The integral sign's precise birthdate is unusually well documented for a mathematical glyph. Gottfried Wilhelm Leibniz drew the elongated S in an unpublished manuscript dated 29 October 1675, deliberately abbreviating the Latin word summa—sum—while preparing notation for what he would later publish as the calculus. The stroke was a stylized long s, a typographic form already familiar to seventeenth-century printers but rarely loaded with mathematical content. Leibniz needed an operator that spoke continuity rather than discrete addition, and the long s carried exactly that visual cadence: vertical, unhurried, suggestive of motion through a range rather than terminal closure.
The notation's adoption was neither universal nor immediate. Isaac Newton's parallel work on fluxions used an entirely different symbolic apparatus, and the priority dispute that consumed the next half-century left European mathematicians divided between competing traditions. Continental mathematicians gradually converged on Leibniz's glyph because it composed cleanly with differential notation, leaving the dx beside the integrand without orthographic collision. By the late eighteenth century the long s had crowded out every alternative across continental Europe, and English mathematicians eventually capitulated when the analytical methods proved indispensable.
What no contemporary anticipated was the glyph's elasticity. Leibniz introduced it to sum infinitesimals along a real interval, yet the same stroke now indexes Lebesgue integration over measure spaces, contour integrals across complex domains, and stochastic integrals against Brownian paths. The mark survived because it abstracts well: it does not promise a specific computational recipe so much as a commitment to accumulate against some announced measure.
The integral glyph elongates the Latin lowercase s into a tall, narrow form that the visual cortex parses as continuous motion rather than discrete punctuation. Unlike the summation roof, which the eye reads as architectural closure, the integral's S-curve hints at a stream the reader could follow from bottom to top without lifting attention. That fluid silhouette aligns precisely with the operator's mathematical content: integration is the limit of an aggregation refined past every finite bound.
Compared to ordinary differentiation glyphs, the integral's extreme vertical proportions create what typographers call a display character: a mark too tall to nest comfortably inside body text, demanding its own line and its own breathing room. That spatial assertion is a design feature rather than a flaw. It forces readers to slow down, anchor on the limits of integration, and treat the expression as a substantive computational object rather than an inline aside.
The integral's semantic layers have accreted across three and a half centuries. In the Leibniz-Newton synthesis it indexed an antiderivative—the inverse of differentiation, recoverable through symbolic manipulation. By Bernhard Riemann's 1854 habilitation thesis it had been recast as the limit of partition sums whose mesh refines to zero, providing the first rigorous account of which functions actually admit integration. By Henri Lebesgue's 1902 dissertation it had migrated again, this time onto measure spaces where the underlying notion of length was itself generalized, opening the door to integration of functions Riemann's framework could not handle.
For working analysts, each layer remains active. Calculus students meet the antiderivative reading; engineering programs emphasize Riemann partitions and numerical quadrature; probability theory and functional analysis operate exclusively in Lebesgue terrain. The glyph compresses all three traditions into a single mark, which means context determines which interpretation a particular the integral actually invokes. Reading the operator therefore requires reading the company it keeps.
Contemporary infrastructure renders the integral sign across LaTeX-compiled papers, Jupyter notebooks, computer algebra systems, and the symbolic execution layers of machine-learning frameworks that implement automatic differentiation. Probabilistic programming languages expose the glyph in their notation for expectations and marginal likelihoods, treating integration as a primitive operation that compilers attempt to evaluate symbolically or estimate via Monte Carlo when closed forms refuse to appear. The mark's presence in such pipelines reflects a deeper continuity: every modern statistical inference engine ultimately discharges its outputs into an integral somewhere in the chain, even when the screen never displays the glyph.
Hydrologists measure the volume of water passing a fixed station by integrating instantaneous flow rate over time. The river never asks the gauge to enumerate every drop; the gauge integrates because enumeration would be operationally impossible and analytically uninformative. The integral sign performs the same compression on every quantity that varies continuously: instead of cataloguing every infinitesimal contribution, it certifies that the total has been properly weighted across the relevant range.
Shift to forensic accounting: an auditor reconstructs annual revenue not by inspecting every transaction but by accumulating against a sampling frame chosen to bound estimation error. The integral encodes that disciplined aggregation in mathematical notation, replacing the auditor's sampling plan with a measure-theoretic specification that any qualified reader can audit in turn.
Every integral sign hides a choice about what counts as the relevant measure. Public-health models that integrate mortality over population strata implicitly accept that two deaths weighted differently produce different policy conclusions, a methodological move with profound distributional consequences. Climate-economy models integrate damages across decades using discount rates that compress future suffering into present-value totals, a choice that economists describe as technical and ethicists describe as generational arithmetic disguised as calculus.
The remedy is not to abandon integration but to surface the measure. Sensitivity analyses that report results under alternative weighting schemes, dual-criterion presentations that show both integrated totals and unweighted distributions, transparent documentation of the underlying measure space—these are the professional habits that prevent the glyph from becoming an ethical alibi. The integral sign is innocent of moral commitments; the measure invoked under it carries every value judgment the analyst would prefer not to argue explicitly.
The integral sign rarely appears alone in serious analytical work. It pairs constantly with the partial differential when integration occurs over multidimensional regions, with the n-ary summation when discrete and continuous accumulation must be reconciled, and with the infinity symbol whenever the interval of integration extends without bound. Each pairing teaches a different discipline about composition: the partial reminds the reader which variable carries the differential, the summation indexes families of integrals, the infinity marks where finite limits surrender to asymptotic regimes.
In probability theory the integral attaches to the differential of a measure—dμ, dF, dP—rather than the simple dx of elementary calculus, a notational shift that signals the underlying generalization without changing the glyph itself. That same robustness lets the operator survive across Riemann-Stieltjes, Lebesgue-Stieltjes, and Itô integration, each context loading the mark with a different procedural meaning while preserving its outward shape.
Working analysts distinguish between the integral as a symbolic object, as a numerical estimator, and as an existence claim—three readings that beginners conflate. A symbolic integral may be exactly computable yet practically irrelevant when the integrand has no closed-form antiderivative; a numerical integral may converge rapidly under Gaussian quadrature yet diverge catastrophically under naive trapezoidal rule on the same function. The glyph hides which approach the writer intends, which is why publication standards in computational mathematics increasingly demand explicit declaration of numerical scheme alongside the symbolic expression.
A subtler observation involves swap conditions. Interchanging the order of integration in iterated integrals requires Fubini's theorem; interchanging integration and limit requires dominated or monotone convergence; interchanging integration and differentiation requires uniform regularity. The integral sign never warns the reader which theorem authorizes a given step, leaving the entire burden of justification on accompanying prose. Mathematical referees consequently spend disproportionate effort verifying that swap arguments quoted in proofs actually meet their hypotheses.
Pragmatism reinforced by linguistic anchoring. The long s was already familiar to seventeenth-century printers, ensuring his manuscripts would reproduce reliably without commissioning new type. The character also abbreviated summa—Latin for sum—giving the operator an immediate mnemonic for readers trained in scholastic Latin. Borrowing established typography over inventing fresh notation has historically proven the faster path to adoption: the Bitcoin sign's late entry into Unicode followed roughly the same logic three centuries later. Leibniz traded notational originality for typographic reliability, and the trade paid off across every subsequent printing technology.
In elementary calculus, the integral typically denotes Riemann integration over an interval of the real line, with the differential dx implicitly carrying the standard Lebesgue measure of length. In measure theory, the same mark generalizes to the integral f dμ over an arbitrary measure space, where μ might assign weights radically different from length—counting measure on integers, probability measure on event spaces, Hausdorff measure on fractal sets. The glyph remains identical; the underlying machinery shifts from partition refinement to measurable-function approximation by simple functions, a foundational reformulation that resolves convergence pathologies the Riemann framework cannot handle.
Yes, when the context leaves the integration domain or measure ambiguous. A probabilist reading the integral f dP interprets the operator as expectation against a probability measure; a measure theorist may read the same expression as Lebesgue integration against a non-probability measure; a physicist may interpret it as a path integral over field configurations under a formal measure that mathematicians regard as undefined. The professional habit is to declare the underlying space, the σ-algebra of measurable sets, and the specific measure before any manipulation occurs. Without that disclosure ritual, integral notation becomes a polite fiction in which everyone nods at the same glyph while computing incompatible numbers.
| Nom du symbole | Integral |
| Version Unicode | 1.1 |
| Unicode | U+222B |
| Bloc Unicode | |
| Catégorie générale | Math Symbol (Sm) |
| Code CSS | \222B |
| Code hexadécimal | 0x222B |
| Code HTML | ∫ |
| LaTeX | \int |
| Symbole | ∫ |
| Encodage URL (pourcent UTF-8) | %E2%88%AB |
| Raccourci Discord / Slack | :integral: |
| Nom oral / lecteur d’écran | Integral |
| UTF-8 | E2 88 AB |
| UTF-16 | 222B |
| UTF-32 | 0000222B |
1\documentclass{article}2\usepackage{pifont}3\int4\end{document}Vous pouvez saisir le symbole integral sur la plupart des appareils modernes avec les méthodes suivantes :
Alt + 8747 on the numeric keypad, or insert via Character Map (search "integral").
Edit → Emoji & Symbols, search "integral", or Unicode Hex Input then 222B.
Ctrl + Shift + U, type 222b, then Enter (layout-dependent).
Paste from this page or use third-party math keyboards with Unicode palettes.
Paste from this page or select ∫ from extended math symbol panels.
1span.int::before { content: "\222B"; }1<span>∫</span>La représentation du symbole Integral dans différents langages de programmation se trouve dans le tableau ci-dessous :
| Langage | Représentation |
|---|---|
| JavaScript / TypeScript | '\u222B' or String.fromCodePoint(0x222B) |
| Python | '\N{INTEGRAL}' or chr(8747) |
| Rust | '\u{222B}' |
| C / C++ | UTF-8 source or wchar_t with U+222B |
| Go | string(rune(0x222B)) |
| Ruby | "\u222B" |