The Arts: Methods and Tools | Quomodo Scimus : IBTOK

areas of knowledge - the arts

Methods and Tools of the Arts

A conservator can now describe how Van Eyck built a jewel out of paint, and a scanner can show exactly what he painted underneath what he painted. Neither one tells you which finished object gets to count as the work.

The face nobody wanted found

THE PROVOCATION

For over sixty years, the Lamb of God had four ears, and almost nobody asked why. A partial cleaning in 1951 exposed a second set beneath the surface, and the anomaly simply became normal: an odd detail nobody investigated further until the current restoration team did.

In December 2019, conservators working on the Ghent Altarpiece finished removing centuries of overpainting from a single face: the Lamb of God, the small mystical animal at the centre of the whole composition. Infrared reflectography and X-ray imaging had already shown them what lay underneath. What they revealed was not a cleaned version of the face millions of visitors already knew. It was a different face, forward-looking, with strikingly human eyes, replacing the gentler, more animal profile generations had grown up with.

The reaction, once photographs of the restored Lamb went online, was immediate and largely hostile. People called the new face unsettling, and more than one commentator joked online that the Lamb now looked like it had seen a ghost. Restorers and art historians were not surprised by the science. Several had already argued that the sixteenth-century painter who added the earlier, more placid face had softened Van Eyck’s original because it unsettled contemporary taste even then. The tools revealed two historically real faces. Conservators then had to choose which one the public would encounter as the work.

That choice, more than any single fact the tools revealed, is what this page is about. The arts use a wider range of methods and tools than any other area of knowledge: a painter’s hand, a conservator’s scanner, a composer’s borrowed instrument, a camera, and now a machine that has never held a brush. What each of them can do, and what each of them cannot, decides not just how well we come to know a work of art, but which work we end up believing we are looking at.

This page follows three distinct, cumulative questions. Big Idea 1 asks what kind of knowledge artistic skill actually is. Big Idea 2 asks whether a technical tool can ever recover a work neutrally, without also remaking it. Big Idea 3 asks whether a work can be known at all apart from the place and history that made it. Each builds on the one before it: a maker’s own hand, then the instruments used to examine what that hand produced, then the place and history that gave the finished object its meaning in the first place.

This page is focused on one source far more than any other page on this site, and that deserves a word of explanation. I have taught the Ghent Altarpiece for decades in the history of the Renaissance, and I was living in Belgium myself when the restoration project began. It is due to finish later this year (2026). As I said in the very first sentence of the the very first lesson, this is my TOK website, and this page is where that shows most clearly.

Below is a photograph of the unrestored altarpiece (open view), I highly recommend a visit to the Closer to Van Eyck website to see the progress on the restoration.

Big idea 1 - Embodied skill - the craft nobody could fully explain

Van Eyck's procedural knowledge

The Ghent Altarpiece helped launch the Northern Renaissance through Jan van Eyck’s extraordinary handling of oil paint. Although Vasari wrongly claimed that Van Eyck invented the medium, he correctly recognised the novelty of its effects. Van Eyck layered costly pigments such as ultramarine and red lakes, then applied translucent oil glazes with exceptional precision. These layers reflected light like the surfaces they represented, creating the illusion that jewels, fabrics and flesh possessed real depth and material presence rather than being merely painted.

Ask any student in this course to draw a face from memory. Most will produce something recognisable as a face. Almost none will produce something that looks right, and few will be able to say exactly what is wrong with it or how to fix it. That gap, between knowing what a good likeness looks like and knowing how to make your own hand produce one, is procedural knowledge: knowing how, as distinct from knowing that. Drawing is the one art form every student in this course already has direct, personal experience of, whether or not it ever became a serious ambition, and the gap it exposes is the same gap the rest of this Big Idea follows through two very different, deliberately complementary cases.

But how is procedural knowledge taught or learnt? How did you learn to tie your shoe laces? How would you teach someone to tie their shoe laces without showing them how to do it? As a thought experiment, try to envisage the set of instructions you would write on how to tie shoe laces with words alone and no pictures.

Van Eyck's own skill did not appear from nowhere. Painters of his period normally learned through workshops and prolonged practical training, although almost nothing is known about where or under whom Van Eyck himself trained. The production of a masterpiece in. a guild is not an abstract theoretical exam. It is proof that the knowledge has been embodied. That kind of training is procedural in the fullest sense: judgement absorbed into the hand and the eye through repetition and correction, not knowledge that could simply be stated and handed over. This is exactly the shape of tacit knowledge Polanyi describes in Lesson 1, and the clearest modern example of that same kind of learning still working today is the guru-shishya parampara. What actually separates Van Eyck's case from Hindustani classical music is not that one worked alone while the other works together. It is that Van Eyck's particular chain of transmission broke. The guild that trained him is gone, and nobody alive learned his specific procedure from a living teacher in an unbroken line. The guru-shishya parampara's chain has never broken. One tradition's knowledge has had to be reconstructed centuries later from the object it produced. The other's is still being passed on exactly as Van Eyck's once was.

For centuries, later artists and historians could not fully account for how Van Eyck achieved his effects. His contemporaries suspected some secret alchemical process, and writing more than a century later, Giorgio Vasari went further still, crediting Van Eyck with inventing oil paint itself, a claim modern conservators have shown to be false, since oil paint existed long before him. What conservators at Cambridge's Hamilton Kerr Institute have since described, working directly on the panels, is a procedure rather than a secret: layer after layer of transparent oil glaze, using pigments including ultramarine ground from Afghan lapis lazuli and reds drawn from madder root or crushed cochineal beetles, built up to reproduce the way light behaves inside a real jewel or real skin, not just its outline or its hue. In their own words, the technique "isn't about capturing a perception with the brush... it's about recreating the object itself by simulating in paint the inner structure of its materials." Even now, exactly how Van Eyck judged the thickness and order of each glaze remains something conservators can describe only in general terms, not a formula they could hand a student.

The guru-shishya parampara shows the same kind of transmission still working, unbroken. In this teacher-student lineage at the centre of Hindustani classical music, a student learns a raga's specific ornamentation, timing and tone by listening to and imitating a teacher over years, not by reading a score. Standard notation can capture which notes to play. It cannot capture the microtonal bend of a phrase or the exact moment to hold a silence, which is precisely the part a student has to learn by ear and by repetition, in real time, from a living teacher, rather than by reading a recipe or reconstructing a lost one.

My Methods in Language page traces the same gap through another of my favourite artists, Dylan Thomas's own craft notebooks. Polanyi's claim from Lesson 1, that we know more than we can tell, is the reason all three cases belong on the same page: it is a fact about what skill itself is, not about painters, musicians or poets individually. None of this settles how much of a maker's skill could, in principle, be written down given enough effort, and how much genuinely cannot be. Van Eyck's procedure and the guru-shishya parampara sit at opposite ends of the same problem, one lost and painstakingly reconstructed, the other alive and still being passed on, and neither can be reduced to a formula a reader could simply follow.

Big idea 2 - Technical recovery - does the tool find the work, or make a new one

The sheep with four ears.

For over sixty years, the Lamb of God had four ears, and nobody investigated why. A partial 1951 cleaning had exposed a second set beneath the surface, and the anomaly simply became normal. In this extract from a 2019 lecture, Hélène Dubois, who directed the Ghent Altarpiece's restoration, explains how her team used X-ray fluorescence to confirm Van Eyck's original face underneath: human, forward-facing eyes, deliberately unsettling. Her reading is that Van Eyck wanted that stare to confront viewers with the sacrifice, not softened for comfort.

The Ghent Altarpiece’s current restoration campaign began in October 2012 and is still not finished. The first phase, running to 2016, treated the eight exterior panels, removing varnish that had yellowed and cracked over centuries. The second phase, running from 2016 to 2020, turned to the lower register of the open altarpiece, including the central Adoration of the Mystic Lamb panel itself. A third and final phase, covering the upper register, has been underway since 2022 and is expected to finish later this year, having already uncovered further sixteenth-century overpainting on the draperies of John the Baptist, Mary and the Godhead. Restorers used infrared reflectography to see through paint to the underdrawing beneath, X-radiography to reveal each panel’s full structure, and pigment analysis, including macro X-ray fluorescence and scanning electron microscopy, to identify individual pigments layer by layer, separating later additions from what Van Eyck himself had painted. Working on the exterior panels first, restorers discovered that around 70 percent of their visible surface had been covered by later overpainting, some of it added as early as the sixteenth century.

By the twentieth century, almost nobody realised how much of what visitors saw on the closed altarpiece was never Van Eyck’s at all. On the open section and the central panel of the Adoration of the Lamb, stripping back the overpainting on parts of the background townscape exposed areas where no original paint survived, leaving only the panel or its preparatory layers. There, the tools found nothing to recover, only the panel itself. Conservators filled and reconstructed the damaged architectural passages, drawing on the surviving fragments and on advice from historians and archaeologists specialising in medieval Ghent. See below the macrophotography of before, during and after the recent restoration. How much of what we now see was produced in the 21st century? To see for yourself go to the fantastic Closer to Van Eyck website.

Imaging of the central Adoration of the Mystic Lamb panel revealed evidence of an earlier composition beneath the Fountain of Life: a simpler natural spring, attributed to Hubert van Eyck. The spring does not appear clearly in any single scan. Researchers identified it by combining infrared imaging, X-ray fluorescence and close study of the paint layers. Jan later painted the monumental fountain over Hubert’s earlier design while completing the altarpiece. Conservators did not remove Jan’s fountain. The spring remains hidden beneath it and is known only through technical evidence and scholarly reconstruction. Here, new tools increased knowledge without changing what visitors see.

The Lamb of God’s new face, described in the provocation above, is the opposite case. There, conservators did decide to remove four centuries of overpainting and expose a face nobody living had ever seen in its original state. The same restoration project produced two different outcomes from the same kind of imaging. With the spring, knowledge stayed beneath the surface, available to scholars but invisible to any visitor standing in front of the panel. With the Lamb, the same kind of evidence led to a visible, physical change to what the public actually sees. Whether evidence should lead to that kind of intervention, or should stop at reconstruction and leave the visible work alone, is not something the tools themselves can answer. That decision is made collaboratively, by conservators and the advisory committees overseeing the project, not settled by the instruments that supplied the evidence for it.

The theorist Salvador Muñoz Viñas has argued that this whole way of talking about restoration does not survive close examination. Conservators often describe a treatment as having revealed an object’s "true condition," but that claim only makes sense if the object was already assumed to have a false condition to begin with. There is no neutral procedure that arrives back at a single, original truth waiting beneath the overpaint. What conservators actually do, on his account, is negotiate which version of an object will count as authoritative, weighing expert judgement against how the public and other stakeholders have already come to know the work.

They Shall Not Grow Old

Peter Jackson’s They Shall Not Grow Old transforms First World War footage through digital restoration, colourisation, reconstructed sound and corrected speed. The soldiers appear immediate and recognisably human rather than distant figures from damaged silent film. Yet much of this closeness is constructed. Where the original cameras recorded no colour, sound or continuous movement, technology supplies a plausible version. The film therefore raises a central question: does restoration recover the past, or create a convincing present-day interpretation of it?

Of course the role of preservation and restoration is not limited to paintings of the northern Renaissance. We have encountered similar issues in history, most notably with the use of digital technology in the restoration of old newsreel footage from World War 1. In Peter Jackson’s They Shall Not Grow Old, the soldiers no longer move in the jerky, distant way associated with early film. They appear in colour, at natural speed, surrounded by voices, footsteps and artillery. The war seems suddenly closer. But this closeness is partly constructed. Jackson’s team repaired damaged film, stabilised it and corrected its speed. They also added colour and sound, used lip-readers to reconstruct speech, and created intermediate frames that the original cameras never captured. Historical research made these additions plausible, but it could not recover information that had never been recorded.

Literature faces an equivalent problem. The manuscripts of T. S. Eliot’s The Waste Land preserve a much longer poem, heavily cut and reshaped by Ezra Pound. The drafts reveal an earlier stage of the work, just as imaging reveals Hubert’s spring beneath Jan’s fountain. But readers still have to decide which object counts as the poem: Eliot’s first draft, the version created through Pound’s intervention, or the whole history of revision. Shakespeare’s King Lear presents another case. It survives in two significantly different early printed versions. For centuries, editors combined them into one apparently complete play, even though Shakespeare may never have written that combined text. Like a conservator reconstructing missing paint, the editor used surviving evidence to produce a coherent whole that had not existed in precisely that form before.

Across the arts, restoration does not simply return a lost original. It produces a present version from surviving traces, specialist knowledge and informed judgement. The central question is therefore not only whether a reconstruction is plausible, but whether the audience can still see where the evidence ends and the reconstruction begins.

Big idea 3 - Reproduction, place and aura - can a work be known apart from where it was made

The Ghent Altarpiece was built to match a particular light. Every painted highlight and shadow, on the donors kneeling in the closed panels, on Adam and Eve in the open ones, on the jewels set into the angels’ robes, is painted as though lit from one specific direction: the actual angle of sunlight through St Bavo’s Cathedral’s own windows. One angel’s jewel reflects, in paint, the literal shape of the window that lights it. The altarpiece was designed to be inseparable from the one room it was built for.

Ghent Altarpiece - Aura

In this two-minute extract, the Ghent Altarpiece is shown reacting to real light: shadows and highlights on statues, donors, angels and jewels are painted to match sunlight through St Bavo's Cathedral's actual windows. Adam, lit from the left, has a highlight in his eye; the shaded Eve does not. One angel's jewel even reflects the exact shape of the window lighting it. The clip ends on the altarpiece's present location, removed from that chapel, now displayed elsewhere behind bulletproof glass.

Walter Benjamin wrote about this in his 1935 essay "The Work of Art in the Age of Mechanical Reproduction." By then, photography and cheap printing meant a painting could be reproduced thousands of times over, in books, on postcards, in newspapers, seen by far more people than could ever stand in front of the original. Benjamin's claim was not that reproduction shows a viewer less. A photograph can show more of a painting's surface, in sharper focus, than most visitors ever see with the naked eye. Just look at the images above. His claim was that something else is lost regardless of how much detail survives.

Even the most perfect reproduction of a work of art, he wrote, is "lacking in one element: its presence in time and space, its unique existence at the place where it happens to be," and it is exactly this, he argued, that "withers in the age of mechanical reproduction." He called it the work's aura. The Ghent Altarpiece tests this claim in an unusually literal way. Its panels were painted to react to one specific room's light, and a reproduction can now show a viewer every detail of that light, in perfect focus, on a screen anywhere in the world. What it cannot show is the altarpiece actually standing in St Bavo's, lit by St Bavo's own windows, which is the condition Benjamin thought its aura depended on.

I said this page was personal.

I have visited this memorial to Benjamin, Dani Karavan's Passages, at Portbou myself a number of times. This is the town where Benjamin died. A rusted steel corridor is cut into the cliff and descends toward the sea, ending at a glass pane looking straight out over the water he was trying to cross.

Standing in it, you understand something no book quite gets across: some places hold their history in a way no reproduction of them can, which is exactly Benjamin's own argument, tested here on the man who made it.

John Berger’s own version of argument shows why a photograph is not a neutral copy. Opening the first episode of Ways of Seeing with Botticelli’s Venus and Mars, he shows how a printed reproduction can isolate one detail from the whole composition and change what it means: "Reproduction isolates a detail of a painting from the whole. The detail is transformed. An allegorical figure becomes a portrait of a girl." The painting in the National Gallery still shows Mars asleep after Venus has exhausted him, an allegory about love disarming war. The same detail, printed alone on a page or a postcard, shows only a young woman’s face. A camera does not simply record a work. It crops, enlarges and relocates it, and each of those choices gives the image a new context and, on Berger’s argument, a new meaning.

Photography did something further still: it changed what painters believed painting was for. Once a camera could produce an exact likeness in a fraction of a second, painters had less reason to compete with it on those terms. Photography was one of several technological and cultural changes that encouraged painters to move away from detailed representation and investigate what painting could do that a camera could not. A group working in and around Paris in the 1870s began painting light, colour, movement and their own subjective percept

John Berger on Aura

In this four-minute extract from Ways of Seeing (1972), John Berger argues that a painting's meaning was once tied to one place: altarpieces belonged to their churches, icons to their shrines, seen only by those who travelled to them. Reproduction changes that. Using Botticelli's Venus and Mars and the example of the icon, Berger shows how a painting's image now travels everywhere, surrounded by our own rooms, while its meaning becomes, in his own words, transmittable, 'information of a sort.'

Rather than reproduce the harbour with photographic precision, Monet records a fleeting experience of mist, light and colour. Photography helped free painting from the task of exact representation, allowing artists to investigate how the world appears to a particular observer at a particular moment. A separate, humbler invention made the shift practical: the collapsible metal paint tube, patented in 1841, let painters carry ready-mixed, vivid colour outdoors for the first time, working directly from the changing conditions in front of them rather than reconstructing them from memory indoors. Whether a new tool that reproduces art only changes how existing work is known, or changes what artists believe art is for, is the question this Big Idea leaves open.

Methods and Tools in the Arts compared

If the tools that recover a work also decide which of several real paintings it gets to be, what is left of the idea of the original?

Natural Science also depends on instruments that see further than the human eye, infrared imaging, X-ray, mass spectrometry among them. There, an instrument’s reading is checked against a single physical fact it is trying to measure, a particle’s mass, a molecule’s structure, and better instruments converge on one answer. The tools used on the Ghent Altarpiece converged on several answers at once: Hubert’s spring, Jan’s fountain, four centuries of overpainting, all physically real paint on the same panel. A scientific instrument narrows down what is true. An art-historical scan can multiply the number of things that were once true.

History also argues over which layer of the record is genuine, and Trouillot’s four moments of silence, examined in Core Lesson 4, describe exactly how a later account can bury an earlier one without destroying it outright. A historian revising the archive adds a new interpretation alongside the old one, and both remain available to be checked against each other. Stripping four centuries of overpainting from the Lamb of God did something a historian’s revision cannot: it physically removed one real version of the object to expose another.

Mathematics is furthest from this problem, not closest to solving it. A proof holds or fails regardless of where it is written down or who is holding the page, and Gödel’s incompleteness theorem, on Methods and Tools in Mathematics, sharpens that standard rather than escaping it. The Ghent Altarpiece cannot be checked anywhere else. Its meaning, on Benjamin’s argument, depends on standing in one specific building, lit by one specific set of windows, which no proof would ever need.

Human Science raises a related problem from a different direction. Methods and Tools in Human Science argues that a subject who knows they are being studied can change simply by being observed. The tools examined on this page do something closer to the reverse: an X-ray or a solvent does not make the Ghent Altarpiece behave differently, but it can multiply how many different, physically real versions of the same object turn out to exist underneath the one you thought you were looking at.

Next: Ethics in the Arts

Think further: questions and resources

Van Eyck’s contemporaries assumed his technique was a kind of alchemy, and Vasari, writing decades later, credited him with inventing oil paint outright, which modern conservators say is simply false. If five centuries of expert study can still only describe what Van Eyck did rather than fully explain how he decided to do it, is that a gap science will eventually close, or a permanent feature of embodied skill?
The tools used to strip overpainting from the Ghent Altarpiece revealed at least three different real paintings on the same panel: Hubert’s original spring, Jan’s Fountain of Life, and centuries of later restoration. If recovering the earliest layer erases a version that millions of people actually venerated for four centuries, which layer has the strongest claim to being the work?
Walter Benjamin argued that reproduction strips a work of its aura, its unique existence at the place where it happens to be. The Ghent Altarpiece’s panels were painted to match sunlight through one specific cathedral’s windows, and the altarpiece no longer stands there. Does a copy occupying the intended space, rather than the real panels displaced from it, have a stronger or weaker claim to the work’s aura?
Photography took over the job of producing a likeness that painting had always done, and painters responded by abandoning likeness altogether. If a new tool can take over what an art form used to be for, does that free the art form, or does it just relocate the question of what counts as skill?
Berger argues that reproduction can isolate a detail from a painting and change its meaning entirely, turning an allegorical figure into a portrait of a girl. If the same reproduction technology can also reveal details of the Ghent Altarpiece invisible to any single visitor standing in St Bavo’s, does it produce new knowledge of the work, or a different work altogether?
Return to the provocation. The same scientific tools that revealed the Lamb’s original face are the tools generations of visitors trusted to tell them what they were looking at. If a public used to one face can be shown scientific proof that a different face is the real one and still find the new one harder to accept, what does that suggest about what knowing a work of art actually requires, beyond evidence?

Films
For more see my 10 films for the TOK journey page.

🎬 WATCH — Ways of Seeing Episode 1 (1972)

John Berger directed by Mike Dibb BBC

Berger shows how a printed reproduction can isolate a single detail from Botticelli’s Venus and Mars, the girl’s face, and strip it of the allegorical meaning it carries in the full composition: the direct source for this page’s Big Idea 3. A little later, roughly minutes seven to eleven, he turns to the National Gallery’s Virgin of the Rocks and the fourteen pages of its catalogue entry devoted to proving the painting genuine: the source for a different argument on Perspective in Arts, about the authority experts hold over what counts as a real work.

🎬 WATCH — Tim’s Vermeer (2013)

Teller

Tim Jenison, a Texas inventor with no painting training, spends eight years testing whether Johannes Vermeer used optical devices, a camera obscura and a comparator mirror, to achieve photorealistic effects a century and a half before photography existed. Whether Vermeer actually worked this way remains disputed among art historians. Jenison’s real contribution is to test this page’s Big Idea 1 question from the opposite direction: if a tool can substitute for the years of embodied training a painter’s guild would have demanded, is the resulting skill the same kind of knowledge, or a different one produced by different means? My students can see the film here.

🎬 WATCH — They Shall Not Grow Old (2018)

Peter Jackson