ContentsThe library

Matrices as Maps

The Shadow and the Thing

Last timeOne Map After Another

Every map reaches a limited set of outputs and destroys a definite set of inputs, and the two sizes add up to the input dimension, which is the whole bookkeeping of rank.

A map can only produce outputs that are mixtures of its columns, because that is

what multiplication does. If a matrix has three columns, every output it will ever

produce is three numbers worth of mixing applied to those three particular

destinations. Nothing else is reachable, no matter what input you feed in.

That observation sounds modest and is not. It means the set of possible outputs is

fixed by the matrix before any input arrives, and it can be much smaller than the

output space looks. A map from the plane to the plane whose two columns point in

the same direction can only ever produce outputs along that one direction. The

output space is a plane; the reachable set is a line inside it. Half the apparent

freedom was never there.

What can be reached

Call that reachable set the image of the map. It is always flat and it always

contains the origin, because feeding in the zero vector gives zero out. Its size

is the number of genuinely different directions among the columns, which is

usually smaller than the number of columns and never larger.

FIG 1Where the input space goes
Two flat sets attached to one matrix. The reachable set lives on the output side and answers what can be produced. The destroyed set lives on the input side and answers what has been forgotten. Everything in this lesson is about how these two sizes are related.

What is destroyed

Now the other side. Ask which inputs the map sends to zero. There is always at

least one, the zero vector itself, and that case is uninteresting. The question is

whether a non-zero input is sent to zero.

If one is, the map has lost information, and the argument is short. Suppose some

non-zero vector is sent to zero. Take any input at all and add that vector to it.

By additivity the output is unchanged, because the added piece contributed

nothing. So two different inputs produce the same output, and from the output

there is no way to tell which one you started with. The map is not reversible, and

no cleverness in the arithmetic will make it so.

Running the argument the other way is just as clean. If two different inputs land

on the same output, their difference lands on zero. So confusion between inputs

and non-zero vectors sent to zero are the same phenomenon viewed twice.

The lesson stops here

6 more paragraphs to go

You have read the opening. The rest of the argument, the problems that check whether it landed, and the lines worth keeping at the end all come with a plan.

The first lesson of every course in the library reads the whole way through, free, so you can see exactly what the rest of them are.

See the planThe contents

This is the reading half

Starting the course gives you your own copy of it. Every idea on every page has problems standing under it, marked with a reason rather than a tick, and any sentence you do not believe can be opened and argued with. None of that can happen on a page nobody owns.

The contents

The rest of this course

  1. 01Not a Grid, a Verb
  2. 02Each Column Is a Destinationopening only
  3. 03Why the Multiplication Rule Looks Like Thatopening only
  4. 04The Shadow and the Thingyou are here
  5. 05Undo, If You Canopening only
  6. 06The Directions That Only Stretchopening only
  7. 07Every Map Is a Rotation, a Stretch and a Rotationopening only
  8. 08When You Cannot Hit the Targetopening only

Read alongside