1

Independent Landscape

The first independent landscapes in the history of European art were painted by Albrecht Altdorfer. These pictures describe mountain ranges and cliff-faces, skies stacked with clouds, rivers, stream-beds, roads, forts with turrets, church steeples, bridges, and many trees, both deciduous and evergreen. Tiny, anonymous figures appear in some of the landscapes. But most are entirely empty of living creatures, human and animal alike. These pictures tell no stories. They are physically detached from any possible explanatory context – the pages of a book, for example, or a decorative programme. They are complete pictures, finished and framed, which nevertheless make a powerful impression of incompleteness and silence.

These landscapes are small enough to hold in one hand. They were meant for private settings. Extremely few have survived: two were painted not directly on wood panel like ordinary narrative or devotional paintings, but on sheets of parchment glued to panel; three others were painted on paper. Altdorfer also made landscape drawings with pen and ink, and he published a series of nine landscape etchings. Altdorfer signed most of his landscapes with his initials. He dated one of the drawings 1511 and another 1524, and one of the paintings on paper 1522. In Altdorfer’s time, drawings and small paintings were seldom signed or dated. A signed landscape, given its unorthodox content and ambiguous function, is doubly remarkable.

The Landscape with Woodcutter in Berlin, painted with translucent washes and opaque bodycolours on a sheet of paper 20.1 centimetres high and 13.6 wide, opens at ground level on a clearing surrounding an enormous tree (illus. 1). At the foot of this tree sits a figure, legs crossed, holding a knife, with an axe and a jug on the ground beside him. From the trunk of the tree, high above his head, hangs an oversized gabled shrine. Such a shrine might shelter an image of the Crucifixion or the Virgin; in this case we cannot say, for it is turned away from us. The low point of view and vertical format of this picture reveal little about the place. Is this a road that twists around the tree? Are we at the edge of a forest or at the edge of a settlement? The tree poses and gesticulates at the centre of the picture as if it were a human figure, splaying its branches into every corner. And yet it is truncated by the black border at the upper edge of the picture.

1 Albrecht Altdorfer, Landscape with Woodcutter, c. 1522, pen, watercolour and gouache on paper, 20.1 × 13.6. Kupferstichkabinett, Berlin.

2 Hercules Segers, Mossy Tree, c. 1620-30, etching on coloured paper, 16.9 × 9.8. Rijksprentenkabinet, Amsterdam.

To this iconographic austerity corresponds a great simplicity of means. Many effects are achieved with few tools. The picture is imprecise, shambling, genial, candidly handmade. The branches of the tree sag under shocks of calligraphic foliage, combed out by the pen into dripping filaments. This pen-line emerges from under layers of green or brown wash at the margins of the tree-trunk and the copse at the left, in licks and tight curls, legible traces of the movement of the artist’s hand. Contours do not quite coincide with masses of colour. Behind the tree-trunk, a row of trees was outlined in pen, but never coloured in. The brittle buildings tilt to the right. The picture is touched by a slightly manic spirit of exaggeration, which makes the woodcutter’s passivity and the silence of the place even more uncanny. These are rough-hewn and generous effects. A curtain of blue wash creates a breathy, sticky ether that fades into white near the profile of the mountains. The mountains – blue, like the sky – are overlaid with a ragged net of white streaks. Masses of leafage, by contrast, are made tangible by a speckling of bright green, so unreal that it conspicuously sits on the surface of the picture rather than in it. Paint adheres to paper like a sugary residue. All the picture’s technical devices are disclosed, yet they look unrepeatable.

Such a picture as Altdorfer’s Landscape with Woodcutter – as rich as a painting, as frank and gratuitous as a drawing – is not easily located either within an art-historical genealogy or within a cultural context. No Netherlandish or Italian artist of the early sixteenth century produced anything quite like it. Some German contemporaries and followers of Altdorfer did make independent landscapes. Wolf Huber of Passau, it appears, even established a kind of trade in landscape drawings. Altdorfer published some of his own ideas about landscape in his brittle, spidery etchings, and indeed through these etchings he exerted an impact that can be traced for several generations. But by and large this is not a success story. The independent landscape staked only the feeblest of claims to the surfaces of wood panels, which in southern Germany around 1500 were still largely the territory of Christian iconography. Nothing like Altdorfer’s mute landscape paintings, with their elliptical idioms of foliage and atmosphere, would be seen again until the close of the sixteenth century, in the forest interiors of Flemish painters working in the wake of Pieter Bruegel: Jan Bruegel the elder, Jacques Savery, Lucas van Valckenborch, Gillis van Coninxloo.1

Not until the tree studies of Roelandt Savery, Hendrick Goltzius, and Jacques de Gheyn II did anyone again twist a trunk or attenuate a branch with such single-minded fantasy.2 But the most apt comparisons of all are the moody etchings of Hercules Segers, an eccentric and still dimly perceived giant of early seventeenth-century Dutch landscape, pupil of Coninxloo and powerful inspiration to Rembrandt, working a full hundred years later than Altdorfer. Only a single impression survives of Segers’s Mossy Tree, an etching on coloured paper of a nameless, spineless specimen that might have been plucked from a fantastic landscape, perhaps from one of Altdorfer’s etchings (illus. 2).³ Indeed, the affinity between Segers’s prints and those of Altdorfer’s earliest followers was remarked on as early as 1829 by J.G.A. Frenzel, the director of the print cabinet in Dresden.⁴

Altdorfer’s independent landscapes, as strange and unprecedented as they were, left no discernible traces in contemporary written culture. They are mentioned in no letter, contract, testament or treatise. The earliest documentary reference to a landscape by Altdorfer dates from 1783; that picture has since disappeared. One of the surviving landscapes can be traced back to a late seventeenth-century collector, but no further. Thus we do not even know who originally owned or looked at these pictures.

What does an empty landscape mean? Christian and profane subjects in late medieval painting were often staged in outdoor settings. Many pictures juxtaposed and compared earthly and heavenly realms, or civilization and wilderness. Many stories about saints revolved around spiritual relationships to animals or to wilderness. German artists, and Altdorfer in particular, addressed these themes also through rejuvenated pagan subjects, such as the satyr, or through autochthonous mythical characters, such as the forest-dwelling Wild Man. Literary humanists – many of whom kept company with artists – debated the historical origins of Germanic culture and its peculiar entanglement with the primeval forest. But none of these themes is simply illustrated by Altdorfer’s landscapes. These pictures lack any argumentative or discursive structure. They make no move to articulate a theme. Instead, they look like the settings for missing stories.

Deprived of ordinary iconographic footholds, one might well wonder whether the early independent landscape is not simply an image of nature. But German pictures in this period, particularly small and portable pictures on paper or panel, had limited tasks: they told stories; they articulated doctrine; they focused and encouraged private devotion. To ask the landscape of the sixteenth century to be a picture ‘about’ nature in general is to impose a weighty burden on it. We would probably not think to do it had not some Romantics elevated the landscape into a paradigm of the modern work of art. Schiller, for example, expected the landscape painting or poem to convert inanimate nature into a symbol of human nature.5 The difficulty of the early landscape is that it looks so much like a work of art.

Both Leonardo da Vinci and Albrecht Dürer, Altdorfer’s older contemporaries, had something to say about nature. Nature for Dürer meant the physical world. This nature ought to curb the artist’s impulse to invent and embellish, to pursue private tastes and inclinations. ‘I consider nature as master and human fancy [Wahn] as a fallacy’, wrote Dürer.⁶ His erudite friend Willibald Pirckheimer, in a humanist ‘catechism’ of 1517, an unpublished and only recently discovered manuscript, numbered the ‘striving for true wisdom’ among the highest virtues: ‘Observe and study nature’, he advised; ‘inquire into the hidden and powerful workings of the earth’.⁷ Cognition of nature, in Pirckheimer’s catechism, became an ethical duty. In his book on human proportions, published posthumously in 1528, Dürer warned that ‘life in nature reveals the truth of these things’. Nature provided examples of beauty and correct proportions, and therefore the controlling principles of art: ‘Therefore observe it diligently, go by it, and do not depart from nature according to your discretion [dein Gutdünken], imagining that you will find better on your own, for you will be led astray. For truly art is embedded in nature, and he who can extract it, has it.’⁸ For Dürer, the study of nature was a discipline, and nature itself the foundation of an aesthetic of mimesis. ‘The more exactly one equals nature [der Natur gleichmacht],’ he stated in two different treatises, ‘the better the picture looks’.⁹ But Dürer never suggested that nature should become the basis for an independent landscape painting, and indeed he did not make any such objects. In his drawings he left first-hand accounts of his investigations into the morphologies of animals, plants and minerals. But these were private experiments and memoranda, not public works. His watercolour and silverpoint landscape views registered the impressions of a traveller. Dürer’s eye was too hungry, and often he did not spend enough time on the paper to finish the drawing. Instead, a change of sky or a suddenly perceived riddle of foliage would distract him from the discipline of pictorial organization, and the work would never be concluded. Water-colours like the Chestnut Tree in Milan¹⁰ or the Pond in the Woods in London (illus. 74) break off in the middle – appealingly, to our eyes. When Dürer ran out of patience, or light, he simply stopped adding layers of wash. Dürer was charmed by the random datum and the ephemeral impression. His Fir Tree in London, rootless, floats on perfectly blank paper (illus. 3).¹¹ Pedantry would have killed the tree, robbed the boughs of their spring and splay. Instead Dürer wielded the brush with a draughtsman’s agitated hand. He animated the fir with a mass of short, wriggling strokes overlaid on a base of yellow and green washes; the strokes escape around the edges like fingertips. Such a transcription of a particular tree manifests neither the vagaries of personal inclination nor the stable certainties of intelligible nature. Dürer’s watercolour studies, even if completed, were never offered to the world as pictures.

3 Albrecht Dürer, Fir Tree, c. 1495, watercolour and gouache on paper, 29.5 × 19.6. British Museum, London.

Leonardo, whose thinking was more involuted, escaped this impasse. Science or knowledge, for Leonardo, would reveal the principles and dynamic processes of nature – an ideal or ‘divine’ nature – while painting would accurately represent the visible data – the works of nature. But those principles of dynamic or generative creativity uncovered in nature could then serve as a model for the inventive and fictive capacities of the true artist. Knowledge of nature distinguished the artist not from the unreliable fantasist, as it did for Dürer, but from the merely reliable transcriber of the physical world. The painter who adjudicates among the data and recom-bines them in his work is ‘like a second nature’.12 Leonardo apostrophized the ‘marvelous science of painting’: ‘You preserve alive the ephemeral beauties of mortals, which through you become more permanent than the works of nature.’¹³ These thoughts anticipate the classic defences of poetry of the later sixteenth century, where fiction is prized over history precisely because fiction need not tell the truth. The painter, wrote Leonardo, can produce any landscape he pleases, for ‘whatever exists in the universe through essence, presence, or imagination, he has it first in the mind and then in his hands’.¹⁴ Such a landscape Leonardo might have painted! But he never did; that is, he never painted a mere landscape, one with a frame around it. Leonardo had new ideas about how to paint, not what to paint. He painted the same kinds of pictures that painters before him had: narratives, figure groups, portraits with landscape behind them. Leonardo’s drawings of plants and geological phenomena are even less like complete pictures than Dürer’s. A tree study at Windsor Castle, dating from c. 1498, a tightly controlled spray of dusty red chalk, illustrates a manuscript preparatory to an eventual treatise on painting (illus. 4).¹⁵ The specimen – a birch, or a locust? – supplements a terse set of observations: ‘The part of a tree which has shadow for background, is all of one tone, and wherever the trees or branches are thickest they will be darkest, because there are no little intervals of air.’¹⁶

4 Leonardo da Vinci, Tree Study, c. 1498, red chalk on paper, 19.1 × 15.3. Royal Library, Windsor Castle.

Albrecht Altdorfer, on the other hand, left no nature studies at all. For most painters in this period, there was really no need to venture out under an open sky. Apart from a few drawings in pen and ink, Altdorfer’s landscapes are indoor affairs. He was largely indifferent to the measurable or nameable attributes of the natural object. Sometimes, in the margins of his ordinary sacred paintings, Altdorfer described identifiable flora, for example plants with medicinal or symbolic significance.17 But so many of his trees are monsters or fictive hybrids.¹⁸ Moreover, he left no writings on nature or art; indeed, none on any topic at all. Nor is there any documentary or even circumstantial evidence that reveals what or whether Altdorfer thought about nature.

Altdorfer is known to us only through his works and through the silhouette of a public career. Like a number of other German and Italian artists of the period he managed to convert his talent into local social and political standing. He died in Regensburg on 12 February 1538, probably in his mid-fifties, at the crest of a highly visible and public career. His red marble tombstone in Regensburg’s Augustinian church, where he was provost and trustee, was recovered by accident in 1840 during excavations. The inscription described him not as a painter, but as the ‘honourable and wise Herr Albrecht Altdorfer, Baumeister’ (illus. 5).¹⁹ As the superintendent of municipal buildings he had overseen the construction of several commercial structures, perhaps even designed them; they all stand today. He also shored up the fortifications of the city. In his day Altdorfer was one of the outstanding political figures in Regensburg. He had sat in the Outer Rat or Council since 1517 and in the Inner Council since 1526, and had occupied various lesser public offices. In 1535 the city sent him as an emissary to King Ferdinand in Vienna. He owned three different homes, two of them at the same time, as well as vineyards outside the city.

What did Altdorfer look like? A portrait of an architect, formerly in Strasbourg but since destroyed, with an airy, capricious landscape behind him, was at one time judged to be a self-portrait by Altdorfer, or a portrait of him by a colleague.²⁰ But the only indisputable revelation of Altdorfer’s person is a painting on parchment by his pupil Hans Mielich (illus. 6). This painting is found on the second page of the Freiheitsbuch, a kind of constitution of the city, dated 1536, and depicts the assembly of the Inner Council on the occasion of the presentation of this very book to the mayor.²¹ Altdorfer is usually identified by his coat of arms, the fourth from the bottom on the left-hand side, as the figure in the black beret and fur-lined cloak. But in a chalk drawing by Joachim von Sandrart, the seventeenth-century academician and historian of German art, Altdorfer wears a forked beard.²² Sandrart had the drawing engraved and published together with the biography of Altdorfer in his Teutsche Academie of 1675 (illus. 7).²³ Both Sandrart’s drawing and a contemporaneous engraving by Mathias van Somer evidently derive from a common source, perhaps an old painted portrait belonging to the city. Sandrart was in Regensburg in 1653 and 1654 and later painted an altar for the abbey of St Emmeram. Mathias van Somer worked in Regensburg from 1664 to 1668. Their versions of Altdorfer’s head actually match a different councillor in Mielich’s miniature – the third from the bottom on the left, the one gesturing with his hands.²⁴ Finally, a man holding a scroll and looking out of the background of a panel of Altdorfer’s St Florian altarpiece, who had already struck some as a possible self-portrait, wears a similar forked beard.²⁵

5 Fragment of the tomb of Albrecht Altdorfer, 1538, marble. Stadtmuseum, Regensburg.

Altdorfer was not born to such status. Like Dürer, Lucas Cranach, Hans Burgkmair and Hans Holbein, Altdorfer was the son of an artist. Altdorfer won public attention not by undertaking huge and time-consuming public painting projects, as artists of his father’s generation might have done. Nor did he attach himself to a princely court, as Cranach did, although, like Dürer, he did do occasional work for the Emperor and other potentates. He eventually made a giant painted altarpiece and many large devotional panels. But at the start Altdorfer made his name with intimate, modestly scaled works in unconventional media and with eccentric subject-matter. He painted small, clever, devotional panels; he made finished drawings of profane and historical subjects; he published his ideas in engravings, woodcuts and etchings. No contemporary reactions to Altdorfer’s work are recorded. Yet we can infer that he appealed to a narrow and sophisticated audience with his esoteric subject-matter, his oblique treatment of traditional subjects, his visual wit, and his technical ingenuity and self-confidence. Most important, following Dürer’s lead, he signed and dated his works. Altdorfer put vivid testimonials to his own talents, linked by a distinctive style and signature to his person, directly into the hands of individual amateurs and collectors. His monogram, which mimics Dürer’s, was actually carved on the centre of his tombstone, no longer the mere corroboration of style, but an emblem in its own right.

6 Hans Mielich, Presentation of the ‘Freiheitsbuch’ before the Assembly of the Inner Council, 1536, miniature, from the Freiheitsbuch, 41 × 27.5. Stadtmuseum, Regensburg.

7 Philipp Kilian, Portrait of Albrecht Altdorfer, engraving after a drawing by Joachim von Sandrart, in Sandrart’s Teutsche Academie (Nuremberg, 1675). Houghton Library, Harvard University, Cambridge, Mass.

By the end of his career Altdorfer was dividing his time between politics, architecture and a few prestigious painting projects. The small and intimate works, including the landscapes, played an ambiguous role in his career, just as they did in Dürer’s. They were the foundation of his reputation, and they effectively preserved it, especially the small engravings with pagan or erotic subjects. Yet Altdorfer himself surely imagined they would, in the end, be overshadowed by his buildings and by his larger paintings. One of these late, public pictures is Altdorfer’s grandest and most celebrated work, and indeed one of the most remarkable of all Renaissance paintings, the Battle of Alexander made for Wilhelm IV, the Wittelsbach Duke of Bavaria, in 1529 (illus. 8).²⁶ Over the next decade Wilhelm assembled a cycle of eight paintings of Antique heroes, all in vertical format; they hung in the Residenz in Munich together with a cycle of heroines in horizontal format. Wilhelm drafted all the outstanding Bavarian and Swabian painters of the day; Altdorfer and Hans Burgkmair were the first. The commission was so important to Altdorfer that he declined a term as mayor of Regensburg in order to complete it. The picture represents the victory of Alexander the Great over the Persian monarch Darius at Issus in Asia Minor in the year 334 BC, as recounted by the first-century historian Curtius Rufus. The texts inscribed on the floating tablet and in various banners lofted by the armies were very likely composed by Aventinus, the Bavarian court historian, who was resident in Regensburg from 1528. The lower half of the picture describes Alexander’s pursuit of Darius’s chariot, slicing a path through a bristling, twitching carpet of soldiers. Darius throws a panicked glance backward, exactly as he does in the famous floor mosaic from the Casa del Fauno at Pompeii, which was not to see the light of day for another three hundred years (illus. 10). The upper half of the Battle of Alexander expands with unreal rapidity into an arcing panorama comprehending vast coiling tracts of globe and sky (illus. 9). It is as if the momentous collision of armies exploded outward into three dimensions, and was only then projected back onto a planar surface. For the picture represents an historical pivot, the expulsion of a force from the East, a muscular resistance to the left-to-right flow of history. The sun outshone the moon, just as the Imperial and allied army successfully repelled the Turks – massed under the sign of the Crescent – from the walls of Vienna in October 1529, the very year this picture was painted.

8 Albrecht Altdorfer, The Battle of Alexander and Darius at Issus in 334 BC, 1529, oil on panel, 158 × 120. Alte Pinakothek, Munich.

9 Details from The Battle of Alexander.

10 Details from The Battle of Alexander.

The Battle of Alexander is the foundation of Altdorfer’s modern fame. As late as 1740 a German writer attributed the painting to Dürer.27 But in 1800 Napoleon’s Rhine Army carried it from Munich to Paris, and the work and its real author entered art history. In 1814 the victorious Prussian troops who commandeered Napoleon’s residence at Saint-Cloud as their headquarters supposedly found it hanging in the Emperor’s bathroom.²⁸ Napoleon must have discovered in Altdorfer the same sincere vigour that he savoured in his favourite poet, another ‘northern Homer’, the supposed primitive, Ossian. In the summer of 1804, the Romantic essayist and critic Friedrich Schlegel, one of thousands of post-war German pilgrims to the Louvre, saw the Battle of Alexander and marvelled. ‘Should I call it a landscape, or a historical painting, or a battle piece?’, Schlegel wondered. ‘Indeed this is a world, a small world of a few feet; immeasurable, uncontrollable are the armies which flow against each other from all directions, and the view in the background leads to the infinite. It is the cosmic ocean . . .’. He ended by pronouncing the picture a ‘small painted Iliad’.²⁹

Schlegel’s question introduces one of the recurrent themes of landscape painting in the West: the pretension to epic. Landscape painting magnifies outdoor setting at the expense of subject-matter. But as outdoor space expands, it offers itself again as the setting for still grander subjects. The far-flung spaces of the Roman ideal landscapes of the seventeenth century, whose leading painters were Claude Lorrain and Nicolas Poussin, pulse with the aura of their mythological and historical subjects, no matter how minuscule those subjects appear. The rhythmic landscapes of Rubens portend vast narratives even when those narratives are physically absent. The give and take between the intimate and the epic landscape is encapsulated already in Altdorfer’s career. Altdorfer’s forest landscapes, such as the Berlin watercolour of the single tree, generate mood by relaxing temporal structures. They distil mood from narrative. Altdorfer’s empty landscapes carried to a logical conclusion the pattern of extreme negligence towards narrative already manifested in a picture such as the St George and the Dragon in Munich, dated 1510, where man and monster are dwarfed by feathery bolsters of foliage (illus. 96). But even within Altdorfer’s work, the peculiar achievements of these pictures – their greenness, their textures – were constantly being absorbed back into epic, where they entered into restive tension with story and character. Landscape pressed itself into the foreground, literally or thematically, in all the largest and most imposing panels of his middle years: the lurid, nocturnal outdoor scenes from the Passion cycle of the St Florian altarpiece (illus. 50); the Crucifixion in Kassel, with figures like implacable totems before an immense, bulging panorama (illus. 11); the Two St Johns in Regensburg, rooted in the earth like fleshy fungi (illus. 12); the billowing, overgrown Christ Taking Leave of His Mother in London (illus. 13). The trifling, intimate, hand-held landscape can hardly stand comparison to such great-souled paintings as these. And yet, once planted within an epic, landscape could surreptitiously eat away at it from the inside, until the day when Schlegel could momentarily wonder whether the Battle of Alexander was actually a ‘landscape’.

The purpose of this book is, in effect, to agree with Schlegel that it is not. In his true landscapes Altdorfer isolated certain pictorial qualities associated with the settings of narrative and made them the stuff of their own pictures. He prised landscape out of a merely supplementary relationship to subject-matter. This moment of isolation tends to get obscured when it is assumed that it was nature that furnished the principle of isolation. For a picture that retains nature while rejecting its antithesis, culture, can choose either to include or to exclude the human figure. The human figure need not disrupt nature; it can participate in and imitate natural processes, or it can share the painter’s and the beholder’s openness to nature. As a result, landscape in the West has been conspicuously tolerant of insect-like staffage or ‘filler’ figures, of embedded ‘beholder’ figures, and even of stories. In some cases the artefacts of culture themselves are assimilated to nature and thus granted safe passage within the landscape picture. Ruins or cemeteries in Ruisdael, for instance, or a watermill in Constable, are treated as if they were nature. In this tradition, therefore, it has not been strictly necessary to exclude subject-matter in order to win a reputation as a landscape painter.

Altdorfer’s principle of exclusion was not the divide between nature and culture, but rather the divide between setting and subject. The emptiness of his landscapes is the hole once filled by the acting human figure. The clue is their verticality: these pictures derive structurally from small pictures with narrative and hagiographical subject-matter.

To recover a sense of the radicality and open-endedness of Altdorfer’s invention, one needs to abandon any presumption of a pre-existent idea about nature, of a Naturgefühl or ‘feeling for nature’, or, indeed, of any primary experience of nature that his landscapes were meant to preserve. The idea of nature is so protean and problematic that a hypothesis about the origins of landscape that can afford to set it aside automatically becomes, by any standard of theoretical parsimony, highly attractive. But this is not so easy to do. Any interpretation of Altdorfer is burdened by the dense and prestigious later career of landscape painting. For many historians who were witnessing the closing phases of this career, landscape painting had come to compensate for the great dissociation of mind and nature which modernity seemed to suffer under. Landscape in the West was itself a symptom of modern loss, a cultural form that emerged only after humanity’s primal relationship to nature had been disrupted by urbanism, commerce and technology. For when mankind still ‘belonged’ to nature in a simple way, nobody needed to paint a landscape. When poets, wrote Schiller, can no longer be the custodians of nature, ‘and already in themselves experience the destructive influence of arbitrary and artificial forms, or have had to struggle against that influence, then they emerge as the witnesses and the avengers of nature. Either they are nature, or they will seek lost nature’. Landscape painting restored, momentarily, an original participation with nature, or even – in its greatest Romantic apotheoses – re-established contact with the lost sources of the spiritual. For Schiller, treading the fragile bridge between aesthetics and morality constructed by Kant in the Critique of Judgement, it was actually better to search for nature than to possess it. ‘Nature makes [the poet] one with himself; art divides and splits him in two; through the Ideal he returns to oneness . . . The goal which man strives after through culture is infinitely preferable to that which he arrives at through nature.’30 This potent model of cultural loss and redemption through art governed the Romantics’ perception of their own situation. Moreover, it went on to pervade interpretations of the Renaissance. For a number of modern art historians, who came to see the Renaissance as a kind of prefiguration of the Romantic crisis, Altdorfer straddled a threshold of intellectual history. His works preserved a twilit image of a pre-modern animated cosmos, an organic totality infused with a divine spirit or linked by correspondence with a divine macrocosm.³¹ There was indeed plenty of interest in such a cosmology in Altdorfer’s time. Versions of pantheism, with ties to pagan cults, neo-Platonism and the occult, simmered in the writings of natural philosophers like Paracelsus, literary humanists like Conrad Celtis, and even radical Protestants. Altdorfer’s landscapes have thus been understood as the belated epiphanies of a lost unified consciousness, talismans charged with authentic prelapsarian energy, perhaps demonic, perhaps simply pious and popular, or völkisch.

11 Albrecht Altdorfer, Crucifixion with Virgin and St John, c. 1515, oil on panel, 102 × 116.5. Gemäldegalerie, Kassel.

12 Albrecht Altdorfer, The Two St Johns, c. 1515, oil on panel, 173.2 × 233.6. Stadtmuseum, Regensburg.

Sociological interpretations of the origins of landscape have been aimed directly against this sort of neo-Romanticism. Such interpretations dismiss the reunification with nature proposed by landscape painting as an aesthetic fiction. Art, in other words, should not deceive us into thinking we are actually making contact with nature; in pictures the culture merely constructs a nature. But materialist art history is often tinged with its own nostalgia, not for the pre-modern and pre-technological art of nature, but for a pre-bourgeois life of nature. Matthias Eberle, for example, following Schiller, Marx and Adorno, sees the Renaissance landscape as the expression of an urban or bourgeois consciousness of a new distance and detachment from the land. ‘As long as nature stands in direct connection with human life’, he explains, ‘the connection between nature and human history does not need to be questioned’.32 Landscape is the result of the subject’s conversion of nature into an object. Eberle’s own piety for nature remains intact. He accepts the reality of the original pre-modern oneness with nature, and thus laments the alienation that drove the painter to paint a landscape.

13 Albrecht Altdorfer, Christ Taking Leave of His Mother, 1520(?), oil on panel, 141 × 111. National Gallery, London.

To dispel these various nostalgias for nature, one must preserve the critical insight that pictures themselves actually generate ideas about nature, and at the same time refrain from dismissing that ‘secondary’ image of nature as the phantasm of the aesthetic ideology. There is no use searching for nature either among the relics of a pristine and more authentic art, or in sensual and empirical life. Once the process of figuration is uncoupled from a prior natural object, one in fact loses confidence that nature will ever manifest itself anywhere but in that figure. Otto Benesch raised this possibility, an important complication of his own neo-Romantic argument about Altdorfer’s cosmology, with the phrase malerische Welteroberung (‘conquest of the world through painting’), which suggests the active part that picture-making took in the genesis of ideas about the cosmos. The phrase All-Belebung des Bildorganismus (‘total enlivening of the picture-organism’) in turn implies that the concept of the organic unity of the physical world was encouraged by the analogy with the pictorial composition, with the formal coherence of the depicted landscape.33 Equally, one could argue that the vigour and elusiveness of the artist’s calligraphic style sharpened perceptions of the dynamic and the kinetic in the physical world. When Benesch, and following him Franz Winzinger, wrote that the early landscape transformed the beholder’s religious piety into piety for nature, they implied that the rhetoric of the picture itself was capable of supplanting Christian iconography and installing a new subject.³⁴

This hyperbole should not obscure the essential insight into the apriority of figuration that lies behind it. Ernst Gombrich, both in Art and Illusion and in his influential essay on Renaissance landscape, evoked with uncommon pungency the power of pictorial formulations to shape perceptions of real landscape.³⁵ The Venetian Lodovico Dolce attested to the overwhelming power of the image to displace the perception of the original, when he described a patch of landscape in a picture by Titian – adapting a commonplace of ancient writing on art – as so good that ‘reality itself is not so real’.³⁶ More than a century later the French critic Roger de Piles observed that the bad habits of painters ‘even affect their organs, so that their eyes see the objects of nature coloured as they are used to painting them’.³⁷ Verbal representations of landscape in the Renaissance, too, can hardly be trusted as simple registers of optical experience. Many apparently straightforward landscape descriptions were shaped by earlier literary treatments or by classical literary topoi. This has been amply demonstrated, for example, for the landscape descriptions of Aeneas Silvius Piccolomini (Pope Pius II), which figured so memorably in Jakob Burckhardt’s account of the Renaissance ‘Discovery of the World’.38 Even Columbus’s simple descriptions of the landscape of the New World, which Tzvetan Todorov read as testimony to an ‘intransitive’ admiration of beauty,³⁹ were moulded and constrained by an ideal of idyllic or paradisaical beauty.

The present study is aimed directly at this gap between the physical world and its image, between place and picture. In this respect it differs from some other accounts of the rise of landscape in the Renaissance, such as Otto Pächt’s derivation of the Franco-Flemish calendar illustration out of the northern Italian nature study.⁴⁰ Altdorfer did not ground his landscapes in any fresh scrutiny of natural data. Instead, he worked from the frame inward. His landscapes began with strong gestures of style – flamboyant strokes of the pen, ostentatious inflammations of colour, extravagances of scale. This interpretation of the pictures will in turn begin with an historical account and localization of these gestures.

Svetlana Alpers in her extraordinary study of the visual culture of seventeenth-century Holland, The Art of Describing, showed how pictures peculiarly cast the world.⁴¹ Any historical analysis of a pictorial culture, she argued, needs to be routed through an understanding of the special conditions of picture-making. The Northern, ‘descriptive’ painter in particular insisted on freedom from the expectations and exigencies of texts and their readers. Altdorfer certainly resisted texts, if anything even more so than Alpers’s Dutch painters a century later. His work is equally poorly served by hermeneutic methods developed for the study of Italian narrative art. But Altdorfer was not a descriptive artist either. He did not seek meaning in the surfaces and textures of objects around him. Like many German painters he had another ambition: self-manifestation in the picture, through idiosyncratic line and generally through sharp-edged disregard of the criteria of optical verisimilitude. Michael Baxandall initiated one line of historical inquiry into the German ‘florid style’ in his Limewood Sculptors of Renaissance Germany.⁴² Baxandall compared the sculptors’ gratuitous elaboration of drapery lines to the preoccupation of the Meistersinger with melodic inventiveness and originality. The stylistic ambition, which almost literally ‘flourished’ in the decades immediately preceding the German Reformation, in the generation of Dürer and Altdorfer, cuts across the dichotomy of description and narration. It reveals another dimension of Northern art altogether. Description and narration are terms derived from Neoclassical literary theory.⁴³ Neither mode acknowledges, or has any particular use for, the so-called ‘deictic’ utterance, the utterance that calls attention to the circumstances of its own physical production. In linguistics, a deictic sign is a grammatical element (an adverb, a verb tense) which places the source of the utterance in some spatial or temporal relationship to the content of the utterance. A fictional character or a narrator engages in deixis by embedding information about the context of a statement within the very form of the statement.⁴⁴ Norman Bryson has complained that modern professional art history, uncritically acquiescent to received Christian and Neoclassical theories of the image, ignores the deictic reference.⁴⁵ Those theories, he argues, were excessively concerned with an image that simulated optical impressions of space. They legislated an effacement of the picture surface, and then set the fictive space to the task of narration. Bryson proposed, and has practised, an alternative history of the somatic gesture preceding the picture: painting as a performing art, as it were. But Bryson overstated his grievance. It is true that art-historical scholarship has been excessively enthralled to the theologians and the rhetoricians. The paintings themselves, on the other hand, are not such assiduous propagandizers against the deictic interpretation. Nor has the best writing on Northern art been negligent. The ambitious German Renaissance artist deliberately packed his pictures with deictic markers. And no theorist was exhorting Altdorfer to abolish the material picture surface.

14 Albrecht Altdorfer, Castle Landscape, c. 1515, pen and pale red ink on parchment, from The Prayer Book of the Emperor Maximilian, size of page 27.7 × 19.3. Bibliothèque Municipale, Besançon.

The project of distinguishing a Northern from a Mediterranean practice of painting provides one historiographical framework for this study. Another is the recently renewed recognition of the persistence and charisma, in late medieval and Renaissance painting, of the icon. The Christian imago, the portrait of the sacred personage, was the institutional source and structural model of every independent painted panel. In a series of radically original inquiries into fourteenth- and fifteenth-century panel painting, Hans Belting has explicated the rhetoric of the early modern image as a set of responses to the prescribed and remembered functions of the Christian image. In an essay on Giovanni Bellini’s Pietà in the Brera, for example, Belting shows how the medieval functional and formal categories imago and historia became the field for an overwrought professional rivalry among the outstanding painters of the late fifteenth century.46 Bellini and Mantegna were competing for the attention of the most exacting northern Italian collectors. Belting recapitulates the successive fine-tunings of the pictorial categories: garrulous narrative was trimmed to an iconic core; the imperative and static image was then in turn reanimated in order to pull it back from iconicity. Out of this dialectic emerged a class of pictures whose consanguinities with the most ancient Christian images were at once more conspicuous and less binding than ever before.

A parallel track has been laid out by Joseph Leo Koerner in his work on Dürer’s frontal Self-portrait in Munich.47 Koerner shows how immemorial Christian concerns about the authenticity of sacred portraits framed that portrait, ceremoniously inscribed and dated 1500. Frontal portraits of Christ, in medieval tradition, all derived from an authentic original fabricated by direct physical contact with Christ’s face. All such portraits shared in the prestige of the acheiropoeton, or the image made ‘without hands’, without the unreliable intervention of a human artist. Dürer thus commits a momentous double hubris: in the Self-portrait he places his own face, clear-eyed and bearded, in place of Christ’s; and he competes with the perfection of divine fabrication by scrupulously concealing the trace of his own hand in the paint surface. Paint in this picture is a perfect match for the textures of hair, skin and fur; there is no visible remainder that can be attributed to ‘style’. But this tour de force of self-concealment, which explicitly refers to the virtuosity of the Netherlandish pioneer of oil technique, Jan van Eyck, is itself the most spectacular possible boast and self-advertisement. Dürer submerges the trace of his brush hand; but then he raises his left hand – which is reversed in the mirror and thus looks like the right hand in the finished self-portrait – into the picture field, an emblem of the real subject of the portrait, the source of Dürer’s fame, his performance.

The icon held the Christian beholder in a dialogue. It placed the beholder before the image of an other. The independent painted panel, descendant of the icon, became a principal venue for the Northern artist’s presentation of self because it was the place where the beholder was most keenly and unavoidably aware of the artist’s presence. That presence was always threatening to interrupt the dialogue between mortal and immortal. Thus the history and structure of the independent panel, and the theology of the Christian devotional image, become major contexts for Altdorfer’s landscapes.

It is not difficult to situate Altdorfer within a contemporary discourse about the function and structure of the devotional image. He was living through the most tempestuous crisis of the Christian image since the eighth-century Byzantine iconoclasm. The utility and propriety of traditional religious imagery had become the subject of vituperative public debate in southern German cities. Like other German artists, Altdorfer was pressed into a political role. In 1519 and 1520 he supervised the production of devotional images commemorating a miracle performed by the Immaculate Virgin in Regensburg. The images became the focus and the fuel for a massive pilgrimage. Thousands worshipped and purchased Altdorfer’s woodcuts and metal badges. The credulity and superstition of these pilgrims, and the fervid commerce of images, appalled many observers and helped to shift opinions in the direction of Lutheran and Zwinglian reform. The first iconoclasm provoked by the Reform took place in Wittenberg in 1521; many similar episodes, orderly and violent, followed. And by the 1530s, long after the Marian pilgrimage had been discontinued, Altdorfer as a city councillor was actually encouraging the introduction of Lutheranism to Regensburg. Traditional Christian iconography was eventually either discredited or placed on the defensive in southern Germany. It was left to the secular works – pagan and historical subjects, portraits, so-called ‘genre’ painting, and landscape – to assert the innocence of painting.

The independent image was thus by no means to be taken for granted. The professional painter trying to pursue a career in this stormy climate had to secure his audience, and one way of doing this was to address that audience directly. Style – the heteroclite and unrepeatable mark – fixes the beholder’s attention. The deictic trace implores the beholder to stay within the frame, to resist turning elsewhere for a narrative or a context that will justify the picture. This is essentially what Altdorfer was doing in his own time: authenticating the image with a rhetoric of personality that was only legible within the image. With his style he began to shift the project of self-presentation away from a sheer display of technical prowess towards a more complex manifestation of interpretative nuances: the capacity to make fictions, the assumption of a distinctive authorial tone. The act traced by style became an act of judgement as much as a physical stroke. But style still pointed back to a moment of execution, and the pointer has stuck to the work. Altdorfer’s most aggressively stylish pictures, including the independent landscapes, have also proved unusually intractable to historical explanation. Iconography, format and function of a picture are more manifestly susceptible to external or material pressures. The artist’s characteristic contribution to the picture, on the other hand, is irreducible and exempt from any possible causal explanation. Casting that contribution as a somatic gesture rather than as an intellectual design does not simplify matters. It is style that makes Altdorfer’s landscapes so distinctive in tone and texture, so different from anything else around them, and indeed so different from one another. Most contextual interpretations – sociological, for example, or intellectual-historical – have quickly run up against their own limits.

This book will instead direct questions from within the history of German painting, and particularly from within Altdorfer’s career. It begins by asking where landscape belonged physically and what formats and forms landscape took. It then asks what a picture was and how it could ever be self-sufficient; it tries to explain the structural and evolutionary relationships between independent landscape and pictures with subject-matter; it addresses the connection between the incipient category ‘work of art’ and a stable concept of the frame. All this ‘philological’ groundwork is laid out in the rest of this chapter and in the next. Chapter Three reads the paintings and watercolours as representations, carried out on more than one semantic level, of a native or regional landscape. The forest appears in these pictures simultaneously as a potential refuge from orthodox but idolatrous cult practices, and as the mythical setting of heathen idolatry. The landscapes themselves, finally, refer formally to older Christian images and thus complicate any claim that landscape as a genre may be making to doctrinal innocence. Chapter Four, on the pen-and-ink landscape drawings by Altdorfer and his colleague Huber, analyses the tension between the task of topographical description and the stylistic and fictional imagination. These drawings, but even more so the etchings discussed in chapter Five, mark the establishment of a secular culture of amateurism and collecting. Altdorfer’s landscape etchings were mechanical reproductions of pen drawings. The trace of the artist’s hand is visible but no longer present in the work. Moreover, the etchings reintroduced, through their horizontal format, an epic dimension to landscape. These etchings were the only real bridge linking Altdorfer’s experiment to the next generation of German landscapists, and ultimately, through Bruegel, to the future of landscape.

Where landscape could appear

‘Independent’ is really a negative description: it tells us more about what Altdorfer’s landscapes are not than about what they are. The independent landscape is, first of all, a complete picture not physically connected to any other picture. It is neither an element of a decorative scheme, such as the painted wall of a villa, nor part of an illuminated manuscript. It appears neither on the shutters of an altarpiece or a portrait, nor as the mere background of a narrative composition or a portrait. The independent landscape makes a clean break with the topological conventions of the dependent landscape. But it is not an unpredictable break: it follows already visible fault lines.

15 Albrecht Dürer, Shepherds Making Music, miniature on parchment, pasted into a printed edition of Theocritus, Venice, c. 1495-6, size of printed page 31.1 × 20.2. National Gallery of Art, Washington, DC, Woodner Collection.

Landscapes in the late Middle Ages appeared above all on walls, either in tapestries or mural paintings. By the fifteenth century painters could confirm the antiquity of the practice by consulting Pliny’s Natural History. The first-century Roman fresco painter Studius (or Ludius) had painted – in Lorenzo Ghiberti’s paraphrase – ‘landscapes [paesi], seas, fishermen, boats, shores, greenery.’48 But nearly all the works inspired by such texts have perished. An inventory of 1492, for example, reveals that the Medici owned some large Flemish paesi on canvas.⁴⁹ But tapestries, paintings on cloth, and domestic frescoes were particularly vulnerable to depredations of climate and shifts in taste. This is a chapter of the history of European art that will never be written.⁵⁰ Among the best surviving mural landscapes are the festive and hunting scenes at Runkelstein (Roncolo) near Bolzano, and the cycles of the months in the Torre Aquila at Trent, both dating from c. 1400. The frescoes at Trent, visible through an illusionistic loggia, were commissioned by Bishop George of Lichtenstein, probably from a Bohemian painter (illus. 16).⁵¹ Horizons are kept high to make room for as much earthly business as possible: mowing, raking, scythe-sharpening, fishing, love-making, all in the month of July. Hartmann Schedel, the learned author of the Nuremberg Chronicle, once described a series of murals he saw in a monastic library in Brandenburg in the 1460s.⁵² A scene labelled ‘Agriculture’ looked like a Garden of Love; ‘Hunting’ was set in a dense grove; both scenes fell under the category of Mechanical Arts. The pretext of these long-lost frescoes was didactic, but the treatment, according to Schedel’s ekphrases, was highly pictorial. Swiss and Alsatian tapestries, meanwhile, described the domestic and public misadventures of the sylvan Wild Man.⁵³ All these scenes retained some narrative or allegorical dimension. They described social types from opposite ends of the spectrum of wealth and elegance, and they described human activities, both work and leisure. Good painters did not disdain the decorative landscape. Vasari described the loggia that the Perugian Bernardino Pinturicchio painted for Pope Innocent III in the 1480s as tutti di paesi. . . alla maniera de’ Fiamminghi, ‘which, as something not customary until then, pleased very much.’⁵⁴ Some walls were simply decked with trompe-l’oeil foliage. Leonardo himself, in 1498, painted a pergola of tangled vines and branches on a vaulted ceiling in the Castello Sforzesco at Milan; on the wall he exposed a miraculous cross-section of rocky subsoil intercalated with roots and a hollow-eyed cadaver.⁵⁵

16 Bohemian master(?), July, c. 1400, fresco, Torre Aquila, Castello del Buonconsiglio, Trento.

Landscapes in illuminated manuscripts usually replicated the types and themes of the wall scenes. But others were truly empty. This extraordinary material has never been properly surveyed. Already in the 1320s the Parisian miniaturist Jean Pucelle had evacuated his spare, schematic landscapes on the calendar pages of a Book of Hours, illustrating the passage of the Seasons not by the cycle of human labours, but by the clothing of the earth itself.56 A Bruges master, perhaps the Master of Mary of Burgundy, painted in a Book of Hours of around 1490 a series of rural landscapes with signs of the zodiac in the skies.⁵⁷ A rural landscape is painted on the first page of a fragmentary Prayer Book in the manner of Alexander Bening from the early 1490s.⁵⁸ The Prayer Book of James IV of Scotland in Vienna contains twelve half-page landscapes, again with zodiac signs, executed by a painter of the Ghent–Bruges school between 1503 and 1515.⁵⁹ Otto Pächt has traced at least some aspects of this sudden Franco–Flemish fascination with the earth’s surface back to naturalistic northern Italian illustrations of herbals and bestiaries.⁶⁰

Manuscripts furnished various pretexts to paint empty landscapes. In Bartholomeus Anglicus’s De proprietatibus rerum and other high-medieval encyclopaedias, the sections on materials and elements, birds, trees, the seas, the earth and the provinces of the world were frequently illustrated with schematic landscapes. Many of these were carried over into the illustrations of the earliest printed versions, like the woodcut from the section of Bartholomeus Anglicus on ‘Water and its Ornaments’ in an English edition of 1495 (illus. 17).⁶¹ Herbals or medical handbooks also depicted depopulated places. A manuscript Tacuinum sanitatis in Liège, originally produced in Lombardy in the late fourteenth century, offers a spare and empty landscape under the heading ‘Snow and Ice’, alongside the customary wheatfields and orchards.⁶² In topographical treatises and travel descriptions, columns of text were increasingly interrupted by pictorial maps. The most spectacular were the woodcut urban panoramas in Bernhard von Breydenbach’s Peregrinatio in Terram Sanctam, printed in Mainz in 1486, and in Schedel’s Nuremberg Chronicle of 1493, with 1,809 woodcuts printed from 645 different blocks (illus. 36).⁶³ Some of the simplest topographical woodcuts were also the most picturesque, for instance the view of Castle Langenargen on the Gaisbühel, recently built, in Thomas Lirer’s Swabian Chronicle, published in Ulm in 1486 (illus. 18).⁶⁴ Around 1508–9 Hartmann Schedel pasted a miniature painted in bright matt blues and greens into his manuscript copy of the pilgrimage report of the Dominican Felix Fabri. Framed with red bands and captioned in the margins by Schedel himself, the miniature portrays Mounts Horeb and Sinai in the Holy Land (illus. 19).⁶⁵

17 English master, On Water and its Ornaments, woodcut, 15 × 15, from Bartholomeus Anglicus, De proprietatibus rerum, London, 1495. Houghton Library, Harvard University, Cambridge, Mass.

Otto Pächt called attention to a remarkable miniature landscape illustrating a topographical text produced in Bruges c. 1470 (illus. 20).⁶⁶ The scene is nothing more than a vignette of ideal rural existence, complete with a cripple, a mounted knight, windmills and watermills, and a very thin, gabled building, all laid out along trim orthogonals. Detail does not imply specificity: the scene is no portrait of a real place, but is in fact an adaptation of a landscape setting from an earlier narrative illustration. The desire to construct appealing landscapes took root even in the most arid textual environments. As Walter Cahn has pointed out, there is no a priori obligation to sever the medieval descriptive or classificatory impulse from more affective, even if poorly articulated, attitudes towards the outdoors.⁶⁷

18 German master, Castle Langenargen, woodcut, 18.8 × 11.9, from Thomas Lirer, Chronik von allen Königen und Kaisern, Ulm, 1486. Beinecke Rare Book Library, Yale University, New Haven.

19 German master, Mounts Horeb and Sinai, before 1508–9, miniature from a manuscript copy of Felix Fabri’s Evagatorium, vol. 2, owned by Hartmann Schedel, 10.9 × 7.8. Bayerische Staatsbibliothek, Munich.

Some illuminators seized any pretext to frame a landscape. An agrarian panorama painted in Regensburg in 1431, with trapezoidal fields and a heap of prismatic rocks, actually illustrates a moralistic text about larks (illus. 21).⁶⁸ Other book painters exploited the lower margin or bas-de-page. Innovations, in Pächt’s phrase, often appeared first at the point of least resistance.⁶⁹ The bas-de-page opened up when miniaturists converted the lower edge of the frame, wound with schematic leaves or rinceaux, into a ground plane. Suddenly blank page signified space.⁷⁰ Jan van Eyck in the Turin Hours (early 1420s), stretched a Baptism of Christ into a cool river landscape along the bottom of a page.⁷¹ Altdorfer, too, when asked by Emperor Maximilian to make marginal drawings for a printed Prayer Book, took advantage of the lower strip to draw a landscape – a spectral massif of turrets and crenellations on a rocky hill – in pale red ink (illus. 14).⁷²

Occasionally, empty landscapes found their way into the marginal zones of altarpieces. They became part of the diverse apparatus that surrounded the painting at the centre of the altarpiece. A long, low frieze of mountains serves as the predella or base to Lorenzo Lotto’s Assumption altar in the cathedral at Asolo, dated 1506 (illus. 22).⁷³ The landscape measures 25 × 146 centimetres, and includes a chapel on a hill, a walled city with a citadel and, on the lower edge, a row of still-lifes: two quails, a sprig of flowers supporting a butterfly, a crayfish and two partridges. The predella is treated as if it were part of the frame; for Gombrich this explained the simulation of intarsia or inlaid wood in the foreshortened terrain.⁷⁴ This appears to be the only uninhabited landscape in an Italian predella. A restoration report of 1822, however, suggested that the Asolo panel is actually a Baptism of Christ with its middle section removed, and that it originally belonged to another, wider, altar that was not by Lotto at all.⁷⁵

20 Flemish master, Flemish Countryside, c. 1470, 16 × 22, miniature from Cotton MS, British Library, London.

Narrative scenes painted on predellas, like the sabotaged Baptism, were indeed often unified by landscape settings. Horizontal strips of landscape were frequently painted along the lower edge of the altar painting itself. Karel van Mander in 1604 described a work he knew by the fifteenth-century Dutch painter Albert van Ouwater: ‘at the foot (voet) of the altar was a pleasant (aerdigh) landscape in which many pilgrims were painted . . .’.76 The altar is lost, but the scene may have been a detached predella. A sixteenth-century chronicler reported a watercoloured Hell on the voet of the van Eycks’ Ghent altarpiece; no one can agree whether this was a predella or an antependium, that is, a cloth or panel hanging from the altar table itself.⁷⁷ In a Sicilian–Venetian altarpiece of 1489, the Madonna of the Rosary in Messina by Antonello de Saliba, a landscape panel is physically inserted into the bottom of the main composition. This picture within a picture represents a city against a backdrop of mountains. The landscape, or the panel, is watched over by angels holding a propitiatory banderole, and is not spatially coordinated with the rest of the scene at all.⁷⁸

21 German master (Regensburg school), Landscape with Larks, c. 1430, miniature from manuscript of Hugo von Trimberg, Der Renner, size of page 40.8 × 28. University Library, Heidelberg.

22 Lorenzo Lotto(?), Predella from Assumption altarpiece, 1506(?), oil on panel, 25 × 146. Cathedral, Asolo.

23 Albrecht Dürer, Madonna and Child Above a Rocky Landscape, c. 1515, woodcut, diameter 9.5, height 14.9. Museum of Fine Arts, Boston.

‘Internal’ predellas – strips of landscape at the lower edge of an altar painting, inside the frame – often represented the earthly realm in contradistinction to a heavenly event hovering in the sky. Dürer used the device in his Landauer altarpiece of 1511, and Altdorfer in a panel now in Munich, a Madonna and Child among angelic music-makers borne on a cloud above a mountain landscape.79 In Altdorfer’s picture, the emblem of heaven’s eclipse of the world is a fir-tree brutally truncated by the sacred cloud. Dürer around 1515 made a curious woodcut, perhaps an experiment, with a flinty, wedge-shaped landscape at the lower edge and a Madonna and Child suspended above in a circle (illus. 23).⁸⁰ Panofsky wondered whether the woodcut was not printed illicitly from two separate drawings by Dürer.81 Many later collectors disliked the effect and cut the landscape away. Lorenzo Lotto himself introduced a clutch of Germanic thatch-roofed farm buildings into the base of the Assumption at Asolo, in the main panel just above the disputed predella. The strip of landscape with woodsmen in Lotto’s Sacra Conversazione in Edinburgh, bizarrely, runs across the top of the composition.⁸²

The backs of altarpieces and altar covers also became breeding grounds for unorthodox pictorial and iconographic themes. These surfaces were separately framed but still physically attached to the primary picture. The backs of many fifteenth-century German altars – for example the altar by Wolgemut, Dürer’s teacher, at Zwickau in Saxony – are coated with mazy coils of painted foliage.⁸³ The reverse of the Uttenheim Altar of c.1470, south Tyrolean work close to Michael Pacher, combines foliage with a portrait of Christ as the Man of Sorrows, a devotional subject also found frequently on the backs of altars, as if taking refuge from the chaotic grid of the narrative sequences on the front.⁸⁴ Hans Memling painted two tall, bushy landscapes inhabited by cranes and a fox on the backs of a pair of panels representing saints, the wings of a lost small altarpiece.⁸⁵

Around 1511, on the back of the wings of a triptych now in Palermo, the great Antwerp master Jan Gossaert painted a Paradise with Adam and Eve, a subject that could never take over the centre of an altarpiece.⁸⁶ It is as if the outside surface of the closed wings only half-belonged to the triptych as a whole, and incompletely shared its function; the outside enjoyed iconographic license and could thus represent a realm prior to and outside of sacred history. Altdorfer, too, painted a prelapsarian landscape: Adam and Eve in a feathery forest on the outside wings of a triptych, the ‘regiments’ of Bacchus and Mars on the insides. The central image of the triptych – the portrait of a humanist, or even a mythological subject, the core of a possible ‘pagan altarpiece’ – is lost.⁸⁷ Most remarkably of all, Gerard David of Bruges painted at some point in the 1510s two vertical forest landscapes on the outsides of the wings to a Nativity altarpiece (illus. 24).⁸⁸ In the left panel David depicted a cottage and an ass, in the right panel two oxen drinking at a pond. No one has been able to explain these landscapes. If this is Paradise, where is the first couple? Perhaps the wings represent the resting place of the Holy Family on the Flight into Egypt, the episode that immediately preceded the Nativity. The ass and the ox belonged to the conventional bucolic audience of the Nativity; sometimes they were differentiated as symbols of Old and New Testament, although not conspicuously in the Nativity behind these wings.⁸⁹ Perhaps David liked the effect of an empty, leafy space waiting at the end of an arched, sun-dappled, forest-like nave, as if he were inviting a beholder to enter a numinous grove. Esther Cleven has recently proposed that David, who habitually painted tall, billowing trees in the backgrounds and margins of his altars and devotional panels, was offering the landscapes as a kind of signature or trademark. The wings referred well-informed beholders to other pictures, including various versions of the Rest on the Flight, by David.⁹⁰

24 Gerard David, exterior wings of Nativity altarpiece, 1510s, oil on panel, 90 × 30.5. Rijksmuseum, Amsterdam, on loan to the Mauritshuis, The Hague.

Angelica Dülberg has recently assembled a huge repertory of eccentric subjects and motifs – still-lifes, emblems and devices, allegorical figures in landscapes – painted on the backs and on the covers of fifteenth- and sixteenth-century portraits.91 A portrait by Memling in New York of a young woman standing at a window, for example, was attached to an upright landscape now in Rotterdam with a pair of horses and a monkey.⁹² At least one of Lorenzo Lotto’s small painted allegories was a portrait cover (see p. 58). A seventeenth-century inventory describes a lost work by Dürer with a ‘landscape, a hunting-scene, lovely foliage’ and two women holding the arms of the Pirckheimer and Rieter families, painted on parchment pasted on panel. The panel may originally have covered a portrait diptych of Willibald Pirckheimer and his wife Creszentia Rieter that was executed sometime between 1495 and 1500.⁹³ Portrait covers were places where iconography could run wild.

But most landscapes in this period simply appeared in the backgrounds of other paintings. The word landscape in Renaissance texts most frequently refers to an area within a panel painting, such as the background of a narrative composition or a portrait. Pinturicchio, for example, agreed in a contract of 1495 to paint ‘in the empty part of the pictures – or more precisely on the ground behind the figures – landscapes and skies.’⁹⁴ An early altar contract from Haarlem stipulated that ‘the first panel where the angel announces to the shepherds must have a landscape’.⁹⁵ Dürer referred routinely to the ‘landscapes’ in two sacred pictures in letters of 1508 and 1509 to his patron Jakob Heller.⁹⁶ A contract for an altarpiece in Überlingen on Lake Constance from 1518 specified that the Landschaften in the scenes on the wings be painted in the ‘best oil colours’.⁹⁷ In Leonardo’s notebooks, paese refers to countryside itself as an object of imitation, but at least in one case means the landscape within the picture.⁹⁸ Dürer, too, according to Willibald Pirckheimer’s afterword to the Four Books on Human Proportion (1528), had planned to write about landscape.⁹⁹

Painting backgrounds was a specialized skill. Workshops collected model drawings of landscape motifs which were later copied into paintings. Nuremberg and Bamberg workshops in Altdorfer’s time, for instance, preserved accurate drawings and watercolours of local buildings and skylines and inserted them in the backgrounds of altar paintings (see pp. 243–9 and illus. 69, 146–50). Italian amateurs of painting had long admired paesi ponentini: ‘Western’, or Flemish, landscapes.¹⁰⁰ The antiquarian Ciriaco d’Ancona saw a painting by Rogier van der Weyden in the 1450s and marvelled at the ‘blooming meadows, flowers, trees, and shady, leafy hills’, as if Mother Nature herself had painted them.¹⁰¹ Some Italian workshops employed Netherlandish specialists to paint landscape backgrounds. According to Vasari, Titian kept several ‘Germans’ (presumably Flemings) under his own roof, ‘excellent painters of landscapes and greenery’.¹⁰² Flemish landscape backgrounds were so admired that some Florentine artists copied them directly into their own works. Botticelli enlarged the background of Jan van Eyck’s Stigmatization of St Francis, a tiny devotional panel that survives in two versions, in Turin and Philadelphia, and adapted it to his Adoration of the Magi in London.¹⁰³ In the 1490s the Florentine Fra Bartolommeo copied a landscape from a Madonna and Child by Hans Memling (illus. 25), directly into his own Madonna and Child.¹⁰⁴ One Italian miniaturist actually detached Memling’s landscape from its composition and presented it as a picture in its own right (illus. 26). The Flemish vignette, with half-timber watermill and a serpentine brook, becomes in the Zibaldone or ‘commonplace book’ of Buonaccorso Ghiberti (grandson of Lorenzo) an illustration to a discussion of the casting of bells. The copyist, probably Buonaccorso himself, merely subtracted Memling’s swans and added a pair of smelting pots in the right foreground as the iconographical pretext.¹⁰⁵ The landscape backgrounds of Dürer’s engravings proved the richest lode of all. Marcantonio Raimondi, Cristofano Robetta, Nicoletto da Modena, Zoan Andrea, Giulio Campagnola and Giovanni Antonio da Brescia all transplanted trees and buildings from Dürer into their own engravings.¹⁰⁶ Even Vasari relished the ‘most beautiful huts after the manner of German farms’ in Dürer’s Prodigal Son engraving.¹⁰⁷

The next step for some Flemish painters was to let the setting dominate the picture, and indeed to make setting into the basis for the classification of the picture. Joachim Patenir of Antwerp, emboldened by the Italian taste for Northern rusticity, began as early as the 1510s to expand the backgrounds of his paintings out of all proportion. Patenir and many followers in Antwerp over the next decades painted horizontal panoramas of fantastic landscape seen from a bird’s-eye view. A painting like the Rest on the Flight in Madrid technically still had subject-matter. But it violently reversed the ordinary hierarchy of subject and setting (illus. 28).¹⁰⁸ Indeed, in at least one case – the Temptation of St Anthony, also in Madrid – Patenir painted the landscape but left the foreground figures to his friend Quentin Massys.¹⁰⁹ Patenir converted a workshop speciality into a livelihood and a reputation. Dürer met Patenir in 1521 while travelling in the Low Countries and referred to him in his journal as ‘the good landscape painter’.¹¹⁰ Moreover, the Italian enthusiasts of Netherlandish painting did not hesitate to describe pictures which still had recognizable subject-matter as landscapes. The term was even extended to earlier pictures. The Venetian patrician and amateur Marcantonio Michiel saw in 1521 an entire group of tavolette de paesi attributed to Albert of Holland, presumably the early Dutch painter Ouwater.¹¹¹ These were not independent landscapes by any means, not even panoramic landscapes with diminished subjects, but probably small devotional panels with notably green backgrounds, much like the moody portrait of St John the Baptist in the Wilderness (illus. 27) by Geertgen tot Sint Jans, a younger contemporary of Ouwater, also of Haarlem.¹¹²

25 Hans Memling, Madonna and Child with Two Angels, c. 148os(?), oil on panel, 57 × 42. Uffizi, Florence.

26 Italian master, Casting of Bells, c. 1472–83, miniature, from the zibaldone of Buonaccorso Ghiberti. Biblioteca Nazionale, Florence.

27 Geertgen tot Sint Jans, St John the Baptist in the Wilderness, c. 1475, oil on panel, 42 × 28. Gemäldegalerie, Berlin.

Much innovation in the history of art involves excavating subjects and ideas from hidden or marginal places and throwing them out onto a public surface, a prestigious surface. Many of the painted proverbs of Hieronymus Bosch, for example, enjoyed a clandestine prior existence in the margins of manuscripts or among the carved misericords on the undersides of choir-stalls.113 The mechanics of the relocation of the landscape – the promotion of the landscape from the margin to the centre – will be the subject of chapter Two. This promotion was accomplished first by assigning the landscape a name, and thus an invisible and virtual frame, and then by sealing it off from texts and other pictures with a physical frame.

28 Joachim Patenir, Rest on the Flight into Egypt, oil on panel, 121 × 177. Museo del Prado, Madrid.

Landscape and text

Max J. Friedlander, who in 1891 published the first book on Altdorfer, later wrote that ‘in the Netherlands we can, if put to it, trace the germination and efflorescence of landscape as an historical process – at least we are impelled to make the attempt’. But, he conceded, ‘faced with south German art, the historian lays down his arms’.114 There are a number of reasons for this. Interpretation of the Netherlandish landscape has been facilitated by its affinities with conventions of mapmaking.¹¹⁵ Artists used to make symbolic world maps by dividing a circle into three horizontal bands standing for sea, land and sky. Pisanello, in the mid-fifteenth century, put such a shorthand map on a bronze medal – once again, the verso of a portrait. The circle reduces the vast globe, symbol of the worldly dominion of the sitter, Don Iñigo d’Avalos, to a wavy sea, a strip of cities and pointed mountains, and a starry sky (illus. 29).¹¹⁶ The tiniest world map of all, another aerial – urban – aquatic abbreviation, appears in a Dutch manuscript of 1431 (illus. 30).¹¹⁷ This is a medallion painted on the imaginary sarcophagus of Darius described in the vernacular legend of Alex ander the Great. The text reports that Apelles, no less, ‘installed the round world’ on this tomb, and ‘made and engraved all the cities that were there, and all the rivers, and all the people that lived there; and those languages that were spoken there, and the great water inland; and also all the islands of the sea and what they are called’. On the manuscript page the medallion measures less than two centimetres across. The fifteenth- century Flemish panel painting would clear considerably more room for its microcosms.

Both the steady, earthbound eye of the urban topographer and the imaginary vantage-point of the pictorial map implied an intellectual grasp of terrain, the knowledge of the geographer or the cartographer. This was not a textual function; on the contrary, the topographical picture exploited a capacity of images that language could never rival. But it was a function textual culture could acknowledge. Literary humanists praised and justified painting as an instrument of knowledge: in 1336 a lettered Venetian referred to a cache of papers containing a picture or plan (une figure) of Florence by Giotto;118 almost two centuries later Raphael undertook a monumental pictorial catalogue of Roman antiquities. The humanist Bartolommeo Fazio described in 1456 a ‘circular representation of the world’ painted by Jan van Eyck for Philip the Good, Duke of Burgundy, which was presented to King Alphonso of Aragon. In this work, ‘more perfect than any of our age’, one distinguishes ‘not only places and the lie of continents, but also, by measurement, the distances between places’. The Eyckian mappemonde was probably a bird’s-eye panorama of a slightly convex terrestrial surface, and may have been overlaid with Ptolemaic longitudinal and latitudinal coordinates.¹¹⁹

This is not to say that the Germans neglected cartography. Martin Behaim of Nuremberg constructed the very first globe – his Erdapfel, or ‘earth-apple’ – in 1491–2.¹²⁰ Dürer, in one of the drafts for the introduction to his treatise on painting, claimed that ‘the measurement of the earth, the waters, and the stars has come to be understood through painting’.¹²¹ Dürer may have encountered the thoughts of Leonardo, who had boasted in his own unpublished treatise that painting, ‘by virtue of science, shows on a flat surface great stretches of country, with distant horizons’.¹²² In 1515 Dürer published a woodcut projection of the globe and a pair of star charts for the imperial astronomer Johann Stabius.¹²³ Veit Stoss, the Nuremberg wood sculptor, showed his carved and painted mappa, possibly the earliest full-relief landscape model, to the calligrapher and art-historical chronicler Johann Neudörfer.¹²⁴ Literary culture paid tribute to this notational capacity by comparing its own verbal descriptions to pictures. Humanists in Dürer’s and Altdorfer’s time conventionally held up the panel painting as a paragonal format for descriptions of the universe, a more adequate medium for this task than language. Conrad Celtis, travelling scholar and poet laureate, compared Apuleius’s cosmographical treatise De mundo, which he edited in 1497, to a small painting or sculpture: ‘they learn, as it were from a small panel or sculpture of words, how and by whom this universe is put together and maintained in its form.’125 Johann Cuspinian, the Viennese poet and physician, promised in his university lectures c. 1506 on Hippocrates ‘to describe, just as the painters do, the entire world on a small panel’.¹²⁶ Celtis, finally, in the dedication to the Amores, his geographical love poem published in Nuremberg in 1502, promised that ‘our Germany and its four regions’ would be visible ‘as if painted on a small panel’.¹²⁷ In this case Celtis actually designed for his book, through the agency of Dürer’s workshop, four woodcut panoramas based on astrological calendar scenes. The walled city at the centre of the woodcut introducing book Two – the account of Celtis’s romance with ‘Elsula’ – is Regensburg itself (illus. 31).¹²⁸

But, in general, the exemplary pictures to which the humanist formulas refer never existed. These passages belong to the general humanist justification of descriptive language under the Horatian motto ut pictura poesis. They sound like disingenuous apologies for a rhetorical topos (the ekphrasis tou topou, or description of place¹²⁹) or a literary form (the cosmographical treatise) that, in fact, required no apology. They reflect either a general epistemological or pedagogical preference for the sensory over the intellectual, grounded in some version of Nominalism or anti-scholasticism,¹³⁰ or the more ancient preference for sight over the other senses, formulated by both Plato and Aristotle and invoked by Leonardo and Dürer in their writings on painting.¹³¹

29 Antonio Pisanello, portrait medal of Don Iñigo d’Avalos, verso with ‘world map’, c. 1448–9, bronze, diameter 7.7. British Museum, London.

30 Dutch master, Tomb of Darius, miniature from a manuscript Historie des Bijbels, Utrecht, 1431. Bibliothèque Royale, Brussels.

The acknowledgement of a common purpose with textual culture helped legitimate the new Netherlandish landscapes of the 1510s and 1520s. It was, after all, not hostility toward subject-matter which distinguished these pictures, but rather the generosity toward terrain displayed by the horizontal format and the high horizon. Horizontality was an essential element of the novelty of the sixteenth-century Netherlandish landscape. Portable panel paintings were almost always vertical in format before 1520. Patenir’s landscapes were among the first small horizontal panels of any sort. Horizontality was implied in the very word ‘landscape’, which at least in Italian (paese), German (Landschaft), and Dutch (landschap) meant terrain or countryside itself before it meant painted landscape. (English broke the pattern by absorbing the Dutch word in its secondary, pictorial meaning as early as the 1580s, and only later extending it to actual terrain.) Paracelsus, the sixteenth-century German physician and natural philosopher, used the word in this sense in his treatise on images: ‘There are other kinds of chiromancy – a chiromancy of herbs, one of the leaves on trees, one of wood, one of rocks and mines, one of landscapes through their roads and water-courses.’132 (Paracelsus, incidentally, called the specialist who is supposed to read the palm of landscape the cosmographus.) The horizontal format offered the landscape as a surface to be read. It encouraged discursive speculation about man’s relationship to the surface of the earth, about the relationship between place and event, about the local and the universal, about time, travel and work. There was no obvious advantage in expunging subject-matter altogether, for the diminutive human figure, embedded in a magnified world, only multiplied and complicated meaning. The horizontal format permitted a smoother introduction of epic themes and grandiose moods into landscape. The Antwerp landscapes flowed directly into the main tradition of European landscape, through Bruegel to Rubens and beyond.

Most of Altdorfer’s landscapes are vertical in format. They are hardly the tabellae that the humanist cosmographers had in mind. Because the horizons are narrow and low, the pictures are uninformative about terrain. Whereas the high horizon of the Netherlandish panoramic landscape implied a detached and omniscient narrator, Altdorfer’s pictures not only imply but sometimes actually represent a real physical standpoint for the beholder of the scene. His points of view are more plausible and empirically more familiar than those of Patenir. This is the basic structure of most of Altdorfer’s narrative paintings, prints and independent drawings, indeed even when they are horizontal in format. In narrative pictures, the rhetorical effect of the low point of view is obvious: the painter offers himself as eyewitness to an event and invites the beholder to share that privilege. It is not so clear, however, what purpose was served by a low angle on open terrain, or within an empty forest.

That emptiness is the absence of a narrative. Altdorfer’s landscapes do not merely fail to tell a story: they are also empty of subject-matter itself, the raw material of narrative. They are narrative compositions from which subject-matter has been removed, or perhaps allowed to bleed into setting and dissipate. This statement needs to be elaborated, for ‘subject-matter’ is a literary term. Subject-matter is something subjected to a telling. Subject-matter is thus more or less equivalent to Aristotle’s logos or argument, the primary imitation of action that serves as the raw material for plot or mythos.133 Classical rhetorical theory acknowledged an equivalent hierarchy between materia, or raw material, and res, or conceptual content.¹³⁴ Medieval and Renaissance epic drew its plots from an unformed stock of characters and events it called ‘matter’, for example the ‘matter of Britain’ (the Arthurian legends) or the ‘matter of Troy’. Spenser in the Faerie Queene (III. iv. 3) apostrophized: ‘thee, O Queene, the matter of my song’; Milton used the same formula in Paradise Lost (III. 412–3): ‘thy Name / Shall be the copious matter of my song’.

In fifteenth- and sixteenth-century German altarpiece contracts, the word materi refers to the subject-matter of narrative sequences and, by extension, to the pictures themselves: ‘the materi of St Wolfgang’, for example, in a contract with Michael Pacher from 1471, or the four carved materie on the wings mentioned in a contract with Veit Stoss from 1520.¹³⁵ Nearly all paintings had subject-matter. Dürer justified his art by pointing out, in a draft of his planned treatise on painting from 1513, that ‘paintings are employed in the service of the Church, and by them the Passion of Christ and many other good images [Ebenbild] are shown’.¹³⁶ If the German painter was to make a name for himself, he would have to achieve it as an effective expositor of narrative material. Thus it is hard to overstate the strangeness of the utter absence of subject-matter in Altdorfer’s landscapes.

The independent landscape was derived from a particular kind of picture with subject-matter. Many of Altdorfer’s landscapes shared their format, scale and medium with prints, drawings, watercolours and small painted panels. These pictures can sensibly be called narratives, although they did not all represent action. Some represented a single figure, usually a saint or other holy personage, identifiable as the protagonist of some familiar narrative by attributes and setting. Others represented figural types from pagan or secular culture without necessarily telling stories about them, for example satyrs or peasants. (This sort of subject-matter, however, seldom appeared on panels; it survived instead in more frail media, such as engravings.) But even pagan material, more often than not, was shaped into narrative, or at least the narrative qualities of such pictures were considered essential to them. Dürer, for example, in a letter to Pirckheimer of 1506, referred to Italian pictures with pagan subjects generically as historien.¹³⁷ The fundamental characteristics of such a picture were, first, that it was contained by a frame and shaped into a coherent and self-sufficient composition; second, that it was not merely an element in a larger decorative scheme; and third, that the composition rested on a distinction between subject-matter, embodied in the human figure, and setting, which included any adjacent attributes of the figure as well as its physical environment.

31 Workshop of Albrecht Dürer, Elsula Alpina, woodcut, from Conrad Celtis’s Quatuor Libri Amorum, Nuremberg, 1502. Houghton Library, Harvard University, Cambridge, Mass.

32 Lorenzo Lotto, A Maiden’s Dream, c. 1505, oil on panel, 43 × 34. Samuel H. Kress Collection, National Gallery of Art, Washington, DC.

Close in structure to Altdorfer’s landscapes, and perilously indifferent to the conventional equilibrium between subject and setting, were the pagan and pastoral cabinet pictures by contemporary Venetian painters. These works cluster around the person and aura of Giorgione. Some of these pictures broach themes from literary pastoral, either by adapting scriptural, hagiographical or classical subject-matter to pastoral formulas and moods, such as the Landscape with SS Roch, George and Anthony Abbot in London, possibly by Giorgione himself;138 or by directly illustrating poems, such as Andrea Previtali’s four panels illustrating an eclogue of Tebaldeo.¹³⁹ But others decline to tell stories. Giovanni Bellini’s Orpheus¹⁴⁰ simply encloses a mythological personage in lush forest setting. Lorenzo Lotto’s Maiden’s Dream and Allegory in Washington are allegories with pagan overtones.¹⁴¹ For its dense, feathery trees, the Maiden’s Dream (illus. 32) of c. 1505 tapped German prints, or perhaps even paintings. The woman’s pose repeats figures by Dürer or the German–Italian engraver and painter Jacopo de’ Barbari. Still other works celebrated secular pleasures by staging fictional concerts or love-making in the locus amoenus, for example the famous Concert in the Louvre,¹⁴² or the Edinburgh Three Ages of Man¹⁴³ by Giorgione’s prodigious pupil, Titian. Although the Venetian pictures never quite expelled the human figure, they exhibited an unprecedented defiance or indifference toward the ordinary functions of painting. The Venetian cabinet pictures flourished accordingly in marginal zones and spaces, adjacent to the primary contexts for painting. A lost series of eighteen Ovidian panels by Giorgione were evidently set into a wall, in a band running just below the ceiling.¹⁴⁴ Other panels were attached to furniture. Lotto’s Washington Allegory was the cover of a portrait of Bernardo de’ Rossi, a humanistically-minded cleric.¹⁴⁵ The Maiden’s Dream may, in turn, have belonged to a portrait of Rossi’s sister Giovanna.¹⁴⁶ The dreamy and allusive moods of the Venetian cabinet picture easily spilled over into pen drawings and prints, first in the engravings of Giulio Campagnola, later in the woodcuts and drawings of Titian and Domenico Campagnola.

The intense debate about the meaning of the Venetian cabinet pictures gathers around a single painting by Giorgione, the Tempest (illus. 33) in the Accademia in Venice, a disquieting conjunction of nameless characters in a landscape.¹⁴⁷ A halbardier looks across a brook toward a semi-nude, nursing mother. A bolt of lightning illuminates a city and a geometrical pile of classical ruins. Many readings have been put forward, from Mars and Venus, to Adam and Eve, to Fortitude and Charity; none has prevailed. The figures look as if they might be involved in some drama, and the setting looks as if it might be invested with symbols. But nothing is quite clear; prior iconographic traditions are at once invoked and distorted.

The argument that Giorgione’s Tempest was never meant to be ‘read’ was expounded most memorably by Lionello Venturi and Creighton Gilbert, and has taken on various formalist and anti-iconographical colourings.¹⁴⁸ The latest, and fiercest, assault on this position was launched by Salvatore Settis, who has denied that any picture of the period could have dispensed with subject-matter.¹⁴⁹ Even landscape, for Settis, is subject-matter. This debate has been hobbled by imprecise terminology. Giorgione’s Tempest is surely not devoid of subject-matter. That subject-matter is a minor and secular subject like gypsies or wild families, or pastoral, or even conceivably a minor mythological subject, and not a scriptural subject. But the raw material has not been shaped into a story. The durable appeal of the picture consists precisely in its incomplete and enigmatic presentation of subject-matter, and its simulation of more orthodox and informative pictures. In arguing that the picture represents Adam and Eve and therefore does have a subject, Settis defines ‘subject’ too narrowly: the picture does not need Adam and Eve to have subject-matter. Settis then defines subject much too broadly when he claims that a landscape, even when ‘empty of symbols’, is also a subject. By subject here he must mean ‘content’ in some general sense. A landscape might initiate a train of thought, for example, about the relationship between man and nature; so might the Tempest, or indeed any of the Venetian cabinet pictures. This train of thought would be facilitated by any Virgilian or Ovidian thematics, for in the poetry such questions are addressed explicitly and profoundly. But this kind of thematic meaning is not equivalent to subject-matter.

The Venetian painters cultivated semantic elusiveness, a certain fugitive quality more familiar from lyric poetry. Indeed, the pictures were often called poesie, and some were directly inspired by poems. It is as if the pictures wished to retreat from the crass and earnest plenitude of expository or hortatory painting, back toward a state of deliberate incompleteness. The tempered glow of a sward of green or a lost contour replaced legibility. Indeterminacy of meaning was encoded at the level of the pictorial sign. Roland Barthes, in a suggestive discussion of semantic bruit or noise, once characterized all art as ‘clean’, that is, free of interference, functional to the last detail; this is what distinguishes it from life.150 But then Barthes conceded that painting, unlike writing, could introduce artificial interference, as it were, as a code for indeterminacy. He was evidently thinking of painterly techniques such as the deliberate dissolution of outlines, smoky or fading boundaries between colour areas, or atmospheric perspective. Venetian brushwork and palette immediately warned beholders either to drop their iconographical antennae, or to keep them raised and savour the vain effort to clear up the muddle. The Florentine Vasari, only two generations later, was stumped by Giorgione’s frescoes on the Fondaco dei Tedeschi, the German merchants’ compound at the Rialto. Giorgione, he complained, had thought only of fashioning fanciful figures and displaying his art. Vasari found the frescoes unreadable. ‘Nor, with all the inquiries that I have made, could I ever find anyone who did understand, or could explain them to me.’¹⁵¹ But plenty of amateurs relished indeterminacy. Federigo Gonzaga in 1524 ordered from the Venetian painter Sebastiano del Piombo not ‘sacred things’, but rather ‘some charming [vaghe] pictures, beautiful to look at’.¹⁵²

33 Giorgione, The Tempest, c. 1507, oil on canvas, 68 × 59. Galleria dell’Accademia, Venice.

Meaning of this sort emerges out of a deliberate detachment from subject-matter. Crucial to Giorgione’s rhetoric of enigma was the overdevelopment of outdoor setting. Not only bits of narrative, but also allegories were often embedded in outdoor greenery. In Lorenzo Lotto’s Allegory or in his Maiden’s Dream, or in Piero di Cosimo’s Allegory in Washington,153 or in the grisaille verso of the portrait diptych of Alvise Contarini in New York,¹⁵⁴ with its shadowy, draped, river-bank figure, landscape amputates and frustrates ordinary trains of thought, dialectical links, and temporal vectors. Emblems and imprese are insulated by landscape, as if by a buffer zone protecting their fragile semantic mechanisms from the world outside. Marcantonio Michiel manifested his perception of such imbalances between setting and subject in his use of the word paese. In Michiel’s notes on Venetian painting, paese can still refer to an element of the picture, for example in the description of the ‘foreground landscape’ in Giovanni Bellini’s St Francis Receiving the Stigmata, now in the Frick Collection.¹⁵⁵ In other entries, however, the term paese becomes the primary identity of the picture: Giorgione’s tela del paese cun el nascimento de Paris (landscape canvas with the Birth of Paris) or, again, the molte tavolette de paesi painted by Albert of Holland.¹⁵⁶ Most remarkable is Michiel’s description in 1530 of Giorgione’s Tempest as a ‘paesetto on canvas with a tempest, a gypsy woman and a soldier’ (proving that if there ever was a story, it had been forgotten only twenty years after Giorgione’s death!).¹⁵⁷ From this point on it becomes possible to describe a whole picture as ‘a landscape’. Lodovico Dolce, for instance, referred in one of his dialogues to a paese by Titian illustrating the adventures of a nymph and several satyrs, which itself alludes to an imaginary pastoral painting described by Jacopo Sannazaro in his hugely popular Arcadia. Dolce had in mind paintings such as the landscape frieze with a reclining shepherd in the Barnes Foundation in Philadelphia, possibly a cassone panel, and possibly by Titian. Setting had come to dominate such pictures.158

And yet the Italians never permitted landscape to swallow up the human figure entirely. In the backgrounds of extremely sophisticated High Renaissance altar paintings, like Fra Bartolommeo’s Carondolet altar in Besançon (c. 1511), unnamed figures in hazy and portentous background landscapes contributed to a sense of meaningfulness without necessarily denoting anything in particular. Although diminished and garbled, subject-matter still exerted a gravitational pull on setting. There was no pressing reason to create a rigorously independent landscape.

Landscape as parergon or by-work

The Venetian paintings were able to uncouple the art of painting from storytelling and exposition because they offered something in exchange. The correlative of semantic indeterminacy was sensory delight. But views and foliage had been associated with pleasure since Antiquity, especially in domestic decorative painting. The main idea was that the harried beholder would seek refreshment or recreation in landscape. The representation of landscape was taken as a substitute for an experience of the out of doors.

Thus in Leon Battista Alberti’s Ten Books on Architecture, written c.1450 and published in 1486, it is the beholder who delights in and is ‘cheered beyond measure by the sight of paintings depicting the delightful countryside, harbours, fishing, hunting, swimming, the games of shepherds – flowers and verdure . . .’. Such painting is ‘suitable for gardens, for it is the most pleasing of all’.159 In his own treatise on painting, Leonardo da Vinci agreed that poetry made travel superfluous: ‘What moves you, man, to abandon your home in town and leave relatives and friends, going to country places over mountains and up valleys, if not the natural beauty of the world . . . ?’ Why not accept the poet’s description and stay at home, avoid the heat and fatigue, escape illness? Better yet, let the painter put before your eyes the very landscapes ‘where you have had pleasure beside some spring, [where] you can see yourself as a lover, with your beloved, in the flowered meadows, beneath the sweet shade of trees growing green’.160 Erasmus, highly suspicious of profane painting, nevertheless condoned decorative landscape: ‘If you are charmed by painting, what is more fitting to the Christian house than the acts of Christ, or the examples of the saints? And if it pleases you to mix these with pleasantries [joci], there are moral tales, there are the innumerable forms of trees, plants, flowers, and animals.’¹⁶¹

Recreation in this tradition stands as a supplement to work, in the sense made familiar from Derrida’s Of Grammatology and other writings. The supplement for Derrida is an amphibious entity that presents itself as a superfluity, a gratuitous excess, and yet whose very existence reveals a deficiency or lack in the object it supplements. The supplemented object turns out to be knowable only through, and in terms of, its supplement. Thus writing offers itself – misleadingly – as a ‘mere’ supplement to the spoken word; or art as a supplement to nature. The disclosure of the mutual dependency underlying the notional hierarchy of object and supplement is one of the central and most characteristic operations of Deconstruction. Recreation in Western pictorial culture, then, is meant to follow work and therefore stand outside it. It is, perhaps, a reward for work completed. But pleasure also prepares one to resume work by restoring or recreating the spirit. The danger is that the interlude of recreation will not only restore the labourer, but ultimately provide a superfluity of pleasure, and even replace work. It is easy to see how the pleasure taken in outdoor scenes, first in the mere representation of a venue associated with retreat and leisure, eventually in the painterly effects achieved in those representations, could generate both enthusiasm for, and resistance to, landscape painting in the Renaissance.

Landscape painting also stood as a supplement to the work of the painter. The most suggestive of all contemporary comments on Renaissance landscape is a remark about Dosso Dossi, the Ferrarese court painter, in a dialogue written by Bishop Paolo Giovio, a historian and art collector:

The elegant talent of Dosso of Ferrara is proven in his proper works, but most of all in those that are called parerga. For pursuing with pleasurable labour the delightful diversions of painting, he used to depict jagged rocks, green groves, the firm banks of traversing rivers, the flourishing work of the countryside, the joyful and fervid toil of the peasants, and also the distant prospects of land and sea, fleets, fowling, hunting, and all those sorts of things so agreeable to the eyes in an extravagant and festive manner.162

In this text of 1527 the painting of landscape is presented as a diversion for the painter, a respite from his proper tasks; or rather, it garnishes his labour with pleasure. Leonardo chided Botticelli precisely because he took landscape too lightly: the younger painter considered landscape a cosa di brieve e semplice investigatione, for would not a sponge thrown against the wall serve equally well as a studied composition? But random stains supplied only the invention. Botticelli painted miserable (tristissimi) landscapes because he neglected to apply himself to finish them.¹⁶³ Landscape is a parergon because it stands alongside both work (labour), and the work (opus). The same ambiguity is preserved in the obsolete English word by-work, which, according to the Oxford English Dictionary, means both ‘work done by the way, in intervals of leisure’, and ‘an accessory and subsidiary work’. Thus the beholder’s pleasure in looking at landscape overlaps the painter’s pleasure in making it. The danger of pleasure, as Michelangelo supposedly observed in a famous passage from the Dialogues of the Portu-guese painter Francisco da Hollanda, was permanent diversion from a primary task, namely the representation of the human figure: ‘In Flanders they paint to attract the eye with things which please . . . They paint draperies, masonry, green meadows, shady trees, rivers, bridges, and what they call landscapes [paisagens], and in addition many animated figures scattered here and there’. ‘I criticize Flemish painting’, the ‘Michelangelo’ character concluded, ‘not because it is entirely bad, but because it tries to do too many different things.’¹⁶⁴ Attention was not only diverted, but strewn across the surface of the picture.

One model for the pleasure taken in private and recreational painting was the humanist ideal of retreat into contemplation and creative languor, and in particular of spiritual refreshment in a rustic haven. Although this dream was rooted in Virgil and Horace, the more immediate models for the humanist contemporaries of Giorgione and Altdorfer were Petrarch and the Florentine neo-Platonists. Conrad Celtis, in particular, cultivated the image of a plein-air poet.¹⁶⁵ The appeal of the retreat from society for the humanist was its paradoxical and patrician reversal of the ordinary hierarchy between productivity and leisure: work – creative work – becomes the fruit of play rather than the other way around. Work is ultimately transformed into an aspect of play. This reversal could only be carried out on the margins of civilization, and so the poet modelled himself after the shepherd. The new cooperation of poet and nature is illustrated by the marginal illumination that Dürer painted and pasted on the first page of the edition of Theocritus owned by his friend Pirckheimer (illus. 15).¹⁶⁶ Two shepherds, tranquil yet alert, make music in a crumpled, rocky valley. The illumination itself, not a proper work but an amicable offering, dances around and below the block of printed text, itself the product of the most sophisticated textual criticism, and one of the early monuments of Greek typography in the West.

Another model for the experience of early landscape painting – another alternative, that is, to the escapism described by Alberti – is provided by the figure of the antiquary. In his intellectual pleasure in the past the antiquary escapes the ordinary experience of time. The association of archaeology and epigraphy with recreation was made explicit in the celebrated excursion on Lake Garda undertaken in September 1464 by Andrea Mantegna, in company with the antiquaries Felice Feliciano and (possibly) Giovanni Marcanova.¹⁶⁷ The pleasure of ruins and inscriptions was indulged alongside, or outside of, everyday activity, embedded in a fantastic round of mock ceremonies and re-enactments. Indeed, Marcanova’s antiquarian pursuits – whether he was actually on the boat with Mantegna at Lake Garda or not – were described by his eighteenth-century biographer as a parergon to his primary work.¹⁶⁸

Yet another model was the pleasure taken in the comic, in the spectacle of mundane or even unworthy characters, ‘men who are inferior but not altogether vicious’, in the definition of Aristotle.¹⁶⁹ Domestic decorative landscape was a primary setting for this subject. Alberti had characterized his ‘garden’ painting as the type of painting ‘depicting the life of the peasants’, distinguishing it both from the type that ‘portrays the great deeds of great men, worthy of memory’, and from the type that ‘describes the habits of private citizens’. These rustic characters were so low in station that in the eyes of the beholder the distinction between their work and their play was blurred. In Giovio’s text, for example, the toils of the peasants are described as ‘joyful and fervid’ (laetos fervidosque). The hard comic possibilities of the situation were exploited especially in the North, where the peasant was brought ever closer to the surface of the landscape and yet made to look ever stranger. In Bruegel, who, in van Mander’s words, ‘delighted in observing the droll behaviour of the peasants, how they ate, drank, danced, capered, or made love’, the boundary between profound contempt and profound sympathy is imperceptible.¹⁷⁰ In early landscape painting, even subject-matter is divided against itself, divided between the serious and the frivolous, as if mirroring the division of the picture as a whole.

Landscape thus stood outside or alongside the proper work or task of painting, the business of treating subject-matter. The veiling of subject-matter by landscape, as in the Venetian paesi, or its expulsion, as in Altdorfer’s landscapes, were exercises played out far from the primary and serious business of painting. For Alberti, the grandest painting, the representations of ‘the great deeds of great men, worthy of memory’, belonged in ‘public buildings and the dwellings of the great’. Recreational landscape painting, on the other hand, was relegated to the gardens, presumably in the form of wall-painting on, or in, garden buildings.

The structural relationship between important and trivial painting is complicated when the two must share the same surface. This problem had been raised already by Pliny. Protogenes of Rhodes, the rival of Apelles, had ‘added’ to his wall-paintings at Athens small depictions of warships ‘in what the painters call parergia’.171 The parergia were either the margins of the fresco, or the background, or perhaps an additional field outside the scene altogether, an ornamental strip. The parergia came to mean both the areas of the picture susceptible to landscape, and the landscape motifs that appeared in those areas. This ambiguity is brought out in an ekphrasis, or description of a picture, from the Hypnerotomachia Poliphili by Francesco Colonna, the fantastic and learned prose romance of the late Quattrocento. The passage describes a frieze of mosaics on the inside of an arched gate, adorned ‘with exquisite parergi, waters, fountains, mountains, hills, groves, animals’.¹⁷²

The accessory landscape was both superfluous to the essential purpose of the picture, and an inextricable part of the picture. In a eulogy to the miniaturist Simon Marmion by a sixteenth-century French chronicler, supplementary details have become as indispensable as skill in drawing: ‘He laboured over and ornamented all his works so well, as much by his ingenious and expert drawing as by his buildings and other accessories [accessories], that one finds nothing incorrect or imperfect.’¹⁷³ Another clever miniaturist, possibly Simon Bening of Bruges, made a schematic image of the double face of landscape setting in an early sixteenth-century manuscript (illus. 34).¹⁷⁴ The scene inside the inner frame, St Christopher carrying the Christ child across the river, is complete; normally it would be surrounded by blank paper. But the painter has let the river flow under the frame and fill an outer frame. The hagiographical scene has been supplemented by a second margin of landscape setting, and it is still complete. This model of the relationship between subject and setting survived for some generations. The lexicographer Thomas Blount, who rated ‘landskip’ among those ‘Hard Words . . . as are now used in our refined English Tongue’, spelled it out clearly in 1656: ‘All that which in a Picture is not of the body or argument thereof is Landskip, Parergon, or by-work. As in the Table of our Saviour’s passion, the picture of Christ upon the Rood . . . the two theeves, the blessed Virgin Mary, and St. John, are the Argument. But the City Jerusalem, the Country about, the clouds, and the like, are Landskip.’¹⁷⁵

34 Flemish master (the Hortulus Master, or Simon Bening?) ALBRECHT ALTDORFER St Christopher, c. 1524, miniature, from Hortulus animae. Österreichische Nationalbibliothek, Vienna.

All these tensions were strung across an invisible boundary between the work and its by-work. There was no symmetry between work and accessory: one was subordinate to the other. The unequal dependency kept subject and setting suspended in a symbiosis. The parergon fluctuated between usefulness and uselessness. Some accessories were overdetermined: the landscape invested with didactic markers (the paysage moralisé), the fragments of still-life that invoked theological principles (disguised symbols). Other accessories were underdetermined: the recreational landscape. Many parerga were both at the same time. In the Christian tradition, the deliberate upending of the hierarchy of centre and margin generated real irony. The ‘work’ of a Christian image was to assert the supremacy of the metaphysical over the terrestrial. Painting had to preach spirit in the language of matter. This paradox fascinated especially Northern painters during and after the Reformation. It surfaced in still-life painting, for example in the gross and sumptuous market scenes of Pieter Aertsen, with sober biblical episodes sequestered in the backgrounds; pictures that state the superiority of sacred narrative to naked description, but perform the opposite.176

Altdorfer’s landscapes are altogether different sorts of objects. They abolish the boundary between subject and setting, as well as the hierarchy, by banishing subject and funnelling setting toward the literal centre of the image. Once subject is gone, it is no longer possible to characterize landscape as either overdetermined or underdetermined. It no longer even makes sense to call it setting. For the primary boundary is now the frontier between work and world. The symbol of this new boundary is the frame.

Altdorfer’s landscapes, in contrast to all the other artefacts we have been thinking about, were independently mounted and framed and entirely portable. The only possible Italian equivalents – framed images truly empty of subject – were the paintings of Dosso Dossi. Paolo Giovio contrasted Dosso’s parerga to his ‘proper works’ (justis operibus). There is, unfortunately, no precedent or corroborating document that would prove that here parerga refers to independent paintings.¹⁷⁷ It is conceivable that by Dosso’s ‘proper works’ Giovio simply meant the subjects of his paintings, to which his rocks, rivers, peasants and vistas were the accessories. Dosso did paint fluid, evocative pictures with esoteric subjects, such as the Three Ages of Man in New York (possibly a fragment) or the Travellers in a Wood in Besançon. There is even a pure landscape, but this is a fragment.¹⁷⁸ The best example of what Giovio meant by a parergon is perhaps the Landscape with Figures on a Country Road recently on the art market in New York – a painting without any discernible subject, and definitely not a fragment.¹⁷⁹ Altdorfer’s and Dosso’s pictures afforded recreation a discrete, even institutional, shelter. Such a painting was an authentic innovation, as far as we know, even over classical art. The landscape paintings described by the third-century Sophist Philostratus, for instance, were panels (pinakes) inserted into the walls of a villa.¹⁸⁰ And the landscapes by the first-century painter Studius, noted by Pliny and Ghiberti (see p. 38), were only frescoes. ‘Among artists’, according to Pliny, ‘great fame has been confined only to those who paint tabulas.’181 Portable panels were a more prestigious field – too prestigious for empty landscape.

Now instability was permanently built in to the structure of the work. Derrida’s ‘supplement’ is officially superfluous; in fact, it fills a lack. And the ergon, contrariwise, is officially self-sufficient; in fact, it needs a supplement. With the independent landscape, that mutual incompleteness became the basis for the structure of the work. Without a subject in sight, the setting remained suspended in its supplementarity. All those suspicious and useless recreational moments were absorbed into the work. The work of art became an object that worked and diverted at the same time. But this duality was no longer a hierarchy. The new frame devalued or even obscured entirely the old internal divisions, the boundaries between work and accessory that once ran across the life of the painter and the surface of the painting. The dissolution of these interior frontiers had ethical implications, for it placed work in an uncomfortably ambiguous and depen-dent relationship to pleasure. The work now displayed the escape from temporal prisons, from responsibility, from community and sociability, from profitable labour, from the imperatives of salvation; it displayed a movement without end, blocked by the frame from any possible end.

The achievement of this frame was first articulated by Kant in his Critique of Judgement. The frame in Kant’s scheme definitively severed the work of art from the world. It disqualified the world as a criterion of evaluation of the work. The work manifested, moreover, an ‘artificial free beauty’ that may be taken not as a poor imitation, but as a metaphor for natural beauty. Derrida’s analyses of Kant in his essays ‘Parergon’, in The Truth in Painting, and ‘Economimesis’, reveal the power and appeal of the scheme. Kant’s ‘pure cut’ between work and world, announced and guaranteed by the frame, makes possible all the various modern interpretative protocols. Formalist, allegorical, propagandistic, expressionist, psychoanalytic, utopian: all these hermeneutics of the work of art have proceeded from a stable notion of the frame. Nor are ‘contextual’ readings exempt. For the very notion of context depends on a clear understanding of where the text ends. There can be no thought of transgression without a boundary to cross.

But in The Truth in Painting Derrida also exposes a latent instability in the Kantian version of the structure of the work. It turns out that the parergon is still alive. The old hierarchy of work and by-work, work and supplement, ought in principle to have been abolished by the ‘pure cut’. But Kant’s remarks on the frame are inconsistent in a way that undermines the accomplishment of the cut. Derrida’s basic purpose was to reveal the precariousness of the modern doctrine of the aesthetic. But his analyses also cast a spotlight backwards in time, and thus prove useful to the present study. For the historical roots of Kant’s uncertainty lie within an early sixteenth-century investigation of the tension between subject and setting.

Derrida first focused on a parenthetical comment made by Kant about the parerga or ornamental accessories to the aesthetic object.182 Kant argued in §14 of the Critique of Judgement that ‘even what is called ornamentation [Parerga], that is, what is only an adjunct, and not an intrinsic constituent in the complete representation of the object, in augmenting the delight of taste does so only by means of its form’. Kant then gave three examples of parerga: the frames of pictures, the drapery on statues, and the colonnades of palaces. ‘But if the ornamentation does not itself enter into the composition of the beautiful form – if it is introduced like a gold frame merely to win approval for the picture by means of its charm – it is then called finery and takes away from the genuine beauty.’ Kant thus defined a penumbral zone whose relationship to the proper work of the work of art was contingent on the character of the effect it exercised on the beholder. If it appealed to taste through its form, then it belonged to the work; if it relied on mere charm or Reiz, then it was extraneous to the work.

Immediately there is trouble. The placement of architecture alongside the plastic arts unhelpfully reintroduces the question of function. It led Kant to insinuate an ontological boundary between the nude sculpted body and its sculpted garment that one is hard-pressed to reconcile with modern aesthetic doctrine. It also encouraged him to raise the possibility that the frame around a picture – the literal framework in carved, painted or gilded wood – might actually and essentially contribute to the work of the picture. It appears to us that modern artists place the ‘pure cut’ clearly on one side of the frame or the other: either the frame is part of the work or, more usually, it is not. Indeed, in the twentieth century the pure cut was taken so seriously that the frame was often jettisoned altogether. For Kant, by contrast, the frame is still a parergon. It supplements the work: it is not supposed to be necessary to the work, indeed frames are more or less interchangeable. Yet the work cannot do without a frame. The frame is two-faced: it looks outward and inward at once. From the point of view of the work, it belongs to the wide world beyond the pure cut. From the point of view of the wide world, however, it is firmly nailed to the work.

For Kant, the pleasurable or underdetermined accessory was still slightly suspect. Kant hesitated, in effect, before the thorough modern symbiosis. The full embarrassment of this hesitation is brought out in a later passage singled out by Derrida: ‘So designs à la grecque [i.e., straight-line designs in labyrinth form], foliage for framework [Laubwerk zu Einfassungen] or on wall-papers, etc., have no intrinsic meaning; they represent nothing – no Object under a definite concept – and are free beauties.’ Here the frame actually serves as an illustration of the essential properties of the aesthetic object, the locus of an artificial free beauty. There is nothing parergal about it anymore.

The recognizably modern frame was in place before any theoretical discourse had addressed it and explicated it. By the time Kant defined the aesthetic object, the frame and the artistic personality who drew or made it were givens, historical faits accomplis. Kant acknowledged the ad hoc and insufficiently theorized nature of this frame when he gave the frame, in this pair of comments from the Critique of Judgement, a ‘thickness’. The frame in these passages is not merely the public marker of a one-dimensional border. By looking past the stable and clean frame, the one-dimensional frame, and instead treating the frame either as the accessory to the work, or as the very emblem of aesthetic gratuitousness, Kant in effect placed himself within the problem that the Renaissance had posed to itself. He raised a question that was in some ways already obsolete in the eighteenth century. But in so doing he suggested the historical source of his own inquiry into the structure of the work of art.

How did the by-work come to take over the work in the Renaissance? To start with, artists began to perform in the marginal zones. The margins are the locus of the work’s own internal commentary on the conditions of picture-making, and on the identity and character of the artist himself. In these penumbral spaces the artist leaves deictic signs, signs that plainly confess the materiality of the picture and unmask any fictions of a non-existent fabricator or a transparent picture plane. The supplement was likewise the locus of the signature. Indeed, the warships which Protogenes painted in the parergia of his frescoes, Pliny reports, served precisely as a signature: they identified him as one who had begun as a mere ship-painter.183 A connoisseur culture, in turn, dedicated to discriminating between one picture and the next on the basis of workmanship or authorial handwriting, tends to follow the lead of the artist and shift its attention from centre to margin.

In Antiquity, the parergon was normally understood as subordinate, by-the-way, trifling. Pliny told endless anecdotes about the virtuosity of painters precisely because virtuosity was considered slightly frivolous and beside the point, just as we might enjoy odd anecdotes about composers or writers even though we are convinced that such curiosities have no bearing on the value of the works. It was subject-matter that mattered. Thus, the consternation of Protogenes when the Rhodians paid too much attention to an ingenious detail in one of his paintings. Protogenes had painted a satyr standing by a pillar with a partridge perched on top of it. Everyone admired the partridge. ‘When Protogenes saw that the ergon [the main work] had become the parergon’, he obtained permission from the authorities to efface the partridge.184

The German Renaissance contributed to the broader process of centring the supplement by aggressively converting portions or aspects of the picture that had once been accessory to its purpose, landscape in particular, into the most conspicuous theatres for pictorial ingenuity and inventiveness. The public personality of the artist, not merely manifested in some abstract design or idea behind the picture, but palpable in its very substance, reassembled and reunified both the picture and the beholder’s experience of it. The experience of the German picture became indistinguishable from the experience of this personality.

The key was a notion of style, or a personally specific contribution to the picture, grounded in execution rather than in invention, selection or judgement. German painters in the first decades of the new century, following Dürer, began to mark their pictures with recognizable personal styles, which often took the form of a distinct ductus or linear mannerism, or even with signatures, usually monograms or initials. The ambitious artist sought not merely to control the outcome of the work, but actually to register and lay bare his control directly in the work. Legible in the stylish line, in principle, are the conditions of its own creation, the time it took to make it, the speed of the hand.

In this scheme the artist is closer to the scribe or the clerk than to the poet. Indeed, in medieval jurisprudence, ‘style’ had drifted far away from its older rhetorical associations and had become a pragmatic means of identifying or authenticating legal documents: the maintenance of a formal or procedural standard protected the document against forgery. It provided the document with a provenance.¹⁸⁵ The historical link to a place and time of execution, in other words, was forged on the mechanical level of legal procedure, or even of penmanship. This was also the function of the clerk’s paraph or flourish, an abstract and unrepeatable signature meant to link the document with a particular hand and to foil counterfeiters.186 This recalls the old sense of style as ‘title or appellation’. Falstaff, for example, exclaimed ‘Ford’s a knave, and I will aggravate his style.’¹⁸⁷ The usage is ancient in English, but survives today only in the phrase ‘self-styled’.

This sense of style as personal mark was also current in Renaissance Germany. The theologian Hieronymus Emser, for example, thought to refute Luther’s attribution of a polemical pamphlet to him by an appeal to style: ‘Except for someone completely blinded like you, no one would judge that to be my style or invention.’¹⁸⁸ Style as an unrepeatable flourish established a link to the artist’s person that was both more direct and ultimately less suggestive than that established by style as a general principle of procedure. Here, style was explicitly associated with the concrete historical actions of a particular individual, and not merely a rhetorical decision that might reflect super-personal principles, or which might have been made at any time prior to the execution of a work. At the same time, because the executive gesture itself is not closely interrogated, the link to person proposed by a pen stroke does not yet beg difficult questions about the nature of personhood.

The self-presentations of German artists of the High Renaissance were extraordinarily strident and blunt. Only Dürer, among the Germans, rivalled the great Italian personalities in sophistication and self-esteem. And yet Altdorfer and Huber, Lucas Cranach and Hans Burgkmair, Matthias Grünewald and Hans Baldung Grien, not to mention such one-dimensional figures as Mair von Landshut or Urs Graf, or the still lesser epigones of Dürer and Altdorfer, nevertheless presumed to sign their pen drawings and to put before the eyes of the world coherent and premeditated linear styles. This phenomenon attracted little attention in Europe at large, except in the person of Dürer, who was marvelled at by Italians as a singularity, a kind of isolated comet. Indeed, Vasari made a point of mentioning that Dürer, even ‘while still quite young’, signed his prints with his own hand.¹⁸⁹ By the end of a generation, the irruption of the signature had subsided, leaving only ripples.

The claim of this book, put in its strongest form, is that the independent landscape was born directly of this radical intrusion of the German artist’s personal authority into his works. Outdoor settings were especially susceptible to the self-aggrandizing and self-advertising devices of the authorial persona. Landscape was a hospitable venue for pungent colouristic effects. Trees and mountain-ranges encouraged eccentric details and calligraphic line. Several Renaissance texts on pictorial randomness actually associate landscape with freedom from rule. In his treatise on painting Leonardo recommended that the artist examine stains on walls or mottled stones, for ‘you will be able to see in these the likeness of divine landscapes, adorned with mountains, ruins, rocks, woods, great plains, hills and valleys in great variety’.190 Vasari later reported that Leonardo’s contemporary Piero di Cosimo had delighted in imagining battle scenes, cities and landscapes in stained walls and in clouds.¹⁹¹ Imagination without study was futile, of course: remember Leonardo’s strictures against Botticelli (p. 65), who frivolously made landscapes by throwing sponges against the wall. The point was that in the landscape the imagination had plenty of leeway. Contours could bend this way or that, forms and motifs could be juxtaposed and rearranged ad libitum, all without impairing verisimilitude.

A swift, slight pen sketch by Dürer can stand as an emblem of this entire phenomenon (illus. 35).¹⁹² Here Dürer seems to be trying out a new quill, taking rapid, natural strokes, loosening his hand in preparation for a more disciplined drawing. His scratchings drift between three semantic possibilities: at some moments they mean nothing; at others they become calligraphic, like fragments of a signature; at still others they blend into the contours and shading of the rock and the bush. One starts to wonder whether there was any landscape before Dürer’s eyes at all. Perhaps linear whimsy and thoughtlessness naturally drifted toward landscape.

Many comparable observations can be made about the Venetian paese as well, in paint as well as in the graphic media. Giorgione’s damp, slow-burning landscapes are inseparable from the contemporary cult of his personality. But setting in the Italian painting was closely bound to its subject. Writing on art in Italy, which was generally at pains to adapt painting to principles of classical rhetoric, prescribed a limited and functional relationship between subject and setting. Alberti, in his treatise On Painting of 1435, gave short shrift to backgrounds, and instead addressed the problems of representing bodies and composing them into an istoria. He only briefly treated the question of drapery and flowing hair as attributes of bodies.¹⁹³ In the sixteenth century Lodivico Dolce and Gian Paolo Lomazzo both placed great weight on the principle of decorum, or appropriateness of means to ends. This was the principal source of resistance to landscape painting in the Renaissance. Lomazzo’s precepts reverberate in the treatise on ‘the arte of limning’ by Nicholas Hilliard, the Elizabethan miniaturist: ‘the landship, weare it never so well painted, it weare not to the purposse or matter, but rediccalious and false . . .’;¹⁹⁴ and in the Graphice of 1612 by Henry Peacham, which deprecates parerga as ‘needless graces’.¹⁹⁵ Northern classicists in particular were defensive on this point, with good reason. The antiquary Franciscus Junius, for instance, in his Painting of the Ancients (1638), raised this warning: ‘If we doe finde in the meane while, that the Artificers hit the true force and facilitie of grace better in these sudden [i.e., unexpected] things than in the worke it selfe, yet must wee never be so inconsiderate in our judgment, as to preferre the by-work before the work’.¹⁹⁶ At the very least, the supplement needed to be recharged with signification. This is the recommendation of Rembrandt’s pupil Samuel van Hoogstraten in his handbook on painting, Inleyding tot de hooge schoole der schilderkonst (1678): ‘The most praiseworthy way of decorating [a work of art] is with accessories [bywerk] which provide concealed elucidation.’¹⁹⁷ In practice, German painters also observed a distinction and even a hierarchy between subject and setting. Fifteenth-century Italian and German painting shared plenty of common sources. But there is no evidence that any Germans of Altdorfer’s generation were cognizant of a rhetorical or neoclassical doctrine of decorum. The only functional analyses of pictorial structure to be heard were reactionary and curt: the protests of the theologians, Catholic and Reform, against the excesses of kunst or ‘artistry’ in the sacred image. Ambitious German artists answered this objection, with which they may well have sympathized, first by cultivating style, which was free of the crass materialist overtones of mere kunst, and which was not so obviously inimical to the purposes of devotion; and second by cultivating profane subject-matter and new kinds of pictures altogether. Thus setting was relatively painlessly isolated from the composition as a whole and made available for other pursuits. Hans-Joachim Raupp has advanced an analogous argument about the peasant genre in German Renaissance art. He suggests, with a touch of deliberate hyperbole, that by the sixteenth century images of rural characters and episodes had more or less been drained of any original semantic charge. Instead, the entire genre served as an open field for the demonstration of artistic ability. In the realm of peasant iconography the artist was emancipated from the constraints of convention and doctrine, and could instead flex his style. Raupp even interprets a famous passage in Dürer’s writings in this vein: ‘An artist of understanding and experience can show more of his great power and art in crude and rustic things, and this in works of small size, than many another in his great work.’¹⁹⁸

35 Albrecht Dürer, Quarry, c. 1510, pen on paper, 6.1 × 4. British Museum, London.

When the German artist applied his pervasive stylistic mark to the picture and began to upset the hierarchy between subject and setting, the boundary around the picture as a whole – the frame – became clearer and correspondingly more important. The frame defined the picture as the work of an artist. Wherever the artist-author manifested himself became the work of art; everywhere else was not-the-work, or the world. By severing the work from the world, the frame made the work truly portable. The frame isolated the work from ordinary objects and from the world in general, just as artists would eventually be distinguished from ordinary people by certain defining myths about them, and by expectations about their behaviour and appearance.

But there is a paradox here. The idea of style as trace – as a metonymy for the artist – is incompatible with the doctrine of the pure cut. The trace inhabits the work like an illicit agent from the world abroad. Thus the integrity of the work of art was compromised at its very inception. For it was exactly the authority of the self-promoting artist that severed the work from the world in the first place. And yet that artist’s liberating performance immediately broaches the new imaginary frontier between inside and outside. The trace makes him present and absent at the same time. And in truth, the doctrine of aesthetic autonomy makes no more room for the deictic gesture than had the ancient orthodoxies of the Christian doctrine of images. In some ways the problem is more pressing here. For the old sacred image could afford to exclude the author: subject-matter alone, the holy personage, provided the work with its raison d’être. But what would justify the image of landscape, if not its author? This question was asked for the first time in the German Renaissance; European art posed the same question to itself repeatedly over the next five centuries. No answer was ever again so spirited and impressive as the first one.