← Language
ES EN FR IT PT DE NL SV
HomeArticlesHistory and curiosities

Sabio Tarot · History and curiosities

By Sabio Tarot

What “Tarot de Marseille” means in deck history

A shop may sell two “Marseille” tarot decks whose colours, names and details do not quite match. That does not necessarily make one a fake. The name describes a recognisable family of decks and a history of circulation and reproduction, not one original, unchanging box signed by a single maker.

The Bibliothèque nationale de France brings together decks called “de Marseille” in Gallica, as part of a much broader playing-card collection. The British Museum holds a later reproduction of Nicolas Conver's 1760 design. Both records help make a visual tradition more specific.

Design family, edition and physical copy

Three levels need separating. A design family shares features: Italian suits, court cards, named trumps and related compositions. A particular edition chooses artwork, colour, words and printing technique. A physical copy is the box someone handled, which may now show wear, repairs or cards mixed from different sets.

The British Museum's record is especially revealing. It catalogues a 78-card deck as a reproduction of the tarot created by Nicolas Conver in 1760, but dates the preserved copy to the late nineteenth century. It adds that the box belongs to a different game and that some information attached to it needs reviewing. If we treated box, design and copy as one thing, we would tell their history incorrectly.

“Restoration”, “facsimile” and “edition inspired by” do not mean the same thing either. A reproduction may adjust colours or lines to make damaged details legible; a facsimile aims for a different relationship with the original's present condition; a reinterpretation may redraw much more. Before choosing a guidebook, see what its publisher says and compare actual cards.

Features that help you recognise the family

In many decks from this family, numbered Cups, Swords, Wands and Coins are made up of suit symbols and decoration. They lack the narrative scenes found on many Rider–Waite–Smith Minor Arcana cards. Trumps in French editions bear names such as La Papesse, Le Pape and La Maison Dieu. These clues are useful, but they cannot by themselves identify the exact workshop, city and year of a copy.

The name “Marseille” should not erase tarot's earlier Italian period either. The earliest known evidence clusters in fifteenth-century northern Italy. The Marseille family is a later branch of tarot history, not proof that the game was born in that city.

Why this matters to someone reading the cards

Suppose a book asks you to observe a person's gesture on the Five of Cups. If your card shows five cups arranged with plant motifs, that explanation does not describe the image in your hand. You may still learn from it as a different interpretive tradition, but know that you are importing a device from another system.

Check four control cards before buying or using a book: the Papess or High Priestess, Justice, Strength and a numbered card such as the Five of Cups. Compare names, numbers and the presence or absence of scenes. Note the edition printed on the box too. This small inventory usually reveals more than a sales label.

The BnF documents variants and a history of changes in French tarot decks. You can learn the family while respecting its differences. A good manual states which edition it uses as a reference and how to adapt if your deck changes a detail.

To compare systems, read Marseille versus Rider–Waite–Smith and how to choose a Marseille edition. The guide to tarot deck types offers an initial map.

A name that needs a surname

If a website describes “the Tarot de Marseille card”, it should say which edition supplied the details. Even when several decks share a broad composition, a missing line, restored colour or translated title can change what there is to observe. The broad name points readers toward a family; the edition name lets them repeat the comparison.

The BnF arranges its card collections by game type and notes historical variants. That organisation counters two oversimplifications: treating all French tarot decks as one design and using “Marseille” to mean any old deck. A card can be historical and French yet differ from the pattern you are seeking. The label becomes precise only when attached to a maker, date and object.

The British Museum example adds another important layer. Its reproduction is based on Nicolas Conver's 1760 design, but the physical copy was printed much later. Its box comes from another set. A review that copied only the inscription on that box would misattribute the deck. A museum record can correct a sales story because it checks the material, image and object history together.

Studying design is different from studying how to read

For design, you can compare two Marseille editions card by card: line, colour range, suit arrangement and legibility of names. To learn to read, you must also ask what method the guidebook's author proposes. The same image can receive different interpretations in books written centuries apart. A faithful graphic reproduction does not automatically make a contemporary explanation “the original one”.

Choose the Five of Swords from two editions. Count the swords, describe how they cross or separate, and note the decorations you can see. Then consult two manuals and separate their rules. Do they draw on what you see, numerology, traditional associations or a scene borrowed from another system? A hybrid interpretation need not be forbidden, but it should be named. If a book markets itself as a Marseille guide yet describes a figure found only in Rider–Waite–Smith, its promise of visual compatibility fails.

For a beginner, the practical conclusion is simple. Choose by an identifiable edition and sample cards, not only by the word “authentic”. Learn first to describe the deck you actually have. Then add a reading tradition, knowing its date and author. That sequence reduces confusion and leaves room to study other variants without treating each difference as a printing error.

Sources consulted