Boy Peeling Fruit by Caravaggio

The Dog Under the Paint:

Caravaggio, Cornacchia, and the Question of Why

​I. Introduction — The Gap

In 1642, Giovanni Baglione, a rival painter, wrote that Caravaggio kept a black poodle named Cornacchia, taught it to perform bellissimi giochi (beautiful tricks), and kept it constantly by his side. In 2025, the art historian Gianni Papi* reported an X-ray of a painting attributed to Caravaggio that revealed the silhouette of a small dog beneath the final layer. The dog’s snout had been repainted into the shadows on a boy’s shirt. The dog came first, then was covered.

The question this essay asks is not biographical in the usual sense and not strictly art historical. It is behavioral: what might lead a man, already famous and already violent, to spend time teaching a dog to walk on its hind legs? The evidence is textual (Baglione), physical (Papi), and comparative (animal-training models). The answer is not a diagnosis. It is a hypothesis.

​II. What Is Known — The Record

Baglione’s account is the only contemporary description of the dog’s training. He was hostile to Caravaggio, which may reduce the likelihood of flattering invention, but it does not eliminate distortion. He describes a black poodle, always present, performing tricks that included walking upright. No other source confirms or denies the training. Mancini and Bellori, other early biographers, do not mention the dog.

Papi’s X-ray provides physical evidence that Caravaggio painted a small dog into an early canvas, then painted over it with the image of a boy peeling fruit. The dog’s shape remains visible beneath the paint layer. Whether the dog in the painting was Cornacchia is inferred, not proven. The X-ray shows presence, not behavior.

The textual source is single and hostile. The physical source shows a form, not a motive. The connection between the painted dog and the trained dog is plausible but not certain.

​III. Why the Question Was Missed

Baglione recorded the dog as a curiosity, not as a clue. Later biographers omitted it. Art historians, including Papi, focused on attribution: the dog beneath the paint matters because it helps identify Caravaggio’s hand. The question of why Caravaggio would spend time training an animal for no practical purpose fell outside the usual disciplinary frame.

That omission is not proof of insignificance. It may simply mean the question was never formulated.

​IV. A Behavioral Model — Why Train Without Utility

Training an animal to perform a non-adaptive behavior requires repetition, patience, and close physical contact. When no external reward follows, the reward must be internal. The trainer watches incremental progress. Each corrected posture, each small improvement, each moment of successful imitation becomes feedback.

The training process also narrows attention. Rivalries, debts, uncertainty, and danger recede. The room shrinks to the relation between hand, voice, and animal body. External instability is still there, but temporarily suspended.

Constant proximity to an animal can also create a form of companionship that is simple, nonverbal, and less exposed to human negotiation. The animal does not argue, flatter, or betray. It responds, resists, learns, and repeats. That makes training a small but repeatable exercise in agency.

In a life marked by instability, such an exercise may matter. It offers a controlled domain in which action produces visible effect. That does not prove Caravaggio experienced it this way. It only shows why the behavior could have been meaningful.

​V. The Convergence

Three points can be stated without exaggeration:

• A man trained a black animal to perform an unnecessary action.
• The same man lived amid instability, conflict, and uncertainty.
• The same man included the animal, at least in an early stage, within his work.

No claim of causality follows from this. No claim that every unstable man trains animals. Only that these elements coexist in this case, and that the coexistence is worth noticing.

​VI. What Remains Unknown

Why Caravaggio taught Cornacchia to walk upright is not answered here. The evidence does not support a conclusion. Baglione’s text is single-source and hostile. Papi’s X-ray confirms a painted dog, not a motive. Behavioral models suggest possible explanations — control, attachment, curiosity, amusement, cognitive narrowing — but they cannot prove them.

The dog under the paint is a trace. The training is unrecorded. The question stays open.

​VII. No Conclusion

This essay does not conclude. It frames the question, presents the evidence, proposes a behavioral model, and leaves the inference suspended. Whether the pattern means anything is left to the reader.

*Gianni Papi, “The First Caravaggio,” 2025. See also: Angelica Villa, “X-Ray Reveals Hidden Dog in Early Caravaggio Work,” ARTnews, June 12, 2025.