Reviewed learning material · Video analysis · EnglishRead the full overview
This video demonstrates how to prove that the limit of the sequence an=(−1)n does not exist by using the contrapositive of the Cauchy Convergence Principle. It defines the negative condition (∃ε>0,∀N,…), selects ε=1, and uses visual examples with specific indices (N=3,6,…) to show that adjacent terms always maintain a distance of at least 2. Finally, it generalizes this finding to conclude that since the sequence is not a Cauchy sequence, its limit cannot exist.
Use the learning inspector for key ideas and moments, or open the reading tabs for the complete notes.
Generated from the video's visuals and explanation; not verbatim speech.
The video aims to disprove the existence of a limit for the sequence an=(−1)n using the reverse proposition of the Cauchy Convergence Principle. The screen displays the formal definition: there exists an ε>0 such that for any N, we can find m,n>N where the distance between terms ∣an−am∣≥ε. A coordinate system plots the sequence points oscillating strictly between y=1 and y=−1.
To satisfy the condition, we fix ε=1. The animation tests various thresholds for N. For instance, if N=3, choosing subsequent indices m=4 and n=5 yields values 1 and −1, resulting in a difference of magnitude 2, which is clearly greater than or equal to our chosen epsilon. This pattern repeats for larger N values like 6,11, and 15. Regardless of how far out we go, picking consecutive integers ensures one term is positive and the other negative, maintaining a constant separation of 2.
Synthesizing these observations into a rigorous proof: For ANY arbitrary integer N, let us constructively choose m=N+1 and n=N+2. Since these are consecutive integers, their parity differs, meaning am and an will have opposite signs (1 and −1). Consequently, ∣an−am∣=2. Because 2≥1 (our selected ε), the negation of the Cauchy criterion holds true. Therefore, (an) fails to be a Cauchy sequence. By the completeness of real numbers, a non-Cauchy sequence implies divergence, proving the limit does not exist.
Knowledge cards
01
Negating the Cauchy condition
While the standard theorem states convergence iff Cauchy, the logical inverse allows us to test for divergence. We look for a 'barrier' distance ε that prevents tail elements from clustering together indefinitely. If such an ε exists alongside infinite pairs violating it, the sequence diverges.
∃ε>0,∀N,∃m,n>N:∣an−am∣≥ε
02
Oscillatory Behavior
The sequence (−1)n represents pure oscillation without decay. Unlike convergent sequences where amplitude shrinks towards zero, here the gap remains maximal forever. Geometrically, the dots form two parallel lines separated by distance 2, making them impossible to fit inside an arbitrarily small interval eventually.
an=(−1)n
03
Choosing Epsilon Strategically
Selecting ε=1 is crucial because it sits below the maximum possible jump size of 2 but above 0. Any value less than 2 would technically work, but 1 provides a clear margin. Setting up this threshold simplifies the arithmetic verification later on.
ε=1
04
Index Construction Technique
Instead of searching randomly, the proof relies on structural properties. Picking neighbors N+1 and N+2 exploits the alternating nature of powers of -1. One index must be odd and the other even, guaranteeing distinct function outputs regardless of where they appear on the number line.
∣aN+2−aN+1∣=2≥1
05
Logical Implication Chain
Failure of the Cauchy property directly maps to failure of convergence in complete metric spaces like R. Showing the sequence isn't Cauchy serves as sufficient evidence that no finite limit point attracts all sufficiently large terms simultaneously.
The reviewed neighboring-term card constructs indices n=N+1 and m=N+2 for every cutoff N. For an=(−1)n, their distance is always 2, so choosing ε=1 disproves the Cauchy condition and hence convergence. The argument excludes every possible real limit, rather than just one guessed value.