Span — definition: Visual walkthrough — Maths deep dive

We build every symbol before we use it. Nothing is assumed.

Step 1 — One arrow, and what "scaling" means

WHAT. We draw a single arrow. Call it $v_{1}$ . An arrow is just an instruction: "go this far in this direction, starting from a home point we call the origin $O$ ."

WHY start here. The whole idea of span is reaching places with your arrows. Before we combine arrows, we must be crystal-clear on the one move that acts on a single arrow: scaling.

To scale means: pick a plain number — call it $c$ — and multiply the arrow by it. Written $c v_{1}$ .

$c v_{1} c = stretch factor a real number, v_{1} = fixed direction the arrow$

$c = 2$ makes the arrow twice as long, same way.
$c = \frac{1}{2}$ makes it half as long.
$c = - 1$ flips it to point the exact opposite way.
$c = 0$ shrinks it to nothing — you land on the origin $O$ .

PICTURE. The dashed line is the set of all the tips of $c v_{1}$ as $c$ sweeps through every real number.

Step 2 — The span of ONE arrow is a whole line

WHAT. Let $c$ run over every real number, and mark where the tip of $c v_{1}$ lands each time. Collect all those landing spots into one set. That set is $span {v_{1}}$ .

$span {v_{1}} = {a landing spot c v_{1} for every real c c \in R}$

Read the vertical bar " $∣$ " as "such that". The whole line reads: "the set of all points $c v_{1}$ such that $c$ is any real number."

WHY it's a full line and not a ray. Because negatives are allowed. If we only used $c \geq 0$ we'd get a half-line (a ray) shooting one way. Allowing $c < 0$ extends it backwards through $O$ . So the span of one nonzero arrow is a complete line through the origin.

PICTURE. Same arrow, but now the entire line is drawn solid — that solid line is the span.

Step 3 — Bring in a second arrow, and what "adding" means

WHAT. Now add a second arrow $v_{2}$ pointing a genuinely different way. The new move is vector addition, written $u + w$ .

To add two arrows: lay the tail of the second at the tip of the first ("tip-to-tail"). The combined arrow runs from the original start $O$ to the final tip.

$u + w u = walk this first arrow, w = then walk this second arrow$

WHY this move. A vector space hands you exactly two tools: scaling (Step 1) and adding (this step). Nothing else. Span will be everything these two tools can produce — so we must see addition geometrically before combining it with scaling.

PICTURE. The tip-to-tail construction; the blue diagonal is $v_{1} + v_{2}$ .

Step 4 — Scale BOTH, then add: the linear combination

WHAT. Combine the two moves. Scale $v_{1}$ by some number $c_{1}$ , scale $v_{2}$ by some number $c_{2}$ , then add the results. This single expression is the linear combination:

$scale v_{1} c_{1} v_{1} + scale v_{2} c_{2} v_{2}$

Every symbol here is already earned: $c_{1}, c_{2}$ are scalars (Step 1), the products $c_{1} v_{1}$ and $c_{2} v_{2}$ are scalings, and the " $+$ " is tip-to-tail addition (Step 3).

WHY it's the master move. Any sequence of stretching, flipping, and adding two arrows, no matter how long, always simplifies down to this one shape $c_{1} v_{1} + c_{2} v_{2}$ . It is the most general thing you can build. See Linear combination.

PICTURE. One example: $c_{1} = 1.5$ , $c_{2} = 1$ . Watch the two scaled arrows form the parallelogram, and the pink arrow lands at the combination.

Step 5 — Sweep the dials: two arrows fill a PLANE

WHAT. Now turn both dials $c_{1}$ and $c_{2}$ over every real value at once and mark every landing spot. That entire collection is the span:

$span {v_{1}, v_{2}} = {c_{1} v_{1} + c_{2} v_{2} c_{1}, c_{2} \in R}$

WHY it becomes a plane. With one dial (Step 2) you got a line. A second, independent dial lets you move sideways off that line too. Two independent directions of freedom = a full flat sheet, i.e. a plane through the origin.

PICTURE. A grid of reachable points as $c_{1}, c_{2}$ vary — they tile out a whole plane.

Step 6 — The degenerate case: two arrows that secretly agree

WHAT. Suppose the second arrow is just a stretch of the first, e.g. $v_{2} = 2 v_{1}$ . Feed this into the combination:

$c_{1} v_{1} + c_{2} (2 v_{1}) = just one new number (c_{1} + 2 c_{2}) v_{1}$

Two dials collapse into one effective dial. So no matter how you turn $c_{1}$ and $c_{2}$ , you never leave the original line.

WHY it matters. Having two arrows does not guarantee a plane. They must point in genuinely different directions — the property called linear independence. If one arrow is already inside the span of the other, it adds no new territory. See Linear independence.

PICTURE. Both arrows lie along the same dashed line; the span stays a line, not a plane.

Step 7 — The always-true anchor: the origin is in every span

WHAT. Set every dial to zero: $c_{1} = c_{2} = \dots = 0$ . Then

$0 \cdot v_{1} + 0 \cdot v_{2} + \dots + 0 \cdot v_{k} = 0$

where $0$ is the zero vector — the arrow of length zero sitting at the origin $O$ .

WHY it forces "through the origin." Because the all-zero choice is always a legal combination, the origin is a member of every span. That's why a span is never a floating line or a tilted plane that misses $O$ — it must pass through it. A span is a point, a line, a plane, or a higher flat, but always through the origin. This is exactly the anchoring property of a Subspace.

PICTURE. All the pictures so far, showing every span pinned to the shared origin dot.

The one-picture summary

The whole story compressed: one dial → line; two independent dials → plane; both dials zero → the origin; a dependent second arrow → back to a line. Every case shares the same origin dot.

Recall Feynman: retell the whole walkthrough plainly

I gave you a magic arrow (Step 1). You can make it longer, shorter, or point it backwards — that's scaling. Marking every place its tip can reach draws a straight line, and because "backwards" is allowed the line goes both ways through home base (Step 2). Then I gave you a second arrow and taught you to add: walk the first, then walk the second from where you stopped (Step 3). Turning a knob on each arrow before you add gives a linear combination — the master move (Step 4). Spin both knobs over every number and you paint a whole flat sheet, a plane (Step 5). But watch out: if the second arrow is secretly just a longer copy of the first, the two knobs behave like one and you're stuck on a line again (Step 6). And no matter what, turning both knobs to zero leaves you standing at home — so every span touches the origin (Step 7). That's the span: everywhere scale-flip-add can carry you, always starting from home.

Active recall

Span of one nonzero arrow, geometrically?

A line through the origin.

Two independent arrows in the plane span what?

The whole plane (all of

R^{2}

Why is the origin in every span?

Choosing all scalars zero gives the zero vector, always a legal combination.

v_{2} = 2 v_{1}

, what is

span {v_{1}, v_{2}}

Just the line through

v_{1}

— the second arrow adds nothing.

When does a new vector enlarge a span?

Only when it is not already inside the current span.

What two operations generate a span?

Scalar multiplication and vector addition.

Connections

Linear combination — the master move built in Steps 3–4.
Linear independence — decides line vs. plane in Step 6.
Subspace — the "flat through the origin" structure Step 7 guarantees.
Basis — the smallest set of arrows whose span is everything.
Dimension — number of independent dials = 1 for a line, 2 for a plane.
Column space — the span of a matrix's columns.
Determinant — its nonzero-ness tests whether two arrows fill $R^{2}$ .