Who Painted the Mona Lisa: The Workshop Evidence Behind Leonardo’s Portrait

Leonardo da Vinci painted the Mona Lisa, also known as La Gioconda, between roughly 1503 and 1519. He began it in Florence, likely working from a panel of poplar wood, and carried it with him for the last sixteen years of his life. The painting now hangs at the Louvre in Paris, where it is catalogued as Portrait of Lisa Gherardini, wife of Francesco del Giocondo.

That short answer hides a stranger story than most biographies admit. The portrait was never delivered, never signed, and never dated by its maker. What survives is a physical object whose layers record decades of revision.

The painting’s layered construction

The Mona Lisa is oil on a single poplar panel measuring about 30 by 21 inches, or 77 by 53 centimeters. Poplar was standard for Florentine panel painting, but it is a soft, dimensionally unstable wood. This one has warped and cracked across centuries, which is why the Louvre controls its environment tightly.

Leonardo built the image in dozens of translucent oil layers rather than the faster tempera method common in his youth. Oil dries slowly, which allowed him to rework edges and glaze shadows over months. That slow drying is the mechanical reason the portrait looks soft rather than drawn.

He applied glazes so thin that some are estimated in the range of a few micrometers to a few tens of micrometers. At that thickness, paint stops behaving like pigment and starts behaving like tinted glass. Light passes into the layer, scatters off what is beneath, and returns to the viewer’s eye. The result is the hazy transition the Italians called sfumato, literally “gone up in smoke.”

The technique is the evidence. X-ray and infrared imaging show underdrawing that does not match the final surface in several places. The position of the fingers, the outline of the veil, and the shape of the chair all shifted during painting. A copyist does not revise. An inventor does.

One counterintuitive finding matters here. In 2004, a team led by Pascal Cotte used a multispectral camera to scan the panel in thirteen wavelengths from ultraviolet to infrared. Their published analysis argued for an additional underpainting visible beneath the surface. Interpretations of that layer remain disputed, but the deeper point is uncontested. What visitors see is the top of a stack, not a single decision.

Quantified context: dates, dimensions, and the documentary record

The most cited date range for the portrait is 1503 to 1506, with continued work afterward. That range comes from a 1503 note by Agostino Vespucci, a Florentine official, who compared Leonardo’s unfinished portrait of Lisa del Giocondo to classical portraits by Apelles. The note is the only near-contemporary document tying the sitter to the painting.

Giorgio Vasari, writing in 1550 and again in 1568, described the portrait in his Lives of the Artists. Vasari never saw the finished picture as far as anyone can prove, and his account contains errors about other works. Historians treat him as a late, secondhand source rather than an eyewitness.

The Louvre’s own conservation records date the panel’s earliest known provenance to the French royal collection. Francis I reportedly acquired it, and it appears in royal inventories from the sixteenth century onward. The painting entered the Louvre after the French Revolution, in 1797.

Physical scale matters more than most explainers admit. At roughly 77 by 53 centimeters, the panel is smaller than a modern poster and slightly larger than a sheet of A3 paper. Its fame is not proportional to its size. Its protective glazing, bulletproof since a 1956 acid attack, adds another layer of distance between viewer and paint.

Here is the honest limit. We cannot date the final brushstrokes to a year, a month, or even a decade with current evidence. Census-style precision does not exist for this object.

The revision timeline

The table shows the core problem. The documentary record is thin at both ends, and the physical record carries the argument.

Edge cases and competing claims

Could someone else have painted it? A minority argument holds that a workshop assistant contributed. Leonardo ran a studio and trained pupils, and workshop practice in Renaissance Florence assumed collaboration. But the handling of the face shows a level of glazing control that no documented assistant of his reproduced. The Louvre attributes the work to Leonardo alone.

Could the sitter be someone other than Lisa Gherardini? Several identities have been proposed, including a mistress of Giuliano de’ Medici and an idealized composite. The 1503 Vespucci note is the strongest single piece of evidence for Gherardini, and it is not seriously displaced.

Could it be a copy? Multiple early copies exist, including one at the Prado in Madrid. The Prado copy, cleaned in 2012, revealed a different background and confirmed that copyists worked alongside Leonardo. Copies prove the composition circulated. They do not make the Louvre panel a copy.

Red flags for anyone evaluating a claim about the painter:

A source that names an assistant without citing conservation imaging or archival documents A date precision of a single year presented as settled fact Vasari quoted as though he saw the finished painting Conspiracy framing around the sitter’s identity that ignores the Vespucci note Claims that the painting is unsigned and therefore unattributable

Unsigned does not mean unattributed. Almost no Italian panel of this period carries a signature. Attribution rests on technique, provenance, and documentary triangulation.

Misconceptions that survive repetition

The most persistent error is that the painting is famous because it is exceptionally beautiful. It is a fine portrait, but its fame is largely modern. It became a mass cultural icon after its 1911 theft from the Louvre, when the empty wall drew crowds and newspapers ran for weeks.

A second error treats the smile as a deliberate optical trick with no basis in paint. The softness comes from glazing, not from hidden symbols. The effect is real, but its cause is chemistry and brushwork.

A third error assumes Leonardo finished it. He did not. The panel left his hands in a state he considered incomplete, and it was never delivered to the sitter’s family. The lack of a final varnish layer in some readings supports this.

A fourth error is the notion that the painting is unique in technique. It is exceptional in execution, not in method. Venetian painters used oil glazes extensively, and Netherlandish artists had refined the approach decades earlier. Leonardo’s contribution was control, not invention.

Plain language: the painting looks alive because thin layers of oil let light bounce around inside the paint.

Practical actions for evaluating claims

If you want to verify an attribution claim yourself, use this sequence. Start with the museum’s own catalogue entry, which lists provenance and conservation history. Then check whether the claim cites imaging studies, archival documents, or both. Absence of either is a warning.

Next, check dates against the documentary anchors. Any claim that predates the 1503 Vespucci note needs extraordinary evidence. Any claim that treats Vasari as an eyewitness needs correction.

Then compare technique descriptions against known Renaissance practice. Claims that Leonardo invented oil painting, or that the painting contains a hidden code visible only to the claimant, fail basic scrutiny.

Finally, treat crowd-pleasing theories with proportional skepticism. A claim that overturns a Louvre attribution should be published in a peer-reviewed conservation or art-history venue before it reaches a headline.

Threshold guidance, reconciled. A single documentary note plus consistent imaging supports standard attribution. A single documentary note alone, without physical evidence, supports a hypothesis. Physical evidence alone, without provenance, supports a technical observation, not an authorship claim. You need at least two independent lines to make a strong case.

What remains uncertain

The honest summary is that the who is settled and the when is not. Leonardo painted the Mona Lisa, but the painting’s final state reflects years of revision that cannot be dated precisely. The panel’s condition, its glazing, and its layered construction all limit what any surface examination can conclude.

The 2004 multispectral work opened questions about underpainting that conservators continue to debate. No published consensus resolves every layer. That is normal for a 500-year-old object, not a scandal.

Any explainer that claims a definitive completion date, a confirmed workshop assistant, or a decoded hidden message is overreaching. The evidence supports Leonardo as painter and a long, uneven execution. It does not support more.

If you are researching the painting for academic work, cite the Louvre’s catalogue and the conservation literature directly. If you are writing for a general audience, say what is known and name what is not. That standard is what separates a real explainer from a thin one.