Iconography: How to Read Symbols in Art

What iconography names in art history

The term does three related jobs, and search results mix them. First, it names the stock of conventional images attached to a subject: the winged sandals that mark Mercury, the lute that can mark harmony or, with a broken string, discord. Second, it names the scholar’s procedure for identifying those images. Third, in older English and in Orthodox usage, it can mean the making of sacred panel icons. This article stays with the first two jobs in Western painting, sculpture and print. Sacred panel technique and saint-by-saint lists belong elsewhere.

Émile Mâle, Aby Warburg and Erwin Panofsky turned a loose antiquarian habit into a teachable method in the late nineteenth and early twentieth centuries. Before them, artists already used handbooks. Cesare Ripa’s Iconologia (Rome, 1593; illustrated 1603) described personifications of virtues, vices and abstract ideas for painters and poets. Vincenzo Cartari’s Le imagini de i dei (1556) and Pierio Valeriano’s Hieroglyphica (1556) fed that market. Encyclopaedia Britannica still cites Ripa as the most famous of those repertories. The books are period tools, not oracles. Ripa invented figures when antiquity supplied none. You use him to recover how a seventeenth-century workshop might have staged “Good Counsel,” not to decode a fifteenth-century fresco he never saw.

A second confusion sits in design blogs and software documentation. There “iconography” means a set of interface glyphs: a magnifying glass for search, a gear for settings. That usage borrows the Greek root and stops there. It has no claim on Panofsky, Warburg or museum cataloguing. If you arrived from a UI style guide, the rest of this page is a different discipline.

Colour can carry convention too, but colour-as-meaning is a separate guide in this series: Colour Symbolism in Art History. The job here is method: how you move from a visible thing to a defensible reading.

Panofsky’s three levels

Erwin Panofsky (1892–1968), trained in Hamburg in the circle around Aby Warburg, set out a three-layer scheme that still structures undergraduate teaching. He published an early version in the journal Logos in 1932 and restated it in the introduction to Studies in Iconology: Humanistic Themes in the Art of the Renaissance (New York: Oxford University Press, 1939). He reprinted and adjusted the scheme in Meaning in the Visual Arts (1955), where the third layer is labelled iconological interpretation. The layers are not three essays you write in order and then forget. They are checks on one another. A wrong description at the first layer wrecks the second. A second-layer type that never existed in that decade wrecks the third.

Panofsky’s famous classroom example is a man in the street raising his hat. At the first layer you see a male figure, a hat, an upward movement of the arm. At the second you recognise a greeting, a convention that has a history (the hat as a sign of status; the lift as a polite reduction of that status). At the third you may read a whole bourgeois world of manners, class and secular civility that the man himself does not announce. The same three moves apply in a museum, with one extra demand: you must show where the convention is documented.

Pre-iconographical description

You start with primary or “natural” subject matter. Name figures, objects, gestures, spatial relations and the medium. A winged seated woman, a sleeping dog, a polyhedron, an hourglass, a bat with a banderole. Oil on oak, 207 × 209.5 cm. Two men, a two-tiered shelf, a smear on the floor that refuses to resolve from the front. This layer needs familiarity with objects and events, and a history of style so you do not mistake a Mannerist elongation for a deformity in the sitter.

The discipline here is refusal. You do not yet say “Melancholy” or “memento mori” or “the Ambassadors.” You say what a visitor with good eyes and no specialised reading would still see. If the picture includes text, copy the letters. If a string is broken, say so. If a flute case has an empty slot, say so. National Gallery wall text on Holbein’s The Ambassadors (1533, NG1314) builds its later argument on those facts: one lute string broken, one flute missing, an arithmetic book open at division. Those are first-layer observations. The political reading comes later.

Photographs lie about scale and surface. If you can, stand in front of the object. Prints exist in multiple impressions; Dürer’s Melencolia I (1514) is in the Metropolitan Museum of Art (43.106.1) and in the British Museum (1912,1220.2), among other cabinets. You describe the plate, not a poster.

Iconographical analysis

The second layer is conventional subject matter: images, stories and allegories. You connect the visible motif to a theme that a period viewer could have named. Mercury wears a helmet and winged sandals and holds a staff; the Uffizi’s entry on Botticelli’s Primavera (c. 1480, inv. 1890 no. 8360) uses those attributes to close the left side of the composition. Flora is Chloris after transformation. Cupid is blindfolded. None of that is visible as biology. It is visible as type.

This layer needs literary sources and a history of types: how a theme was pictured under different conditions. Ovid, the Bible, Ripa, a print series, a contract, a hymn book with readable staves. Panofsky insisted that you correct the texts by looking at pictures, and the pictures by looking at texts. A Judith holding a head is not a Salome if the maid, the bag and the narrative moment match Judith’s story instead. University teaching sites still walk that example through Artemisia Gentileschi. The point is comparative, not pious.

Some pictures have thin second layers. Panofsky noted that certain landscapes, still lifes, genre scenes and non-objective works do not carry stories in the same way. You can still describe them. You should not invent a myth to fill the gap.

Iconological interpretation

The third layer is intrinsic meaning: the work as a symptom of a world view, a religious attitude, a social formation. Panofsky called the required faculty “synthetic intuition,” trained by a history of cultural symbols. The artist may not have intended the symptom. That is the controversial part. Later critics accused iconology of finding Zeitgeist under every bush. The useful remainder is more modest. Once the type is secure, ask what job it does in this room, for this patron, in this year.

The Louvre’s notice on Poussin’s Les Bergers d’Arcadie (c. 1638, INV 7300) summarises Panofsky’s own essays of 1936 and 1955 on the inscription Et in Arcadia ego. Grammar, the presence or absence of a skull, and the difference between the Chatsworth version (c. 1628) and the Louvre version change whether Death speaks in the present or a dead shepherd speaks in the past. That is third-layer work grounded in second-layer facts. It is not a free meditation on “time.”

Keep the third layer last and short until the first two are solid. Gallery visitors who jump to “what it says about us” skip the part the picture can teach.

Attributes, types and the history of a motif

An attribute is an object, animal, garment or bodily mark that identifies a figure or a personification once you know the convention. Winged sandals: Mercury. A caduceus: Mercury again. A caliper in a seated winged woman’s hand, in Dürer’s print, belongs to geometry among the liberal arts, not to a carpenter’s portrait. Attributes are economical. Workshops repeated them so a viewer across a nave or a print-shop counter could name the subject at speed.

A type is a larger pattern: the equestrian ruler, the reclining nude with a small attendant, the Arcadian shepherds at a tomb, the hat-raising greeting. Warburg tracked types as they migrated. In a 1905 lecture on Dürer and Italian antiquity he used the term Pathosformel (pathos formula) for a crystallised gesture of intense feeling that later artists reactivate. His unfinished Bilderatlas Mnemosyne (begun in earnest in February 1927; last photographic campaign after his death on 26 October 1929) pinned 971 items on 63 panels, in the final recorded state held by the Warburg Institute, University of London. Plate 46 gathered the ninfa, a young woman in antique dress who moves through Quattrocento festival and painting. The atlas is a method of juxtaposition, not a dictionary. You look at recurrence across media, including newspaper clippings and adverts. That breadth is the point and the risk. Recurrence is not identity.

Tate’s glossary gives a compact warning with the dove. In Christian painting the dove can stand for the Holy Spirit. In classical myth the same bird points toward Aphrodite or Venus. If you lift the bird out of its picture and file it under a universal “peace,” you have left the method for greeting-card habit.

Types have dates. A personification posed from Ripa cannot explain a work painted a century before Ripa printed. An ICONCLASS notation can retrieve hundreds of “melancholy” images; it cannot tell you which of Agrippa’s three grades of melancholy Dürer had in mind. The Met’s entry on Melencolia I treats Cornelius Agrippa’s De occulta philosophia as a probable key to the numeral I (artistic imagination as the first grade) and still hedges: “almost certainly known to Dürer.” Hedge when the museum hedges.

How to check a source before you trust a reading

A reading is a claim about a convention in a time and place. Claims need sources with dates, languages and reasons to exist. The sequence below is the one curators and cataloguers already use. You can copy it in a library or on a museum site without a doctorate.

Start with the object. Read the inscription, the signature, the inventory number, the support. Holbein dated The Ambassadors on the marble floor and signed it; the National Gallery notes that he seldom signed, which is evidence of pride or of a special brief, not a mystic code. Copy the Latin. Aetatis suae on Dinteville’s scabbard and on de Selve’s book gives their ages in 1533. Then open the current collection catalogue, not a blog that quotes an older catalogue. Susan Foister’s 2024 National Gallery volume is the present scholarly baseline for that picture. The Louvre’s online notice for the Poussin names Panofsky, Mérot and Verdi and records that the notice itself may lag the latest literature. That honesty is part of the method.

Printed repertories and classification systems

Ripa’s Iconologia remains the first stop for Baroque personifications. Use the edition that could have reached the artist: 1593 unillustrated, 1603 illustrated, later translations. Hans-Joachim Zimmermann’s survey of English adaptations stresses the obvious trap. If you cite a nineteenth-century Ripa for a canvas of 1620, you are reading the wrong book. Cartari and Valeriano matter for mythographic types in the later sixteenth century. Emblem books (Alciato and his followers) matter for motto-plus-picture combinations. None of these handbooks governs every picture in Europe. They govern workshops that owned them or shared their habits.

ICONCLASS, devised by Henri van de Waal at Leiden from the 1940s and later published in print and at iconclass.org, is a decimal classification of subjects in Western art. Divisions 1–5 cover general topics; 6–9 cover narrative fields (7 Bible, 9 classical myth); division 0 was added in 1996 for abstract art. A notation lets a database retrieve comparable images. Roelof van Straten argued in 1986 that ICONCLASS stores and retrieves second-layer identifications. Retrieval is not interpretation. Two pictures that share a sitting-figure notation do not share a meaning. Princeton’s Index of Medieval Art is the same kind of finding aid for medieval types. It does not licence pasting a twelfth-century type onto a fifteenth-century panel.

Museum catalogues, libraries and the Warburg habit

Prefer collection pages from the museum that holds the work: Uffizi, National Gallery, Louvre, Met, British Museum, Tate. Heilbrunn-style essays and wall texts are starting points. They compress. For a contested reading, follow the bibliography to a catalogue raisonné or a recent journal article. The Warburg Institute’s library in London was built for this cross-check: photographs, festival books, astrology, survival of antique gesture. You do not need Warburg’s metaphysics of Nachleben to use his habit of putting a print next to a sarcophagus next to a news photo and asking whether the formula is the same.

Two independent recent sources is the bar for dates, attributions and locations. If they disagree, say so. The Uffizi still calls the full meaning of Primavera a mystery while naming the nine figures. That is a model sentence. Name what is secure (Mercury’s sandals; Chloris becoming Flora; 42 plant species identified in the vegetation) and leave the Medici wedding-allegory industry in the notes until a source from the 1480s appears.

Auction essays and undated listicles fail this test. Wikipedia is a trailhead. Verify every date elsewhere. For places to see the works named here, start with Exhibo’s museums directory, then the collection pages.

Six pictures, read as method

The examples below are not a canon of “the most symbolic paintings.” They are six objects on which the method can be watched. They run from a Florentine mythological panel to a felt suit. Christian narrative is present only where a crucifix is a fact on the Holbein (and then only as one object among many). Vanitas still life, memento mori as a cultural history, and colour symbolism have their own pages in this series; this section will not steal them.

Read each example in three passes: what is there, which type it matches, what a responsible third-layer claim would need.

From Botticelli to Holbein

Botticelli’s Primavera (c. 1480, Uffizi, tempera grassa on poplar, 207 × 319 cm) shows nine mythological figures on a flowery lawn in a grove of orange and laurel. Daniela Parenti’s Uffizi text names them from right to left in narrative order: Zephyrus seizes Chloris; Chloris becomes Flora; Venus stands chaste at the centre with blindfolded Cupid; the Graces dance; Mercury, helmeted, with winged sandals, touches a cloud with his staff. First layer: nine bodies, identifiable plants, a wooded stage. Second layer: Ovidian transformation and a classical messenger-god. The painting was recorded by the late fifteenth century in the Via Larga house of the heirs of Lorenzo di Pierfrancesco de’ Medici, hung above a lettuccio; Vasari later saw it at Castello with the Birth of Venus. Third layer: a celebration of love, peace and prosperity, in the Uffizi’s cautious wording. Flower-by-flower Medici cryptography goes beyond what the museum will sign. The method stops where the documents stop.

Dürer’s Melencolia I (1514, engraving, plate about 24 × 18.5 cm) seats a winged woman with a caliper among tools of geometry, a putto, a sleeping dog, a polyhedron, an hourglass, scales, a magic square and a bat bearing the title. The Met groups it with Knight, Death, and the Devil (1513) and Saint Jerome in His Study (1514) as the Meisterstiche, corresponding in older commentary to moral, theological and intellectual virtue, yet “in no way a series.” Second layer: melancholy as a humour, upgraded in Renaissance thought from disease toward creative genius, with geometry as the liberal art under artistic making. Agrippa’s first grade of melancholy is the Met’s preferred explanation of the numeral. Third layer: a spiritual self-portrait of the intellectual artist, which the Met states as a reading, not as Dürer’s caption. Panofsky wrote a monograph on the print; later scholars have argued with him. You can teach the attributes without pretending the argument is closed.

Holbein’s The Ambassadors (1533, oil on oak, National Gallery, London) portrays Jean de Dinteville and Georges de Selve beside a two-tiered stand of instruments. First layer includes a celestial globe, a polyhedral dial associated with Nicholas Kratzer, a terrestrial globe marking Polisy, a lute with a broken string, a hymn book showing two Lutheran pages that do not follow each other in the standard book, an arithmetic manual open at division, a missing flute, a crucifix half-hidden in the curtain, and an anamorphic skull that resolves from the lower right. The Gallery’s current text treats the objects as a visual essay on religious and political discord in the year Henry VIII married Anne Boleyn. That is a second-layer cluster (scientific instruments, music, liturgy, death’s-head) tied to a documented diplomatic stay. The skull is also a display of skill. You do not need to choose between virtuosity and meaning; Holbein used both.

From Poussin to Beuys

Poussin’s Louvre Les Bergers d’Arcadie (c. 1638, oil on canvas, 85 × 121 cm) gathers shepherds and a woman around a tomb inscribed ET IN ARCADIA EGO. Guercino had used the phrase on a painting of 1618 now in the Galleria Nazionale d’Arte Antica, Palazzo Corsini, Rome, with a skull. Poussin’s earlier Chatsworth version (c. 1628) still carries that skull. The Louvre version drops it. Panofsky argued that present-tense grammar (“Even in Arcadia, I [Death] am”) fits the skull pictures, while the Louvre’s quieter group invites an elegiac past (“I too lived in Arcadia”), even though he judged the present tense the only fully correct Latin. The Louvre notice records Alain Mérot’s caution that a Christian current still runs under the Virgilian calm. First layer: four figures, a sarcophagus, an inscription, a landscape. Second layer: pastoral type plus a motto with a short visual history. Third layer: a shift from moral shock to meditation, which you should label as Panofsky’s thesis, not as the painting’s subtitle.

Picasso’s Guernica (1937, Museo Nacional Centro de Arte Reina Sofía, Madrid) was painted for the Spanish Pavilion at the Paris Exposition after the bombing of Gernika. A bull, a wounded horse, a lamp, a fallen warrior, a screaming woman with a dead child: these are first-layer facts. Second-layer identifications splinter. Viewers have made the bull into Spain, into Franco, into brutality, into the artist’s private stock of Spanish beasts. Picasso gave interviews that refuse a single key. The honest iconographic sentence is that the painting uses a short list of repeated animals and lamps from his earlier work, now scaled to a public atrocity, and that no handbook of 1937 fixes those animals the way Ripa fixes “Good Counsel.” Personal iconography, which Tate ascribes to much of Picasso’s oeuvre as autobiographical, still has to be argued picture by picture.

Joseph Beuys built a late-twentieth-century system out of substances. Tate names felt, fat and honey as his iconography of life and society. Felt Suit (1970) exists in multiple examples; Tate holds the type. Felt and fat return in sledges, vitrines and the 1965 action How to Explain Pictures to a Dead Hare (honey and gold on the head; one foot on felt, one on metal). Beuys tied the materials to a wartime story: a crash in the Crimea, rescue by Tatar hosts, wrapping in fat and felt. Tate’s 2005 exhibition text calls that story a legend with “an element of legend.” You may describe the materials as a personal attribute-set. You should not repeat the crash as military fact. Contemporary artists often replace Ripa with a private handbook. The method still applies: list the recurring things, find the artist’s statements and the first exhibitions, separate statement from later myth.

Kehinde Wiley’s Napoleon Leading the Army over the Alps (2005, Brooklyn Museum) restages Jacques-Louis David’s 1801–1805 equestrian type with a Black sitter in contemporary dress against a decorative ground. The second layer is the equestrian ruler type itself, one of the most stable political formulas in Western painting. Wiley’s move is to keep the type and change the body that occupies it. You do not need a new dictionary of objects. You need the history of the type, which Warburg would have put on a panel next to Roman imperial cameos and Baroque state portraits.

How contemporary artists still use the method

Personal systems did not begin in 1970. Tate points to William Blake’s private mythology in the 1790s, still the subject of a specialist literature, and to Picasso’s autobiographical motifs. The difference now is institutional. Artists publish statements, give interviews, stage actions and sell editions with titles that do the old work of attributes. You treat those texts as Ripa’s grandchildren: useful, dated, sometimes self-mythologising.

Kerry James Marshall poses Black figures in academy genres (studio, picnic, portrait) so that the genre, plus a few stubborn props, does the identifying work. Cindy Sherman’s History Portraits (1988–90) put her body into Renaissance and Baroque formulas so the type, not a fictional sitter, becomes the subject. Start with description. A ruff is a ruff. A horse on its hind legs is an equestrian type. Then name the source picture and the date of the restaging.

Materials can be attributes. Beuys’s felt recurs until the viewer expects it, closer to a heraldic colour than to a still-life object. A living artist’s interview is a source. A fan theory is not. Biographical claims about living makers stay factual: exhibition dates, materials, titles.

If you work in a studio, invert the method. Choose a type with a history. Decide whether you want that history readable. Supply an attribute a catalogue could name in one line, or withhold it, knowing that viewers trained on Ripa will hunt anyway.

How to identify iconography when you are looking

Stand at ordinary viewing distance. Name the medium and the size. List the figures and the portable objects. Read every inscription, including tiny ones on books and instruments. Move. Holbein’s skull is a tutorial in the need to change your feet. Then step in for attributes: shoes with wings, a staff with snakes, a broken string, a square of numbers.

Ask whether the same cluster appears in other works of the decade. A single skull in a portrait of 1533 sits in a known Renaissance habit of including a death’s-head; it does not turn the painting into a Dutch vanitas banquet of the 1660s. A dove needs its companions. A horse on its hind legs needs David, or Rubens, or Marcus Aurelius, before it needs your feelings about power.

Write the second-layer name only when you can point to a source. “Mercury” is allowed once sandals, helmet and staff coincide and a catalogue agrees. “The spirit of capitalism” is not a second-layer name. Save it, if you must, for a seminar after the sandals.

In a museum with labels, treat the label as a hypothesis the collection is willing to print. Check the date of the text. Collections revise. The National Gallery’s Holbein page now folds Foister 2024 into public prose; older guidebooks still misidentify sitters, as the 1589 Polisy inventory misidentified de Selve. Disagreement in the record is data.

If you are writing an essay, keep a three-column notebook: visible facts, proposed type, source with year. Anything that cannot sit in a column does not belong in your paragraph.

Common confusions and misuses

Iconography and iconology are related and not interchangeable. In Panofsky’s 1939/1955 wording, iconography classifies images, stories and allegories; iconology interprets intrinsic meaning by synthesis. Many museums now use “iconography” for both. If you write for a course, keep Panofsky’s split. If you write a wall label, “iconography” will be the word the public searches.

Iconography is not the Symbolist movement. Painters in every century used symbols. Symbolism with a capital S names a late-nineteenth-century cluster around Moreau, Redon, Böcklin and their critics. That history is in Symbolism in Art. Mixing the two produces ghost sentences: “Botticelli was a Symbolist because Venus is a symbol.” He was a Florentine painter using classical types.

Iconography is not Orthodox icon-painting as a craft, though the words share a root. Courses in egg tempera and canons of the face belong to religious practice and to Byzantine surveys. This method page will not teach either.

Iconography is not a universal decoder ring. Objects change jobs. Tate’s dove is the teaching example. A skull can mark death, artistic skill, anatomy study or all three. A bull in 1937 is not a bull in a Minoan fresco. If a reading cannot survive a second picture of the same object in the same decade, drop it.

Iconography is not UI design. Interface glyphs trade on recognition speed. They do not have Ripa editions. This article answers the art-historical intent and leaves glyph libraries alone.

Over-reading is the occupational disease. Panofsky’s third layer, Warburg’s atlas and undergraduate enthusiasm all push toward total explanation. The correction is documentary. Uffizi: meaning of Primavera still in part mysterious. Met: Agrippa “almost certainly.” Tate: Beuys’s crash story as legend. Louvre: two translations of one line, both in play. Copy those hedges. They are the method working.

FAQ

These questions repeat the searches that sit around the word “iconography”: simple definition, iconology, worked examples, purpose, Orthodox craft, and whether a living artist can have a system without Ripa. The answers stay inside the method. They do not open a saint list, a MENA inventory, or a survey of the Symbolist movement. If a question is about church types, go back to the sibling named in the opening. If it is about Moreau and Redon, use the live movement essay. If it is about interface glyphs, you are in the wrong discipline. Short answers below; the long versions sit in the sections above.

What is iconography in art, in plain language?

It is the study of what pictures show by convention: subjects, attributes and types, as distinct from brushwork or composition. You describe the visible facts, match them to a known type, and test the match against a dated source. The word also names the set of types itself, as in “the iconography of Mercury.”

How does iconography differ from iconology?

In Panofsky’s scheme, iconography is the second layer (stories, allegories, identifiable types) and iconology is the third (the work as a symptom of a world view). He set the scheme out in 1932 and 1939 and renamed the third step more clearly in 1955. Everyday museum prose often uses “iconography” for both layers. If your tutor wants the split, keep it. If you are reading a label, expect the broader usage.

What is a concrete example of iconography?

Mercury in Botticelli’s Primavera is identifiable from helmet, winged sandals and staff, as the Uffizi states. The same method names Dürer’s seated woman as Melancholy by her tools, title-bat and posture, and names Holbein’s smear as a skull once you stand at the correct angle. Each identification is a second-layer claim with a paper trail.

Why do art historians bother with this method?

Because style description cannot tell a Judith from a Salome, or a greeting from a random arm movement. Patronage, liturgy, diplomacy and teaching all depended on viewers recognising types. Catalogues still depend on it: ICONCLASS notations, Index of Medieval Art headings, and collection searches for “caduceus” only work if someone has done the identification.

Is iconography the same as Christian icon painting?

No. Shared Greek roots explain the overlap in search results. Panel icons, canons and Orthodox theology are a religious and technical tradition. This guide is a Western art-historical method for reading attributes and types. Western church subjects belong to the sibling article on Christian iconography.

Can a contemporary artwork have iconography if there is no Ripa for it?

Yes. A personal or political system of repeated materials and types is still an iconography in Tate’s sense. You document the recurrences, the titles, the first exhibitions and the artist’s statements. You do not grant those statements the status of uncheckable revelation. Beuys’s felt is a clear case; so is Wiley’s use of the equestrian formula.

Where should I look first if I am stuck in front of a picture?

The object, then the museum’s current catalogue page, then a period handbook that could have existed for the maker, then a classification browser such as ICONCLASS for comparanda. Two recent independent sources for any date or attribution. For the works in the table, the Uffizi, National Gallery, Louvre, Met, British Museum, Reina Sofía and Tate pages are the first URLs to open.

Does every object in a painting “mean something”?

No. Some objects are tools of the trade, still-life bravura, or studio leftovers. Panofsky already carved out landscapes, still lifes and genre scenes that may lack a second layer of story. The method includes the right to stop. A claim without a type and a source is decoration.