Definition · SequenceA sequence is a function whose domain is the positive integers or a finite initial segment 1, 2, …, N of them. Its outputs are its terms, written a₁, a₂, a₃, …, and aₙ is the n-th term or general term. A sequence that stops is finite; one that does not is infinite.
The subscript is an index, not a multiplication and not a value: a₃ is the third term, whatever number that happens to be. In the sequence 7, 7, 7, … every term is 7 and a₃ = 7; in 10, 20, 30, …, a₃ = 30, not 3.
A sequence is not a set. Order is part of it and repetition is allowed: 1, 0, 1, 0, … has two distinct values and infinitely many terms, and 1, 2, 3 and 3, 2, 1 are different sequences built from the same three numbers.
Example · Three sequences given explicitlyaₙ = 3n - 1 gives 2, 5, 8, 11, 14, … . bₙ = (-1)ⁿ gives -1, 1, -1, 1, … , which is how an alternating sign is usually produced. cₙ = 1/n gives 1, 1/2, 1/3, 1/4, … , whose terms shrink toward 0 without ever reaching it.
Definition · Recursive definitionA recursive definition of a sequence has two parts: one or more initial terms, and a recurrence giving aₙ in terms of earlier terms. Both parts are needed. The recurrence alone does not identify a sequence.
That last sentence is the one to take seriously. The rule aₙ = 2aₙ₋₁ for n ≥ 2 is satisfied by 5, 10, 20, 40, …, by 3, 6, 12, 24, … and by 0, 0, 0, … . It says how to move, not where to start, and a definition missing its initial term is not incomplete in a small way — it names infinitely many different sequences.
Example · The same sequence, both waysa₁ = 5 with aₙ = 2aₙ₋₁ for n ≥ 2 gives 5, 10, 20, 40, 80, … . So does aₙ = 5·2ⁿ⁻¹, because doubling n - 1 times from 5 is what that formula does. “Geometric Sequences and Series” shows this is the general shape of a geometric sequence.
Example · Two recursions worth knowingFibonacci: F₁ = 1, F₂ = 1, Fₙ = Fₙ₋₁ + Fₙ₋₂, giving 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, … . It needs two initial terms because the rule reaches back two. Factorial: 0! = 1 and n! = n·(n - 1)!, giving 1, 1, 2, 6, 24, 120, … . “The Binomial Theorem” uses the second one throughout.
From a recursion to a formula, and the check that is not a proof
There is no general method for turning a recursion into an explicit formula. What there is, at this level, is a procedure: compute several terms, look for a pattern, write a candidate formula, and test it on terms you did not use to build it.
Testing is not proving. The sequence 1, 2, 4, 8, 16, … looks settled after five terms, and here is a sequence that starts that way and then does not continue: mark n points on a circle, join every pair by a chord with no three chords meeting at a point, and count the regions the disc is cut into. For n = 1 through 5 the counts are 1, 2, 4, 8, 16. For n = 6 the count is 31.
So a formula that matches the first five terms of a recursion is a candidate, not a fact. Establishing that it matches for every n requires mathematical induction, which this path does not cover: it is the subject of the Induction and Recursion course on the Discrete Mathematics path. Within this course, say "this formula agrees with the first six terms" and mean exactly that.