The Algebra of Limits.
Limits don't break when you do arithmetic to them. Add, multiply, divide, compose, and the limit follows along, almost always. Almost.
§ 1 · Why this matters
Limits behave like numbers, until they don't.
Most of the limits you'll need to compute are not exotic. They're built out of simpler pieces: a sum here, a product there, a function plugged into another function. The good news, and the content of this chapter, is that the limit operator distributes through that structure.
If you know the parts, you know the whole. Take the limit of each ingredient, then assemble. That principle is what makes limits computable instead of mysterious, and it's what every later result, from derivatives to integrals, quietly leans on.
§ 2 · The five limit laws
The same two functions, used five different ways.
Below are the canonical functions we'll keep coming back to: f with $\lim_{x \ o 2} f(x) = 3$, and g with $\lim_{x \ o 2} g(x) = 1$. Every law just says: the limit of an expression built from f and g equals the same expression built from L and K.
Why does this work?
Think of the limit as a commitment to a value. If f has committed to L and g has committed to K, then the running product f(x) · g(x) can do nothing but commit to L · K. The arithmetic on the inputs becomes the same arithmetic on their destinations.
§ 3 · A worked example
Substitute. Simplify. Done.
The five laws together justify direct substitution for any rational function at a point where the denominator is nonzero. Here's a typical calculation, end to end.
The first equality is the Quotient Law. The numerator and denominator are polynomials, and limits of polynomials are computed by substitution, that's the Sum and Product laws applied repeatedly. The denominator's limit is 4, which is nonzero, so the Quotient Law is in play and we're allowed to write what we wrote.
§ 4 · Composition, the hand-off
Two machines in series.
Composition asks something different. We have $f \circ g$: feed x into g, take the output, feed it into f. As x approaches c, what does the chained output approach?
provided f is continuous at K.
That clause, “f is continuous at K,” is the whole game. Without it, the second machine can betray you, even if both f and g have perfectly fine limits in isolation.
The hand-off doesn't care which side you came from.
A common worry: if x → c from the left, but g is decreasing, then g(x) approaches K from the right. Does that break the chain? It doesn't, f's left and right limits both equal L when it's continuous, so a directional flip on the inside is harmless.
Try it. Drag x, then toggle the direction of g:
Both sides, both planes.
The diagram below makes the whole symmetry visible at once. The moss-colored stream tracks x → 2⁻; the slate stream tracks x → 2⁺. Both streams transit through the bend, both arrive at L.
§ 5 · When the chain breaks
Three cases, only one of which fails.
The continuity hypothesis isn't decoration; it's load-bearing. Here are three scenarios that look almost identical but behave very differently.
See it for yourself.
The same three cases, now interactive. Switch between them and drag x. Watch the readout: in cases A and B the output settles, in case C it refuses to.
§ 6 · Summary
What to remember.
Limit laws
For arithmetic combinations, sum, difference, constant multiple, product, quotient (denominator nonzero), power, root, the limit of the combination equals the combination of the limits. This justifies direct substitution for any rational function at a point where the denominator doesn't vanish.
Composition
For $f \circ g$, the limit hand-off works provided f is continuous at K = $\lim g$. If f isn't continuous there, the composition's limit may still exist, but only if g's outputs avoid the discontinuity near c.
The mental model
A limit is a commitment to a value. Arithmetic on commitments produces the obvious commitment. Composition transmits commitments through machines, as long as the receiving machine doesn't have a hole at the point being delivered.