One of the first academic copies Picasso made from the flat is the left eye of a classical figure from two different angles. He was about 11 when he made it. And here too the plate he was copying was probably of low quality, though it was not from Bargue's Cours. Picasso's study is distinguished by very simple, generic outlines for nose, brow and especially the eyelids — combined, once again, with a very limited indication of form by means of gradations of light and shade. That early study, like Picasso's copy of the Torse Antique, was a portent of things to come. We encounter its echo in Picasso's stylised oil portrait of Gertrude Stein (1906) — with its very generic, clearly delineated facial features — and in his later, more overtly neo-archaic and "sculptural" pastel drawing of a woman's head (1921; owned by the Fondation Beyeler in Basel). In his schematic treatment of the eyes in both pictures, Picasso simply refined the defects in the technique he had learned as an 11-year-old.

Photography, then, did not just infect Picasso's ersatz classical training; its anti-classical ramifications, particularly evident in the Torse Antique plate, were born out in his Modernist work. It also contributed to Picasso's failure to follow the form rather than the shade in the Vollard portrait. A photograph, after all, is produced by reflected light, and in recording visual phenomena through gradations of light, shade and shadow it typically fails to capture nature's geometric depth and complexity, not least where human anatomy is concerned. As the Greeks were the first to realise, it is incumbent on art to reveal that complexity by freeing itself from the pictorial constraints of human vision, which photography simply mimics, in order to create a more complete, more emphatically spatial reality.

As the byproduct of reflected light, rather than an entity whose geometric structure logically antecedes and transcends the realm of optical phenomena, the human figure loses its classical content, not only in the Bargue primer but in the entire programme of late-academic instruction in which Picasso so conspicuously excelled. The figure becomes a classical simulacrum. Michelangelo's famous red-chalk studies for the Libyan Sibyl in the Metropolitan Museum provide a good idea of what I mean by classical content. Michelangelo was concerned with geometry above all — with the relationships between the complex shapes that define the structure of the figure. He too relied on outline, light and shade. But he used them to model those shapes with utmost precision so as to reinforce the figure's structural and spatial character.

Nowhere in Michelangelo's studies does the use of shadow spare the artist the trouble of precisely articulating the anatomical areas it covers. But this short-cut, which is very much a part of the Bargue method, crops up in Picasso's work — on the bony backs of the Two Women Seated at a Bar, for instance. An early example is his academic study of a plaster cast of the reclining Ilissus from the Parthenon (also in the British Museum) done when he was a 13 or 14-year-old student in Barcelona. Here the figure's left flank is largely blanketed by shadow.

In the original statue, the figure's left shoulder is a magnificent form of great complexity. This form Picasso reduced to a dumb, almost spherical shape, in line with the Bargue method. The thorax of the Ilissus rotates on the pelvis, and Picasso incorrectly rendered its lower right rib, creating a rather weird overhang that casts a deep shadow line on the belly. He thus disrupted the unity of the figure while simplistically and inaccurately treating the figure's abdominal and lower right flank (or external oblique) muscles as a single mass. Picasso heavily outlined the figure's left pectoral to set it apart from the shadowed area of the adjoining flank. But that outline distorts the pectoral's form. Even allowing for the possibly inferior quality of the cast Picasso was working from, none of this points to "mastery of the conventions of classical representation".