Dr. Kim H. Veltman
VI Art and Representation
1. Introduction
2. Objects
3. Functions
4. Narrative
5. Contexts
6. Conclusions
Perspective is inevitably associated with Renaissance painting: with Masaccio's Trinity, Leonardo's Last Supper and Raphael's School of Athens. However, if we are to assess the implications of perspective for representation and art, it is necessary, at the outset to emphasize how it affected both a variety of media, notably, marquetry, fresco, sculpture, as well as painting and a variety of objects including doors, altars, furniture, walls, ceilings as well as painted panels. It will be useful also to explore connections between perspective and different goals or functions of art to show that various goals essentially precluded the possibility of perspective and thereby to suggest new reasons why perspective occurred so late in the history of representation and specifically in a European context. We shall claim that there were special links between perspective, literacy and the narrative function in particular; that this interplay between perspective and narrative was by no means straightforward; that it had consequences which inevitably undermined a strict narrative sequence, leading to new contexts of art which we now associate with high renaissance, mannerism and baroque art.
Ultimately perspective had artistic consequences which went far beyond this, transforming the architecture of buildings, landscape gardens and indeed the whole environment; changing the meaning of metaphor and illusion in ways that provoked new avenues of artistic freedom and creativity. These will be the concerns of the final two chapters. Here we shall limit ourselves to the consequences of perspective on representation and art in a stricter sense.
In an earlier chapter we have already noted that the development of perspective in painting was part of a larger movement in European culture which affected space in architecture (pl. 4.1-3) and a whole variety of media, including bronze (pl. 4.4), manuscript decoration (pl. 4.5), marble (pl. 5.2), stone (pl. 5.7), drawing (pl. 6.1-2, 4) as well on painting. Wood was another medium. Indeed, the art of marquetry or intarsia was among the most important early contexts for perspective. Vasari reported that "the method of joining small pieces of coloured wood together to make perspectives"1 was introduced in the time of Brunelleschi and Uccello. In his life of Paolo Uccello, Vasari described at further length:
When Paolo showed his intimate friend, Donatello the sculptor, mazzocchi with projecting points and bosses, represented in perspective from different points of view, spheres with seventy-two facettes like diamonds and on each facette shavings twisted round sticks with other oddities upon which he wasted his time,the sculptor would say: "Ah, Paolo, this perspective of yours leads you to abandon the certain for the uncertain, such things are only useful for marquetry, in which chips and oddments, both round and square and other like things, are necessary.2
Marquetry soon became much more than a play of chips and oddments in creating regular and irregular solids. In rare cases, such as Urbino (pl. 74.1) or Gubbio marquetry served as an elaborate wall decoration, simulating furniture, painted emblems, instruments, windows and even squirrels. In Italy marquetry became particularly popular in the choir stalls of churches such as Todi (pl. 74.2), Verona, Bergamo, where the representation of man-made objects, books, chalices, instruments (pl. 75.3-4) frequently alternated with man made scenes, usually idealized views of cities. The practice also spread to furniture, which gained importance in Germany and the Netherlands. In Augsburg, Lorenz Stoer (1567), produced a book of perspectival examples specifically for craftsmen in wood (pl. 75.1-2), involving various semi-regular solids and ruins. Idealized ruins became a significant theme in its own right (pl. 75.4).
In his Lives of the Artists, Vasari discussed the inlaid chests of Giuliano da Maiano in the sacristy of S. Maria del Fiore and the marquetry portraits of Dante and Petrarch by Benedetto da Maiano in the Palazzo Vecchio. He also mentioned the marquetry of Giusto and Minore, Guido del Servellino, Domenico di Mariotto, Battista del Cervelliera, Giovanni da Verona, Baccio D'Agnolo, Francesca de Salviati and Fra Damiano.3 Even this was but a small sample of activity in the field. The chronicler, Benedetto Dei, listed 84 workshops around 1470 devoted to marquetry in Florence alone.4
Bronze was another medium. In the case of the doors of the Baptistery in Florence, Vasari noted how Ghiberti "introduced a building in perspective with great effect"5 (pl. 4.4). In the case of Donatello's bronze reliefs on the high altar of San Petronio in Padua, Vasari recorded that they were:
6executed with such judgment that masters of the art have been struct dumb with admiration in beholding them when they have considered their beautiful and varied composition comprising such a number of remarkable figures placed in diminishing perspectives.
Modern tourists tend to be slightly more blasé.
Marble was yet another medium. Donatello's famous sculptured relief of Saint George and the Dragon on the facade of Or San Michele in Florence at once comes to mind. Indeed perspectival images were used in different media for almost every object. Uccello, according to Vasari, even made "a number of small pictures in perspective for the sides of couches, beds and other things"7 in many houses in Florence.
Certain objects, particularly altars, played a special role in the development of perspectival representation. In the Italian tradition, altars were typically dominated by a central saint, often the Virgin Mary seated on a throne. By the late thirteenth century artists became concerned with rendering these thrones spatially. In the altars of Cimabue, Duccio and Giotto a clear progression can be traced which led, via Daddi, Simone Martini and Fra Angelico to famous examples such as Domenico Veneziano's Saint Lucy Altar or Mantegna's Saint Zeno Altar. The throne, in turn, became an excuse for spatial treatment of an entire context as seen in Piero della Francesca's Brera Altar (pl. 7.1) and altars by Bellini, Cima da Conegliano, Cosme Tura and a host of others. As in the case of niches, vaults and windows mentioned in our opening chapter (p. ), thrones were another of those basic forms which became part of a cumulative heritage of spatial ingredients.
Meanwhile the Franciscan movement had focussed new attention or episodes in the lives of Christ and the Saints which had various consequences for the history of altarpieces. In rare cases, such as Duccio's Maestà (pl. 2.1), such episodes dominated the back of an altar. In the case of the Master of Saint Cecilia, six scenes from the life of the saint flanked a central panel with an enthroned Madonna. As the fourteenth century progressed another alternative became standard, namely, to continue having a dominant central figure flanked by two, four or more saints, beneath which were small predellas or panels with scenes from the life of Christ or a saint. These panels are of particular interest for our purposes for they contain important cases of perspectival space. An early example was Simone Martini's Saint Louis of Toulouse (Naples, Museo di Capodimonte, c.1317) which includes five scenes from his life in these panels below the main figure. A little over a century later there was Gentile da Fabriano's Altar of the Adoration of the Magi (now Florence, Uffizi, 1423 with a predella now in Paris, Louvre) and his Quaratesi polyptych (now Florence, Uffizi with its predellas in the Vatican and Washington, National Gallery).
More dramatic were the predellas beneath Fra Angelico's Louvre Altar (1434-1435), Carlo Crivelli's Polyptych of Massa Fermana (1468), or the altar of Matteo di Giovanni (San Sepolcro) which originally had Piero della Francesca's Baptism (now London, National Gallery) as its central panel. Piero's own Polyptych of Saint Anthony (Perugia, Galleria Nazionale dell 'Umbria, 1460) transformed this tradition with three spatial scenes beneath and a subsequent perspectival Annunciation (1470) above the three main panels.
In terms of perspectival effects, Renaissance altars are of special interest because they highlight differences between approaches to space South and North of the Alps. While the Italians were intent on representing space, artists in the North were equally concerned with reconstructing space. In some cases, such as the Mengelberg Altar in the cathedral at Cologne, this amounted to a three dimensional sculpture of a scene. More often an extraordinarily detailed wooden carving was involved, as in Tilman Riemenschneider's Altars, such as that in the Hergottskirche at Creglingen (c. 1505-1510), the Agilophas Altar in the cathedral at Cologne (c. 1520), Hans Bruggeman's Bordesholmer Altar in the cathedral at Schleswig (1621), or the magnificent altars in Lübeck and Stockholm.
Art historians have frequently characterized the represented space of Italian altars as Renaissance, while describing the constructed space of Northern altars as Gothic, i.e., primitive and behind the times compared to Italy. To insist on the contrast would lead us to overlook cases such as Michael Pacher, who was involved in both sculptured and painted altarpieces and to ignore a whole series of altars which combined sculpted and painted elements such as the Deocarus (11436-1437) and Saint Roch (c. 1490) altars, both in St. Lorenz in Nürnberg.
There is a further reason for seeing the fascination with reconstructed and represented space as two facets of a single development, for it offers insight into a phenomenon connected with Flemish painting, namely, the tendency to represent figures in paint as if they had been reconstructed in sculpture, both with respect to isolated individuals (e.g. pl. 76.1-3) and series of figures (e.g. pl. 8.1, 3-4). Had there not been such sculptural masterpieces as precedents, the perspectival breakthroughs in painting would have been virtually impossible. In this context the whole paragone debate, which opposed the merits of painting and sculpture, obscured an historical interplay which had been essential for the development of painting.
There are striking parallels in the development of choirs South and North of the alps, and it may well be that a similar revision of thought is necessary in this case. Perhaps we need to see the intarsia representations of Dante and Petrarch by Giovanni and Giuliano da Maiano in the Palazzo Vecchio, which Vasari attributed to Botticelli, as closely related to the sculptured busts found in cathedrals such as Ulm and Vienna. Perhaps Italian marquetry and Northern misericords should again be seen as two facets of a single fascination with three dimensional space. If so, we might be closer to understanding why Ghiberti, Donatello, and indeed so many of the early perspectivists were both sculptors and artists. Sculpture provided the reconstructions or models which made feasible their representation in space.
The full implications of these suggestions are more dramatic than may be immediately apparent. We need to return to the claim made at the beginning of our essay, where we drew attention to parallels between developments in Romanesque constructed, architectural space and reconstructed space (pl. 4.1-2) which the Gothic period developed (pl. 4.3) in cases such as the West facade of Notre Dame in Paris, and which the Renaissance subsequently represented (pl. 4.4).
If we look more closely at the portals of Notre Dame we see that this constructed spaced is peopled with realistic three-dimensional sculptures which,-- we are tempted to say who--, are the direct ancestors of the sculpted portrait busts in the choirs of Ulm and Vienna, as they are of the figures on the facade of Or San Michele: Nanni di Banco's Martyrs, Ghiberti's St. John the Baptist and Donatello's St. George. In short, attention to individual, sculptural realism, which Auerbach vividly characterized as creatural realism,8 went hand in hand with exploration of constructed, reconstructed and ultimately represented space.
To state the problem in these terms may seem altogether obvious. Yet to accept the proposition is to recognize the constructions and reconstructions of Northern Gothic architecture and sculpture as essential ingredients for the representations which the so-called Italian Renaissance made famous. From this many things follow. Renaissance perspective would have been unthinkable without Gothic art. The Gothic North was not behind Italy: it was exploring other dimensions of the same spatial problems of creatural realism. Which is why perspective spread so quickly to the North. As we have already mentioned, Van Eyck, Bellini and Alberti were contemporaries, as were Filarete, Fouquet and Memling. So too were Piero della Francesca and Dirk Bouts. Perspective was a European phenomenon, and in this context it is neither mysterious or paradoxical that it very quickly became as much, and soon, primarily a Northern affair.
Of course, some will argue that all such comparisons between North and South are very misleading: that there is a world of difference between Van Eyck's paintings where lines converge towards a central area and the mathematical precision of a vanishing point as specified by the legitimate construction. Yes, the point is well taken if early Northern practice is compared with early Italian theory. But the opposition disappears if both regions are compared in terms of practice. For instance, closer inspection of inlaid marquetry panels such as those at San Sepolcro connected with the school of Piero della Francesca, reveals that their perspectival constructions are imprecise from a strictly technical standpoint. Very often the perspectival lines were drawn by hand rather than constructed by ruler. This happened equally in other media. The foreshortened squares in the floor of Raphael's School of Athens (cf. pl. 11.5) were hand drawn, and from a strictly geometrical point of view were quite incorrect. The floor patterns further back were grossly distorted. Such irregularities or so called errors were not simply due to difficulties inherent in the medium. They were so universal that perspectivally perfect works of art were virtually non-existent in the Renaissance.
Artists ignored technical perspective in its narrow sense because they obviously soon discovered that they could achieve extraordinary spatial effects such as we still experience in seeing the School of Athens (pl. 11.5), even if the lines were slightly crooked and only approximately right. Which helps explain why Dürer, who went to such trouble to learn perspective, only used it technically in two or three of his hundreds of works.9 And this could explain why Van Eyck never even bothered to try. What is important for our purposes is the slightly paradoxical fact that, while perspective in its narrow technical sense was virtually ignored by Renaissance artists,--Vasari even complains that Uccello was wasting his time in trying to get it right,10--perspectival effects nonetheless had a seminal influence on representation of the time.
Thus far we have outlined these effects on a whole range of media including marquetry, fresco, bronze marble and painting, as well as noting its significance in the case of particular objects such as altars. The most remarkable examples of Renaissance perspective have yet to be considered, namely, those on the walls of churches. They were usually frescoes. Some dealt with a symbolic theme in isolation as in the case of Masaccio's Trinity (Florence, Santa Maria Novella, c. 1427-1428). Some were portraits of famous individuals, such as Uccello's Sir John Hawkwood (1436) or Castagno's Niccolo' da Tolentino (1456) both in the Cathedral at Florence. Some presented events, as a single scene as did Uccello in his Battle of San Romano (London, National Gallery, c. 1445), although in this case he made two other versions (Paris, Louvre and Florence, Uffizi).
Yet, in a sense, all of these were exceptions for, as in the case of altars, the great examples of perspectival wall frescoes involved story telling and cycles of images and once again, the roots of this practice lay firmly in the mediaeval tradition. To be sure, as Gombrich has shown,11 mural painting had its origins in early Antiquity and elements thereof evolved in an almost unbroken tradition. Of particular interest for our purposes, however, are developments of narrative sequences evident as early as the ninth century in the Reichenau, and the ceiling at Hildesheim in the eleventh which led, by the late twelfth century to that remarkable seequence of 167 scenes from the old and new testaments in the cathedral at Monreale.
It is this tradition which leads via Giotto in Assisi, Padua and Florence to the great perspectival cycles of the Renaissance: Masaccio in the Brancacci Chapel (Florence, Santa Croce); Mantegna in the Ovetari Chapel (Padua, Eremitani Church, 1454-1457); Piero della Francesca in San Francesco in Arezzo (c. 1452-1457); Ghirlandaio in the Sassetti Chapel (Santa Trinita, 1483-1486) and the Capella Magiore of Santa Maria Novella (1485-1490) and, of course, Raphael in the Stanze of the Vatican (c. 1509-1516). We shall wish to examine more closely the connections with narrative which generated these prime examples of Renaissance perspective, but by way of context a detour is necessary.
One of the enduring contributions of Sir Ernst Gombrich has been to demonstrate that the history of art cannot be seen as a single, simple line of development and must be studied as a series of parallel and even conflicting goals: that the magical function of so called primitive art, must be separated from that of ornament, and mimesis. If this approach is taken in combination with the idea of different levels of literacy it offers further distinctions between these goals of representation which, we shall suggest, are useful in understanding why perspective occurred when and where it did. Indeed of the six basic functions to be considered (connecting, ordering, imitating, matching, mixing, exploring) we shall show how three precluded, one discouraged and two functions encouraged the use of perspective, and then only under special circumstances.
Connecting
So called primitive art12 had a function of connecting a totem in a community with a magical or sacred world beyond it. This connecting function meant that a totem actually became a given deity rather than being a simple representative thereof. A sculpture suited this function much better than a painted representation. Because it served as a bridge between the everyday world of the tribe and a magical world beyond, it had to be sufficiently life-like to be recognizable by its viewers: i.e. anthropomorphism was an inherent part of the system. Yet a fully realistic statue would have linked it too firmly to the present world and thrown into jeopardy its connecting function with a world beyond: i.e. a restricted anthropomorphism was also built into the system. In such a context perspectival realism would have been more of a threat than a goal.
Since these tribal communities were pre-literate there were no canonical texts concerning the shape and meaning of the statue. And in the absence of these sacred texts to establish a sense of community, the sacred statues acquired proto-canonical functions themselves and forged this sense of community directly. Hence any serious deviation in outward appearance was a threat to its connecting function because it introduced the risk that a specific god would not even be recognized.
Since polytheism was the rule in pre-literate societies these principles usually extended to a number of gods. As the number of gods increased, the powers which could be attributed to a given god decreased. Such considerations meant that there were natural controls to keep in check an indefinite extension of these sacred images. In this context the pantheon of images which would have been possible through perspective was necessarily a threat to be avoided rather than a goal to be sought. The connecting function precluded an interest in this-worldly, perspectival space, focussing attention instead on a totem which would ensure contact with a world beyond.
Ordering
A second goal which emerged among primitive tribes involved ordering, producing patterns and ornament beginning with simple regular lines and evolving to ever more complex geometrical shapes. In pre-literate societies these patterns were usually restricted in number and had sacred connotations such that they shared partly in the connecting functions of totems. In some cases, these patterns were applied directly to the totems, such that, both the connecting and ordering functions were contained. A gradual distinction between the two functions was inevitable, however. For whereas the connecting function effectively depended on a pre-literate society, the ordering function did not. The advent of literacy simply extended its repertoire as Sir Ernst Gombrich has so eloquently shown in a Sense of Order.13
Some patterns could even be given spatial characteristics. The menander fret could, for instance, be given a three dimensional effect through a clever use of light and shade. Yet although some sense of depth was possible, systematic treatment of space was not. Hence ordering was another goal which discouraged perspective in its full sense.
Imitating
The primitive mind which saw images as connecting with a magical world beyond, believed in an identity of image and god. A next stage in civilization denied this identity and recognized that the two were separate: that the image was at best an imitation or representation of the god involved. If this distinction between the two mentalities was logically simple, psychologically the distance between them was enormous and occurred gradually during the period between c. 4000 and c. 500 BC.
The shift from connecting to imitating was closely linked with the emergence of literacy in the cultures of Akkad, Babylon and Egypt. Thus far the connecting function had been limited to sacred images. Now it spread to sacred texts and those who controlled them. In Egypt, for instance, the Sacred Book of the Dead became a repository of these magical connections as did the pharaoh. This posed new problems for the production of images. On the one hand, if an image of the pharaoh was to function as a (living) image rather than a representation, it had to become fully realistic and lifelike. On the other hand, this very realism undermined the statue's connecting function, linking it with an other worldly realm. Instead of being recognized as an immortal figure, it now risked being seen as representing an all too mortal figure. One protection against this was to control viewing conditions: placing the statue in a dark place, laden with other-worldly atmosphere under ambigous conditions. Which is precisely what the Egyptians did. The statue in the doorway of the Mastaba of Mereruka at Saqqara comes to mind.14
These principles, designed for a supposedly immortal pharaoh, were inevitably extended to others in his midst. By c. 2580 B.C. these included members of the royal household in the form of reserve heads at Giza (now Cairo, Egyptian Museum). By c. 2400 B.C., they included mortals, such as seated scribe from Saqqara (Paris, Louvre). As this repertoire of mortal images increased, the need to recognize them a representations rather than (living) images became more acute. The crisis or so called revolution came in Greece. In Sir Ernst Gombrich's account:
Narrative art is bound to lead to space and the exploration of visual effects.when classical sculptors and painters discovered the character of Greek narration they set up a chain reaction which transformed the methods of representing the human body--and indeed more than that....
For this reason he believes that it was:
It is here, in the context of plays, based on the ancient mythical tales, that the re-enactment of events according to the poet's vision comes to its climax and is increasingly assisted by the illusions of art.15surely no accident that the tricks of illusimistic art, perspective and modelling in light and shade, were connected in classical antiquity with the design of theatrical scenery.
According to this account, an interplay between literature and art sparked the Greek revolution in art, introducing a form of imitation which amounted to matching objects in the visual world, i.e. perspectival representation. Thereafter the Renaissance was little more than "the return to the classical ideal of the convincing image."16 As will be clear from our opening chapter, in our view the situation was more complex. Mimesis or imitation meant at least five different things. We shall examine each in turn to show that none of them was synonymous with matching in the perspectival sense.
A first meaning involved imitation of verbal narrative. If we accept Gombrich's fundamental insight that narrative texts inspired much of Greek art, we must also accept the consequences. Representations of verbal descriptions of visual objects were not direct records of the visual world. They were imitations, via a verbal filter, of Greek literature which, as Auerbach has shown, had no clear sense of reality when compared to the Judaeo-Christian tradition. The resulting art may have had visual effects or the appearance thereof, yet it remained non-visual in terms of its sources. It never aimed at recording visual reality directly and as such was never concerned with perspective.
A second meaning involved imitation of ideal concepts of objects and persons. There is a well known story of the Greek sculptor commissioned to do a statue of Venus who studied different virgins and the combined their features in producing his ideal statue. Here again there was no interest in recording an individual visual record. This was imitation via the filter of a mental visual image based on a universal concept of Venus which, it will be noted, amounted to much the same thing as a verbal filter. Both mental visual and verbal filters were universal and ultimately opposed to the individuality of a perspectival record. Hence this second kind of imitation was equally non visual.
A third meaning entailed imitation of objects in isolation. The same principles which had led in Egypt to detailed images of the pharaoh and isolated members of his court were extended in the second millenium B.C. to isolated animals and birds such that we find in an otherwise primitive scene, a lion or some geese of striking realism.17 This attempt to copy simple objects precisely continued in Greece but usually in terms of statues rather than paintings. It was presumably to this end that Polycleitus developed his famous canon, a statue which served as a model for others. Pliny also recounts the story of illusionistic grapes by Zeuxis which fooled the birds and the illusory curtain by his competitor, Parrhasius, which in turn fooled Zeuxis.18 These again involved isolated objects rather than contexts, and not unlike similar effects in ornament, depended on some simple tricks of light and shade rather than principles of perspective in order to achieve their results. More significantly we are told that this type of realism represented an early stage: that artists first represented objects as they were and later as they appeared.19 Had the Greeks discovered perspective it would have been in the later period. But as will be shown, this period also had goals inimical to perspective.
A fourth meaning involved imitation of objects using optical adjustments. This was a possibility, against which Plato complained in his Sophist (see above pp. ). It entailed adjusting the original proportions of statues in order that they appear correct. Based on a theory of visual angles, this method imposed on an object a mental concept of how it ought to appear. For it sought to integrate effects of size and distance in the object and keep the image constant. This was a goal fundamentally different from Renaissance perspective, which began with the premise that objects remained constant and sought to record visually effects of size and distance on images. For this reason the visual angles method was paradoxically non-visual in the Renaissance sense of the image. It did not visually record, but rather it physically adjusted objects, so that no change in the visual appearances would be noticeable. As we have suggested, (pp. ) the same mentality applied in astronomy. Saving the appearances was more important than the actual causes governing them. Subjective appearances dominated over objective.
An important shift had begun. The primitive mind had projected magical qualities onto its images. The semi-civilized mind projected its theory of appearances onto images while the civilized mind has attempted20 to produce images devoid of these psychological projections. And the domain of study shifted accordingly from an unseen, magical world beyond to a conceptual world of appearances and finally to a perceptual, visual world of objects. As long as psychological projection onto images continued, study of their perspectival aspects could not yet begin in earnest.
A fifth meaning of imitation involved illusionistic effects of stage scenery using optical adjustments methods which were also affected by this problem of psychological projection of a theory of appearances onto objects. But here there was also a deeper problem on which we touched in our analysis of the famous Vitruvian passage at the beginning of our essay (pp. ). The stage settings were illusionistic in a special sense. They involved hypothetical buildings which never have existed in the physical world. Nor could the space they appeared to represent. Unlike perspective which permits a measured relation between pictorial space and real space, here the buildings and spaces produced a fictive world closed onto itself.
Hence mimesis was many things, and the Greek revolution introduced approaches to art as representation which resembled matching. But ultimately these involved imitating distorted by a mental21 visual or a verbal filter. There existed as yet no systematic quest to record the visible world passively, rather than imposing adjustments on it actively.
Matching
The subtle shift from imitating to matching became a conscious programme during the Renaissance when, as Vasari noted, artists:
This encouraged them to make rules for perspective and to get their foreshortening in the exact form of natural relief.22sought to reproduce what they saw in Nature and no more and thus they came to consider more closely and understand more fully.
It is important once more to stress how gradual was this process. If, for example, we consider some of the chief themes open to artists we could list at least eight basic visual themes in the natural world : portraits, human figures, persons at work, persons at war, persons at play, animals, landscapes, man-made objects, and four other verbal-visual sources deriving from literature ( myth, literature, religion, history). Most images in the Renaissance and the chief instances of perspective were inspired not by the visual themes (1-8), but by religion, and specifically, the Bible, and a few books on lives of the saints. Or to put it slightly differently, matching could involve the visual world; the visual world illustrating verbal sources, i.e. manuscript illustration; and the visual world illustrating implicit common verbal, as in books such as the Bible which were so much a part of a general cultural heritage that their basic themes could be taken as implicit and requiring no explanation. Of these, it was the final of these alternatives which inspired the most striking cases of Renaissance perspective. To understand this we must return to the problem of narrative.
In terms of narrative, it was precisely the best known stories which generated the classic examples of perspective. For instance, in the life of Christ, it was particularly the Annunciation (e.g. pl. 83.1-4) or the Last Supper which became topics of perspective, although other themes included the Birth, Adoration of the Magi, Flight into Egypt, Preaching in the Temple, Marriage at Cana, Flagellation, Crucifixion, Emmaus (pl. 85.2) and the Resurrection. Why should the best known stories become the most perspectival ones?
A contrast with conditions of primitive connecting and Greek imitating is instructive here. In pre-literate societies the statue of a god, as an object which members of a tribe had in common, helped to define the group's communality. As already mentioned, this limited potential variations since deviations from the norm involved risk that the statue would no longer be recognized. With the advent of literacy, this changed. Texts recorded the characteristics of a given god or Deity, thus providing a corpus of what persons knew and had in common, a sense of communality, and since this burden no longer lay with the image, it could now be varied. The more famous a story became through texts, the more liberties could be taken with its representation. Perspective was a key to varying images. Hence the best known themes also became the best examples of perspective.
With respect to the Greeks, it will seem that we have contradicted ourselves. For if Greek narrative precluded perspective, why then should Biblical narrative involve perspective? As Auerbach has shown23 the two traditions had fundamentally different approaches to reality. The Homeric tales were fictions guided by rhetorical ends of story telling, conflating myth and history, leaving no clear relation to reality. The Biblical stories, by contrast, were based on a belief in creatural realism, and were historical, such that their temporal and spatial coordinates were usually clear. The interpretation of Biblical narrative given by the Franciscan movement stressed this creatural realism. The birth of Christ was not merely treated as a story: a real child was laid in a manger and local peasants re-enacted the role of Mary, Joseph and the shepherds.
This was fundamentally different from the Greek theatre (cf. p. 2), which had developed impossible spaces setting it apart from the physical spaces of real architecture and everyday life. In the Franciscan movement, the story of life became a direct extension of the story of Christ and the narrative space in Christ's story could, and implicitly had to be extended into the space of real life. As Christian artists of the latter middle ages explored this narrative space these connections with physical space became ever more explicit until the possibility, even the necessity of matching pictorial and physical space became explicit also.
Both primitive connecting and Greek imitating had been constrained by magical and ideal considerations, which acted as filters limiting art to universals of invisible and verbal worlds. The new concept of matching opened the horizons of artistic representation to the particulars of the visible world, which expanded even more through the prospect of varying.
Varying
In the case of the Annunciation, this process of varying has begun even before the rules of perspective had been formally established, as is evidenced by Pietro Cavallini's Annunciation (Rome, Santa Maria in Trastevere), or Ambrogio Lorenzetti's version (Siena, Accademia, 1344), generally accepted to be the first painting in which all the lines of the tiles converged to a single vanishing point. After Alberti's first treatise (1434), and particularly after the advent of printing in the 1450's, variation increased in scale. Some examples, such as the unknown fifteenth century painter in Santa Maria Novella, continued to produce rough empirical versions. Fra Angelico produced several variants using an open colonnaded space (e.g. Madrid, Prado), thus developing a form used earlier by Nicolo di Pietro Gerini (New Haven, Yale University Collection, 1375); or another with a portico opening into a garden (Florence, San Marco), a theme which Domenico Veneziano (Cambridge, Fitzwilliam, pl. 7.5) also explored. Sometimes the scene was inside on a regular pavement, as in the anonymous Annunciation in the Gardner Collection. Sometimes it was outside on such a pavement, as in the version by Francesco di Giorgio and Naroccio di Landini in the Yale Collection. At other times it was outside in a green garden as in the versions by Filippo Lippi (London, National Gallery) and Leonardo da Vinci (Florence, Uffizi).
Crivelli, by contrast, developed a spatial example from Bellini's Sketchbook (pl. 83.1-2) in his Annunciation (London, National Gallery, pl. 83.3) which was at once symbolic of Christ's coming and at the same time a record of a papal grant by Innocent III to the citizens of Ascoli Piceno concerning certain rights in self government, which reached the town on the feast of the Annunciation, 25 March 1492. He thus combined information from a biblical text, a sketchbook and historical record. More complex textual sources called for a more complex picture, which required complex spatial arrangements made possible by perspective.
Any attempt at classifying the full range of variants on the Annunciation would be a large book in itself. For our purposes it will suffice to note how every region developed its own variants on a subject. In Florence, Annunciations inside homes were the exception (e.g. Pollaiuolo's version in Berlin, Staatliche Museum, pl. 13.1). By contrast, Flemish versions were normally indoors: sometimes in living rooms, as in Robert Campin's version in the Metropolitan, sometimes in bedrooms, as in Rogier van der Weyden's version in Munich, Alte Pinacothek (pl. 12.2), or in the apses of churches, as in Jan van Eyck's version in Berlin, Staatliche Museum. In Germany, Annunciations were also frequently in bedrooms, as in Durer's woodcut, and in churches as in Grünewald's Isenheim Altar (Colmar, Musée d'Unterlinden, 1510-1515), but with very different uses of space. Meanwhile, other Flemish versions had combined elements of the living room, bedroom and church interior in a single, rather unlikely space as, for instance, in the Annunciation attributed to Henri met de Bles (Cambridge, Fitzwilliam Museum). Variants of this composite spatial arrangement became popular in Spain as witnessed by Alejo Fernandez' version (Seville, Museo de Bellas Artes) or in Berreguete's Annunciation (Burgos, Cartuja de Miraflores).
This tradition of using perspective to create unexpected variants of a familiar theme was further developed in the seventeenth century, by which time varying went hand in hand with explorations of scale. In the case of Saenredam, for instance, nine of the eighteen construction drawings for his famous interiors involved a single church, St. Bavo, in Haarlem (e.g. pl. 20.5) which was further studied by De Witte (pl. 20.6), while Berckheydye depicted its exterior from different points of view (pl. 22.1-23.1).
In terms of narrative, varying had a two edged effect on the story-telling process. On the one hand, it made a theme such as the Annunciation immensely rich in its many representations. On the other hand, in focussing so much attention on a key theme, it undermined, and even prevented interest in other elements of the story. Perspective which grew out of narrative thus posed a threat to a story's continuity. This was not.only due to varying. It was caused also by a second feature of perspective which we shall term emphasizing.
Emphasizing
Perspective emphasized scenes in particular ways. It exaggerated the geometry of the man-made environment, thereby drawing a viewer's eye into a spatial scene, while at the same time reducing individual figures therein to a diminutive size. This was no problem in the case of idealized cities (e.g. pl. 96.3), but proved inconvenient in a Christian tradition which focussed on Christ, Mary and various saints. A compromise thus ensued. Individual figures continued to dominate the main panels while perspectival scenes relating to their lives were relegated to the predellas, as was seen in our discussion of altars (pp. ). Once the laws of perspective began to be understood in the 1430's, artists gradually discovered means of keeping figures in the foreground of perspectival settings. Domenico Veneziano's Annunciation (Cambridge, Fitzwilliam, pl. 9.5) was an early example. Piero della Francesca's Flagellation (Urbino, Galleria Nazionale dell Marche, c. 1460-1470) marked an important next step leading to the most famous cases of the high Renaissance: Leonardo's Last Supper (Milan, Santa Maria delle Grazie, c. 1495-1497) and Raphael's School of Athens (Vatican, Stanze, 15-15, pl. 11.5).
In cases such as the Last Supper, there were psychological factors which combined to augment this process of emphasizing. Just as in portraits where eyes looking out of the picture continue to follow a viewer as they move to the side, perspectival pictures with alleys, corridors, rooms or any regular spatial features also follow a viewer as they move to the side.24 For this reason we can look at perspectival settings in theatres and movies, which are an extension of perspectival principles, from a number of seats. Michael Kubovy, who has recently explored this phenomenon, has termed this the robustness of perspective.25
Artists such as Leonardo obviously realized that the Last Supper would work even though its vanishing point was at a height above that of any ordinary observer. Indeed, precisely because it could be looked at without undue distortion from anywhere within the refectory of Santa Maria delle Grazie, was a major reason why it was worth emphasizing this painting to the exclusion of others. The same applied in the case of Bramante's fictive arch in Santa Maria presso San Satiro in Milan (pl. 5.4), and Tullio Lombardo's scenes from the life of St. Mark in Venice (fig. 5.3). The fictive depth involved might be small, as in Piero della Francesca's Brera altar (pl. 7.1), or large, as in Masolino's version of Herod's palace at Castiglione d'Olona. The effects remained the same. And, as in the case of the varying function, the emphasizing function of perspective focussed attention on key episodes of a narrative thus serving also to undermine the continuity of a story. Yet a third factor contributed to this process.
Relating
In representing a story with many episodes painters were faced with a problem of individuating scenes. Frames were of some help, but these could not give many clues concerning the order in which scenes were to be read. Here perspectival treatment of certain features helped to relate scenes while at the same time separating them. We have already alluded to this problem in a quite different context (pp. ) using the example of Duccio's Maestà (Siena, Museo del Duomo, 1288, pl. 2.1), but it will be useful to consider it again in more detail. On the reverse side of the altar, the story begins in the bottom left hand corner with Christ's entry into Jerusalem, moves in an up-down sequence towards the right, then returns to the upper left hand corner again criss-crossing its way to the far right. Three scenes with Christ and his Apostles (Washing of the Feet, Last Supper and Meeting with Apostles ) all share one type of spatial interior with beams of the ceiling converging towards a central axis. Three scenes with Caiphais and the priests occur in an interior with a type of oblique parallel projection. A similar oblique parallel method applied to an awning supported by columns connects scenes with Pontius Pilate in the bottom right and top left. These proto-perspectival elements thus relate separate scenes and help us to follow their sequence.
In the Scrovegni Chapel at Padua (1304-1307), Giotto uses the same principle. An oblique view of an open fronted house serves for both the Annunciation to Saint Anne and the Birth of the Virgin. Similarly, a temple with a niche serves as a continuation between three scenes: the Ceremony of the Rods, Prayer for the Miracle of the Rods and Marriage of the Virgin. This function of relating separate scenes in a complex narrative explains why a few proto-perspectival elements became stock images, which improved empirically, while other architectural elements remained spatially awkward and unconvincing. And as we noted earlier it was precisely these stock images which were consolidated and standardized by the early perspective treatises (cf. pp. , pl. 2.2-3, 3.1-6).
Relating took on many forms. In his Profanation of the Host (Urbino, Galleria Nazionale delle Marche, pl. 78.1), Paolo Uccello used two vanishing points going in different directions in order both to separate and relate the two scenes. The same principle was used in the Munich manuscript of Boccaccio (pl. 78.2), in the organization of the Teatro Olimpico at Vincenza (pl. 78.3) and in the gardens at Versailles (pl. 78.4). Hence scenes with different vanishing points could be implicitly related by means of perspective. Scenes physically separated from one another were also explicitly related by means of a single vanishing point. Giotto's Annunciation in the Scrovegni Chapel in Padua (1304-1306) was an early attempt in this direction. Masaccio--and Masolino?--developed this idea in the Annunciation in San Clemente in Rome, while Foppa used it dramatically in his Annunciation in S. Eustorgio in Milan.26 Nor was the principle limited to Annunciation scenes. Parronchi has suggested that Ghiberti used it on the doors of the Baptistery at Florence27 and has shown convincingly that Masaccio employed it a relating the Distribution of the Goods with Saint Peter Curing the Sick in the Brancacci Chapel (Florence, Santa Croce, 1426-1427).28 Once familiar, the method was used in more subtle ways. Spatially analogous scenes were related without their sharing a common vanishing point as, for instance, in Piero della Francesca's Annunciation and Dream of Constantine in Arezzo.
Raphael developed these principles of relating in his famous juxtapositions of sacred and profane scenes in the Stanze. Here the situation was complicated by typological and symbolic considerations. The mediaeval period had seen an increasing fascination with parallels between the old and new testaments with minor references to relevant pagan figures such as the sibyls. This inspired the ceiling at Hildesheim in the eleventh and the great rose windows at Chartres, Paris and York in the thirteenth. In the next centuries the pagan element29 gained in significance to the point that Raphael in the Stanze was challenged with finding parallels between Christian and Antique themes such as the Church fathers versus the school of Athens. In these and other great cycles it was no longer a question of telling complete stories, but rather of choosing key episodes in stories which could be balanced by others.30
Hence all three basic functions (varying, emphasizing and relating), which made perspective so powerful, had the same effects. While focussing attention on key episodes in a narrative, they simultaneously undermined the continuity of the story telling process. Indeed as perspective provided more complex frameworks for the organization and comprehension of such scenes, their narrative order became less significant and sometimes disappeared. This helps to explain what would otherwise be two contradictory trends in the history of narrative cycles from the time of the Bayeux Tapestry (Bayeux, Town Hall, 1073-1083) and the mosaics at Monreale (1182), to the frescoes of Giotto in the Scrovegni Chapel (1304-1306) and Raphael in the Stanze (1507-1513). As the treatment of space improved, the number of scenes diminished. Monreale had 167, the Scrovegni had 53, the Sistine Chapel had 23 scenes. The sequential order of the story telling process also decreased in clarity. To attain a deeper understanding of these phenomena requires examination of contexts and frames.
Perspective brought with it a tendency to reduce a number of independent episodes and include these with a single spatial context, as is strikingly illustrated in Memling's Seven Joys of Mary (Munich, Alte Pinakothok, c. 1480), which integrated no less than seventeen episodes beginning with the Annunciation and ending with the Assumption of the Virgin. Similarly, in his Treatise of Painting, Leonardo recommended that one:
31must place the first plane at the eye level of the beholder of the scene and on that plane represent the first scene in large size and then, diminishing the figures and buildings on various planes, as you go on, make the setting for the whole story.
These tendencies towards a single spatial context, containing several temporal episodes are of particular interest because they call into question the oppositions between painting and poetry articulated by Lessing. In his Laokoon, he suggested that painting and poetry used completely different means and signs in achieving imitation; that painting used figures and colours in space, while poetry used tones in time.32 Lessing elaborated on these basic oppositions. Painting, he claimed, was concerned with bodies, poetry with actions;33 painting with a totality, poetry with parts;34 painting with space, poetry with time.35 It is instructive that his examples drew constantly on Greek art and poetry and indeed might hold if restricted to a comparison between Greek sculpture and poetry. But they do not hold for the whole of art. Indeed, the many scenes integrated into a single spatial context as practiced by Memling, and recommended by Leonardo, show that perspective removed these oppositions and introduced different actions of a body, different parts in a whole and different times in a single space as important new dimensions of representation, which take us directly to some of the richest aspects of art in and since the Renaissance.
For instance, the first of these, the ability to represent different actions of bodies prompted Leonardo to make a list of eighteen basic actions which could be painted36 and led him to explore kinematic sequences of men at work and play, a principle which has since inspired the development of motion pictures, television and video. The ability to represent different parts in a whole was equally fundamental in its consequences. For it explains why photographic details of Renaissance paintings can function as if they were photographs of complete paintings. This principle also makes possible the game of imposing imaginary frames in galleries and observing how each of these functions as independent pictures. Art dealers, who sawed off sections of old masters, and then sold these whole-sale were, of course, taking the game a bit too far.
The principle which makes these games possible is intimately connected with problems of particulars and universals. A perspectival painting, i.e., a painting which has a context, is based on particulars, is comprised of individual features and has the astounding feature that its "parts" also function as wholes. Nature has this same feature which is why we can take any scene, add different lenses to our concern and each time come up with an independent picture. Note the connection between particulars, individuals and independence. Note that these are also a key to creating new frames, focussing on details and changing scales which are three ways of describing this open process.
It is rather important to realize that none of this is possible as long as universals govern representation, as tended to be the case in Greece. Given universals, the goal of representation is perfection, literally putting an end to, a totality, a perfect totality. To remove any part of a totality is to destroy its perfection and to remove its aesthetic potential. (Or at least in theory, although some art historians will assure us that gods and goddesses are aesthetically the richer through amputation of arms, legs and other parts.) Hence a commitment to universals generates only parts dependent on a totality, which remain impersonal, static and without a temporal dimension. By contrast, a commitment to particulars leads to individuals independent of the whole, which can be personal, dynamic and with a temporal dimension. Because universals limit attention to the perfection of a totality, any change of frames would leave out some important part of that totality; any focus on a detail would leave the totality out of focus and any change of scale would make no difference: which is why photographs or slides of a three inch statuette or a six foot classical statue produce exactly the same effect. When the emphasis is on totality there is no context and no way of inferring scale. Indeed universals, with their commitment to perfection, produced an approach to representation which effectively denied the importance of size, scale, context, frames and time, i.e., precisely those features which perspective made the central concerns of Renaissance art and science.
Mention of time brings us to the third of Lessing's oppositions which we need to consider briefly before returning to the connections between perspective and frames. Lessing's claim that poetry dealt with time, while painting dealt with space overlooked the ways in which perspective introduced spatio-temporal dimensions into painting. The most obvious examples involved episodes in the lives of saints as in the Memling painting mentioned above. But there were also much more subtle examples as in Carpaccio's St. George and the Dragon in the cycle devoted to that saint (Venice, Scuola di San Giorgio).
In the foreground of the painting, just left of centre, we see a scene with a snake looking at a toad which looks in turn at a lizard. In a second scene, further back, we do not see the toad, while the lizard looks at the decomposing body of a woman obscuring most of the snake except for its tail, which forms an unexpected necklace for the corpse. In scene three, the toad reappears just right of the centre near the corpse of a man while the snake lurks beneath the corpse's left foot. In scene four, the snake devours the toad, while the lizard looks on. The subject may be unappetizing, but it suggests that Lessing's claims about time, parts and actions in painting were undigested.
There were other subtle ways in which spatio-temporal dimensions came into play. As we have shown the matching function led to a natural extension of the represented space of the painting into the physical space of the environment where it was painted. Hence, as was noted in an earlier chapter (p. 31, cf. pl. 13.1-4, 26.1), local townscapes inevitably entered as backgrounds in religious paintings. As these background townscapes became more pronounced they brought into focus unexpected anachronisms, for events in the life of Christ which had occurred fourteen or fifteen centuries earlier now stood in the foreground of a contemporary scene. By the late fifteenth century, when Ghirlandaio did his cycle on the life of Saint Francis (Florence, Sassetti Chapel, 1483-1486) ,he depicted the saint literally in the squares and streets of Florence. Here, of course, the anachronism involved, only a few centuries but even so Ghirlandaio did nothing to remove it.
Perhaps there were problems in learning to see spatio-temporal dimensions in paintings, just as it took a long time before painters became aware of problems of shadows caused by the sun at different times of day in their landscapes. We might have expected the writings of Machiavelli, Guicciardini and other historians to introduce a greater historical consciousness which would remove such anachronisms. Instead, the anachronisms persisted throughout the seventeenth, eighteenth and into the nineteenth centuries, until first the camera, and then the impressionists focussed attention on scenes limited to a specific place and time: Paris on a rainy afternoon, Arles on a sunny morning, etc. But long before this the anachronisms had taken on a subtler form. For as the contemporary background slowly moved forward to dominate even the foreground, the historical event retreated quietly into the background. By the mid seventeenth century with Claude we (almost) need to be told that the four figures standing in a landscape involve the story of Jacob and Laban (London, Dulwich Art Gallery); a principle that applies equally to mythological scenes such as his Coastal landscape with Apollo and the Cumean sibyl (Private collection, 1665).
These shifts introduced by perspective deserve much more attention. For it is usually assumed that the development of secular art was largely due to a rejection of the religious tradition. We are suggesting that the reverse was true: that it was paradoxically the Christian tradition of creatural realism that combined with perspective to create frameworks for matching which extended biblical narrative into the physical world and made nature first a background topic and gradually a dominant theme in the history of representation. We have shown how contradictions between and combinations of spatial and temrporal dimensions played a central role in these developments. And Lessing's desire to maintain the simple polarities of painting-space versus poetry-time led him to overlook this, and indeed other fundamental contributions of Renaissance art.
Frames
To understand this properly we need to return to the problem of frames. For the whole phenomenon we have been describing of townscapes slowly coming into the foregrounds of religious paintings is very much a question of frames and fully analogous to a zoom lens which focusses on what had been a background detail, frames it and then increases its scale until it dominates the entire scene. Which is also why perspectival representation leads ineluctably towards a photographic image, where framing is almost the name of the game. We shall show that these connections between a play of perspective and frames go back at least to the time of Giotto in the the Scrovegni Chapel at Padua (1304-106), but before doing so we need to refine an earlier claim.
We have stressed that perspective applies not only to painting but to other arts such as architecture and sculpture involving various media including bronze, marble and wood. However, in another sense, perspective has special applications to painting because in this medium it imitates effects also produced by sculpture and architecture, and for this reason, Chastel,37 has justifiably opposed (painted) perspective to sculpture and architecture, or perhaps more accurately, painted perspective vies and equivocates with effects which sculpture and architecture create, thus encroaching upon and/or playing with the frames they impose or suggest.
This did not always happen. In the case of altars, for instance, it played only a small role. As Heydenryk, in his history of frames has noted, Italian afterpieces imitated architectural features and effectively became cross sections of Gothic churches,38 while in the North "the elements of a frame were invariably emulations of architectural elements but no effort was made to create a logical architectural structure as had been done in Italy."39 The advent of perspective affected the contents of altars (cf. above pp. ) and meant that various ornamental patterns on their frames which had previously been sculpted, were now painted. But it had little substantive impact on the function of altar frames. By contrast, in the case of frames in the fresco cycles,40 perspective had an enormous impact.
In the Scrovegni Chapel at Padua, Giotto explored the potentials of using proto perspectival effects to replace, or rather match (see below pp. ) architectural structures in his concealed chapels or coretti on the east wall. But while there was play of boundaries between architecture and painted architecture, there was effectively none between architecture and painted narrative, where each scene was neatly separated from the next by clear cut frames. Giotto experimented with both problems separately in the same building. The early renaissance pursueed both experiments, discovered and formalized the perspectival principles underlying them. The high renaissance integrated the two experiments into a new synthesis as becomes clear if we turn to Michelangelo's Sistine Ceiling (1505-1508).
As in the Scrovegni Chapel, there is a narrative cycle. But whereas Giotto's scenes maintained a certain uniformity in size, Michelangelo plays with their scale. In the central portion four large scenes alternate with five smaller ones which are flanked, in turn, by ten medallion-like scenes. In the corners there are four further scenes making a total of twenty-three episodes from the Old Testament. Then there are the forebears of Christ in the triangular niches and in the semi-circular niches below these. But the complexity of the Sistine Ceiling begins, in a sense, with the six sibyls alternating with six prophets in painted architectural cubicles enclosed by column-like painted sculptures and topped by painted nudes.
The cubicles function as if they were part of a wall, the orientation of which keeps changing as we move through the chapel and constantly contradicts, or rather plays with the curvature of the actual ceiling. The nude figures seated on top of the columns require us to read the surface three dimensionally, while we remain aware of the ceiling as a flat surface. If we look higher than the nudes our eyes are drawn into an orientation 90 degrees to the side, and if we look higher still, we need to shift our orientation a total of 180 degrees if we are not to read the second set of nudes as falling down.
Perspective continues to play a role in the actual scenes, as in the dramatic positioning of Haman on the cross. But its main function is now in the spaces between scenes, in, with, amongst the frames, provoking a complex interplay between painted, painted sculptural and painted architectural elements which, while continuing to separate the scenes, also integrate them into a new kind of systematic whole. Perspective now creates spatial illusions only seemingly to subvert them, playing with and on them to increase the potential for polyvalent readings of different scenes theoretically separated yet, systematically related. These polyvalent readings are encouraged by the nudes and other figures whose arms and feet continually reach and step into the neighbouring spaces. At the same time they are held in check by the painted architectural features which maintain some clear linear boundaries between the scenes.
The mannerist period worried less about keeping these boundaries fixed. Already in the Gallery of Francis I at Fontainbleau, boundaries between painted, painted sculptural and painted architectural spaces were rendered more ambiguous a) by increasing the extent to which figures reach out beyond their given frame into adjacent spaces and b) by deliberate introduction of actual sculptural and architectural elements which overlap with their painted equivalents. This encroachment of figures and overlapping of art forms and media went ever further until it became difficult (e.g. pl. 77.1), and ultimately impossible (e.g. pl. 77.2), to know where one stops and the other begins. Hence if the high renaissance discovered frames and their media as realms of perspective, mannerism used perspective to play with frameworks, anamorphically distorting them in the process. Baroque art went further, playing with the whole distinction between the forms of frames and their contents. By the late seventeenth century, as baroque art moved towards rococo, the actual architectural spaces were manipulated and integrated in order to intensify this playful destruction of distinctions between frames and paintings, between form and content.
The experiments had begun with Giotto's concealed chapels which played with distinctions between painted and real architecture. By the 1580's this play between painted and architectural reality had become a major challenge for masters of perspective, particularly in terms of di sotto in su paintings, which involved illusionistic ceilings such as that of Scamozzi and Sansovino in the Marciana library in Venice (pl. 72.1). This soon found published equivalents in treatises ranging from a single illustration in Danti's commentary (1583, pl. 72.2), to the 75 examples in a now almost forgotten collection by Has, which appeared in the same year (1583, pl. 73.3), which provided a context for Pozzo's illusionism in Il Gesù (1691-1694, pl. 73.1) and also prefigured uncannily the twentieth century work of Escher (pl. 73.4, cf. 73.3).
Such examples bear witness to an extraordinary shift in the applications of perspective, from the content of paintings to their form. For we have shown that perspective had basic consequences for both the spaces in paintings and the spaces of the frames between paintings. The first of these concerns led to shifts from isolated objects to scenes in context, to scenes in the context of a secular background and to secular scenes in a specific place and time and ultimately to new distinctions between these realistic spatio-temporal scenes and other more symbolic ones, where time and sometimes even space were not factors. Meanwhile, the second of these concerns began with the varying, emphasizing and relating functions of perspective, led renaissance artists to discover connections between perspective and the framing process, and to focus attention on the spatial forms of containers of paintings. As they did so, attention to the contents within the containers dwindled, or rather it shifted from narrative episodes to symbolic moments in the narrative which, in the next generations were reduced to symbols almost without moment until, finally, the capacities of perspective were focussed on form without content (e.g., pl. 73.1).
Together these two developments transformed the whole of representation. They removed the oppositions between bodies and actions, totality and parts, space and time which Lessing had seen as separating painting and poetry. Indeed they revealed the limitations of Greek concepts of perfection and introduced new horizons of aesthetics in terms of size, scale, focus and frameworks.
All this was only the beginning. By the mid-sixteenth century perspective had spread beyond the spaces of paintings and their frames, outside the buildings that contained them into the streets and gardens. In the seventeenth century perspective gradually transformed the use of physical space partly so that this too could be rendered more perspectivally, a process of transformation which gradually spread to the whole environment. This will concern as in the next chapter before we turn in the final section to show how perspective also transformed the landscapes of the mind cultivating new concepts of freedom and imagination, inspiring new developments in art which still continue.
Last Update: August 4, 1998