Attribution and the Limits of Connoisseurship

Connoisseurship is the oldest tool in the attribution toolkit: an expert's trained eye, developed through years of close study of an artist's documented work, applied to a work of uncertain authorship. It remains genuinely valuable — no technical test can substitute for a scholar's familiarity with how a particular hand modeled a fold of drapery, resolved a passage of foliage, or handled the transition between light and shadow. But connoisseurship also has well-documented limits, and treating it as sufficient on its own is a common source of attribution disputes that later prove costly.

The first limit is disagreement among experts, which is common enough to be unremarkable rather than scandalous. Two scholars working from the same photographs or even the same object in front of them can reach different conclusions, particularly for a workshop-produced period, a prolific artist with many followers and copyists, or a work in poor condition where surface evidence has been compromised by past restoration. Connoisseurship is a form of expert testimony, not a measurement, and like any expert testimony it can be wrong, incomplete, or influenced by factors — market pressure, institutional reputation, prior published opinion — that have nothing to do with the painting itself.

The second limit is that connoisseurship alone cannot reliably distinguish a masterful period copy or a close contemporary follower from an autograph work, particularly for artists whose workshops routinely produced variants and replicas under the master's supervision or in his style. This is precisely where technical examination adds real value: pigment analysis can rule out a material anachronistic to the claimed date; dendrochronology or canvas analysis can establish a plausible date range for the support; infrared reflectography can reveal an underdrawing consistent — or inconsistent — with an artist's known working method. None of these tests proves authorship on its own, but each can eliminate specific possibilities that connoisseurship alone cannot.

The soundest attributions, in our experience, are the ones built from multiple independent lines of evidence that corroborate each other: a stylistic judgment supported by technical findings consistent with the claimed date and origin, and a documentary record — exhibition history, early photographs, dealer correspondence — that places the specific object, not merely "a work of this subject," in the right place at the right time. Where these lines of evidence converge, confidence is genuinely warranted. Where they diverge, or where only one is available, the honest position is a qualified attribution rather than a confident one — and saying so plainly, rather than resolving the ambiguity in the direction a client hopes for, is the entire discipline of the exercise.