Limit (mathematics)

• 1. Introduction

  ›

  A limit, in mathematics, is the value that a function or a sequence approaches as its argument or index approaches some value. The idea sounds almost too simple to matter, yet it sits beneath nearly all of calculus and mathematical analysis. Continuity, derivatives, integrals: each of these is defined through limits. Hankel, writing in 1871, traced the modern concept all the way back to Proposition X.1 of Euclid's Elements. That proposition underpins the Method of exhaustion found in both Euclid and Archimedes. So how did a value that nothing in a sequence can ever quite reach become the foundation of an entire branch of mathematics? Who taught us to make a quantity arbitrarily close to a target without ever arriving? And what does it mean to chase infinity itself? Those questions guide everything that follows.

• 2. The Long Road To Rigor

  ›

  Grégoire de Saint-Vincent gave the first definition of a limit in his 1647 work Opus Geometricum, calling it the terminus of a geometric series. He described it as "the end of the series, which none progression can reach, even not if she is continued in infinity, but which she can approach nearer than a given segment." In the Scholium to his Principia of 1687, Isaac Newton wrote of ultimate ratios as "limits... which they can approach so closely that their difference is less than any given quantity." Bruce Pourciau argues that Newton understood limits more deeply than he is usually credited, and even supplied the first epsilon argument. Bernard Bolzano, in 1817, developed the basics of the epsilon-delta technique to define continuous functions. His work stayed unknown to other mathematicians until thirty years after his death. Augustin-Louis Cauchy formalized the limit of a function in 1821, and Karl Weierstrass followed; together their work produced what became the (ε, δ)-definition of limit. The familiar notation, with an arrow placed below the limit symbol, was invented by John Gaston Leathem in 1905. G. H. Hardy popularized it in his 1908 textbook A Course of Pure Mathematics.

• 3. When A Sequence Settles

  ›

  A sequence of real numbers converges to a limit when its terms give an arbitrarily good approximation of a number, once finitely many initial terms are discarded. Pick any error window you like, however small, and only finitely many terms will fall outside it. More precisely, there must exist a value such that for each positive real number, some positive integer marks the point beyond which every term lies within that distance. When such a value exists, it is unique, and it is called the limit. If no such value exists, the sequence is divergent. A sequence can instead tend to infinity, which happens when all but finitely many terms exceed any given bound; the positive integers do exactly this. A sequence can also tend to negative infinity when all but finitely many terms fall below any chosen lower bound. In both cases the sequence is said to have an infinite limit, yet it is not convergent, because positive and negative infinity are not real numbers.

• 4. Limits In Abstract Spaces

  ›

  A metric space carries a distance function, and a sequence in it converges to an element when, for any chosen threshold, the terms eventually sit within that distance of the limit. The space of n-dimensional real vectors offers a concrete case, where the Euclidean distance serves as a suitable measure between points. Topological spaces push the idea to its most abstract form, replacing distances with open neighborhoods. There the limit, if it exists, need not be unique; uniqueness is guaranteed only when the space is Hausdorff. Sequences of functions bring their own subtleties, and functional analysis exists partly to identify useful notions of convergence on function spaces. Pointwise convergence can misbehave: one can build a sequence of continuous functions whose pointwise limit is discontinuous. Uniform convergence, measured by the maximum difference between functions across all arguments, behaves better, because the uniform limit of continuous functions is itself continuous. Lp spaces and Sobolev spaces stand as prominent examples of function spaces carrying their own notions of convergence.

• 5. Approaching A Point

  ›

  Suppose f is a real-valued function and c is a real number; the limit of f of x as x approaches c is the value L that f can be made as close to as desired by making x sufficiently close to c. Formally, given any error threshold, there must be a margin within which every qualifying x forces f close to L, which is the basis of the (ε, δ)-definition for functions. An equivalent definition runs through sequences: for every sequence in the domain of f, the image sequence carries the same limiting behavior. One-sided limits split this picture in two, distinguishing a left-handed limit from below from a right-handed limit from above. The positive indicator function shows why this matters: at zero it has a left-handed limit of 0 and a right-handed limit of 1, so its limit there does not exist. Limits can also reach toward infinity, both in the argument and in the value of the function. Tending to infinity in the argument can be treated as a reciprocal tending to zero, and the infinity is normally signed.

• 6. What Limits Build

  ›

  An infinite series is formalized as the limit of a sequence of partial sums; if that limit exists, it gives the value of the sum, and otherwise the series is divergent. The Basel problem stands as a classic example of such a sum. Series carry a complication that sequences do not, because their convergence can depend on the ordering of terms. A series that converges under every ordering is called unconditionally convergent, which proves equivalent to absolute convergence; otherwise it is conditionally convergent. The Riemann series theorem delivers a startling consequence: by reordering a conditionally convergent series, the partial sums can be steered to any real number at all. Power series extend this machinery, often treating the variable as a complex number, with the set of converging values forming a circle whose radius is the radius of convergence. Limits also define continuity at a point, since a function is continuous at c when its limit there equals its value. The derivative, too, is defined formally as a limit, capturing the rate of change as the increment shrinks toward zero.

• 7. Cauchy And Completeness

  ›

  Convergent sequences of real numbers are always Cauchy sequences, meaning that for any arbitrarily small error, the terms eventually all lie within an interval of that diameter. The appeal of this property is that it depends on the sequence alone, never on the limit. For real numbers the two ideas coincide, because every Cauchy sequence converges. In general metric spaces, convergent sequences remain Cauchy, but the converse fails. The rational numbers furnish the classic counterexample: decimal approximations to an irrational value, truncated at each decimal place, form a Cauchy sequence that has no rational limit. A space where every Cauchy sequence does converge is called a complete metric space. Beyond whether a sequence converges at all, the order of convergence measures how fast it approaches its limit, with the governing constant known as the asymptotic error constant. This quantity matters in numerical analysis, particularly in error analysis. Some limits resist computation entirely; there exist limit expressions whose modulus of convergence is undecidable, and the limit lemma in recursion theory shows undecidable problems can be encoded using limits. Convergence tests such as the ratio test and the squeeze theorem can confirm a limit exists, even when they cannot reveal its value.