Otto Demus Byzantine Art and The Spatial PDF
Otto Demus Byzantine Art and The Spatial PDF
Otto Demus Byzantine Art and The Spatial PDF
Ian Verstegen
When attention began to turn toward the new or second Vienna school, it is
understandable that there was a more inclusive approach to which theorists
belonged.1 Thus, Fritz Novotny’s work on Cézanne was included in Christopher
Wood’s Vienna School Reader, even though he had been a product of Stzrygowski’s
Institute and not (the descendant of) Alois Riegl’s.2 Later, it was questioned to what
degree he ought to be counted in this group.3 The same issue faces another student
of Stzrygowsky: Otto Demus.
In the following, I will follow Wood in a qualified way to show how Demus’
characteristic approach to Byzantine art emerged indeed from typical issues
descending from the work of Riegl. Due to thaws between the two institutes,
Demus’ teaching with Hans Sedlmayr, and his socialisation with Viennese-trained
art historians in London during the Second World War, one can see that Demus had
become thoroughly ‘Viennese’ in his outlook well before he returned to Vienna and
in 1963 to occupy a chair.
As I will argue, this makes Demus’ work fit quite well with the second
Viennese school of Hans Sedlmayr and Otto Pächt and to an extant Johannes Wilde,
to the degree that his work is focused on an understanding of the work of art or
monument as a functional whole. Parts and their relationship are understood, or in
the case of a lost work, intuited by way of reconstruction. Also, technical knowledge
is used to judge initial states of objects so that the universal working of perception
will be accurate based on these same givens.
I propose to explain Demus’ methodology through his most famous
contribution: an exegesis of the structure and function of Byzantine mosaics. I will
1 I am grateful to Thomas Dale for serving as peer reviewer of this article, which resulted in
many helpful suggestions and corrections, as well as Ivan Drpic for having offered helpful
comments on an earlier draft.
2 Fritz Novotny, ‘Passages from Cézanne and the End of Scientific Perspective’, in Christopher
Wood (ed.), The Vienna School Reader: Politics and Art Historical Method in the 1930s, New York,
2000, 379-433; excerpted from Fritz Novotny, Cézanne und das Ende der wissenschaftlichen
Perspektive, Vienna, 1938.
3 Agnes Blaha, ‘Fritz Novotny and the New Vienna School of Art History – An Ambiguous
show that his approach, which has been linked to Riegl in a casual way, can be
aligned much more vigorously. After reviewing Demus’ career, and his
interpretation of Byzantine mosaic decoration, I will pass on to an extension of his
ideas in the Latin West, where I will demonstrate unexpected site-specific elements
that have not been noticed in the literature. Not only does this strengthen Demus’
place within a ‘Viennese’ genealogy, there is much latent material in Demus, Wilde
and their students that allow one to chart the movement from ‘participant’ to
‘beholder,’ giving new insights into the rise of western illusionism. Joining Demus
to the Vienna School also aligns two powerful explanatory systems.
Demus, like Wilde, wrote methodologically sophisticated history but was not
himself like Pächt or Sedlmayr a theorist. He does not even seem to have been as
reflective in his reading as was Wilde. His primary ‘theoretical’ reflection on
Byzantine mosaics is contained in Byzantine Mosaic Decoration.4 The work is unique
in Demus’ oeuvre because if one turns to the contemporary book, The Mosaics of
Norman Sicily, and later works, one finds a largely traditional monograph focused
on physical condition, attribution and iconography.5
Part of the difficulty of tracing Demus’ theoretical interests is compounded
by the fact that while Demus trained in Vienna, he was not in Schlosser’s Institute,
the home of Dvořák and Riegl. Demus indicates in his curriculum vitae of 1927 that
he had heard Schlosser’s lectures.6 Of course, one can argue that he would have
been quite aware of what was going on there, but this cannot explain his eventual
theoretical position, especially since Schlosser was the weakest follower of Riegl
compared to Dvořák and Sedlmayr. Indeed, one can note that during the
preparation of Kunstwissenschaftliche Forschungen, neither Novotny nor Demus was
asked to contribute to the journal.7 Nevertheless, although he was Strzygowski’s
student and the two institutes were relatively discrete, it is quite easy to link many
4 Otto Demus, Byzantine Mosaic Decoration: Aspects of Monumental Art in Byzantium, London,
1948; ‘The Methods of the Byzantine Artist’, The Mint, 64–77.
5 Otto Demus, The Mosaics of Norman Sicily, London, 1949.
discussion of Otto Pächt’s apparent influence on Fritz Novotny, ‘Fritz Novotny and the New
Vienna School of Art History’.
2
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
8 This connection is tentatively made by Robert Nelson, ‘To Say and to See: Ekphrasis and
Vision in Byzantium’, in R. Nelson, ed., Visuality Before and Beyond the Renaissance: Seeing as
Others Saw, New York, 2000, 157-58.
9 Demus, Byzantine Mosaic Decoration, xiii.
https://www.univie.ac.at/geschichtegesichtet/o_demus.html
11 Ernst Diez and Otto Demus, Byzantine Mosaics in Greece, Hosios Lucas and Daphni,
Cambridge, 1931.
12 Otto Demus, Die Mosaiken von San Marko in Venedig, 1100-1300, Vienna, 1935, 13, citing
3
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
Fritz Novotny, review of Gerstle Mack, Paul Cézanne, Kritische Berichte, 1937.
16 Otto Pächt, ‘Die historische Aufgabe Michael Pachers’, Kunstwissenschaftliche Forschungen,
1, 1931, 95-132.
17 See Jonathan Petropoulos, ‘Declaration and Expert Report’,
https://artstolenfromfritzgrunbaum.files.wordpress.com/2008/03/dbm-005862-5966-
petropoulos-expert-report-03-25-20082.pdf; consulted February 5, 2015.
18 Dorothea McEwan, Fritz Saxl: Eine Biographie. Aby Warburgs Bibliothekar und erster Direktor
Perspectives on the Beginnings of Comparative Literature in the United States, eds. Lionel
Gossman and Mihai Spariosu, Albany, 1994, p. 52; Helmust Koenigsberger, ‘Fragments of an
Unwritten Biography’, in Peter Alter, eds., Out of the Third Reich: Refugee Historians in Postwar
Britain, London, 1998, 106.
4
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
synoptic decoration with Demus. In particular, Demus shared his ideas on optical
correction in Byzantine mosaic decoration with Wilde, who exchanged his
knowledge of the evidence of the survival of such tactics in Venetian Renaissance
painting, and indeed Demus notes drawings by the Venetian painter, Sebastiano del
Piombo, that show corrections for proper viewing in a cupola, brought to his
attention by Wilde.21
What is obvious is that by 1945, when Demus sat down in London to finalise
Byzantine Mosaic Decoration shortly before his return to Vienna, he hardly obtained
some new – perhaps Anglo-Saxon – perspective connected to his emigration. In
Vienna during his student days, then as an instructor and then with the Viennese
exiles in London and internment, he was absorbing a structuralist point of view. To
someone aware of Wilde’s emerging work on Michelangelo (and Sedlmayr’s
discussions of Borromini and Pächt’s of Pacher, or the Schools of late Gothic
painting), it is clear that Demus is responding to the same idea of context-
dependence, a sense of the whole in which he sees the artist responding.
However, there is a historical difference that can be discerned by studying
the platform that Demus develops in Byzantine Mosaic Decoration. First, it is worth
reflecting on what especially Otto Pächt (and John White) brought to the study of
late-Gothic painting. They were uniquely able, on Riegl’s precedent, to understand
the art in a non-anachronistic way, avoiding any retrospective fallacy. Riegl had
argued that Byzantine art was a necessary weigh-station (Durchgangsphase) on the
way to modernity. What would it mean to take a late medieval point of view, with
its emphasis on pictorial surface pattern (rather than the so-called Albertian
window), and move it even earlier? This would require moving back from linear to
proto-perspective to a kind of what Demus would call negative or preventive
perspective, using optical corrections. This is Demus’ answer to the long-standing
question over ‘inverted’ perspective.22 What all of these approaches have in common
is the belief that these early forms of art can induce depth without consistent
geometry.23
Demus took for granted the immanence theory, according to which divinity
resides in the Byzantine image.24 Even if many scholars have questioned whether
Pictures’, Perception, 10, 1981, 305-12; Rudolf Arnheim, ‘Inverted Perspective and the Axiom
of Realism’; J. B. Deregowski, D. M. Parker, M. Massironi, ‘The perception of spatial
structure with oblique viewing: an explanation for Byzantine perspective’? Perception, 23,
1994, 5-13.
24 This seems to be a lingering heritage of Stzrygowski, who associated the east with magic.
More recent work generally argues that the icon was a kind of sign, which signified divinity
5
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
there is magical emanation from the prototype to the image, the crucial element for
Demus was that these elements bear similarity. The situation that the Byzantine
viewer is faced with, is maintaining an unbroken relationship with the cult image.
As Demus showed, the artists were quite sophisticated in responding to this feature
– another element of Viennese theorising. They used optical or empirical means to
‘correct’ images. In the words of Alberto Perez-Gomez and Louise Pelletier,
‘undistorted presence had ontological priority over any sort of distorted
appearance, regardless of medium or diverse artistic objectives.’25
The type of optical correction investigated by Demus is a form of empirical
perspective no different from the skenographia or scaenographia used in antiquity
from Philon (3rd c. BCE), to Vitruvius (1st c. BCE) and Proclus (5th c. CE) to make
high-up sculptures, stage scenery and architectural elements appear correct from a
distance.26 For Proclus, scaenografia is ‘showing how objects can be represented by
images that will not seem disproportionate or shapeless when seen at a distance or
on an elevation.’27 In a Byzantine context, after the iconoclastic controversy, a
predominantly sculptural discourse was shifted to painting and mosaic. There is no
record of this practice; however, an echo of this discourse is given in John Tzetzes,
12th century text recounting the sculptural competition of Pheidias and Alkamenes
(well over a millennium after the fact!), where he notes Pheidias’ successful optical
corrections, which on the ground looked grotesque but when seen upon an elevated
column were perfect.28 One of Demus’ examples is the twelfth-century apse mosaic
in Torcello Cathedral (Fig. 1).29 All the figures appear the same from the nave of the
rather than inhering within it; Charles Barber, ‘From Transformation to Desire: Art and
Worship after Byzantine Iconoclasm’, Art Bulletin, 75, 1993, 7-16.
25 Alberto Perez-Gomez and Louise Pelletier, Architectural Representation and the Perspective
On Order in Architecture, Cambridge, 1996, 222-224. Robert Ousterhout has kindly confirmed
for me that the technical term from Greek antiquity for optical correction, alexemata, is not
found in Byzantine criticism.
29 For other examples, see Washing of Feet, Hosios Loukas, Katholikon, 11th century (Demus,
Byzantine Mosaic Decoration, 32) and Pantokrator, Monreale, after 1180, (Demus, Byzantine
Mosaic Decoration, 65).
6
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
church, but when examined from within the apse itself it can be seen that the lateral
figures are extremely wide to subtend the same visual angle from the nave.
1 2
Figure 1 Apostles, 12th century, Cathedral, Torcello (from dust cover of Otto Demus, 1948)
Fig. 2 Annunciation, c. 1000, Church of Dormition, Daphni (photo: Ktiv, from Wikimedia Commons)
30 To my knowledge, Demus does not use the German terms Raum-Ikonen (or Bildraum,
below). These are Hans Belting’s renderings probably confirmed in discussion with Demus
himself. However, I cite them to return them to their German language context.
31 Demus, ‘Methods of the Byzantine Artist’.
7
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
But it must be recognised that a particular task creates this kind of special space.
One first must desire to keep the depicted objects and personages life-sized, with
human presence, for them to remain ecologically salient. Then, it is relatively easy
for the artists to imagine how to preventatively alter the figures to preserve their
presence. The figures, we might say, are ‘unbounded,’ and this is the primary
difference between these mural figures and those panel paintings and miniatures
that Pächt (and White) studied.32
How has Demus’ thesis fared? Demus peppers his text with numerous
illustrations, but other examples have been observed and noted by other scholars.
For example, of spatial icons, one can note the Life of Peter (mid 12th century) in the
Cappella Palatina, Palermo, or later the Last Judgment in the Chora Paraklession
(1321 CE).33 In Palermo, saints occupy space across architectural units, and in the
Chora, the whole interior becomes a grand symphony of activity. Of optical
correction, one can note the Crucifixion in the Tokali kilise (10th century), where the
arms of the crucifixion are manipulated to appear correct against the curvature of
the vault.34
This system, to repeat, operates on direct lines of sight that are not
pictorialized. Figures speak to figures. Figures are stretched or curved to appear
straight. This is all possible with the empirical observer. It is not geometry of the
drafting table but ad hoc, of sight lines in situ.35 Many subsequent authors have
further stressed the transitive relationship between mosaics and viewers, light,
colour, sound, and the bodily communication with the image. For example, Liz
James drew attention to light and colour, saturation and the way in which interior
ambience could overwhelm the senses.36 Bissera Pentcheva accepted Demus’ ideas
but wished to extend them into touch and kissing, stressing the performative notion
of icon and church-ensemble activation.37 Finally, Alexei Lidov specifically describes
the ‘spatial icon’ – outlined with no reference to Demus - with his theory of
heirotopy, the baptism of holy space with the combination of image, liturgical
accoutrement, and the liturgy itself.38
‘Working Practices’, Mosaics in the Medieval World: From Late Antiquity to the Fifteenth Century,
Cambridge, 2017.
36 Liz James, Light and Colour in Byzantine Art, Oxford, 1996.
38 Alexei Lidov, Hierotopy: Spatial Icons and Image-Paradigms in Byzantine Culture, Moscow,
2009.
8
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
39For similar view expressed today, see Charles Barber, Figure and Likeness: On the Limits of
Representation in Byzantine Iconoclasm, Princeton, 2002; Glenn Peers, Sacred Shock: Framing
Visual Experience in Byzantium, University Park, 2004; Rico Franses, ‘When all that is Gold
Does Not Glitter: on the Strange History of Looking at Byzantine Art’, in Antony Eastmond
and Liz James, eds, Icon and Word. The Power of Images in Byzantium, Farnham, 2003, 13–24.
40 Antony Eastmond, ‘The Limits of Byzantine Art’, in Liz James (ed.), A Companion to
42 Hans Bloemsma, ‘Venetian Crossroads. East and West and the Origins of Modernity in
Twelfth-Century Mosaics in San Marco’, Journal of Intercultural Studies, 31, 2010, 299-312;
‘Byzantine Art and Early Italian Painting’, in Byzantine Art and Renaissance Europe, eds. A.
Lymberopoulou and R. Duits, London, 2013, 37-60.
43 Thomas Dale, ‘From ‘Icons in Space’ to Space in Icons: Pictorial Models for Public and
Private Ritual in the Thirteenth-century mosaics of San Marco in Venice’, in Hierotopy. The
Creation of Sacred Space in Byzantium and Medieval Russia, ed. Alexei Lidov, Moscow, 2006, 1-
18.
9
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
(for instance in the Karii Djamii) in the Paleologan era.44 If Byzantine corrections had
arisen due to curved surfaces, Romanesque painting, instead, follows the axial
nature of the cruciform church to ‘suit the rhythm of walking.’45
This leads us to wonder about Demus and Pächt working together at the
University of Vienna after 1963. Did they informally theorise the transition from
Byzantine practice to western medieval practice? Recall that Pächt had emphasised
the pattern requirement of late Medieval painting in spite of its putative new
illusionistic function. The works functioned as flat surface patterns before they
became records of images in depth. It is easy to hypothesise what one Otto might
have said to the other Otto: empirical adjustments to figures in spaces never really
disappeared; in smaller works, this empirical perspective gave way to more rule-
based geometrical procedures, but it is in murals that we should expect to see strict
continuity.
Indeed, usually, when Demus cited an Italian example of Byzantine mosaic
decoration, he cited it for its demonstration of either the spatial icon or negative
perspective. 46 Recall that his discussions with Wilde suggested that optically
corrective workshop techniques survived into the Renaissance in Italy. Scaenagrafia
had existed since antiquity and continues today with any form of building where a
viewer cannot move to obtain a clearer view of an artistic object. Sight lines are
extremely simple features found in urban design when one seeks to make an
element of a city visible in another space. Recalling the citation that Demus makes of
Wilde’s observations on Venetian dome painting, we can understand quite easily
that the problem of the negative perspective never really goes away.
To follow Otto Pächt’s example, if the surface-pattern aspect of a painting
never loses its relevance, we might follow Demus and also say that in murals, the real
presence of the depicted forms never loses its relevance. This means that instead of
looking at the morphing of Byzantinizing practice into a Giottesque practice, we
should seek out the older formal operations in the newer art, alongside those that
are newer. However, although Demus’ overall approach of a transitive image has
been followed by many people, his spatial mandate has not. In the remaining part of
the paper, I propose to do this with a perhaps unexpected source, but one that
shows how exactly corrective practices persisted in Italy in the thirteenth and
fourteenth centuries. This involves moving from the ‘spatial icon’ of centralised
Byzantine churches to similar effects in the predominantly basilical churches of the
Latin west.
44 Otto Demus, ‘The Imprint of the System on Later Art’, Byzantine Mosaic Decoration;
Romanesque Mural Painting, New York, 1970, 132: ‘The slight asymmetry and the axial
orientation of Greek domical compositions are replaced by a strict geometrical scheme,
resulting, not in a 'spatial icon', but in an architectural structure built of figures’.
45 Demus, Romanesque Mural Painting, 13.
46 Thomas Dale, Relics, Prayer, and Politics in Medieval Venetia, Princeton, 1997, 65.
10
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
Fortunately, Johannes Wilde, and his student John White, made some
fundamental observations on the Upper Church of Assisi and dugento and trecento
painting, in general, that might reflect some of Demus’ ideas (or, better, a shared
approach). Demus’ system of mosaic decoration is predicated on the assumption of
the church as an ‘all-embracing optical unity,’ and improved visibility of figures
from privileged points of view. Wilde’s insights were gained by thinking of the
Great Council Hall of Florence or the Sistine Chapel in the Vatican as perceptual
wholes.47 We know from John White’s work that Wilde lectured on San Francesco in
Assisi, lectures that were never published. In those lectures, the basic point of which
White adopted for his account of the church’s layout, is that the program in Assisi
be seen from a single viewing position for each of the entire bays and individual
scenes.48 The two are both sensitive to the observer.
To bring the boundary of eastern and western practice under closer scrutiny,
and also suggest hidden continuities between them, I want to examine more closely
the Upper Church of San Francesco in Assisi.49 For the topic we are investigating,
the famous question of the paternity of the frescoes – Giotto or not – is irrelevant.50
What is essential is the aesthetic principles at work. Dominique Raynaud has
recently written, ‘these frescoes…await an interpretation able to reconcile empirical
evidence with conceptual minimalism.’51 In fact, since White we have not answered
the fundamental question: what was the purpose of the Master of the Legend of St.
Francis’ space? In the same way that his buildings do not explain Brunelleschi’s
perspective – the ‘self-isolating solidity’ and ‘portrait quality of the buildings’52 are
not fundamentally different from the supposed panels of the Baptistery – so too we
47 Johannes Wilde, ‘The Hall of the Great Council of Florence’, Journal of the Warburg and
Courtauld Institutes, 7, 1944, 65–81; ‘The Decoration of the Sistine Chapel’, Proceedings of the
British Academy, 1958, 61-81; c.f. Ian Verstegen, ‘John White’s and John Shearman’s Viennese
Art Historical Method’, Journal of Art Historiography, 1, 2009, 1-IV/1, 15 pages.
[https://arthistoriography.files.wordpress.com/2011/02/media_139132_en.pdf]
48 John White, Art and Architecture in Italy, 58-59. On Assisi, see John White, The Birth and
in Assisi, New York, 1962; Hans Belting, Die Oberkirche von San Francesco in Assist, Berlin,
1977; Bruno Zanardi, Federico Zeri and Chiara Frugoni, Il cantiere di Giotto: Le Storie di San
Francesco ad Assisi, Skira, 1996.
50 For discussions of Giotto’s authorship of the Assisi frescoes, see Alastair Smart, The Assisi
problem and the art of Giotto, a study of the Legend of St. Francis in the Upper Church of San
Francesco, Assisi, Oxford, 1971.
51 Dominique Raynaud, Studies on Binocular Vision: Optics, Vision, and Perspective from the
Thirteenth to the Seventeenth Centuries (New Studies in the History and. Philosophy of Science and
Technology), Cham, Switzerland, 2016, 194.
52 White, Art and Architecture in Italy, 144.
11
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
must figure out what is not different in these forms.53 Brunelleschi’s interests led to
perspective, just as the Master of the Legend of St. Francis led to Giotto’s proto-
perspective, but what was latent that only appears to be progressive?
As I will argue, it is the empirical perspectival desire to modify forms for
clarity. This is ultimately Byzantine and survived in various schools – Roman, Pisan,
Venetian – that fed the visual forms of the Master of St. Francis. By the late
thirteenth century, we have to take account of a cultural divergence between Latin
West and Byzantine east that was occurring throughout the Dugento. This
divergence concerns the diffusion of Arab optical texts,54 especially at the papal
court and Franciscan schools, and the rise of methods of practical geometry.55
Christopher Lakey has joined optical science to in-situ optical adjustment to
dugento and trecento monument.56 However, where Lakey sees a calculation in a
modern sense, I would urge an empirical adjustment, further stressing also
similarities to current Byzantine practice.
The most compelling account of early perspective knowledge is provided by
Reynaud, who has convincingly argued that the first systems of proto-perspective
introduced in Italy are binocular, reflecting Arab optical science. Rather than a
steady progress, he finds a bricolage of methods introduced, perhaps the most
startling being the accurate diminution in the coffering of the scene of the Recovery
of the Wounded Man of Lerida in Assisi.57 This correct geometric form was grafted
on the ceiling, with no sense of a complete perspective scene. Thus, in the late
thirteenth century, there was indeed a new comfort with geometry and a renewed
observation of the world, but no consistent sense as to how this would be applied in
painting.
volgarizzamento del ms. vat. 4595 e il ‘Commentario Terzo’ del Ghiberti’, Rinascimento, 5,
1965, 17-49; Dominique Raynaud, L'hypothese D'Oxford: Essai Sur Les Origines De La
Perspective, Paris, 1998; Frank Büttner, Giotto und die Ursprünge der neuzeitlichen
Bildauffassung, Darmstadt, 2013.
55 For practical geometry in Italy, see Annalisa Simi, ‘Celerimensura e strumenti in
manoscritti dei secoli XIII–XV’, in Itinera mathematica. Studi in onore di Gino Arrighi per il suo
90º compleanno, edited by R. Franci, P. Pagli and L. Toti Rigatelli, Siena, 1996, 71-122;
Francesca Cecchini, ‘Artisti, committenti e perspectiva in Italia alla fine del duecento’, in La
Prospettiva. Fondamenti teorici ed esperienze figurative dall’antichità al mondo moderno, edited by
R. Sinisgalli, Fiesole, 1998, 56-74; ‘‘Le misure secondo l’apparenza’. Ottica e illusionismo
nella cultura del duecento: tracce figurative e testimonanzie letterarie’, Micrologus, 4, 1998,
167-185.
56 Christopher Lakey, ‘From Place to Space: Raumkästen and the Moving Spectator in
Medieval Italian Art’, in The Public in the Picture / Das Publikum im Bild, eds. Beate Fricke and
Urte Krass, Zürich, 2015, 113-136.
57 Raynaud, Studies on Binocular Vision.
12
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
58 Hans Jantzen, Über den gotischen Kirchenraum und andere Aufsätze, Berlin, 1951; Robert
Oertel, ‘Wandmalerei und Zeichnung in Italien. Die Anfange der Entwurfszeichnung und
ihre monumentalen Vorstufen’, Mitteilungen des Kunsthistorischen Institutes in Florenz, 5, 1940,
217-314; R. Oertel, ‘Wende der Giotto-Forschung’, Zeitschrift furKunstgeschichte, 11, 1943/44;
W. Schöne, ‘Studien zur Oberkirche von Assisi’, Festschrift Kurt Bauch, Munich, 1957; c.f.
Über das Licht in der Malerei, Berlin, 1954, pp. 32-36, 256-65. Sedlmayr, now no longer in his
structuralist phase, highly praised Schöne: H. Sedlmayr, ‘Über Farbe, Licht und Dunkel.
Bemerkungen zu: Wolfgang Schöne, ‘Über das Licht in der Malerei’, Hefte des
Kunsthistorischen Seminars der Universität Munchen, 5, 1959, 29-51; Martin Gosebruch, Giotto
und die Entwicklung des neuzeitlichen Kunstbewusstweins, Cologne, 1962; c.f., Christian-Adolf
Isermeyer, Rahmengliederung und Bildfolge in der Wandmalerei bei Giotto und den Florentiner
Malern des 14 Jahrhunderts, Ph.D. diss., Göttingen, 1937.
59Jantzen, Das niederländische Architekturbild; Wolfgang Schöne, ‘Zur Bedeutung der
Schrägsicht für die Deckenmaleri des Barock’, in Festschrift für Kurt Badt, ed. Martin
Gosebruch, Berlin, 1961, 144-172; Puttfarken, Masstabsfragen: Über die Unterschiede zwischen
grossen und kleinen Bildern, Ph.D. diss., Hamburg, 1971; The Discovery of Pictorial Composition:
Theories of Visual Order in Painting, 1400-1800, New Haven, 2000. Importantly, the gestalt
psychologist Johannes von Allesch, who had been so important to Sedlmayr and Pächt,
positively reviewed Schöne’s book: Göttingische Gelehrte Anzeigen, 210, 1956, 180-206.
60 On Padua, see Oertel, ‘Wende der Giotto-Forschung’, 16ff; Isermeyer, Rahmengliederung
13
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
many spaces were discrete and isolated choir screens into gender and lay
divisions.61
Similarly, Puttfarken has shown the bias toward treating pre-modern
frescoes as easel pictures. Thus, Giotto became in some of the works I have already
cited the ‘founding father of the European picture and its inherent order.’62
Puttfarken corrects that the figures therein were related to individually (not too
distantly from Demus’ spatial icon), with no sense of modern pictorial composition.
But surely to understand a bias toward the tableau is not to deny the perception of
pattern altogether. Indeed, the Arena Chapel is not a perfect case because judging
the whole and individual pictures is hard to separate, and Puttfarken accepts any
approach that sees the unity of the Arena Chapel. Also, there is something
uncannily tableau-like in the individual scenes of the Arena chapel, which White
recognised. Therefore, in spite of a warning toward modernist bias, it is permissible
to proceed, not to deny perception, but as a welcome opportunity to historicize it.
Subsequent technical and liturgical studies have only confirmed the
fruitfulness of conceiving the Upper Church of San Francesco as unified. Because
the church functioned as a papal basilica, there was to be no visual obstruction of
the apse with its papal throne. Even if there was some separation in other
Franciscan churches – as is clear from the Upper Church’s frescoes themselves (the
Crib at Greccio prominently shows a rood screen, well over head height and
obstructive of frescoes) the unsuitability of a large altarpiece on the high altar where
the pope could officiate even removed this visual obstacle.63 It is safe to say, then,
that both from the point of view of archaeological and liturgical evidence the
intuition given by the paintings is correct: the designers intended to give a unified
prospect to the frescoes.64
61 Jacqueline Jung, ‘Beyond the Barrier: The Unifying Role of the Choir Screen in Gothic
Churches’, Art Bulletin, 82, 2000, 622-57, and ‘Seeing through Screens’.
62Puttfarken, Discovery of Pictorial Composition. For an allied indictment of such
in Pre-Tridentine Umbria’, Journal of the Warburg and Courtauld Institutes, 64, 2001, 1-54; c.f.
Janetta Benton, ‘Perspective and the Spectator's Pattern of Circulation in Assisi and Padua’,
Artibus et Historiae, 19, 1989, 37-52, esp. 37-38.
64 The most sustained analysis by Ursula Arendt (Der Franziskuszyklus in der Oberkirche von
San Francesco in Assisi. Überlegungen zur Abfolge der Szenen, PhD dissertation, University of
Vienna, 2008) posits a rood beam, to support a crucifix, but not a screen. She uses the traces
of the beam in the first and last frescoes to aid in the dating of the series.
14
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
Part, and perhaps a crucial part, of White’s conclusions were derived from the
perspective of the corbel frieze around actual half columns in two scenes. What I
want to suggest is that successive interpretations, which have paid such close
attention to the corbelling above the scenes, has in one way advanced the discussion
but in another held it back (Fig. 3). A slavish geometric (and not merely space-
producing) point of view is the culprit. The work of White consistently separates the
question of geometry from space. Indeed, it is only by forgetting about the Alberti-
like elements of the frescoes that one has any hope of discovering deeper space-
producing tactics.
Figure 3 Master of the Legend of St. Francis, Bay 2, scenes 4-6, San Francesco, Assisi (image source: Web Gallery of
Art; image in the public domain)
65 Benton, ‘Perspective and the Spectator's Pattern of Circulation in Assisi and Padua’, 41.
66 Benton, ‘Perspective and the Spectator's Pattern of Circulation in Assisi and Padua’, 42.
15
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
70 As Judith Field remarks of the slightly later, apparently ‘correct’ pavements in the
Lorenzetti’s works, ‘in the case of pre-Brunelleschian pavimenti we are, moreover, dealing
not with heavyweight mathematicians, but with craftsmen happy to follow established rules
of thumb’; The Invention of Infinity: Mathematics and Art in the Renaissance, 1997, 40.
71 White, Art and Architecture in Italy, 140.
16
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
spaces, both adjacently and cross-aisle, White insists that, ‘the details of the painted
architecture in each bay recede in parallel toward its centre, which becomes a
central focal point.’72 As an example, we may look at the second bay, and scenes six
through eight, showing the St. Francis before the Crucifix, St. Francis Repudiating his
Father and the Dream of Innocent III (Fig. 3). The central scenes feature architectural
and figural symmetry. The outside scenes have a ‘visual equipoise.’ The ruined
church of San Damiano on the left and the Lateran Basilica on the right visually
rhyme, and each have a matching red band which runs ‘to the centre at identical
angles to the lower border.’73 In short, ‘these three very different narratives were
intended to be seen as part of one united, balanced pattern, emphasized by the
perspective unity of the architectural framework.’
Given Edgerton’s choice between converging and divergent perspective,
there is little opportunity to appreciate these effects. But they certainly relate to a
centralised viewer. Although Edgerton shows how the lower modillion borders are
‘corrected’ even in the first bay where they are most off centre above, and this
therefore proscribed where the viewer ought to stand, he does not accept the
corollary that other cues ought to contribute to a unified spatial percept. I shall
return to these scenes below and sketch out an expanded reading of their spatial
effect on the viewer.
suggested this could be one source of Byzantine ‘inverted’ perspective. When compared to
17
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
Figure 4 (left) Master of the Legend of St. Francis, Vision of the Thrones, Assisi (image source: Web Gallery of Art;
image in the public domain)
Figure 5 (right) Master of the Legend of St. Francis, Vision of the Thrones, Assisi, from a computer screen at an angle
so that the altar top becomes a parallelopipedon (photo: author)
Another early example, the Dream of Innocent III (Fig. 6), again in the right-
hand position of the (second) bay, the pope is shown sleeping in an open box,
whose front face is again parallel to the picture plane. Now, orthogonals recede back
from the corner, providing an extremely robust illusion. For as soon as this line
approximates the normal geometry of walls seen from the side, it appears to be a
wall transverse to the viewer. In other words, from the centre of the bay, it is easy to
regard this painted wall as a real wall oriented at 90 degrees from the painted wall
surface.
Demus’ discussion of the goals of Byzantine art, it leads us to search for other site-specific
examples where a viewpoint might correct the geometry.
18
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
Taken together, these two examples show in different ways the desire to use
empirical perspective to remediate the view from the anchored viewer’s position.75
But I want to suggest that this seemingly revolutionary impulse is quite old and in
the case of the Vision of the Thrones indistinguishable from Byzantine practice. Yet
the effect is new in the Dream of Innocent III, and in a number of other frescoes, such
that the combination of the Byzantine style of optical correction combined with the
new observation of reality and use of radical angles to create space, creates a new-
old kind of image, which has not been seen before.
One of the ways in which White’s fine ‘conceptual minimalism’ blocks an
effective bridge between Byzantine and Renaissance perspective can be seen in his
attitude to so-called inverted perspective. Calling it a ‘mythical monster,’ he argues
that it is an unnecessary theoretical apparatus because its results can be explained
by the use of hieratic scale. In stating that there is ‘variations in figure scale are
neither dependent on any spatial relationship within the composition nor upon the
relationship of the scene as a whole to the observer,’ we can see that our analysis of
the Vision of the Thrones already questions this.76 More precisely, when continuing
with the notion of space in the fourteenth century, White misses what has been
called here an essentially Byzantine ‘scenographic’ function – the optical
preservation of the form.
Interestingly, White’s discussion touches on Gaddi’s Baroncelli chapel and
the Northerners, Giusto, Altichiero and Avanzo. While he stresses the ‘onlooker’s
inclusion in the event’ in Altichiero’s work, he sees this as essentially forward-
75 For the use of empirical perspective in architecture, see Marvin Trachtenburg, The
Dominion of the Eye: Urbanism, Art, and Power in Early Modern Florence, Cambridge, 1997.
76 White, The Birth and Rebirth of Pictorial Space, 103.
19
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
looking.77 What is important to note is that these are very large frescoes. As Wilde
confirmed with Demus, all frescoes to a degree partake in this enveloping function
and so to compare Altichiero to Masaccio is correct. But on the other hand, when
compared to one of those smallish altarpieces of the quattrocento that incorporates
perspective with mathematically graduated spatial recession, we can see that the
motivation is quite different. Not unlike a Byzantine artist, Gaddi or Altichiero
improve the image, which is made clearer for someone occupying the central traffic
pattern.
This kind of improvement is certainly found in Assisi too. And while the
lines of influence to Assisi generally point from Rome and Cavallini, and not Venice
(where this practice may have lived on), it will become clear that such concerns
were not unknown to the Master of the Legend of St. Francis. I have talked about
‘bounded’ and ‘unbounded’ forms, but another way to understand the detachment
of the figures from the wall surface is that rather than being seen as through a
window, as in fifteenth-century perspective, the figures protrude. Indeed, Schöne
notes how the architectural enframing of the scenes of the Lower Church in Assisi
seem to protrude through the use of ‘positional light’ (Standortlicht).78 The Cimabue
and the Master of the Legend of St. Francis both rely on powerfully protruding
forms such that we can, ironically (yet in a spirit close to White’s writings), state that
trompe l’oeil is primitive in motivation. The medieval impetus is geared toward
radical presencing.
Another way to bring further specificity to this discussion is to note that
Demus’ paradigm of optical correction is a form of mild anamorphosis.
Anamorphosis is defined as an optical device that involves manipulated stimuli that
appear in a different guise form a privileged viewpoint. Traditional perspective,
instead, uses plane that is contrariwise suggestive of a space further back. We need
not insist that the apostles that we see in Torcello cathedral are ‘protruding,’ to see
that oblique illusions as found in Assisi have more in common with Torcello than
traditional perspective.
For an example, already discussed briefly by Schöne’s student Puttfarken,
we may turn to the Peruzzi chapel, painted by Giotto and his workshop from c.
1326-1330.79 It has been noted that it provides a ‘corrected’ view for an oblique
viewer in the main transept area at the threshold of the chapel. Puttfarken writes,
‘What is happening here can best be described as a slight heightening of the sense of
continuity between viewer and the scene viewed, which endows the figures in the
painting with an enhanced sense of presence, accompanied by a corresponding
briefly by Elkins in this regard but somewhat misleadingly in this regard in a discussion of
centers of projection, which I am arguing the artists were not aware of.
20
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
reduction in the visual effect of the bounded surface.’80 This notion of presence is not
far from Demus’.
Puttfarken derived his ideas from Wolfgang Schöne, who instead
investigated ceiling frescoes. Already aware of the idea of ‘perspective Schrägsicht’
in Netherlandish church interiors from his teacher Jantzen, Schöne went out of the
picture into real space to understand the oblique viewer (and not viewpoint). The
result was an important reflection on ceiling painting, ‘Zur Bedeutung der
Schrägsicht.’ Schöne’s results, in turn, were expanded by his student Puttfarken in
his dissertation and other works, particularly on Titian’s Pesaro Madonna.81
Subsequent research by Julian Gardner and John Shearman has emphasised
physical barriers like chancel gates or presbytery steps, which spurred the
development of a ‘threshold instinct’ among painters.82
The psychological basis of oblique forms, just as overhead forms as found in
Byzantine vaults, lies in anamorphosis, seeing a percept because of spatial
manipulation.83 Recently, the subject has come to the attention of perceptual
psychologists, and the phenomenon can be appreciated for an extremely wide-
spread and important issue for perceiving art work, and in Deregowski’s hands
even suggesting a basis for Byzantine inverted perspective.84 The power of the
illusion in The Feast of Herod in the Peruzzi Chapel is given by the receding
orthogonals appearing now transverse (orthogonal) to the viewer from outside the
chapel (Fig. 7).85 Looking at the oblique image, one can see that the receding
orthogonals of the canopy are roughly parallel to the true transversal lines of the
stained glass window.
Kunstgeschichte, 34, 1971, 89-113; Shearman, Only Connect; c.f., Ernst Kitzinger, ‘The
Threshold of the Holy Shrine: Observations on the Floor Mosaics at Antioch and Bethlehem’,
in Kyriakon. Festschrift Johannes Quasten, P. Granfield and J. A. Jungman, eds., vol. 2,
Aschendorff, Muenster, Westfalen, 1970.
83 Rudolf Arnheim, ‘Perception of perspective pictorial space from different viewing points’,
Leonardo, 10, 1977, 283-288; Kim Veltman, ‘Perspective, Anamorphosis and Vision’,
Marburger Jahrbuch für Kunstwissenschaft, 21, 1986, 93-117.
84 Deregowski et al, ‘The Perception of Spatial Structure’; Ian Verstegen, Cognitive Iconology:
21
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
86 On the chapel, see Michael Kohnen, ‘Die coretti der Arena-Kapelle zu Padua und die
ornamentale Wanddekoration um 1300’, Mitteilungen des Kunsthistorischen Institutes in
Florenz, 48, 2004, 417-423; on the perspective, see briefly Elkins, Poetry of Perspective.
87 On the chapel, see Gardner, ‘The Decoration of the Baroncelli Chapel in Sta Croce’.
88 On Guariento, see Gian Lorenzo Mellini, ed., Altichiero e Jacopo Avanzi, Milan, 1965; on
22
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
appearance of a chapel askance, he could take into account flat decoration on the
wall for a purposeful improvement of such an image seen from the same angle.
Figure 8 (left), Guariento d’Arpo, Vestition of St. Augustine, c. 1360-65, Chiesa dei Eremitani (SS Giacomo e Filippo),
Padua (photo: the author)
Figure 9 (right), Altichiero, Death of St. Lucy, 1379-84, Chapel of San Giorgio, Padua (photo: the author)
Fig. 10 Giotto, so-called coretti, Arena Chapel, Padua (image source: Web Gallery of Art; image in the public
domain)
Indeed, we can look to the fresco of Christ before the Doctors in Padua for such
a precocious acknowledgement of the spectator.90 The Scrovegni chapel is six
registers long down the nave. This fresco appears on the left; thus for a centralised
viewer standing at the centre, the left wall is opened up so as to accommodate an
off-center viewer. This brings us back to Assisi. New ideas were circulating, and a
23
Ian Verstegen Otto Demus, Byzantine art and the spatial icon
slow ‘divergence’ with eastern practice was taking place. I have instead emphasised
continuities to work in a Viennese manner, free of anachronism. Oblique views in
trecento frescoes are not fundamentally new, but refinements of old techniques of
visualization.
Otto Demus shared with his Viennese compatriots an intense sensitivity to
the synchronic composition of works of art. His genius lay in expanding the field of
analysis from the painting (Pächt) or building (Sedlmayr) to seeing one variety of
mural decoration – the Byzantine – as cast upon the canvas of the whole church. He
discovered this formative principle (Gestaltungsprinzip) by cultivating a sensitivity to
the historical tradition under study. By looking laterally at Demus from the point of
view of other Vienna School writers, we can anticipate certain connections between
others’ treatments of medieval art and to revisit the question of the relationship
between eastern (Byzantine) and western practice in exciting new ways.
24