It would be hard on the face of it to find dustier-sounding topics in art than those suggested by Peter Burke’s The Connoisseur: A Cultural History and Shira Brisman’s The Goldsmith’s Debt: Conceptions of Property in Early Modern Art. And yet the questions lurking at the heart of both books could hardly be more timely: What counts as originality, what value do we place on it, how do we protect that value, and for whose benefit?

The stories told by Burke and Brisman are both rooted in the swirl of ideas and adaptive behaviors that animated early modern Europe (the era formerly known as the Renaissance), as people worked to square humanistic ideals about individual creativity with new technologies and new ways of accumulating wealth and power. Sound familiar?

The goldsmith of Brisman’s title is Christoph Jamnitzer, a maker of ornate tableware and madcap ornamental prints in Nuremberg at the turn of the seventeenth century. Christoph’s grandfather was the famous Wenzel Jamnitzer, goldsmith to the stars (in the form of four Holy Roman Emperors) and best known today as the designer of the eloquent, cosmologically profound engravings of Platonic solids collected in the Perspectiva Corporum Regularium—the pride of many a great print collection, math department bulletin board, and Pinterest page.

Christoph was not so illustrious. Brisman’s is the first book dedicated to him, and even her interest lies less in his artistic genius than in what his works have to say about the social and economic situation within which they were made. An art historian at the University of Pennsylvania, Brisman studies the intersection of Northern Renaissance art production and early modern political theory. When she looks at the phenomenon of “welcome cups”—fancy lidded goblets of gold and silver shared ceremonially by hosts and their guests—she is seeking evidence of how material forms and social customs reflect each other.

For the people who commissioned them, the cups were showpieces of wealth and instruments for affirming social bonds, functioning as diplomatic tributes, security deposits, wedding gifts, or intergenerational heirlooms. For the people who made them, they were showpieces of competitive invention and technical skill. In addition to their extravagant detail and frequent wit, they were feats of metallurgical engineering and improbable balancing acts—a golden apple supported by a thin wisp of branch, a giant upside-down pear hoisted aloft by a small golden peasant. In one Jamnitzer cup, Atlas shoulders a celestial sphere; in another, Hercules bears a terrestrial globe.

Brisman sees such creations as allegorical props aimed at reassuring (and occasionally challenging) the powerful about their status in an uncertain world. Awkwardly but inexorably, European societies were moving away from structures predicated on land-based wealth (local, visible, and familiar) to ones better suited to the fluid and often mysterious movements of capital. With its long tradition of international trade, Nuremberg was at the vanguard of those changes; its institutions and its inhabitants had to wend their way through a tangle of customary feudal obligations and new legal scaffolding for property and exchange. In practice, this meant that someone like Jamnitzer (unlike a serf) had the right not only to his labor but to what she calls the “property of skill.” His patron, meanwhile, had the right to physically control the commissioned object, the city council had the right to govern Jamnitzer’s craft, and the emperor had the right to endow his creations with legal protection against copying.

In retrospect it’s easy to view this stage of European development as a kind of wobbly bridge to capitalism, but for the people on the bridge, of course, the end point was unclear. Only the wobble was certain. The game played by gravity-defying welcome cups, Brisman argues, was to invoke this instability without collapsing. Elaborate and improbable yet trustworthy, they could be seen as embodying an aspirational org chart, shoring up a rickety social order with élan. Despite appearances, the structure holds; grab the goblet and take a swig.

In keeping with much recent art history, Brisman’s argument relies on the concurrence of objects and social phenomena rather than written evidence attesting to the intentions and perceptions of the people involved. Did participants feel jollier about the state of things after drinking from a cup that required squeezing a peasant between finger and thumb? Probably, though that might have been the alcohol.

She is on firmer documentary footing when she turns to Jamnitzer’s New Book of Grotesques (1610), a volume of sixty etchings accompanied by a trio of texts explaining the artist’s aims in prose, poetry, and legalese. The etchings are “grotesque” not in the sense of horrid but in that of ancient Roman grotto decor, with its mash-ups of foliage, fantastical creatures, and architectural doodads. To that aesthetic baseline Jamnitzer added Seussian botany and Boschian trippiness: ornamental scrollwork sprouts heads and lobster claws; coffee pots grow wings; rabbits emerge from spiderwebs. The action may be scatological (a dramatically pooping putto) or mock heroic (a jousting rat astride a snail shell with a dragon’s head, equine hindquarters, and a peacock’s tail). Every creature seems cobbled together from a loose parts bin. In these one-man exquisite corpses, Brisman sees a parallel to the period’s composite notions of property.

Each of Jamnitzer’s three texts comes to its task with a different brief: one is a dedication to a local aristocrat, one a poem addressed to (among others) “young people who want to learn something,” and one an imperial privilege protecting the book from piracy. In the dedication, Jamnitzer adopts a jocular, feisty voice and invites comparison between his own daring and that of Christopher Columbus. The privilege, the closest thing to copyright at the time, forbade copying the book in its entirety for five years. Jamnitzer’s poem, on the other hand, explicitly invites young artists in need of inspiration to treat his designs like a Renaissance version of clip art. But he signs off with the couplet:^*

Art remains art; in the end just mist.
But all’s not in vain: the name persists.

Written in lively, self-aware doggerel, the poem is nonetheless for Brisman “a performance of political speech.” After appealing for aristocratic patronage and asserting his control over the market for his product, Jamnitzer offers “what he has in excess”—his wacky whatnots—to those in need. (In the end, the print scholar Madeleine Viljoen has shown, almost no one took him up on his offer.)

Jamnitzer’s Grotesques came into being at roughly the same time that Europeans began to codify the “collective practice of judging works of art” that is the subject of Peter Burke’s book. While acknowledging that China’s culture of connoisseurship was already centuries old by this point, here he is concerned with the skills and values that developed in service to old master painting. A new interest in discussing the styles and merits of individual artists had been on the boil since Raphael and Michaelangelo. But in the years around 1600 those conversations began to solidify into something like a discipline.

Burke, a distinguished cultural historian, is particularly interested in connoisseurship as a form of knowledge. For connoisseurs, it was important to experience an artwork visually (color, form, organization, and all the rest of it) and also to know things about it. He borrows the French terms connaissance and savoir to distinguish these two modes. A knowledgeable connoisseur should be able to recognize the subject of a picture, the methods and materials employed, and most importantly, who made it. Identifying the maker isn’t usually a big problem with new art, but—Jamnitzer’s hopes notwithstanding—names can get lost with astonishing rapidity. Inspired by the practices of collectors of classical antiquities, lovers of painting set out to identify individual styles and hands at work.

Identifying the artist mattered not just as a guarantee of quality in the manner of a brand name but also because once people start valuing artworks as creative personal expressions, what is seen and understood in the artwork is partly determined by external information about its making. A doglike squiggle drawn by a toddler and one drawn by a professional artist are evaluated differently because the choices and intentions of the makers are different. Knowing where something came from orients viewers toward what it might mean, which in turn allows them to pass judgment on how effectively it performs its assumed task. Judgments about quality and judgments about attribution have never been fully independent of each other.

What had started as a pastime of the few began to spread into the wider culture. Household inventories began to note the names of artists, rather than just subjects, in their descriptions of artworks. Inspired by Vasari’s Lives of the Artists (1550), authors such as Karel van Mander (whose own Schilderboek was published in 1604) worked to document the achievements of artists of their own place and time. Across Europe, academies were founded to teach the mechanics of art and to theorize about its functions and qualitative hierarchies. (Mimesis was important, but the highest art yoked it to messages of philosophical values and moral improvement.)

Burke dates the English use of the word “connoisseur” to 1714. Five years later the painter Jonathan Richardson published a treatise that Burke identifies as “the first reasoned account of the processes of connoisseurship,” in which he endows the “science of a connoisseur” with the power to reform public manners and increase wealth. Writers of all stripes took on questions of how to tell good art from bad, original from copy. Diderot invented art criticism as a literary genre (albeit in privately circulated manuscripts). Sir Joshua Reynolds delivered his Seven Discourses. Philosophers—Hume, Edmund Burke, and Kant—weighed in. Peter Burke’s list of the eighteenth century’s “most important texts on judging art” runs almost three pages.

The Romantic emphasis on individual genius, meanwhile, encouraged parsing artworks with ever finer gradients of originality. Old master drawings, once thought secondary to finished paintings, assumed new value for their nearness to the spark of creativity, especially as paintings were more likely to contain the handiwork of assistants. Collectors grew increasingly squeamish about distance from the artist’s own hand. Concerned by the suggestion that a painting he was considering might not be the original Andrea del Sarto copy of Raphael’s portrait of Leo X but a mere copy of del Sarto’s copy, the first Duke of Chandos complained that “a picture ought to be as tenderly handled as a lady…the least question upon them casts a stain upon their Reputations which is hardly ever washed off.” In 1760, when Oliver Goldsmith observed that the title of “connoisseur” was “the safest passport into every fashionable Society,” he was already flagging an obnoxious pretension.

Connoisseurs strove to recognize material properties that separated the real from the imitation. Burke cites a satirical dialogue by Joseph Addison on collectors who would study a coin by listening to its sound, smelling it, and licking it “to distinguish the age of it by taste.” Personal access to the genuine article was a prerequisite of proper connoisseurship, whether via the grand tour, private collections, or the novel institutions of the public gallery or museum. Nonetheless, in Burke’s history as in Brisman’s, prints play an important role. Reproductions of paintings had been a part of the print market almost from its start in the fifteenth century, but steady improvements in verisimilitude made them increasingly useful to those wanting to compare compositions existing in different places, or to remind themselves of images once seen in person. Burke doesn’t go into this, but the limitations of various print techniques dictated what types of qualities could be discussed and compared, and thus what it was that the trained eye learned to see. Accurate color reproduction was a very late arrival, but even the differences between linear engravings and velvety mezzotints affected how different generations thought about what Raphael or Rubens had been up to. Later, photography and railroads would change this again. Connoisseurship has always been technology dependent.

The nineteenth century brought what Burke calls “the age of the professionals.” The new discipline of art history pumped up the savoir half of the equation, embedding objects within their original cultural settings, and organizing artists and objects into expansive taxonomies and narratives of development. In light of such grand ambitions, the task of attribution began to look minor—less like intellectual inquiry and more like data collection. Even Bernard Berenson, perhaps the most famous professional connoisseur of all time, lamented having spent so much time on what he called the “who-is-who-sity” of attribution.

Everything, however, still relied on attribution—not just the grand historical narratives of art history but the rapidly expanding international art market. Political and economic shake-ups had loosened rafts of old master art from the collections in which they had been moored for centuries. Dealers and buyers were desperately in need of reliable (or at least believable) attributions. Supplying them became a competitive sport. Giovanni Morelli, one of a number of medically trained connoisseurs who pop up in Burke’s book, made a name for himself with a method of looking that emphasized close attention to meaningless detail to ferret out originals from imitations. Workshop assistants would have adhered closely to the designs and methods of their masters but would not imitate every odd tic of the wrist. Likewise it is easier for a forger to mimic an earlier painter’s style and compositions—art students were taught to do exactly that—than to notice that painter’s habitual way of depicting a knuckle. (Freud, ever alert to the potential significance of unconscious acts, recognized a methodological cousin in Morelli.)

Berenson’s approach was more holistic and more secretive. Disinclined to give much credence to documentary evidence such as provenance, he relied on perception. “Arduous training” was important, he acknowledged, as was having a good visual memory, but in the end he put his decisions down to a “sixth sense” or “God’s gift” rather than a teachable method. The eminent Berlin museum director Wilhelm von Bode, for his part, described Morelli as a “quack doctor” and argued for prioritizing the documentary record over the pronouncement of any expert eye, though he himself possessed one and was prone to pronouncements as gnostic as those of Berenson.

The field was riddled with conflicts of interest. Bode supplied certificates of authenticity that benefited dealers in exchange for gifts to his museum. Berenson, more baldly, received both a lucrative retainer from the dealer Joseph Duveen and a cut of paintings sold with his attribution. Burke deftly elucidates the pleasures and moral pitfalls of this unquantifiable yet potentially remunerative form of expertise, from Vasari’s praise for Michelangelo’s success in passing off his own sleeping Cupid as a classical antique to the forger Eric Hebborn, who claimed never to sell his fakes to nonexperts as “there is no satisfaction in deceiving the neophyte.”

Burke’s book is old-fashioned in its terminology, its amiable pace and style, even its typography, which is apt given the degree to which those seventeenth-century values and techniques held firm well into the nineteenth century. New players subbed in for old ones with no substantive change to the rules of the game. The twentieth century, however, scatters some stumbling blocks, leaving Burke unsure of how the old methods of attribution apply. Abstract painting, he writes, “lacks Morelli’s famous clues to authorship,” yet while this might be true in the limited sense of no knuckles to look at, it is untrue in the meaningful sense of visible tics and habits of a particular hand. When he raises the possibility that Duchamp’s paradigm-shifting Fountain might actually have been the brainchild of Elsa von Freytag-Loringhoven, he sees it as a possible downgrading of the work. But since the value of Fountain lies in its power to make traditional notions of authenticity ridiculous, the Freytag-Loringhoven story is just a further wrinkle in an object prized for its wrinkliness.

Burke’s penultimate chapter is titled “Crisis.” Over the past several decades, as contemporary art has come to occupy an ever larger slice of the art market, the need for attribution skills has diminished. He connects this to the art world’s global turn—the increased attention now being paid to art from other places—though it is more likely a result of the dearth of great old master material to sell. With so much of it already locked up in museums, any aspiring Isabella Stewart Gardner or Henry Clay Frick has to look elsewhere. In that way, the “crisis” is the product of connoisseurship’s past success.

Today science can answer many more questions about objects that are in need of attribution, though as any conservation scientist will tell you, their findings may be able to prove that something could not be from a given artist’s hand but not that it certainly is. Still, there are people out there, actively employed, with trained eyes and deep knowledge, making judgment calls about historic works of art. The preferred term now is “expert,” which feels less redolent of pompous chicanery and more inclusive of objective science: connoisseurs had loupes and dusty archives; experts have OptiVISORS and macro X-ray fluorescence scanning. (Hugo Chapman, the head of prints and drawings at the British Museum, once said he would “rather gouge my eyes out with a rusty penknife than describe myself as a connoisseur.”)

And then there is AI. Burke acknowledges that AI systems “may be able to carry out stylistic analysis more rapidly and perhaps more thoroughly than human traditional connoisseurs.” He quickly follows this observation by pointing out that those systems are trained on accepted attributions—mostly derived from human connoisseurship—and paraphrases the twentieth-century art historian Erwin Panofsky’s observation about X-rays: “We may say that AI…remains methodically no different from the use of spectacles or a magnifying glass.”

This is simply not the case. A magnifying glass (like an X-ray) enhances human vision; it doesn’t make judgments. Optically and conceptually it is transparent. Deep learning systems do make judgments, and those judgments arise not from following rules or criteria supplied by humans but from the system’s own internal representations—a “black box” that may be uninterpretable by us and unexplainable by the system. Berenson may have been unforthcoming, but at least he was operating with an anatomy familiar to the rest of us.

Burke’s swift dismissal of AI has the look of someone whistling past the graveyard. Having defined connoisseurship as “the analysis of style for the purpose of attribution,” he might feel that outsourcing attribution negates the whole enterprise. Yet in doing so, he bypasses the opportunity to think more abstractly as well as more pragmatically about the future of connoisseurship, close looking, the assignment of authorship, and the parsing of originality.

The goal of attribution, like that of reading an MRI or X-ray, is to arrive at a truth that exists independent of the act of looking. But connoisseurship was predicated on the act of looking as something of value in itself—an experience, a form of learning, a means of training the eye for both pleasure and purpose. Those things continue whether we win or lose at attribution. It has been a decade since an unassisted human Go player was able to beat DeepMind’s AlphaGo system, yet millions of people still play the game avidly. (Humans also continue to lift weights, run races, and participate in spelling bees.)

When it comes to the discussion of relative quality in art there’s no competition. Though LLMs can articulate (or at least feign) such judgments, they aren’t better at them than we are. Such opinions, even from the most eminent connoisseurs or critics, have never been absolute. They’re conversation starters, jumping-off points for looking harder, thinking harder. The grandees of the 1863 Paris Salon rejected Manet’s Luncheon on the Grass and gave a gold medal to Alexandre Cabanel’s soft-core, putti-packed Birth of Venus. We may think they were fools while also recognizing that periodic reshufflings of the deck help us see in new ways.

The traditional visual training that went into attribution was useful because it forced people to pay attention to the nonobvious—to bypass the storyline and attend to a knuckle, to ignore the flayed satyr and focus on the brushstrokes. That kind of close looking has not been central to art history education for decades. If you want to take the optimistic view, it is possible that AI systems could be useful sparring partners, expanding our ability to see what we’re looking at: when a system spits out an answer that seems counterintuitive to human experts, we will need to decide whether it is just wrong or is detecting real properties that traditional connoisseurship had missed.

The career of Christoph Jamnitzer previews a different set of concerns about AI. With its elliptical three-part argument for giving away his art while retaining credit, his Book of Grotesques closely echoes the sentiments of thousands of artists who upload their work to Instagram, hoping to generate a following while also hoping to prevent replication without their names attached. In addition to its other qualities, Jamnitzer’s book is a kind of corkscrew monument to a basic problem of intellectual property: like later copyright law, the print privilege treated creative expression as if it were a stable piece of physical property, but the natural behavior of ideas is to travel and proliferate. The poem announces his willingness to encourage that proliferation. It’s a tension that has only grown more pointed: today the equivalent of Jamnitzer’s privilege legally lasts seventy years past the death of the creator, even though practically everything is up for grabs the minute it enters the world.

It seemed only apt to ask Claude Sonnet 4.6 for its view of the situation. The anxiety today, it wrote,

is no longer whether a craftsman’s inventions will be copied by a rival goldsmith down the street, but whether they’ll be ingested wholesale by a system that has consumed the entire visual output of human history and can reproduce the style without reproducing any specific thing…. The privilege said: don’t copy this book. It had no language for: don’t learn from this book in a way that makes copying unnecessary.

Learning in a way that makes copying unnecessary, however, is exactly what good artists (and good forgers) always do—examine, ingest, reformulate. Raphael copied Michelangelo, Michelangelo copied the Romans, the Romans copied the Greeks, and everybody added something novel along the way. Jamnitzer had it right: art ideas have always been just mist—borderless, particulate, condensing here, dispersing there. One way to think about the history of connoisseurship is not as a history of recognizing originality but as one of identifying the pieces of originality that mattered to a particular group of people at a particular time. The same might be said of intellectual property protections. Insofar as AI necessitates a fresh reconsideration of these matters, it is a reminder that we have never really gotten off the wobbly bridge.