Space Is a Grid You Build

By the end of this section you will see space itself as something you build, one ruler at a time, and you will know exactly what a single vector can — and cannot — reach. This chapter takes the formal definitions of span, independence, and basis from earlier chapters and rebuilds them as a hands-on act of construction: you hand the grid one vector, watch what it can reach, hand it another, and watch the reach grow. Every idea in the chapters ahead — rank, dependence, change of basis — is really a question about how much grid a set of vectors manages to build.

What You Will Be Able to Do

The Big Picture: One Ruler, One Line

Before a machine — or you — can talk about “position,” it needs a ruler. Give it exactly one: the vector $\hat{\imath} = (1, 0)$. That single arrow is now the only unit of measurement available. Every position you can describe has to be built by scaling it:

Plug in a few numbers and watch what happens. Each choice of $x$ slides the ruler out to a new point on the same line:

No matter which real number you pick for $x$, the result $V$ lands somewhere on the same straight line through the origin — the line $\hat{\imath}$ itself lies on. That entire line, every point of it, is called the span of $\hat{\imath}$: the complete set of places you can reach using only that one ruler and ordinary scaling.

This is the constructive definition of a basis vector: $\hat{\imath}$ is a basis for its line precisely because every point on that line is $x \cdot \hat{\imath}$ for exactly one $x$ — one ruler, one recipe, one reachable line. Nothing off that line is reachable yet. In the next section you will hand the grid a second ruler and watch that restriction break open into an entire plane.

The Refusal: A Second î Buys You Nothing

Before you get a second ruler, it is worth trying the tempting shortcut: what if you just used two copies of the ruler you already have? Ask the grid to add “another î, pointing vertically this time.” The grid accepts the request — and then refuses it.

The refusal is not a rule someone imposed from outside; it falls straight out of what $\hat{\imath}$ already is. The vector $\hat{\imath}$ has one fixed coordinate pair, $(1, 0)$, full stop. You cannot hand it a new direction by asking nicely — a “second î” is still the vector $(1, 0)$, because that is the only vector the name refers to. Scaling that duplicate by any $x$ sweeps out exactly the same line the original already swept.

The refusal, precisely. A dependent vector cannot help you build any new part of the grid. Two copies of $\hat{\imath}$ are one direction, not two — the line stays a line.

See It: Build the Line, Then Try to Break It

The widget below is the same Grid Builder used throughout this chapter, pinned to just the first two stages so you can focus on one ruler and the refusal before the plane opens up. Stage 1 lets you scrub $x$ in $V = x \cdot \hat{\imath}$ and watch the point slide along the line; Stage 2 offers the “Add another î” button and shows you the refusal card.

What to look for: everything off the bright highlighted line sits under a dark veil — the grid is deliberately shading out every point one vector cannot reach yet, so the gap between “one line” and “the whole plane” is visible, not just implied. Watch the rank chip: it reads rank 1 through both stages — a second copy of î never moves it. When you press Add another î, notice the duplicate arrow drawn dashed, lying exactly on top of the original: same line, same reach, nothing new.

Common Misconceptions

“A longer î reaches more”

Why it is tempting: a longer arrow looks like it covers more ground, so it is natural to think stretching $\hat{\imath}$ out to length 10 would reach farther than the unit-length original. Correction: the span of a nonzero vector is already the entire line through it, in both directions, forever — stretching or shrinking the vector you started from changes nothing about which points are reachable, only which scalar $x$ gets you there. Example: the point $(6, 0)$ is reachable from $\hat{\imath}$ with $x = 6$ just as surely as it is reachable from a length-2 ruler $(2, 0)$ with $x = 3$ — same line, same span, different recipe.

“A second copy of î eventually adds a new direction”

Why it is tempting: adding a second vector to a set usually does grow what you can reach, so it is easy to assume any second vector will do the same. Correction: growth only happens when the new vector points somewhere genuinely different. A duplicate of $\hat{\imath}$ is still the vector $(1, 0)$ — it has no freedom to point anywhere else — so it sweeps the identical line no matter how many copies you add. Example: ten copies of $\hat{\imath}$ still only reach the 1-dimensional number line; the rank stays exactly 1.

Summary

Practice Problems

Beginner

B1. Compute $V = x \cdot \hat{\imath}$ for $x = 4$ and for $x = -1.5$. Are both points on the same line?
Hint: plug each x into $(x, 0)$ and compare.

B2. Is the point $(0, 3)$ reachable using only $\hat{\imath}$? Explain using the word “span.”
Hint: what is the second coordinate of $x \cdot \hat{\imath}$ for every choice of $x$?

Intermediate

I1. A classmate suggests adding a second, vertical copy of $\hat{\imath}$ to reach $(0, 3)$. Explain, using the refusal, exactly why this does not work.
Hint: what is the actual coordinate pair of $\hat{\imath}$, no matter how you draw it?

I2. If you scaled $\hat{\imath}$ by every real number $x$, would you eventually cover the point $(2, 0.0001)$? Why or why not?
Hint: that point is off the x-axis by a tiny amount — does $x \cdot \hat{\imath}$ ever leave the axis?

Concept Map

This section built the smallest possible grid — one ruler, one line — and then tried, and failed, to cheat past its limit with a duplicate. That failure is the whole point: a grid only grows when a new vector points somewhere genuinely new.

Next up, §2 — Earning the Plane hands the grid a second ruler that actually points somewhere new, $\hat{\jmath} = (0, 1)$, and watches the reachable set jump from a line to the entire plane.