if I write ‘F(a)’, what I mean is look at the element ‘a’ in the set for ‘C’. Then look at what element in the set for ‘D’ the functor matches with ‘a’. If I write ‘F(b)’, what I mean is look at the element ‘b’ in the set for ‘C’. Then pick out whatever element in the set for ‘D’ gets matched to ‘b’. If I write ‘F(f)’, what I mean is to look at the morphism ‘f’ between elements in ‘C’. Then pick out whatever morphism between elements in ‘D’ that that gets matched with.
Here’s where I’m going with this. Suppose my morphism ‘f’ matches ‘a’ to ‘b’. Does the functor of that morphism, ‘F(f)’, match ‘F(a)’ to ‘F(b)’? Of course, you say, what else could it do? And the answer is: why couldn’t it match ‘F(b)’ to ‘F(a)’?
No, it doesn’t break everything. Not if you’re consistent about swapping the order of the matchings. The normal everyday order, the one you’d thought couldn’t have an alternative, is a “covariant functor”. The crosswise order, this second thought, is a “contravariant functor”. Covariant and contravariant are distinctions that weave through much of mathematics. They particularly appear through tensors and the geometry they imply. In that introduction they tend to be difficult, even mean, creations, since in regular old Euclidean space they don’t mean anything different. They’re different for non-Euclidean spaces, and that’s important and valuable. The covariant versus contravariant difference is easier to grasp here.
Functors work their way into computer science. The avenue here is in functional programming. That’s a method of programming in which instead of the normal long list of commands, you write a single line of code that holds like fourteen “->” symbols that makes the computer stop and catch fire when it encounters a bug. The advantage is that when you have the code debugged it’s quite speedy and memory-efficient. The disadvantage is if you have to alter the function later, it’s easiest to throw everything out and start from scratch, beginning from vacuum-tube-based computing machines. But it works well while it does. You just have to get the hang of it.
So, tensors. They’re the things mathematicians get into when they figure vectors just aren’t hard enough. Physics majors learn about them too. Electrical engineers really get into them. Some material science types too.
You maybe notice something about those last three groups. They’re interested in subjects that are about space. Like, just, regions of the universe. Material scientists wonder how pressure exerted on something will get transmitted. The structure of what’s in the space matters here. Electrical engineers wonder how electric and magnetic fields send energy in different directions. And physicists — well, everybody who’s ever read a pop science treatment of general relativity knows. There’s something about the shape of space something something gravity something equivalent acceleration.
So this gets us to tensors. Tensors are this mathematical structure. They’re about how stuff that starts in one direction gets transmitted into other directions. You can see how that’s got to have something to do with transmitting pressure through objects. It’s probably not too much work to figure how that’s relevant to energy moving through space. That it has something to do with space as just volume is harder to imagine. But physics types have talked about it quite casually for over a century now. Science fiction writers have been enthusiastic about it almost that long. So it’s kind of like the Roman Empire. It’s an idea we hear about early and often enough we’re never really introduced to it. It’s never a big new idea we’re presented, the way, like, you get specifically told there was (say) a War of 1812. We just soak up a couple bits we overhear about the idea and carry on as best our lives allow.
But to think of space. Start from somewhere. Imagine moving a little bit in one direction. How far have you moved? If you started out in this one direction, did you somehow end up in a different one? Now imagine moving in a different direction. Now how far are you from where you started? How far is your direction from where you might have imagined you’d be? Our intuition is built around a Euclidean space, or one close enough to Euclidean. These directions and distances and combined movements work as they would on a sheet of paper, or in our living room. But there is a difference. Walk a kilometer due east and then one due north and you will not be in exactly the same spot as if you had walked a kilometer due north and then one due east. Tensors are efficient ways to describe those little differences. And they tell us something of the shape of the Earth from knowing these differences. And they do it using much of the form that matrices and vectors do, so they’re not so hard to learn as they might be.
That’s all prelude. Here’s the next piece. We go looking at transformations. We take a perfectly good coordinate system and a point in it. Now let the light of the full Moon shine upon it, so that it shifts to being a coordinate werewolf. Look around you. There’s a tensor that describes how your coordinates look here. What is it?
You might wonder why we care about transformations. What was wrong with the coordinates we started with? But that’s because mathematicians have lumped a lot of stuff into the same name of “transformation”. A transformation might be something as dull as “sliding things over a little bit”. Or “turning things a bit”. It might be “letting a second of time pass”. Or “following the flow of whatever’s moving”. Stuff we’d like to know for physics work.
“General covariance” is a term that comes up when thinking about transformations. Suppose we have a description of some physics problem. By this mostly we mean “something moving in space” or “a bit of light moving in space”. That’s because they’re good building blocks. A lot of what we might want to know can be understood as some mix of those two problems.
Put your description through the same transformation your coordinate system had. This will (most of the time) change the details of how your problem’s represented. But does it change the overall description? Is our old description no longer even meaningful?
I trust at this point you’ve nodded and thought something like “well, that makes sense”. Give it another thought. How could we not have a “generally covariant” description of something? Coordinate systems are our impositions on a problem. We create them to make our lives easier. They’re real things in exactly the same way that lines of longitude and latitude are real. If we increased the number describing the longitude of every point in the world by 14, we wouldn’t change anything real about where stuff was or how to navigate to it. We couldn’t.
Here I admit I’m stumped. I can’t think of a good example of a system that would look good but not be generally covariant. I’m forced to resort to metaphors and analogies that make this essay particularly unsuitable to use for your thesis defense.
So here’s the thing. Longitude is a completely arbitrary thing. Measuring where you are east or west of some prime meridian might be universal, or easy for anyone to tumble onto. But the prime meridian is a cultural choice. It’s changed before. It may change again. Indeed, Geographic Information Services people still work with many different prime meridians. Most of them are for specialized purposes. Stuff like mapping New Jersey in feet north and east of some reference, for which Greenwich would make the numbers too ugly. But if our planet is mapped in an alien’s records, that map has at its center some line almost surely not Greenwich.
But latitude? Latitude is, at least, less arbitrary. That we measure it from zero to ninety degrees, north or south, is a cultural choice. (Or from -90 to 90 degrees. Same thing.) But that there’s a north pole and a south pole? That’s true as long as the planet is rotating. And that’s forced on us. If we tried to describe the Earth as rotating on an axis between Paris and Mexico City, we would … be fighting an uphill struggle, at least. It’s hard to see any problem that might make easier, apart from getting between Paris and Mexico City.
In models of the laws of physics we don’t really care about the north or south pole. A planet might have them or might not. But it has got some privileged stuff that just has to be so. We can’t have stuff that makes the speed of light in a vacuum change. And we have to make sense of a block of space that hasn’t got anything in it, no matter, no light, no energy, no gravity. I think those are the important pieces actually. But I’ll defer, growling angrily, to an expert in general relativity or non-Euclidean coordinates if I’ve misunderstood.
By the way the term “general covariance” comes from two pieces. The “covariance” is because it describes how changes in one coordinate system are reflected in another. It’s “general” because we talk about coordinate transformations without knowing much about them. That is, we’re talking about transformations in general, instead of some specific case that’s easy to work with. This is why the mathematics of this can be frightfully tricky; we don’t know much about the transformations we’re working with. For a parallel, it’s easy to tell someone how to divide 14 into 112. It’s harder to tell them how to divide absolutely any number into absolutely any other number.
Quite a bit of mathematical physics plays into geometry. Gravity physicists mostly see as a problem of geometry. People who like reading up on science take that as given too. But many problems can be understood as a point or a blob of points in some kind of space, and how that point moves or that blob evolves in time. We don’t see “general covariance” in these other fields exactly. But we do see things that resemble it. It’s an idea with considerable reach.
I’m not sure how I feel about this. For most of my essays I’ve kept away from equations, even for the Why Stuff Can Orbit sequence. But this is one of those subjects it’s hard to be exact about without equations. I might revisit this in a special all-symbols, calculus-included, edition. Depends what my schedule looks like.
The true but unenlightening answer first: a tensor is a regular, rectangular grid of numbers. The most common kind is a two-dimensional grid, so that it looks like a matrix, or like the times tables. It might be square, with as many rows as columns, or it might be rectangular.
It can also be one-dimensional, looking like a row or a column of numbers. Or it could be three-dimensional, rows and columns and whole levels of numbers. We don’t try to visualize that. It can be what we call zero-dimensional, in which case it just looks like a solitary number. It might be four- or more-dimensional, although I confess I’ve never heard of anyone who actually writes out such a thing. It’s just so hard to visualize.
You can add and subtract tensors if they’re of compatible sizes. You can also do something like multiplication. And this does mean that tensors of compatible sizes will form a ring. Of course, that doesn’t say why they’re interesting.
Tensors are useful because they can describe spatial relationships efficiently. The word comes from the same Latin root as “tension”, a hint about how we can imagine it. A common use of tensors is in describing the stress in an object. Applying stress in different directions to an object often produces different effects. The classic example there is a newspaper. Rip it in one direction and you get a smooth, clean tear. Rip it perpendicularly and you get a raggedy mess. The stress tensor represents this: it gives some idea of how a force put on the paper will create a tear.
Tensors show up a lot in physics, and so in mathematical physics. Technically they show up everywhere, since vectors and even plain old numbers (scalars, in the lingo) are kinds of tensors, but that’s not what I mean. Tensors can describe efficiently things whose magnitude and direction changes based on where something is and where it’s looking. So they are a great tool to use if one wants to represent stress, or how well magnetic fields pass through objects, or how electrical fields are distorted by the objects they move in. And they describe space, as well: general relativity is built on tensors. The mathematics of a tensor allow one to describe how space is shaped, based on how to measure the distance between two points in space.
My own mathematical education happened to be pretty tensor-light. I never happened to have courses that forced me to get good with them, and I confess to feeling intimidated when a mathematical argument gets deep into tensor mathematics. Joseph C Kolecki, with NASA’s Glenn (Lewis) Research Center, published in 2002 a nice little booklet “An Introduction to Tensors for Students of Physics and Engineering”. This I think nicely bridges some of the gap between mathematical structures like vectors and matrices, that mathematics and physics majors know well, and the kinds of tensors that get called tensors and that can be intimidating.
. 
… that’s not even a line. That’s some kind of spiral. Two spirals, really. Kind of wild. 
, and a morphism that matches 
. And we can combine them into another morphism that matches 
. So we have a set of things, and a set of things we can do with those things. And the set of things we can do is itself a group. 
nebusresearch 
