Concept

Magic Square

A grid of distinct numbers whose every row, column, and main diagonal sums to the same constant — a mathematical curiosity that Renaissance magic read as a planetary seal.

A magic square is a square grid of numbers arranged so that every row, every column, and both main diagonals add to one and the same total — the so-called magic constant. For a normal square of order n, filled with the whole numbers from one to n², that constant is fixed by the size alone: it is n(n²+1)/2, which for the smallest interesting case comes out to fifteen. That smallest case fills a three-by-three grid with the digits one through nine so that every line sums to fifteen, and it is essentially unique — there is only one such arrangement, appearing in eight guises once its rotations and mirror images are counted as the same figure. The next size up, the four-by-four with constant thirty-four, breaks open into eight hundred and eighty distinct solutions, a count first established exhaustively by the French mathematician Bernard Frénicle de Bessy (d. 1674) in his Des quarrez ou tables magiques, published posthumously in 1693. By order five the number runs into the hundreds of millions. The construction and enumeration of such squares is a genuine branch of combinatorial and recreational mathematics, wholly independent of any use ever made of them — the arithmetic would hold whether or not a single talisman had ever been cut.

The oldest grid: the Luoshu

The object is very old and travels widely. The earliest known example is the Chinese Luoshu — the “Luo River Writing” — the three-by-three arrangement of one through nine that Chinese cosmology ties to markings borne on the shell of a turtle that rose from the waters of the Luo for Yu the Great, the sage-king who tamed the floods. In that tradition the figure is not a curiosity but a chart of the cosmos in miniature: its nine cells map the eight directions around a center, each number weighted as a count of yin or yang dots, the odd (heavenly) and even (earthly) numbers alternating around the rim. It belongs with the Hetu, the companion “Yellow River Chart,” in the diagrammatic cosmology that grew up around the I Ching and the Five Phases. The historical record draws a careful line here: the dot-diagrams in the precise forms transmitted today are of Song-dynasty provenance, entering the divinatory tradition through the Daoist adept Chen Tuan and the early Neo-Confucians and reaching their canonical shape with Zhu Xi (1130–1200). Their claimed descent from high antiquity was the tradition’s own self-understanding, and the tradition itself eventually corrected it: Qing-dynasty evidential scholars, above all Hu Wei in his Yitu mingbian (prefaced 1706), traced the diagrams to Daoist and Song sources rather than to the turtle of legend. The numerical figure of the three-by-three, however, is genuinely ancient in China and the recognized ancestor of every later square; the order-three square of the Arabic mathematicians is simply the Luoshu turned a quarter-turn.

Building the higher orders: Arabic and Indian mathematics

Squares of order higher than three are the work of medieval Islamic mathematics. The polymath Thābit ibn Qurra (836–901) — a native of Harran, the city whose star-temple religion gave the medieval world a line of astronomer-translators — wrote on them in the ninth century. The first known squares of order five and six appear around 983 in the Rasāʾil Ikhwān al-Ṣafāʾ, the encyclopedia of the Brethren of Purity compiled at Baghdad, where the figures are presented frankly as small models of a harmonious universe. From the start the Arabic tradition ran on two tracks at once: a sober combinatorial science of construction, and a magical one. The square of three appears in the eutocic literature — squares set above a laboring woman to ease a delivery — in terms close enough to Indian sources that scholars take that medical-protective use to mark the very origin of the magical square. The talismanic line culminates in the Shams al-Maʿārif, the great book of letter- and number-magic ascribed to the North African Sufi Aḥmad al-Būnī (died 1225), who described a bordering method for enlarging squares; the attribution of that encyclopedia to al-Būnī himself is now treated as uncertain. This is the same matrix of learning surveyed under Islamic astrology, alchemy, and astral magic, the world in which a philosopher like al-Kindī argued that the rays of the stars bind every part of the cosmos into a single field of influence — a doctrine that gives a talismanic figure something to capture.

In India the square is at least as old and as well worked. A four-by-four of constant thirty-four is carved into the Parshvanath Jain temple at Khajuraho, the Chautīsā Yantra, dated to the tenth century — and not a plain magic square but a “most-perfect” one, in which the broken diagonals and every two-by-two block also total thirty-four. The fullest medieval mathematical treatment is Nārāyaṇa Paṇḍita’s Gaṇitakaumudī of 1356, whose closing chapter on bhadragaṇita, the “mathematics of auspicious figures,” lays out sixty rules and seventeen worked examples for constructing squares of every parity, along with magic triangles, rectangles, and circles. The figure passed westward through this Islamic and Indian learning, and entered Europe by way of Byzantium: around 1300 the Constantinopolitan scholar Manuel Moschopoulos, a grammarian of the Palaiologan revival, wrote the first systematic Western treatise on their construction — sorting his methods by whether the order is odd, singly even, or doubly even, and pointedly leaving the mysticism of his sources aside, treating the squares as nothing but numbers in a grid. The thread of Greek-East learning that carried so much else into the Latin West carried this too.

Dürer’s square and the Renaissance turn

By the high Renaissance the magic square stood at the exact seam where serious mathematics met learned magic, and a single image fixed it there. In 1514 Albrecht Dürer set a four-by-four square into the upper right of his engraving Melencolia I, behind the brooding winged figure surrounded by the unused instruments of geometry and craft. It is a genuine and elegant square — every row, column, diagonal, the four central cells, and each corner block sum to thirty-four — and Dürer worked the date into it: the two central cells of the bottom row read fifteen and fourteen, 1514, the year of the engraving and of his mother’s death, flanked by cells that on one reading spell his own initials. The square belongs to the engraving’s whole apparatus of measured order set against the melancholy temperament, the saturnine cast of mind that the Renaissance bound to genius and to the slow, cold planet. That a four-by-four sat at Dürer’s elbow is itself telling: in the magical scheme then taking shape, the four-by-four was the table of Jupiter — the benefic whose influence was thought to lift and counterbalance the heaviness of Saturn.

The planetary seals: Agrippa’s seven squares

That magical use was systematized in the early sixteenth century. Heinrich Cornelius Agrippa, in the De occulta philosophia libri tres (completed in draft by 1510, printed in full at Cologne in 1533), gathered the scattered currents of Renaissance learned magic into a single three-tiered cosmos, and in its second book — the book on number, harmony, and the heavens — set out seven squares of ascending order and assigned each to one of the seven classical planets, in the descending planetary order of the ancients. Saturn takes the three-by-three (constant fifteen); Jupiter the four-by-four (thirty-four); Mars the five-by-five (sixty-five); the Sun the six-by-six (one hundred eleven); Venus the seven-by-seven (one hundred seventy-five); Mercury the eight-by-eight (two hundred sixty); and the Moon the nine-by-nine (three hundred sixty-nine). The Hebrew term for these planetary tables is kamea — a word borrowed from the practice of Jewish protective amulets, and the point at which the squares touch the number-mysticism of the Kabbalah. Agrippa’s seven tables fold the abstract grid into a dense web of correspondence: each planet carries a numerical seal traced as a path through its square, and the names of its governing intelligence and its spirit, written in Hebrew and resolved into figures by gematria, are mapped onto the same grid to yield their sigils. The architecture is Pythagorean to its root — the conviction, descending from Pythagoreanism, that number is the hidden armature of reality, so that a perfectly balanced numerical figure is a participation in cosmic order rather than a mere notation of it.

The operative claim built on this is straightforward to state and bounded here to its architecture. Practitioners of astral and ceremonial magic held that a square, engraved on the metal proper to its planet and made at an hour when that planet was strong, would draw down and concentrate the planet’s virtue — so that the figure became a talisman, a made object charged toward a specific end: Jupiter’s square cut in tin for increase and favor, Saturn’s in lead for endurance and the binding of harm. This is the same logic of capture treated under astral-talismanic and Renaissance astro-magic: the heavens pour a continuous influence downward, and a rightly made figure is a vessel shaped to receive a particular stream of it. The claim is the practitioners’; the squares themselves are only arithmetic, true in any cosmos. What the magical tradition adds is not a different mathematics but a different reading of the same one — the intuition that a configuration returning the same answer however it is read is a visible token of an order that runs deeper than the figure.

A word square apart: the Sator

A different figure is often filed alongside these, though it is not numerical at all. The Sator square is a five-by-five grid of five Latin words — SATOR · AREPO · TENET · OPERA · ROTAS — that reads the same across, down, backward, and bottom-to-top: a palindrome built into two dimensions rather than a sum. It is the oldest of the protective squares of the West by a wide margin. Examples survive scratched into the walls of Pompeii — one fragmentary, found in the 1920s, and one complete, cut into a column of the great exercise-yard, the Palaestra — and so they predate the eruption of Vesuvius in 79 CE, the complete one very likely earlier still. The square then recurs for more than a thousand years across the Latin and Christian world, scratched on church walls, carried as a charm against fire and madness and the bites of mad dogs, set into amulets from Britain to Cappadocia. Its meaning is genuinely unsettled. The central word tenet forms a cross of its own through the middle, and the twenty-five letters can be rearranged into the words PATER NOSTER — the opening of the Lord’s Prayer — twice over in the form of a Greek cross, with a spare A and O left to read as the alpha and omega; that cryptogram is the strongest argument for a Christian hand. Against it stands the awkward fact that arepo is otherwise unattested in Latin, making any clean translation strained, and that the Pompeii dating leaves uncomfortably little room for Christians in the town. The proposed readings run from a flat agricultural sentence about a sower and his wheels to elaborate Mithraic, Jewish, and Christian reconstructions, and none of them has carried the field.

The mathematics and its study {#research}

The magic square is one of the rare objects that has been continuously alive in both the mathematical and the magical literature, and the two strands can be traced through their primary texts and the modern scholarship on them. The combinatorial heart of the subject — the magic constant n(n²+1)/2, the eight hundred and eighty squares of order four, the enumeration that defeated hand calculation above order five — is mapped in the St Andrews history of mathematics, whose survey of the figure from the Luoshu through the Islamic and Indian mathematicians to the European recreationalists is freely available at mathshistory.st-andrews.ac.uk/HistTopics/Magic_squares/. The integer sequence of distinct normal squares by order — 1, 0, 1, 880, … — is cataloged in the On-Line Encyclopedia of Integer Sequences as A006052, with its references back to Frénicle’s 1693 enumeration. The mathematics of Dürer’s particular four-by-four, and the physical and art-historical reading of Melencolia I, are worked out in close detail in a study deposited on the open preprint server arXiv, physics/0602185.

For the magical reception the indispensable primary witness is Agrippa’s own text. The Cologne 1533 Latin folio of the De occulta philosophia, with its plates of the seven planetary tables in Hebrew numerals, is long in the public domain; a scholarly transcription of the 1651 English Three Books of Occult Philosophy — the version through which the squares entered the Anglophone occult tradition, and which Francis Barrett would later lift wholesale into his Magus of 1801 — is hosted by Joseph Peterson’s Esoteric Archives at esotericarchives.com/agrippa/. The single best modern study of how that system actually works as a system, rather than a compilation, is Christopher I. Lehrich’s The Language of Demons and Angels: Cornelius Agrippa’s Occult Philosophy (Brill, 2003). The Chinese strand — the Luoshu, the Hetu, and the question of their true age — is set out with the Qing evidential critique in Richard J. Smith’s Fathoming the Cosmos and Ordering the World: The Yijing and Its Evolution in China (University of Virginia Press, 2008), the standard account in English. For the Byzantine hinge, P. G. Brown’s translation of The Magic Squares of Manuel Moschopoulos makes available the treatise that first carried the construction methods, stripped of their mysticism, into Western learning.

What the various squares share, in the end, is a habit of mind rather than a single doctrine. A figure that is perfectly balanced — that returns the same answer however it is entered, that holds its constant against every attempt to read it crooked — looks like order caught in the act of being itself, and that appearance of an order deeper than the marks on the page is exactly what the talismanic tradition reached for when it cut the figure into metal at the planet’s hour. The grid does not need the magic to be beautiful, and it does not need the metal to be true. It is the place where a sum and a seal happen to wear the same face, and where the question of whether symmetry is power or only symmetry was asked again in every culture that drew one.