Three Explanations: Which Solves the Column Paradox?

Three competing theories attempt to resolve the contradiction. The two-version hypothesis suggests Leonardo painted two distinct portraits. Compositional evolution proposes a single painting revised over decades. Copyist invention argues Raphael and later artists added columns for conventional framing. None fully accounts for all evidence, and scholars remain divided.

Theory 1: Two-Version Hypothesis

The Claim: Leonardo painted TWO distinct portraits—an earlier version (c.1503—1506) with full columns, and a later version (c.1513—1517) without, now in the Louvre.

Evidence Supporting

Documentary contradictions: Vasari describes a work commissioned by Giocondo, unfinished; de Beatis saw a finished portrait for the Medici. These may be different paintings.
Lomazzo 1584: mentions “della Gioconda E di Mona Lisa” (AND, not OR)—suggesting two works, and Lomazzo knew Francesco Melzi, Leonardo’s heir.
Raphael’s 1504 sketch shows full architectural columns not visible in today’s Louvre painting.
Precedent exists: Leonardo created two versions of the Virgin of the Rocks and multiple versions of the Madonna of the Yarnwinder.
Advanced glazing (30—40 layers) appears unprecedented in Leonardo’s work before 1508, yet the Mona Lisa shows this technique.

Evidence Against

No historical inventory or estate document mentions two Mona Lisas or separate paintings by Leonardo.
De Beatis saw a singular portrait in 1517, described as if a single work of art.
The Prado copy (c.1503—1516) reproduces the Louvre composition with minimal columns, not an “earlier column version.”
Mainstream Leonardo scholars (Kemp, Zöllner, Syson) reject the two-version hypothesis.
Occam’s Razor: the hypothesis requires multiple unsupported assumptions and missing historical evidence.

Theory 2: Compositional Evolution

The Claim: Leonardo began with full columns (as Raphael saw in 1504), then progressively minimized them over 14 years of continuous revision before the painting’s completion.

Evidence Supporting

Leonardo’s documented habit of constant revision is well-known: the Adoration of the Magi, Virgin of the Rocks, and Salvator Mundi all show extensive pentimenti and reworking.
Raphael saw the painting one year into its creation; Leonardo had 10—15 more years to modify it.
Removing architectural columns fits Leonardo’s documented artistic philosophy of eliminating hard borders and external frames in favor of sfumato and psychological intimacy.
Spolvero marks and preparatory cartoon evidence prove Leonardo worked from detailed preliminary designs, supporting iterative change.

Evidence Against

The Prado copyist (1503—1516) shows minimal bases, not intermediate stages with partial columns—no progression toward removal is visible.
Infrared and X-ray examination have not revealed hidden full columns beneath the surface paint layers.
The 2004 comprehensive examination: no evidence of earlier full architectural elements overpainted or removed.
If Leonardo progressively removed columns, why did 16th—17th century copies consistently include them? Where did the column tradition come from?

Theory 3: Copyist Invention

The Claim: Leonardo never painted full columns. Raphael and later copyists added or expanded them for conventional architectural framing, a Renaissance aesthetic convention.

Evidence Supporting

Martin Kemp (Oxford) argues that Raphael may have interpreted Leonardo’s “hinted columns” (the partial bases visible today) as full architectural elements.
Renaissance aesthetic conventions strongly favored architectural frames and classical orders in portraiture.
Leonardo increasingly eliminated backgrounds in his later works: Lady with Ermine and Ginevra de’ Benci feature minimal background elements.
Most copies with full columns date to the 16th—17th century, a time gap that allows interpretation and embellishment by independent copyists.

Against

Raphael drew from direct personal observation in 1504, not from copying existing copies or drawings.
A preparatory sketch made for personal study, not for a client, would have no reason to include “conventional framing” invented by Raphael himself.
Louvre curator Vincent Delieuvin acknowledged in 2019 that Raphael saw something architecturally distinct from the current painting.
Early copies (16th century) show too much consistency in column representation for independent copyist invention.
Most major Leonardo scholars do not believe Raphael invented the columns; they believe he documented something real.

Scholar Positions

Leading Leonardo experts hold diverse views on the mystery:

Martin Kemp (Oxford): Favors Theory 3—copyist interpretation and Renaissance convention.
Mona Lisa Foundation: Favors Theory 1—the two-version hypothesis.
Vincent Delieuvin (Louvre Curator): Acknowledges the paradox openly, stops short of definitive position; notes Raphael saw something different.
Frank Zöllner (Goethe University): Rejects Theory 1; skeptical of two-version claim.
Luke Syson (National Gallery): Rejects Theory 1; emphasizes documentary evidence against it.

No Theory Is Complete

The two-version hypothesis explains Raphael’s observation but cannot account for the Prado copy’s composition. Compositional evolution explains Leonardo’s artistic logic but lacks physical evidence of hidden full columns. Copyist invention explains later copies but struggles to explain Raphael’s direct observation. Each theory solves part of the paradox while failing to resolve the whole.