AP Calculus AB and BC glossary
Nonremovable Discontinuity
Also called: Non-removable discontinuity
A nonremovable discontinuity is a break that cannot be repaired by redefining the function at that one point. Where f is defined on both sides of c, that happens exactly when the two sided limit at c fails to exist; at an endpoint, the relevant one sided limit is what has to exist.
The repair test is direct. To make continuous at you would define to be , so that limit has to exist first. When the one sided limits disagree you have a jump, when either one runs off to infinity you have an infinite discontinuity, and when the values keep oscillating, as does at , there is no value to assign. None of the three can be repaired. Some texts reserve the name essential discontinuity for the narrower case where a one sided limit itself fails to exist, infinite or oscillating, which excludes jumps.
Compare the case that does work. For the limit at is , so defining closes the hole and the repaired function is continuous. A vertical asymptote offers you nothing to define: no single number sits at the end of an output that grows without bound.
The two sided test assumes there are two sides. Continuity at an endpoint of the domain is one sided, so for on with and for , the two sided limit at does not exist and the break is still removable: resetting the single value to makes continuous on its domain. Apply the criterion at interior points, and the matching one sided limit at the ends.
The mistake
Trying to patch a jump by splitting the difference. Setting to the midpoint of the two one sided values makes the graph pass through the gap, but now neither one sided limit matches , so the function is still discontinuous and the discontinuity is still the same kind.
Appears in: Unit 1: Limits and Continuity