Introduction to an "Iconography of Mediaeval Architecture"
Author(s): Richard Krautheimer
Source: Journal of the Warburg and Courtauld Institutes, Vol. 5 (1942), pp. 1-33
Published by: The Warburg Institute
Stable URL: http://www.jstor.org/stable/750446
Accessed: 04-06-2017 18:47 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/750446?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
The Warburg Institute is collaborating with JSTOR to digitize, preserve and extend access to Journal of
the Warburg and Courtauld Institutes
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
INTRODUCTION TO AN "ICONOGRAPHY OF MEDIAEVAL
ARCHITECTURE"
By Richard Krautheimer
Since the Renaissance it has become customary to consider architecture asbeing determined by "commodity, firmness and delight" or, to use a less
Wottonian terminology, by function, construction, and design. To view
architectural problems from these angles and from them alone, has become
something like a fundamental tenet of architectural history. Yet the validity
of such a view appears rather doubtful where mediaeval architecture is
concerned.' Obviously there can be no doubt that problems of construction,
design and function, and of the integration of these elements, were of fundamental
importance to mediaeval as well as to later architects. Yet it would
seem that these essentials of architecture as laid down by Sir Henry Wotton,
and before him by Leone Battista Alberti, were differently emphasized and
that in addition to them other elements played a vital part in the mediaeval
conception of architecture. As a matter of fact, no mediaeval source ever
stresses the design of an edifice or its construction, apart from the material
which has been used. On the other hand the practical or liturgical functions
are always taken into consideration; they lead on to questions of the religious
significance of an edifice and these two groups together seem to stand in the
centre of mediaeval architectural thought. Not once, it will be remembered,
does Suger refer to the revolutionary problems of vaulting and design in his
new building at St. Denis. Evidently the design of an edifice or for that matter
the construction were not within the realm of theoretical discussion. On the
other hand the religious implications of a building were uppermost in the
minds of its contemporaries. Time and again Suger discusses the dedications
of altars to certain Saints. Questions of the symbolical significance of the layout
or of the parts of a structure are prominent; questions of its dedication to
a particular Saint, and of the relation of its shape to a specific dedication or
to a specific religious-not necessarily liturgical-purpose. The 'content' of
architecture seems to have been among the more important problems of
mediaeval architectural theory; perhaps indeed it was its most important
problem. The total of these questions would form the subject of an iconography
of architecture. Such an approach would merely return to an old
tradition which as recently as a century ago was still present in the minds of
archaeologists of art ;2 it is during the last fifty years only that this has apThis
article is based on a brief paper read
to the meeting of the College Art Association
held in Chicago in January 1941. I want to
thank my wife, Mrs. Trude KrautheimerHess
for her continuous collaboration in preparing
and writing this essay.
1 Throughout this paper the term "Middle
Ages" will be used so as to cover the whole
period from the 4th to the end of the 12th
century.
2 J. Ciampini, Vetera Monimenta ..., Rome,
1690-99; J. Bingham, The Antiquities of the
Christian Church, London, I708-22; J. Britton,
Architectural Antiquities, London, 1807; J.
Kreuser, Der christliche Kirchenbau2, Regensburg,
I86o-6 I; Otte, Handbuch der kirchlichen
Kunstarchdologie des deutschen Mittelalters5, Leipzig,
1883-85. A very few ecclesiastical
archaeologists have continued this century-old
tradition to the present day; the most
prominent among them are: J. Sauer,
Symbolik des Kirchengebdudes, Freiburg, I924
and F. J. Doelger, Antike und Christentum,
Miinster, 1929 ff.
I
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
2 RICHARD KRAUTHEIMER
parently been superseded by a pu
remarks are not intended to be com
tributions towards a future icono
i. Copies in Mediaeval Archi
An approach towards the discove
of mediaeval men were outstandin
architectural copies which wer
Obviously the relations between t
reveal some of these elements. Of
one another in mediaeval writing
author came to see any resembl
Miracula S. Maximini, for instance
Pr6s was built like the palatine
. . quae Aquis est constituta . . .
Malmesbury makes a similar statem
Robert of Lorraine at Hereford ". . . ecclesiam tereti edificavit scemate
Aquensem basilicam pro modo imitatus suo."2 Since two of these edifi
Aix-la-Chapelle and Germigny-des-Pr6s, still exist and Hereford is reco
in an i8th century drawing3 it is easy to check these examples; yet it is
for a modern beholder to see anything comparable in them. The chap
Aix, with its domical-vaulted octagonal centre-room surrounded by a si
sided ambulatory and by galleries, seems quite different from the squ
church of Germigny with its open central tower, its barrel-vaulted cross a
and its domed corner bays; nor does it seem to resemble the square do
storied chapel at Hereford in which of the nine bays the middle one is
in order to connect the two stories and the remaining eight are co
with groin vaults. One might at first be inclined to say that these state
are based simply on mistakes; but they are made so frequently and with
precision that this explanation seems too easy an escape. For instance
Sophia at Beneventum, an hexagonal structure with two ambulator
compared to the Hagia Sophia at Constantinople;4 the small Ioth ce
church of Petershausen is likened to St. Peter's in Rome ". . . secundum ...
formam basilicae principis apostolorum Romae constructa(m);"5 the I
1J. v. Schlosser, Schriftquellen zur Geschichte
der Karolingischen Kunst, Wien, 1896, in:
"Quellenschriften fiir Kunstgeschichte . . .
herausgegeben von R. Eitelberger," N.F.,
IV (henceforth quoted as Schlosser, "Karol.
Kunst"), no. 682.
2 Willelmi Malmesberiensis, Gesta Pontificum
Anglorum, ed. N. E. S. A. Hamilton, London,
1870 (Rolls Series), p. 300.
3 A. W. Clapham, English Romanesque Architecture
after the Conquest, Oxford, 1934, p. I 12,
from Society of Antiquaries, Vetusta Monumenta
I, London, 747, pl. XLIX.
4 "Translatio Sancti Mercurii" (I Ith century?),
M.G.H., SS. Rer. Langob., p. 576 ff.,
particularly p. 577: ". . . sancte Sophie
basilica, quam exemplar illius condidit
Justinianae;" cf. "Carmen de translatione
duodecim martyrum" (mid II th century),
ibid., p. 575-
5 Vita Gebhardi Episcopi Constantiensis,
M.G.H., SS. X, p. 582, particularly p. 587;
cf. J. Gantner, Kunstgeschichte der Schweiz, I,
Frauenfeld, 1936, p. 134 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 3
century cathedral at Bremen to those of Cologne and Beneventu
could be enlarged considerably, yet time and again the validity o
parison would be questioned. The only justifiable conclusion s
that the mediaeval conception of what made one edifice comp
another was different from our own. Mediaeval men must have had tertia
comparationis utterly at variance with those to which we are accustomed.
In order to understand these different principles it may be advisable
turn to buildings which were definitely copied from clearly established p
totypes.
Among the great number of edifices erected throughout the Middle Ages
with the intention of imitating a highly venerated prototype, one group is
particularly suitable for establishing the nature of a mediaeval copy: the
imitations of the Holy Sepulchre at Jerusalem. They exist not only in great
numbers but also depend on a model which is still relatively well preserved
and can easily be reconstructed in its original aspect. These copies were built
all over Europe from the 5th to as late as the i7th century.2 Yet although the
intention of imitating the Rotunda of the Holy Sepulchre is expressly stated
in many instances, the buildings vary surprisingly from each other; they are
also astonishingly different from the prototype which they mean to follow.
There is, for instance, the small church of St. Michael at Fulda (P1. Ia, e).
It was erected by Abbot Eigil, possibly with the advice of Hrabanus Maurus,
between 820 and 822. Although in its present state the structure is largely
i Ith century, the few extant original parts, the contemporary descriptions
and recent excavations are sufficient to give a fairly accurate idea of the
aspect of the 9th century edifice.3 A small centre-room is surrounded by an
ambulatory; a crypt, covered with an annular barrel-vault on a single short
ionic column extends under the centre-room and is surrounded by a ringshaped
corridor beneath the ambulatory. Eight columns-they were replaced
in the i th century4-carried what appears to have been a dome or an eightsided
domical-vault over the centre-room.5 It is not known whether this
1 Adami gesta Hammaburgensis eccl. pont. lib.
II, cap. 77, Lib. III, cap. 3 (ca. 1075), in O.
Lehmann-Brockhaus, Schriftquellen zur Kunstgeschichte
des II. und 12. Jahrhunderts, Berlin,
1938, nos. 230, 232.
2The subject of these copies has been
treated in a more or less general way by G.
Dalman, Das Grab Christi in Deutschland
("Studien tiber Christliche Denkmdiler," I4),
Leipzig 1922; N. C. Brooks, The Sepulchre of
Christ in Art and Liturgy, ("University of
Illinois Studies in Language and Literature,"
VII, 2), Urbana, Ill., 1921; Anonymous,
"Das Grab des Welterl6sers in seinen mittelalterlichen
Nachbildungen," Der Kirchenschmuck,
XXVI, 1895, pp. 125 ff., 141 ff.,
153 ff.; XXVII, 1896, pp. Io ff., 33 ff. The
latest copy of the Anastasis which has come
to my knowledge was erected at Innichen in
1653; it was copied after 1888 at Potsdam to
serve as a mausoleum for the Emperor
Frederick III; see Kirchenschmuck, XXVII,
I896, p. 12.
3 Candidus, Vita Eigilis, M.G.H., SS. XV, I,
p. 221 ff., particularly p. 230 f.; J. Schalkenbach,
"Die Wiederherstellung der Michaelskirche
zu Fulda," Deutsche Kunst und Denkmalpflege,
1938, p. 34 ff. Only the crypt and
the ground plan of the main floor are 9th
century.
4 The old bases have been found underneath
the I th century ones; four of the
Carolingian capitals have been re-used; see
Schalkenbach, loc. cit.
5 Candidus, Vita Eigilis, op. cit., p. 231:
"... in summo uno lapide istius aedificii
perfectio consummatur... ."
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
4 RICHARD KRAUTHEIMER
original structure had a gallery over the am
the three chapels which now radiate from the
east; the north and south chapels as well as
roof and the long nave, were all added in a r
of the main altar leaves no doubt that the
linked to the Holy Sepulchre,1 and in fact as
conical in shape, rose in the centre of the edifi
Two hundred years after Eigil and Hraban
Fulda, Bishop Meinwerk of Paderborn w
similitudinem s. Jerosolimitane ecclesie"
Helmershausen to Jerusalem to bring from t
"mensuras eiusdem ecclesie et s. sepulgri." "
. . . et mensuras eiusdem ecclesie et sepulchr
the church was built and consecrated in 1036
is at present the Busdorf convent. Excavatio
edifice was an octagonal structure rising f
Three large rectangular chapels radiated from
axes; a fourth chapel, possibly flanked by tw
as an entrance structure. No supports div
building was vaulted or had a wooden roof r
The Rotunda at Lanleff, not far from Cae
the i Ith century.4 It is again a round str
ambulatory; the ambulatory was covered by
supports, each a square pier with four eng
sidioles radiated from the ambulatory. Alth
edifices, the dedication of Lanleff has not b
of these absidioles and their position off-ce
structure also is derived from the Holy Sepu
A fourth building, the church of the H
would seem to date from the first quarter o
was badly restored in 1841, but earlier re
the original state.6 Eight sturdy columns separ
from a centre-room which until 1841 was s
first storey of this tower was obviously the
Evidently the original structure had no choi
I313) and no radiating chapels. Eight twin o
1 Hrabanus Maurus, Carmina, 42, M.G.H.,
P. L. II, p. 209: "In primo Altare.
Hoc altare deo dedicatum est maxime Christo
Cuius hic tumulus nostra sepulcra juvat..."
2Dalman, op. cit., p. 27 f.
3 Vita Meinwerci episc. Patherbrunensis, cap.
209 ff. (second half of I2th century), Lehmann-Brockhaus,
op. cit., nos. 1046-I050. The
excavations have been discussed by W. Rave,
"Die Entdeckung der ursprtinglichen Busdorfkirche
zu Paderborn," Deutsche Kunst und
Denkmalpflege, 1936, p. 221 ff.
4A. Rhein, "Le Temple de Lanleff,"
Congris arche'ol., 81, 1914, p. 542 if.; E.
Viollet-le-Duc, Dictionnaire raisonne de l'Architecture
Franfaise, VIII, Paris, I866, p. 287 f.
The church is at present a ruin; one of the
radiating chapels is preserved, the others have
been restored on the basis of reliable traces;
see Rhein, loc. cit.
5 J. Essex, "Observations on the Origin and
Antiquity of Round Churches," Archeologia,
VII, 1787, p. I63 ff.; Ch. Clarke, "An Essay
towards an History of Temples and round
Churches," in J. Britton, Architectural Antiquities,
I, 1, 1807; Clapham, op. cit., p. 109 f.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
.:ii iiii-lii iiii:-:-:?:::-- -: ii::~:_- !-i::-- ii i~i:iiii iiiii ''ii?:
i??i,, , iii, i : i .:.. iiiiiiilfii ii ii i-i iiiiii :iii iiii'i8:i- ii;'':?: - i-:-';i--"'; :-:'::i-i`i;-:-:- ----:--------- --'-::-
::::::i-:::::::::-:::::iii:i:i::iii::).:::::ii-iiiiiiiiii-i!i:--:::::
i !:illi~ii~i~!
L::-iiii'iii:- iili:i
a-St. Michael, Fulda
(PP. 3, 6)
! .
P :i
! =!
Y
b
(P. 4)
i111' s.\
~I
ii
c-Rotunda, Lanleff ne
(P. 4)
e-St. Michael, Fulda (pp. 3, 7) f-Holy Sepulchre, Cambridge (From Britton, Arch. Antiq
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
,'-' ; ',,'i:" -`l:
\\ !
I"
a-Anastasis,
-,i:iap"ttto-::;::;:19 . cslodch- r i:ol::::iii:::
'::
b-Anastasis, Jerusalem. 1725 (After
Horn, Cod. Vat. lat. 9233, (p. 5)
B
C
cd-Anastasis
in the Sacramentary of
Henry II. Munich, cod. lat. 4456 (p. 14)
e-Anastasis on the Wooden Casket
Rome (p. 15)
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 5
piers and slender intermediary columns, were arranged above
the ground floor, forming a sham gallery.
All four of these structures were intended to represent the Ho
at Jerusalem. But all four are quite different from one anot
round or octagonal, with a single nave or surrounded by a
vaulted or possibly with timber roofs, with one or more absid
or twelve supports. The differences seem considerably to
similarities.
These differences become even more striking when the four edifices are
compared with their common prototype at Jerusalem (P1. 2a, b). This is not the
place to enter into the complicated history of the Anastasis Rotunda.' Yet it
is certain that from 628, when the Rotunda was restored,2 and in all likelihood
from the very outset in 340-350, it was a round structure with an
ambulatory, surmounted by a gallery. Three small absidioles were added to
this ambulatory in the 7th century. An outer ambulatory apparently encircled
the whole building.3 The central room was surrounded by twenty
supports, eight piers in the main axes and three columns in each of the
diagonal axes. In the gallery two columns and one pier rose in the diagonal
axes above the three columns on the ground floor, while in the main axis
two piers corresponded to those below. The arches of the gallery openings
were as wide as those of the lower arcades. It remains uncertain whether
the centre-room was vaulted or had a conical roof similar to that which
existed from the 12th to the early 19th century.
Doubtless there are some general similarities between the 'copies' of t
Holy Sepulchre at Fulda, Paderborn, Lanleff and Cambridge, and their Ear
Christian prototype. Yet these similarities seem to be rather vague to th
modern eye; three of the copies are round, whereas the fourth, Paderbo
was octagonal. Indeed there are more examples of Holy Sepulchres w
polygonal rather than round plans. The 12th century Rotunda of the Ho
Sepulchre at Sto. Stefano at Bologna4 forms an irregular dodecagon (P1.
whereas the church of Sto. Sepolcro at Pisa, built in the middle of the 1
century,5 is a perfect octagon with a very wide ambulatory and unusual
tall piers. In the church of the Holy Sepulchre at Northampton, da
circa I 20,6 the ambulatory is round although divided into eight squa
and eight triangular bays, whereas the centre-room, which was rebu
at a later period, is octagonal. It seems as though circle and polyg
were interchangeable throughout the Middle Ages. For as early as t
4th century Gregory of Nyssa described the plan of an octagonal church
forming "a circle with eight angles" although he apologizes for his somew
1 H. Vincent and V. Abel, Jerusalem
Nouvelle, II, I, Paris, 1914, p. 89 ff.
2 J. W. Crowfoot, Early Churches in Palestine,
London, I937, p. 9 ff. The restoration of 628
has sometimes been assumed to have been a
complete rebuilding. Vincent and Abel's
findings have definitely disproved this hypo-
thesis.
3 R. Krautheimer, "Santo Stefano Rotondo
a Roma e la Chiesa del Santo Sepolcro a
Gerusalemme," Rivista di Archeologia Cristiana,
XII, 1935, p. 51 ff., particularly p. 85 ff.
4 See below, p. 17 ff.
5 M. Salmi, L'architettura Romanica in Toscana,
Rome, n. d., p. 16, fig. 26, pl. 105. G.
Rohault de Fleury, Les monuments de Pise,
Paris, I866, p. 55 f., pl. XVII.
6 Clapham, op. cit., p. 109 f., pl. 23, fig. 35.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
6 RICHARD KRAUTHEIMER
loose terminology.' From then on d
more and more. To Arculph who vi
the octagonal church of the Ascen
"rotunda" and so was the cross-dom
nople.2 Even as late as 1322 Sir Jo
of the Rock "a circular edifice."3 It could almost be said that to mediaeval
eyes anything which had more than four sides was approximately a circle
Nor are semicircle, square and rectangle clearly differentiated: the semicircular
apses of the Anastasis were transformed into squares in Meinwerk
chapel at Paderborn, while in the Arculph-Adamnanus plan of the Rotund
they were given unmistakably as rectangles.4 An approximate similarity
the geometrical pattern evidently satisfied the minds of mediaeval men as
the identity of two forms; survivals of such an attitude could probably b
found to this day.
This 'indifference' towards precise imitation of given architectural shape
prevails throughout these 'copies' of the Holy Sepulchre. The ambulator
around the centre-room is one of their usual characteristics but it is not b
any means indispensable. It is missing not only at Paderborn but also in th
round single-naved chapel of St. Maurice at Constance, which was erected
between 934 and 976 to contain a "sepulchrum Domini in similitudine illiu
Jerusolimitani . . ."5 The three apses off the ambulatory of the Anastasis,
they are repeated at all, are sometimes not only square instead of semicircular
as at Paderborn and perhaps also in the closely related i i th century chap
at the Krukenburg ; they are also frequently arranged in a position differen
from that in the Anastasis. In the church of Lanleff alone they keep the of
centre position of the Holy Sepulchre at Jerusalem. In the chapel of the Ho
Sepulchre at St. Leonard a fourth chapel was added,7 with the result that
the four chapels occupy the main axes of the structure forming a cross.
the churches of Paderborn and of the Krukenburg the place of this fourt
chapel was taken by a longish entrance wing and the same arrangement wa
used when the chapel of St. Michael at Fulda was rebuilt in 1092 (P1 ia
In all probability the fact that, in addition to the three already existin
chapels, a choir was added to the Anastasis in 1 017 by the Emperor Mono
machus inspired this four chapel plan of the copies, although other types may
1 Gregorii Nysseni opera, VIII, 2, Epistulae, ed.
G. Pasquali, Berlin, 1925, p. 76 ff. "... ? x.
0xw yolvatq atEtXelltvoC S ,0. .UX ..0V . 0'[To 7CEpTEspEt q V xvYTa O a ' oxZCymov eywa . C..-
2 Arculph-Adamnanus, De Locis Sanctis, lib.
I, cap. 23; lib. III, cap. 3. I am using for
quotation the edition in "Itinera Hierosolymitana
Saeculi IIII-VIII," by P. Geyer,
Prague, Vienna, Leipzig, 1898 (Vol.
XXXVIIII of the Coipus Scriptorum Ecclesiasticorum
Latinorum), p. 219 ff. (hereafter quoted
as "Arculph").
3 The travels of Sir John Mandeville, The
version of the Cotton MSS. ed. by A. W.
Pollard, London, 1915, P- 54-
4Arculph, op. cit., Lib. I, cap. 2, p. 231.
5 Vita Chuonradi Constantiensis Episcopi,
M.G.H., SS. IV, p. 429 ff., especially p. 432.
6H. Hartung, "Die Kapelle auf der
Krukenburg," Die Denkmalpflege, 1920, p. 27
f.; R. Schultze, "Eine mittelalterliche Rundkirche
im Wesergebiet," Bonner Jahrbiicher,
127, 1922, p. 237 if. The church was evidently
dedicated to St. John the Baptist, but was
never a baptistery.
7 R. Fag6, "L'6glise de Saint Leonard et la
chapelle du Sepulchre," Bull. mon. 77, 19I3,
p. 41 ff. The fourth chapel is a modern
restoration based on old remnants.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
'iiiiii~iiiii
iiiiiiiii
R iii~ iisii
'@I!
a-Sto. Stefanb-Sto.
(P
c-Neuvy-StdThis
content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 7
have exerted a collateral influence. The interpenetration of circul
and cross types is frequently found in Early Christian memorial churc
as Wiranshehir or the octagon church projected and described by
of Nyssa; in the Occident Sto. Stefano Rotondo in Rome, S. Angelo
and the Baptisteries of Santa Severina and Canosa in Southern Ital
quoted.'
The internal supports are sometimes exclusively columns, as for instance
at Fulda, at St. Leonard, at Cambridge, at Northampton and in the church
of S. Giovanni del Sepolcro at Brindisi; at Sto. Sepolcro at Pisa they are
hook-shaped piers with responds, and at Lanleff composite piers with
engaged columns. An alternating rhythm of piers and columns, though
different from the particular pattern of the Anastasis, is found only once,
in the Pisa Baptistery (P1. 3d) which in its original form of I153 ff. was
clearly copied from the Holy Sepulchre at Jerusalem.2 Nor is the gallery
of the Anastasis always repeated even in those copies which have ambulatories
surrounding the centre-room. Frequently, at 'St. Leonard for
instance, and possibly in the 9th century chapel at Fulda, it is omitted altogether;
the clerestory also is frequently wanting, so that the whole edifice is
reduced to a one-storied structure which is quite different from the threestoried
original. If on the other hand a gallery does surmount the ambulatory,
its arrangement differs entirely from that in the Anastasis. Instead of its
complicated rhythm of supports, small twin openings are sometimes arranged
above each arch of the ground-floor; this is the case at Cambridge (P1. If),
at Bologna, and in the present edifice of St. Michael at Fulda (P1. I e) as
rebuilt in 1092. Elsewhere, for example at Neuvy-St.-Sepulchre, all the
gallery openings form a continuous band, supported by plain columns3
(P1. 3c). Finally, in the vaulting patterns of centre-room and ambulatory
all these churches are as different as possible from one another as well as
from the original.
This inexactness in reproducing the particular shape of a definite architectural
form, in plan as well as in elevation, seems to be one of the outstanding
elements in the relation of copy and original in mediaeval architecture.
Indeed it recalls a well-known phenomenon, the peculiar lack of
precision in mediaeval descriptions not only of architectural patterns but of
all geometrical forms. When discussing the elements of geometry, a somewhat
pedestrian but usually precise scholar such as Isidore of Seville becomes completely
vague. A sphere is, in his words, a round figure which is alike in all
its parts; a cylinder is a square figure which has a semicircle on top; a
pyramid-since its name is derived from i7cp-is a figure which tapers like a
1 Krautheimer, op. cit., p. 77 if.
2 M. Salmi, op. cit., p. 16, figs. 27-29, pls.
Io6-og9; Rohault de Fleury, op. cit., p. 56 ff.,
pls. XVIII-XXI; see below, p. 31 f.
3 R. Michel-Dansac, "Neuvy-Saint S 6pulchre,"
Congris archeol., 94, I937, P- 523 if.
The building was founded in 10o42 or 45 "ad
formam sancti Sepulchri lerosolimitani."
According to Michel-Dansac the lower part
of the Rotunda was built about the middle
of the I Ith century, the nave c. 1087, the
upper parts of the Rotunda, including the
gallery, between I12o and 30. J. Hubert, "Le
Saint-S6pulcre de Neuvy," Bull. mon., 9o,
I931, p. 91 ff., dates the whole Rotunda 12th
century without giving any convincing
reasons.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
8 RICHARD KRAUTHEIMER
flame.' Even such an outstanding
unprecise so far as the descript
the other hand the number of
always strongly stressed. A square
within four straight lines: the n
the four lines to one another (w
length and by saying that they st
omitted. The geometrical form
figures.3
This particular attitude suggests a quite different approach as compared
with that of the modern mind to the whole question of copying. Indeed the
lack of geometrical precision is as characteristic as the 'indifference' towards
precise imitation of architectural shapes and patterns. In lieu of this, other
intentions seem to be at the basis of copying architecture in the Middle Ages.
It would seem as though a given shape were imitated not so much for its own
sake as for something else it implied: the connotations of the cross-shaped
ground plan are stressed time and again in mediaeval sources, as has been
frequently pointed out.4 St. Ambrose in 382 was among the first to emphasize
that the cross plan of the church of the Holy Apostles at Milan, which he laid
out, was meant to symbolize the victory of Christ and of His cross. The same
interpretation was given as late as I 122, when the church at Kappenberg was
built.5 Over and over it is emphasized that such and such a church was laid
out "instar crucis"6 or, as at St. Gall in 898 "in honore et modum s. Crucis
*..."7 Yet it does not seem to matter greatly which particular cross shape was
meant, whether it was a basilica plan in the pattern of the Latin crux capitata,
as at Deas ;s whether it was the pattern of the T-cross as at Bamberg cathedral
(11 17);9 or whether a Greek cross plan was referred to as in Arculph's
description of the church at the well of Jacob "quae quadrifida in quatuor
mundi cardines formata extenditur quasi in similitudinem crucis."10 The
term may possibly have been applied even to round edifices with cross chapels,
such as Meinwerk's Holy Sepulchre in Paderborn; a chapel which in 1064
was erected at Schaffhausen is described as having "capellas . . . in modum
crucis per gyrum constructas . . ."11 Occasionally the cross shape refers even
to the pattern in which five churches are laid out within or around a city.12
1 Isidori Hispalensis Episcopi Etymologiarum
sive originum Libri XX, Lib. III, cap. XII, ed.
W. M. Lindsay, Oxford, I9I I.
2 Migne, Pat. Lat. CXXXIX, c. 93 if.
3 Isidore, loc. cit.
4H. Graf, Opus Francigenum, Stuttgart,
1878, "Die Entstehung der kreuzf6rmigen
Basilika," p. 42 ff.; Sauer, op. cit., pp. I Io f.,
291 f., 431.
5 Origo monast. Cappenbergensis, LehmannBrockhaus,
op. cit., no. 669: "constructa
surgit ecclesia instar crucis erecta . . . deinceps
quoque victoriosissimae crucis ac reliquorum
visuntur miracula."
6 Lehmann-Brockhaus, op. cit., passim; J.
von Schlosser, Quellenbuch zur Kunstgeschichte
des abendldndischen Mittelalters, Wien, 1896,
passim; Schlosser, Karol. Kunst, passim.
7 Ekkehardi IV. Cas. s. Galli, Schlosser, Karol.
Kunst, no. 455.
8 Translatio s. Filiberti, cap. 29, Schlosser,
ibid., no. 666.
9 Ebbonis vita Ottonis Bambergensis, Lib. I,
cap. 22, Lehmann-Brockhaus, op. cit., no. I 15.
10 Arculph, op. cit., Lib. II, cap. 21, p. 270 f.
11 Notae s. Salvatoris Scafhusensis, M.G.H., SS.
XIII, p. 727, Lehmann-Brockhaus, op. cit.,
no. 1292.
12 Such a scheme is mentioned in Bamberg
by Adalbertus, Vita Heinrici II, (mid I2th
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 9
Similarly the round (or polygonal) shape of a church evidently
symbolical significance and again it did not make any great differenc
the ground plan of an edifice formed a regular circle or an o
dodecagon or any related pattern. The circle according to St. Aug
a symbol of virtue, an interpretation which he based on Horace
in se ipse totus teres atque rotundus."2 It is pre-eminent amon
geometrical figures and comparable to virtue because of the confo
concordance of its essentials, its "congruentia rationum atque
According to Eigil the circle is a symbol of the Church, never e
containing the sacraments; also it signifies to him the reign of etern
the hope of future life and the "praemia mansura quibus jus
coronantur in aevum."3 Other interpretations of the circle contin
out the Middle Ages down to Dante; the Divina Comedia is full of
ences. Whatever the particular interpretation, however, it is unqu
not so much the precise geometrical shape of a form as its gene
and its implications which count in the opinion of the mediaeva
On the other hand it would certainly be a mistake to assume
bolical interpretations of this kind were always the preponderant
giving a structure a certain shape, to make it for instance round
shaped. Sometimes this may have been so-as in the case of St. Amb
church at Milan; at other times an existing plan may have been
post festum as having some symbolical meaning.4 Usually, however
relations between the symbolical significance of a geometrical patt
ground plan of a building are not so plain. The process is of a m
intricate nature; probably the relation between pattern and s
meaning could be better described as being determined by a
reciprocal half-distinct connotations. Rather than being either t
point or else a post festum interpretation, the symbolical significanc
thing which merely accompanied the particular form which was
the structure. It accompanied it as a more or less uncertain c
which was only dimly visible and whose specific interpretati
necessarily agreed upon. Yet as a connotation it was nearly alway
with the pattern which had been chosen. Its very vagueness e
variety of interpretations given to one and the same form either by
different authors. The situation can hardly be better expressed
by Johannes Scotus Erigena." He speaks about the symbol
number eight, of its relation to Sunday and Easter, to resurrect
generation, to spring and new life. All these different connotatio
are ever present and "vibrate" in him whenever he thinks of eigh
century) Lib. I, cap. 6, Lehmann-Brockhaus,
op. cit., no. 99.
1 Augustine, De quantitate animae, cap. XVI,
Migne, Pat. Lat., XXXII, c. Io51 f.
2Satirarum lib. II, sat 7, v. 86.
3 M.G.H., SS. XV, I, p. 231. Candidus'
interpretation precedes by three hundred
years the similar one of Honorius of Autun,
Gemmae divinae, Lib. I, cap. 147, Migne, Pa
Lat. CLXXVII, c. 590, see Sauer, op. cit.,
p. IIo, n. I.
4 Sauer, op. cit., p. 289 ff.
5 Versus Iohannis Scotti ad Karolum Regem,
M.G.H., P. L. III, p. 550 ff., particularly
v. 45 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
Io RICHARD KRAUTHEIMER
Haec sunt quae tacite nostris in cordibus intus
Octoni numeri modulatur nabla sonorum
Spiritus interior clamat nec desinit unquam
Semper concrepitans, quicquid semel intonat annus
Haec scriptura docet cui rerum concinit ordo.
This brings us to the symbolical meaning of figures and numbers throughout
mediaeval architecture in general and their importance in architectur
copies in particular. Indeed they appear to be prominent among the elemen
which determine the relation between copy and original. At St. Michael a
Fulda, it will be remembered, the centre-room was carried by eight column
at Lanleff by twelve. In fact the number of eight or twelve supports seem
to be almost a constituent element of all imitations of the Holy Sepulchr
throughout the Middle Ages.' The chapel at St. Leonard, the Holy Sepulch
at Cambridge, the one at Northampton, the one at Pisa and S. Giovanni d
Sepolcro at Brindisi all have eight supports; twelve supports are found, in
addition to Lanleff, at Bologna and at the Holy Sepulchre at Augsburg
This seems hard to explain, except on the grounds that after all in any circular
building a number of supports divisible by four is easiest to arrange. Whil
this is undeniably the case, it might be pointed out first, that divisions o
central edifices into seven (Ste. Marie at Rieux Minervois),3 ten (S. Lorenz
at Mantua) or eleven (Neuvy-St.-Sepulcre) bays do occur. An arrangement
of six supports even is quite frequent; it seems to prevail in most Templa
churches.4 Second, within the group of the Holy Sepulchres eight and twel
appear to be the only multiples of four chosen for the number of support
The real explanation may be found in the very fact that they were actuall
intended to reproduce an important feature of the Rotunda at Jerusalem
for it will be recalled that this Rotunda was carried by twenty supports, v
eight piers and twelve columns. Evidently in the 'copies' either the numb
of the piers or that of the columns was chosen and 'imitated' regardless of the
particular shape of the supports. Definite proof of this procedure is found
mediaeval descriptions of the Rotunda at Jerusalem. Arculph in describing
mentions only the twelve columns and completely omits the eight piers in the
main axes. They simply did not exist in his account.5
Obviously the choice of the numbers eight and twelve from the twent
supports which were present in the prototype is again linked to the symbolical
associations of these numbers within mediaeval numerology. The existenc
1 The i i columns in the church of Neuvy
St.-Sepulchre are probably due to a lax
execution of the plan which can also be
observed in other parts of the edifice: half of
the vaults are out of shape and the niches are
almost all different in size. It is likely that the
plan originally was intended to have 12
supports; see Michel-Dansac, op. cit.
2 The chapel is known through a number
of I 7th century drawings; see Dalman, op. cit.,
p. 44 ff.
3 The church at Rieux Minervois rather
than being a Holy Sepulchre, as has been
assumed, was always dedicated to the Virgin;
see J. de Lahondes, "Rieux Minervois,"
Congris Archeol. 73, 19o6, p. 54 ff; M. Young
de Veye, Congre's Archdol. 37, I870, p. II7 iff
4 We quote only a few instances, such as the
Temple in London and the one which existed
in Paris until the I8th century; see Royal
Commission on Historical Monuments (England)
London, IV, London, 1929, P. 137 if.; Violletle-Duc,
op. cit., IX, p. 14 ff-
6 Arculph, op. cit., Lib. I, cap. 2, p. 227
"XII mirae magnitudinis lapidae sustentant
columnae."
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" II
of these numbers in the Anastasis and their numerologic
a reciprocal interrelation to one another. The importanc
bolism in mediaeval thought is too well known to n
Obviously the number twelve was bound to remind any
of the number of the Apostles, particularly when conne
of the Lord; combined as it is of four times three, it linked
four regions of the world with that of the Trinity whose g
the Apostles throughout the world.2 As early as the 4th
columns supporting the hemisphere of the Constantinian
Anastasis reminded Eusebius of the Disciples.3 In the 7t
compared to their number the twelve lamps which hung
the Lord, divided into three groups of four.4 Likewise th
bound to have a particular meaning in association wi
was a perfect number, which generally referred to S
Pentecost, to circumcision and baptism, to regeneration a
-most important of all-to resurrection ; indeed it symb
self. With such connotations in mind it must have b
solution to single out the number of eight supports whic
by the number of piers in the Anastasis. If this number
structure which copied the building over the spot where
the Lord had taken place, it was bound to hold out hope
rection to the faithful.
Of course, the actual number of elements in the prototype may frequently
have stimulated subsequent and divergent interpretations; thus the monk
Candidus in describing the church of St. Michael at Fulda saw in the eight
columns not a symbol of resurrection but of the eight Beatitudes.7 The
same Candidus likened both the one base of the column in the crypt of St.
Michael and the one keystone of the vault to Him who is the Beginning and
the End,8 and one may recall that to the Early Christian period and to the
1 V. F. Hopper, Mediaeval Number Symbolism,
New York, I938, passim; Sauer, op. cit., p.
61 ff.; J. F. Doelger, "Zur Symbolik des
altchristlichen Taufhauses," Antike und Christentum
IV, 1934, p. 153 ff.
2Augustine, In Johannis Evangelium, XXVII,
io, Migne, Pat. Lat., XXXV, c. I6I9 f.:
"... per quatuor cardines mundi Trinitatem
fuerant annunciaturi. Ideo ter quaterni ..."
3 Eusebius, Life of Constantine, Lib. III, cap.
XXXVIII, Nicene and Post-Nicene Fathers, I,
Oxford and New York, I89O, p. 530.
4Arculph, op. cit., Lib. I, cap 2, p. 229:
"in quo utique sepulchro duodenae lampades
iuxta numerum XII sanctorum apostolorum
. lucent ex quibus quattuor in imo illius
lectuli sepulchralis loco inferius positae,
aliae vero bis quaternales super marginem
eius superius conlocatae ad latus dexterum
oleo nutriente praefulgent." According to the
Testamentum Domini Nostri Jesu Christi, I, I9,
ed. I. E. Rahmani, Mainz 1899, p. 23, a
baptistery should be 21 cubits long "ad
praefigurandum numerum ... prophetarum"
and 12 cubits wide "pro adumbrandis iis qui
constituti fuerunt ad praedicandum evan-
gelium."
S Augustine, De Sermone Domini in monte,
Lib. I, cap. IV, 12, Migne, Pat. Lat., XXXIV,
c. I235; idem, Epist. Classis II, Epist. LV,
cap. XVII, 32, Migne, Pat. Lat., XXXIII,
c. 220; see also Hopper, op. cit., p. 85; Sauer,
op. cit., p. 78 f. and Doelger, op. cit., passim,
based on numerous quotations from Early
Christian and mediaeval writers, have discussed
at length the symbolism of numbers in
general and of the number eight in particular.
6 Augustine, De Civitate Dei, Lib. XV, cap.
XX, Migne, Pat. Lat., XLI, c. 462 ff.
SCandidus, Vita Eigilis, M.G.H., SS. XV,
I, p. 231.
8 Candidus, Vita Eigilis, M.G.H., SS. XV,
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
12 RICHARD KRAUTHEIMER
Middle Ages eight was nothing but a
Candidus incidentally made it qui
afterthought when he stated that
represent Christ and the Church
posse figuram."2 In the Vita of Be
three altars of the 8th century ch
that the Church of St. John cont
consisting of seven lamps or seven
as the seven gifts of the Holy Spi
(and the instances given are merel
of certain numbers in architectura
This number consciousness is likew
ments in architectural copies. Measu
ing been brought from Jerusalem fo
the Anastasis or of the Tomb of C
not by any means an isolated case.
in Io63-64 "rotundo schemate in
solimis. Unde et marmor supe
longitudinem 7 pedum quoniam
eiusdem longitudinis existit."1 Som
to give the measurements of the T
hausen where as late as 1492 the s
three intersecting lines on the wall
describing them as representing it
emphasis given to measurements
descriptions of the Holy Sites in
expressly the length of seven feet
measured it with his own hand, "
the approximate height and width
room for nine people-he says "ter
bolical value of the number threethan
a rather tall man. About 8o6 th
circumference both of the outer am
room. 7
I, p. 230: "Cuius tecturae princeps et con
ditor est Christus Jesus, fundamentum scil
columnaque manens semper immobilis .
in quo omnis aedificatio constructa cres
... Quid vero significet hoc, quod in sum
uno lapide istius aedificii perfectio consu
matur idem Doctor insinuat ... . , ut ille
coepit in nobis opus bonum, perficiat us
in diem Christi Jesu, quatenus cuncta
operatio nostra a Deo semper incipiat, et per
eum coepta finiatur."
1 Augustine, Epist. Classis II, Epist. LV,
cap. XVII, 32, Migne, Pat. Lat., XXXIII,
c. 220: ".... ut octavus primo concinat."
2 Candidus, Vita Eigilis, M.G.H., SS., XV,
1, p. 231.
3 Vita s. Benedicti Anian., cap. 26, Schlosser,
Karol. Kunst, no. 574: "tres aras censuit
subponi, ut in his personalitas trinitatis
typice videatur significari. ... In septem item
altaria, in septem candelabria et in septem
lampades septiformis gratia spiritus sancti
intelligitur."
SDe Sanctis Ecclesiae Cameracensis Relatio,
Auctore Monacho Valcellensi (I 2th century),
Lehmann-Brockhaus, op. cit., no. I670.
5 Dalman, op. cit., p. 90 f.
6 Arculph, op. cit., Lib. I, cap. 2, p. 229.
SCommemoratorium de casis Dei vel monasteris,
T. Tobler and M. Molinier, Itinera Hierosolymitana
et Descriptiones Terrae Sanctae, Geneva,
1879, p. 299 ff., especially p. 305This
content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 13
These descriptions make it quite clear that as in the case of other e
the measurements were not by any means reported in toto. In the s
in which only one group of supports or only one of the three storie
Anastasis was chosen for reproduction in its copies, only one or two
ments were selected from a much greater number. The writer
Commemoratorium does not report to his correspondent the measure
the inner ambulatory nor does he indicate the height of any par
building. This selective transfer explains also the strange use of measu
in a building such as the chapel of the Holy Sepulchre at Paderborn
the elaborate statement of the chronicler that the exact measurements of the
Anastasis were dispatched by special messenger from Jerusalem to Paderborn,'
to a modern beholder not only the plan but also the measurements of
the chapel look totaliter aliter. Some measurements however do appear to have
been transferred from the Holy Land and used in Meinwerk's structure: the
interior length of each side of its octagon is 5.80 m. and this corresponds
roughly to the distance of 5-70 m., measured between the outer corners of
the pairs of main piers in the east-west axis of the Anastasis.2 The eight piers
at Jerusalem would seem to have suggested to Meinwerk's messenger an
octagon and the measurements taken between two of those piers were used
as a basis for the construction of the whole plan.
This selective transfer of measurements finds its exact parallel in the way
in which prototypes are generally copied in the Middle Ages. It has been
pointed out before that the model is never imitated in toto. A selective transfer
also of the architectural elements takes place. In the chapel at Constance it
is only the roundness of the Anastasis which is transferred; at Paderborn the
roundness and the radiating chapels are taken over. In the Sto. Sepolcro at
Pisa and likewise in the chapel at Brindisi the roundness, the ambulatory, the
clerestory and eight supports are reproduced. In addition to these elements
the chapels of the Holy Sepulchre at Cambridge and Northampton also took
over the gallery above the ambulatory. Evidently the mediaeval beholder
expected to find in a copy only some parts of the prototype but not by any
means all of them.
Another point will have become apparent in this connection. The parts
which have been selected in these 'copies' stand in a relation to one another
which in no way recalls their former association in the model. Their original
coherence has been discarded. The original unity has been disintegrated and
the elements have been reshuffled, as it were. To take just one instance, the
twin openings of the galleries in the chapels of the Holy Sepulchre at Cambridge
(P1. If) and in Sto. Stefano at Bologna (P1. 3b) seem to reproduce in
an abbreviated form the gallery in the Anastasis. In the Rotunda at Jerusalem
two piers always seemed to flank one column, but it evidently appeared
quite natural in these later churches to reduce the more complicated rhythm
to the usual form of a Romanesque twin opening and to arrange one of these
twin openings above every intercolumniation of the ground floor. It is significant
to observe that, as late as the I7th century, engravings of the Anastasis
1 See above, p. 4. 2 Rave, op. cit., fig. 232 and Vincent-Abel,
op. cit., II, i, pl. XIII and fig. 59-
2
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
14 RICHARD KRAUTHEIMER
occasionally transform the rhythm of the ga
the original pattern was preserved until 18
are entirely changed; the ambulatory, whic
fourth the width of the centre-room, is in
than one-third and sometimes more tha
Northampton, Fulda, St. Leonard). Thus
ambulatory stand in an entirely different r
say the relative vertical proportions of arca
correspondingly re-arranged.
This procedure of breaking up the origina
shuffling these, also makes it possible to
elements quite foreign to the original. Buil
blance to the prototype seem to have exer
copy: at St. Michael at Fulda a crypt exte
similar crypt, also surrounded by a corridor,
at SS. Karpos and Papylos at Constantinopl
laid out ". . q aT- r ou ~ XparZthe crypt seems to have been derived ultim
and merged with the prototype of the Ho
Representations of buildings in mediaeva
to confirm the peculiar relation between
architecture. The methods used in these me
been discussed.3 Like the 'copies' they show
type into its single elements, the selective tr
shuffling in the copy. When in 10 I7 the
Sacramentary of Henry II (P1. 2d), one see
lower part and the interior of the Tomb of t
a series of four openings with four window
gallery of the Rotunda; four windows of th
up. Then the illustration shifts from the i
these clerestory windows shows the roof of t
outside with nine windows, the dome and
representation moves back to the interior
edifice, and above the dome of the Rotunda s
flanked by the sleeping soldiers. The disin
single elements and the reshuffling of the
1 See for instance C. Lebruyne, Voyage au
Levant, Paris 1728, II, 242.
2 A. M. Schneider, Byzanz, Vorarbeiten zur
Topographie und Archdologie der Stadt, Istanbuler
Forschungen, Herausgegeben von der Abteilung
Istanbul des Archiologischen Institutes
des Deutschen Reiches, VIII, 1936, p. I ff.,
pls. 1-3.
3 A. Goldschmidt, "Mittelstiicke Fiinfteiliger
Elfenbeintafeln des VI-VII. Jahrhunderts,"
Jahrbuch f. Kunstwissenschaft, I,
1923, p. 30 ff.; D. Frey, Gotik und Renaissance,
Augsburg, 1929, pp. XXIX and 38, while
discussing the problem extends the principles
of mediaeval representations of architecture
into the Gothic period which to this writer
seems to be dominated by entirely different
rules. On this one point I agree as fully with
the brief remarks in the pamphlet of H.
Rosenau, Design and Mediaeval Architecture,
London, 1934, "Planning and architectural
Design," p. I2 ff., as I disagree with the
author's confused interpretation of earlier
mediaeval representations of architecture.
4Munich, Staatsbibliothek, Clm. 4456,
cim. 6o, see Brooks, op. cit., fig. 15This
content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 15
the procedure which prevails in the relation of copy and proto
architecture. Most of the elements of the prototype are prese
have been entirely re-grouped.
Representations of such completeness, however, are rare. Mo
tions are limited to a few elements essential for identifying the A
wooden casket of the Sancta Sanctorum, for example, gives th
simple tegurium and floating above it the clerestory and the
Rotunda (Pl. 2e).1 The same holds true, mutatis mutandis, for t
representations on ampullae, on ivories and in manuscripts.2
tomb, the circular shape of the whole and the uncommon const
vault were prominent characteristics and therefore sufficient
the Anastasis from any other structure.
Indeed it is these same few conspicuous features which also seem
able for identifying an actual architectural copy of the An
recognizable it has to be 'round' and it has either to contain a
of the tomb or to be dedicated to it. These essential outstanding
be elaborated by adding to them other features such as the am
chapels, the gallery, the clerestory, the vault, a certain numbe
and some measurements. These also are typical features of the
therefore may be carried over into the copies. The model conta
a repertory of uncommon elements from which very few hav
whereas others may or may not be selected.
These considerations also give an answer to the question wi
started. Apparently mediaeval writers felt perfectly justified
buildings with one another as long as some of the outstan
seemed to be comparable. The church at Petershausen, for exa
its entirely different plan, had some features in common with
in Rome; it was turned towards the west and the convent of w
part bore the name of Saint Peter.3 Germigny-des-Pres, in sp
ferences, shared with Aix-la-Chapelle the central plan, arrang
dominating central 'tower' and the dedication to the Mother
Lastly Sta. Sophia at Beneventum had nothing in common with
in Constantinople but the dedication.5
But to mediaeval men the dedication of an edifice was one of its outstanding
characteristics. Of course, the dedication was sometimes acc
panied by a more tangible feature, for example by a reproduction of
Tomb of the Lord if the church was dedicated to the Holy Sepulchre. Su
reproductions were quite frequent. It is sufficient to mention those at S
Michael at Fulda, at Constance,6 at Neuvy-St.-Sepulchre ;7 others are
1 Parker Lesley, "An Echo of Early
Christianity," Art Quarterly, II, 1939, p. 215
ff.
2E. B. Smith, "A Source of mediaeval
Style in France," Art Studies, II, 1924, p. 58
ff.; C. R. Morey, "The painted panel from
the Sancta Sanctorum," Festschrift Paul
Clemen, 1926, p. 151 ff.
3 See above, p. 14, n. 3-
4 P. Clemen, Romanische Monumentalmalerei
in den Rheinlanden, Diisseldorf, 1916, p. 55,
n. III.
5 See above, p. 2.
6 See above, p. 6, n. 5-
7 Hubert, op. cit., see above,
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
16 RICHARD KRAUTHEIMER
preserved at Aquileia' and at Bol
Lanfranco, at Pavia, founded in 1
contained a copy of the Holy Sepulc
et altitudinem" ;2 and the church o
reconsecrated in og99 to the Holy S
the Holy City, and a similitudo of t
At times, however, the name alon
features: the church at Sto. Sepol
nothing in common with the An
sometimes supplemented by the exi
a similitudo, a forma of the vener
sufficient stimulus to arouse all th
nected with the prototype. Someti
an edifice may have formed the ter
the fourth quarter of the I Ith cent
Paul on the Vyvehrad at Prague, he
of stones "ad modum quondam Con
church "ad similitudinem ecclesie Romanae s. Petri .. ."5 The common
element between a church which shared with its prototype only the
or the particular manner of its dedication and an architectural copy pr
was evidently the fact that both were mementoes of a venerated site
difference is rather between a more or less elaborate reproduction; an
might say that the more elaborate one only adds some visual elements t
'immaterial' features, that is to the name and dedication. Both immateri
visual elements are intended to be an echo of the original capable of remin
the faithful of the venerated site, of evoking his devotion and of giving h
share at least in the reflections of the blessings which he would have en
if he had been able to visit the Holy Site in reality.6 When in 1076 at
Hubert an oratory was consecrated under the name of "Jerusalem" it
made quite clear that the chapel was meant to reproduce the Sepulchr
the Lord, and thus to represent it for the devotion of the faithful: "ded
1La Basilica di Aquileia, a cura del Comitato
del ... IX centenario, Bologna, 1933, p. 55 f.
2 A. Kingsley Porter, Lombard Architecture,
New Haven, 1917, III, p. I8o, n. 4. "Ecclesia
S. Sepulchri in qua est forma Sepulchri
Dominici secundum longitudinem, latitudinem
et altitudinem . . .;" Porter's translation,
loc. cit., text, "The church of S. Sepolcro
has the same length, width and height as the
sepulchre of our Lord" is erroneous; see also,
loc. cit., n. 4: "Ecclesia S. Sepulchri ubi est
similitudo et forma Sepulchri Domini."
3 Porter, op. cit., II, p. 648. Neither the
documents nor the plan of the edifice suggest
in any way that bishop Anselm's "rebuilding"
of I099 was intended to imitate the pattern
of the Anastasis.
4J. Supino, L' Arte nelle chiese di Bologna
secoli VIII-XIV, Bologna, 1932, P. 43 f. and
p. 104, n. 29, gives a list of buildings which
according to him shared only the name with
the Holy Site of Jerusalem; see J. Hubert,
"Notes sur 1' histoire de 1' abbaye de Ferrieres,"
Annales de la Societi hist. et archeol. du
Gdtinais, 42, 1934, P- 95 ff., where it is shown
that as early as the 9th century the 7th
century convent of Ferrieres was called
"Bethlehem." Whether the 'copy' was
merely a 'copy in name' remains to be carefully
checked in every single case.
5Chronicon Bohemicorum, auctore anonymo,
(before I380), Lehmann-Brockhaus, op. cit.,
no. 1147-
6 The indulgences to be acquired, however,
were rarely as extensive as those connected
with the original; cf. Dalman, op. cit.,
p. 23.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 17
oratorium unum quod dicitur ad sanctam Jerusalem, eo quod ad modum
dominici sepulchri conditum, ipsam quoque eius formam repraesentet
devotioni fidelium."l With this in mind, it is easy to understand why these
copies of the Holy Sepulchre are sometimes situated in cemeteries, such as
St. Michael in Fulda or the Holy Sepulchre at Cambray; obviously a copy
of the Church of the Resurrection of the Lord was particularly fit to hold out
to the visitor the hope of his own future resurrection. The architect of a
mediaeval copy did not intend to imitate the prototype as it looked in reality;
he intended to reproduce it typice and figuraliter, as a memento of a venerated
site and simultaneously as a symbol of promised salvation.
Indeed "typice" and "figuraliter" are the terms used by a 12th century
chronicler when describing the structures at Sto. Stefano at Bologna as a
reflection of Jerusalem.2 There a whole complex of churches was laid out
with the particular aim of reminding the pilgrim of a number of venerated
sites in the Holy Land, and throughout the centuries, despite many changes
of dedication and pattern, the memento character of these edifices has
remained clearly preserved.3
The present structures (P1. 3a) date mainly from the I Ith and 12th
centuries, but their foundation certainly goes back to a much earlier period.
At present two churches, SS. Peter and Paul to the north, and to the south
the Crocefisso flank an irregular dodecagonal church, consecrated to Sto.
Stefano; east of these churches is a courtyard surrounded by arcaded porticoes
-the so-called atrium of Pilate-which extends towards a fourth edifice,
the church of the Trinity, a shallow structure ending in a series of small
chapels.
To-day only the 12th century polygon of Sto. Stefano (P1. 3b) with its
twelve supports and its galleries above the ambulatory, points clearly towards
Jerusalem. Its shape as well as the I4th century Tomb of Christ in the centre
1 Chronicon s. Huberti Andaginensis, cap. 23
(early I2th century), Lehmann-Brockhaus,
op. cit., no. 1776.
2 The documents and the descriptions of
the buildings have been amply discussed by
Porter, op. cit., II, p. I24 ff. The only major
point on which I find myself in disagreement
with Porter is the date of the church of the
Crocefisso; quite apart from its Romanesque
capitals, it contains, in my opinion, large
parts of a 12th century structure in its south
flank. A more recent but also more arbitrary
discussion of the building complex, accompanied
by excellent illustrations, will be found
in Supino, op. cit., p. 26 ff. Supino's main
thesis is that, when founded in the 5th(?)
century, the complex of Sto. Stefano shared
with the Holy Sepulchre only the name
Jerusalem. Both the 5th and 8th century
buildings, which according to Supino preceded
the I2th century Rotunda at Bologna,
would have been baptisteries belonging to the
adjoining "cathedral" of SS. Vitale and
Agricola, now SS. Pietro e Paolo; the whole
present lay-out of the structures would be
12th century and only this late lay-out would
imitate Jerusalem. This whole thesis is based
on two erroneous assumptions; first that the
original lay-out of the buildings of the Holy
Sepulchre at Jerusalem corresponded to
Heisenberg's reconstruction rather than to the
one of Vincent-Abel which, in its general
features at least, is the only one possible; and
second, that SS. Pietro e Paolo was the
"cathedral" (and thus could have a baptistery)
an assumption which has been disproved
by Testi-Rasponi, see below, p. 18, n. I.
3Since the structures were completely
altered by a thoroughgoing restoration some
50 years ago, it will be preferable to base the
description on older plans, such as the one in
F. Osten, Die Bauwerke in der Lombardei,
Darmstadt 1846-54, and the I6th century
plans, published by Supino, op. cit.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
18 RICHARD KRAUTHEIMER
of it leave no doubt as to its signif
a copy of the tomb is mentioned i
name of the Holy Sepulchre as wel
now the church of the Trinity, w
Holy Cross; one or perhaps two cr
the Atrium of Pilate was then wit
present Crocefisso, was dedicated f
like the chapel which at Jerusalem
stasis. The I2th century dedication
but to the buildings around the A
The group itself was called "Jeru
deed as early as 887 and again in 973
remains found on the site prove th
the present 12th century building
cross-shaped, have been excavat
terminated the Hall of Calvary befo
Whatever may be the exact date a
designed, it is obvious that it reprod
Sepulchre in Jerusalem between 62
remodelled, and the beginning of t
large choir with its ambulatory a
Anastasis by the Crusaders.4 Befor
west, the chapel of Golgotha to the
of the Holy Cross to the south, we
of the lay-out, with porticoes on i
(P1. 2c). It is this pattern which
buildings at Bologna, and it is not
1 The present names are relatively modern
Apart from Sto. Stefano with its reproduct
of the Holy Sepulchre, only the names of t
buildings, the Atrium of Pilate and the Cro
fisso have now some connection with either
Jerusalem or the Passion of Christ. Yet these
present dedications are only residuals from a
large and promiscuous array of names of
devotional stations which in the i6th and 17th
centuries referred to all kinds of venerated
sites scattered all over the Holy City and
indeed over the Holy Land. Within the
complex of buildings, the Valley of Josaphat
(the present Atrium of Pilate), the place of
the denial of St. Peter (S. Pietro in Gallicantu)
and a chapel of the Annunciation were
represented; a Scala Santa and a window,
called the Ecce Homo, were shown inside the
present church of the Crocefisso. On the
other hand the I2th century titles were quite
simple. There were only a few and they all
referred to the buildings of and around the
Holy Sepulchre at Jerusalem. The church of
SS. Peter and Paul did not form part of the
convent of Sto. Stefano before 1200 (cf.
Testi-Rasponi, "Note Marginali al Liber
Pontificalis di Ravenna," Atti e memorie della
R. Deputazione di Storia Patria per le provincie di
Romagna, 191 I, p. 391 ff.). This explains why
it is the only building within the group whose
name had no connection with the sites near
the Holy Sepulchre. See Porter, op. cit.,
particularly p. 129, n. 21; p. 136, n. 51;
p. 138 f., n. 59; P- 141 ff., n. 61 and 63.
2 Porter, op. cit., p. I34 f., n. 47 and 49,
"Sanctum Stephanum quod dicitur Hierusalem";
"ecclesia sancti Stephani q.v. Jeru-
salem."
3 Porter, op. cit., p. 130, n. 29.
4 This addition was made between I Io5(?)
and I 149 when a consecration took place; but
building went on until I169. As a whole the
architectural history of the Crusaders' choir
is far from clear; see C. Enlart, Les monuments
des Croises... II, Paris, 1928, p. 136 ff., and
Vincent-Abel, op. cit.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 19
based on the actual lay-out of Jerusalem prior to the first Crusade or w
it depended on plans or maps which reproduced this earlier state.
Of course the lay-out of the complex at Sto. Stefano departs fro
original like any copy of the Middle Ages. Only a few prominent p
selected: the Rotunda, the courtyard and the hall opposite the Rotu
this latter the sites of the Calvary and of the Invention of the Cross-wh
Jerusalem are separate sanctuaries-seem to have been merged. In
to these elements a measurement seems to have been transferred: the distance
which at Jerusalem separates the Tomb of Christ from Mount Calvary
corresponds approximately at Bologna to that between the copy of the Tomb
in the Rotunda and the centre of the cross-shaped main chapel at the end of
Golgotha hall.1 These few selected elements enabled the pilgrim to visit the
Holy places in effigy and in the very sequence which they have in the prototype.
He could come and venerate here the Tomb of Christ, there His Cross
or the site where the Cross was found. The emphasis is on the commemorative
character of the copy.
This "Jerusalem" at Bologna also seems to illustrate the practice which
was followed in laying out a 'copy.' Obviously the builder of a 'reproduction'
of a Holy Site would try to get the needed data about the original either by
travelling himself or by sending correspondents to the site; or he would rely
on plans, and study descriptions of the prototype. As a matter of fact, plans,
such as the one by Arculph which was copied throughout the Middle Ages,
were evidently of considerable importance, and it is most likely that it was this
or a similar plan which formed the basis from which the general arrangement
of the edifices at Bologna was taken. The wide use made of such plans becomes
evident time and again throughout the Middle Ages. Bede, for instance, who
had never been in the Holy Land, when describing Jerusalem must have
drawn upon a map of the Holy City. He speaks of right and left, above and
below, evidently with a map on his desk.2
It is significant that these plans and descriptions stress the very points
which prevail in actual architectural copies. Arculph, when describing the
Holy Sepulchre mentions first the general lay-out of the Anastasis, its roundness
(its "mira rotunditas"), the three walls which enclose the centre-room
and the ambulatory, the three altars in the inner aisle, the twelve columns
"mirae magnitudinis" and the four doors on either side. He strongly emphasizes
and carefully describes the Tomb of the Lord in all its details. In the
plan which accompanies the description (P1. 2c) the tomb is so oversized that
it almost fills the entire centre-room. All the contiguous buildings are only
sketched in, in his description as well as in his plan. He emphasizes what was
important from the pilgrim's point of view, the Tomb of the Lord. In his
conclusion he makes it quite clear that he is able to give merely a feeble
1 The Commemoratorium de Casis Dei, ToblerMolinier,
op. cit., p. 305, gives the distance at
Jerusalem as 28 dexteri =41.58 m. In reality
the measurement is 4In m.; the distance at
Bologna measures 42 m.
2 Bedae Liber de locis sanctis, I-V, Geyer, op.
cit., p. 301 ff., for instance, p. 306: "In
inferiore ... parte urbis, ubi templum ...";
p. 309: "In hac (sc. valle Josaphat) turris est
regis Josaphat ... cuius ad dexteram de rupe
. . excisa ... domus."
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
20 RICHARD KRAUTHEIMER
reflection of what he has seen: "H
iuxta exemplar quod mihi, ut superi
figuravit cerata, dipinximus; non
pictura, sed ut dominicum monum
tate rotundae ecclesiae constitu
ecclesia vel quae eminus posita d
The difference between such
architecture is obvious. From E
Middle Ages descriptions, depictio
a vilis figuratio, limited to a selec
selection is determined by and th
archic order of their religious im
gradually after the beginning o
association with the analytical
apparent) copies, depictions and d
giving a reproduction of the orig
century on this process becomes
may be changed in a copy or th
details, the relation between the c
portion remains essentially unalte
process of draining the edifice of
means a continuous development
movements, but it grows stronge
early 20oth centuries. Architectural
original significance, a Greek T
Customs House, now Subtreasury)
(New York, Woolworth Building),
York, Main Concourse of Pennsyl
its exactness in reproducing the w
absolute faithfulness, definitely o
the Middle Ages: the content and
2. Baptisteries and Mausol
The previous discussions have led
copies as such. They reveal by im
evidently considered essential to a
these characteristic features were different from what a modern beholder
would consider of fundamental importance. Foremost among these elements i
the principle that any mediaeval structure was meant to convey a meaning
which transcends the visual pattern of the structure. This is so obvious and
it has been so often analysed that no further discussion seems warranted. Nor
1 Arculph, op. cit., Lib. I, cap. II, p. 227 ff.;
especially p. 230.
2 In reviewing Lehmann-Brockhaus' book,
W. Haftmann, Zeitschrift far Kunstgeschichte,
VIII, I939, p. 285 ff., particularly p. 289,
has made some interesting remarks about
copies which lead towards similar conclusions.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 21
does it seem necessary to elaborate the importance of
in mediaeval architecture.
Although the existence of such architectural series is well established,
other questions remain to be answered. Time and again it becomes evident
that certain architectural patterns are related to specific dedications. Churches
dedicated to the Holy Cross are frequently cross-shaped; sanctuaries of St.
Michael are situated on heights, in towers or on hills, and they are frequently
of circular shape; templar churches are round and their roofs are often supported
by six piers; churches of the Virgin also are frequently centrally
planned.' In every case the pattern of the structure is linked to the commemoration
of a particular Saint or of a specific object; or it is related to the
use of the edifice by a specific group such as a religious order. Some kind of
affinity seems to connect architectural patterns and their 'content.' Once
established, the different patterns continue to follow traditional lines. Yet the
question remains as to how these patterns were first associated with the particular
content with which they are subsequently identified. In other words,
in what way and for what reasons did these iconographical types originate?
It is proposed here to investigate one question only: Why are baptisteries
round? The reason which has generally been given for their circular plan is
their alleged derivation from round, vaulted rooms of Roman Baths. These
rooms seemed to offer a clear prototype; they appeared to have the same
shape and to be laid out for a similar use. Despite the difference between an
act of simple cleanliness and a ritual immersion, the similarities have always
been considered strong enough to warrant such a derivation of the baptisteries
from Roman Baths.
But the question arises whether this explanation is actually sufficient for
explaining the origin and the survival of the central type in baptisteries, or
whether additional prototypes could have exerted any influence. Of course
the connection between baptisteries and thermal rooms is undeniable although
it is hardly as plain as has been sometimes assumed. Round rooms do occur
frequently in Roman Thermae, cold water rooms, frigidaria, such as at Badenweiler
and in the Stabian Baths at Pompei, and hot water rooms, caldaria,
such as those in the Thermae of Caracalla and Constantine in Rome.2 Circular
caldaria or frigidaria are, however, relatively rare; the majority of the round
rooms in thermal establishments are either apodyteria, wardrobes and cloak
rooms, or laconica, steam baths.3 As such they neither contained water basins
nor were they used for ablutions of any kind.
The problem is made even more complicated by the fact that the earliest
1 Compare the lists of dedications which
are given in: H. Otte, op. cit.; J. L. Petit,
"Notes on circular churches," Archaiological
Journal, XVIII, 1861, p. oI ff.; H. Bogner,
Die Grundrissdispositionen der zweischifigen Zentralbauten,
Strassburg, I9o6; F. Bond, The
Consecration ... of Churches, London, I9I14. No
far-reaching conclusions have been drawn
from these lists.
2 The most recent and complete collection
of ground plans of Roman Thermae is found
in D. Krenker and others, Die Trierer Kaiserthermen,
"Trierer Grabungen und Forschungen,"
I, I, Augsburg, 1929-
3 See for instance the laconica at El-Djem,
Khamissa, Lambaesis and in the two Thermae
of the Villa of Hadrian at Tivoli, and
the apodyteria at Marienfels and VieilEvreux;
illustrations in Krenker, op. cit.,
passim.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
22 RICHARD KRAUTHEIMER
baptisteries of the 3rd and 4th cen
were square or rectangular with
piscina itself was sometimes round
at Dura-Europos about 231 A.D.,1
early 4th century2 (P1. 4a) and th
Nesactium and Salona.4 The sam
structure underneath the Baptiste
the middle of the 5th century bap
square rooms,6 and this type seems
and throughout the Near East up t
to say that the kinship of these
Thermae rooms is much more evident than that of their later round successors.
Like the early baptisteries the Roman frigidarium was often a square or rectangular
room, sometimes with an apse; a basin for cold water occupied either
the centre of the room or stood at one end, an arrangement which seems to
be repeated literally in the Baptistery of Dura.
It is only from the middle of the 4th century onwards, that this rectangular
type seems to be gradually superseded by baptisteries of circular or octagonal
shape. While these differ widely in their particular pattern, they are all of one
of these two forms or combine a circular shape with eight supports. The second
Baptistery of the Lateran, which replaced the first rectangular one, about
350 A.D., was a round building with eight engaged columns along its interior
walls (P1. 4a).10 Similarly at Salona, between 404 and 420, a polygonal
structure with seven columns close to the wall replaced its rectangular 4th
century predecessor.11 Occasionally, for instance at S. Giovanni in fonte at
Naples (about 400) and possibly in the Baptistery of Giil-bagtsche near
Izmir, the room is square but covered with an octagonal vault supported by
squinches.
The most frequent type, however, is that with four niches arranged in the
corners of an octagon, the lower part of which is sometimes enclosed within
a square. This type is found all over the Christian world from the 5th century
and throughout the early Middle Ages: in Syria (Kalat Siman),12 in Egypt
(Menas Sanctuary),13 in Constantinople (Baptistery of the Hagia Sophia),14
1 C. Hopkins, The Christian Church at DuraEuropos,
Preliminary Report of Fifth Season, New
Haven 1934, pp. 249 if.
2G. Giovenale, Il Battistero Lateranense
(Studi di Antichita cristiana I), Rome, 1929.
3 La basilica di Aquileia, op. cit., p. IO9 ff.,
fig. Io; p. 165 ff., fig. I8; p. 280.
4 R. Egger, Friihchristliche Kirchenbauten im
siidlichen Norikum, Wien, 1916, p. I17, fig. 105.
5 C. de la Croix, "Poitiers," Congris arche'ol.,
70, I903, P- 7 ff.
6 S. Crisogono, second phase, see R.
Krautheimer, Corpus Basilicarum Christianarum
Romae, I, Vatican City, 1937 ff., p. 152; L.
Fortunati, Relazione degli scavi ... lungo la Via
Latina, Rome, I859SS.
Gsell, Les monuments antiques de l'Alge'rie,
Paris, I900-01, II, p. 152 f.-
8 G. Soteriu, A' haoX?ppz v~ flartmxaZq 'EXX&~ao, Athens 1931.
9 H. C. Butler, Early Churches in Syria, 1929,
passim, Athens, 1931; J. W. Crowfoot, Early
Churches in Palestine, London, 1941.
10 Giovenale, op. cit.
11 E. Dyggve, "Salona Christiana," Atti del
Illo Congresso internazionale di Archeologia
Cristiana (Studi di Antichita Cristiana VIII),
Rome, 1934, p. 237 if.
12 Butler, op. cit., p. 156.
13 C. M. Kaufmann, Die Menasstadt, I,
Leipzig, I9I0, figs. 22, 23.
14 E. Swift, Hagia Sophia in Constantinople,
New York, 1940, p. 147 ffThis
content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 23
in Ravenna (Baptistery of the Orthodox; Baptistery of the Arians
the Alpine countries (Riva S. Vitale; Novara; Lomello).2 At t
pattern is slightly more elaborate: in the Baptistery of St. Mary at
at S. Aquilino at Milan and in those of Albenga, Frejus and Mdlas4 r
niches occupy the main axes, semicircular niches the diagonals. Fr
this scheme is enriched by columns in front of the piers which se
niches. At other times the pattern is reduced to a plain polygonal
simple octagonal pattern appears before the middle of the 5th century
maberg5 and some decades later at Grado cathedral; it survives for
in upper Italy in baptisteries such as those at Lenno and Oggione,
the late I Ith century.
Generally speaking it would seem that round or octagonal bapt
were introduced into Christian architecture only after the second h
4th century and that they did not become common until the 5th
Despite the differences in use which were mentioned above, these
patterns may have had some connection with thermal rooms.
plans with or without corner niches are not unusual in Thermae,
regular round types occur much more often; but after all, some ba
were round, such as that of 350 at the Lateran. On the other han
octagonal nor round rooms are in any way limited to thermal arc
indeed vaulted centralized patterns with or without niches an
columns were widespread in antiquity. They occur throughou
architecture, for example as vestibules in palaces and villas such as
Aurea or the Villa of Hadrian at Tivoli.6 They are also frequently
the so-called "Nymphaea" such as the Minerva Medica in Rome.
it has been proved recently that most of them were really diaeta
pavilions, flanked by fountain rooms, but not containing any wate
their plan belongs into the same family as the vestibules and the
rooms. Evidently the round baptisteries form part of a large inte
group of late antique buildings, and while the central rooms o
architecture are among their ancestors, other types may and, ind
likely to have exerted a collateral influence.
This is extremely probable in view of the fact that baptisteries
1 C. Ricci, Tavole storiche dei mosaici di
Ravenna, Rome, 1932; G. Gerola, "I1 restauro
del battistero Ariano di Ravenna," Studien
zur Kunst des Ostens, Vienna, 1923, p. 112 ff.
2 F. Reggiori, Dieci battisteri lombardi minori
("I monumenti Italiani," IV), Rome, 1935.
See also S. Steinmann-Brodtbeck, "Das Baptisterium
von Riva San Vitale," Zeitschriftfiir
schweizerische Archaeologie und Kunstgeschichte,
III, 1941, p. 193 ff.; the issue arrived in
U.S.A. only while I was revising the galley
proof of this paper.
3 Osterreichisches Archiologisches Institut,
Die Marienkirche in Ephesos, ("Forschungen in
Ephesos," IV, i), Vienna 1932, p. 43 ff.
4G. Chierici, "Di alcuni risultati sui
recenti lavori intorno alla basilica di San
Lorenzo a Milano ..." Riv. Arch. Crist., XVI,
1939, P-. 51 ff.; L. Reggiori, op. cit. ;J. Hubert,
L'Art Pre'-Roman, Paris, 1938, p. 2 ff.; G. De
Angelis d' Ossat, "Sugli edifici ottagonali a
cupola nell'Antichit" e nel Medioevo," Atti
del Io Congresso Nazionale di Storia dell' Architettura,
Florence, 1938, p. 13 ff. enumerates a
great number of octagonal baptisteries and
other structures.
5 Egger, op. cit., p. 68.
6 See the list in De Angelis D' Ossat, op. cit.,
p. I7 f.
7 K. Lehmann-Hartleben and J. Lindros,
"II palazzo degli Orti Sallustiani," Opuscula
Archeologica I, 1935, p. 196 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
24 RICHARD KRAUTHEIMER
number of peculiar features which certain
they are surrounded by a low outer ambu
communicates by one or more doors. Suc
in plan, a type which occurs in the 5th c
Ephesus and in those at Giil-bagtsche,
(P1. 4c).1 At other times they are polygonal
at Ravenna and in the 6th century Baptis
Baptistery of Djemila (P1. 4d, e) was encir
with a barrel-vault; its walls are articu
similar ambulatory, interrupted by four
the centre-room, surrounded the dodecag
the ambulatory was covered by a barrel-v
dome. The date of the structure is appar
of these corridors is uncertain, although
they were used for regulating the access
the baptistery and the font.
Interior ambulatories were developed
exterior corridors. Indeed, the Baptistery
may be considered as representing an interm
with outer and inner ambulatory. The bes
Baptistery of the Lateran, the third one o
out in 432-40 and which still forms the nuc
centre-room is supported by eight colum
eight-sided domical vault, surmounting a cle
An octagonal inner ambulatory covered
penetrations surrounds the centre-room,
piscina. A similar interior ambulatory is f
Maggiore at Nocera (5th century), wher
centre-room by a circle of 14 pairs of col
this type occurs at Butrinto at Albania;7 the
ed by two rings of eight columns each, w
unlikely that the building was vaulted
Southern France, for instance at Marseille
niches are arranged in the four corners of t
in Southern Italy four short chapels arran
annular interior ambulatory; its date may
1 Steinmann-Brodtbeck, op. cit., enumerates
a great number of these square ambulatories.
2 D. Frey, "Neue Untersuchungen und
Grabungen in Parenzo," Mitteilungen der K. K.
Zentral-Kommission, 3. Folge, XIII, 1914, pp.
144 ff., '79 ff., especially fig. 31.
3 E. Albertini, "L' archeologie chretienne
en Algerie," Atti del Illo Congresso, op. cit.,
p. 411 f.
4 H. Nachod, "Das Baptisterium von
Canosa," Riimische Mitteilungen XXX, 1915,
p. ii6 ff.
5 Giovenale, op. cit.
8 M. Stettler, "Das Baptisterium zu Nocera
Superiore," Rivista di Archeologia Cristiana,
XVII, 1940, p. 82 ff.
7 L. M. Ugolini, "Il Battistero di Butrinto,"
Rivista di Archeologia Cristiana, XI, 1934,
p. 265 ff-
8 H. Koethe, Friihchristliche Nischen-Rundbauten,
Diss. Marburg, 1928. The date of
Marseilles is certainly 5th century; the
Baptisteries of Riez and Aix may be somewhat
later.
9 P. Laicono, "Sul restauro compiuto al
Battistero di Santa Severina," Boll. d' arte, 28,
1934, P. 174 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
iiiiiidiiliiiiiiiiiiiil
.. . . ......: iiiiiiiiiiiiiii:
iiiiiii-iiiiiiiiiiiiiiii
ia-::iiii--:-i--l iiiiiiiiiii-iiiiiiiii'iiiiiiiiiiii-iiiii i-:
i :::p :::::::.
low:::::
s on-~ii:::::::::-i-?::::-::::::
a-Baptistery, Lateran, Rome
(PP. 22, 24)
b--Baptistery, Lateran, Rome (p. 2c-Bap
d--Baptistery, Djemila, Ambulatory (p. 24)
"- -i~iii:~-i~ii ii ~ii
n
w
u
"" : :
i~i:i i :I~i~i
i ?s~
I
- -~- :,ii
~~Tk --:i _
~---
?~i
1" r
""": d~ ~;g
::::
:s -~--~ , iiil
-:'~::::~~~::"I': ,::I ili~~iiii ~
~iii;:ii~i~i i-::i:~: ,CI~ -~
?~
r k ?~
i ii,, i~W ~WI _~ -
~iljT
~ ~"" -~ii-_
r-
~
~'"'~l~~'~.~;:a,~.~~......., ; . ___~
"
e-Baptistery, Djemila (p. 24) f--Baptistery, Nocera (From J. C. R. de Sai
toresque, 1781)
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
a-Mausoleum, Villa de' Gordiani,
Rome (p. 25)
b-Mausoleum, Villa de' Gordiani,
Rome (p. 25)
c-Mausoleum
(From Barto
d-Sta. Costanza, Rome (p. 26) e-Sta. Costanza, Rome (From S
Rome, fol.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 25
None of the round rooms in thermal establishments shows the combination
of a vaulted centre-room with either an outer or an inner ambulatory.1 As a
matter of fact the outer ambulatory is quite rare in Roman secular architecture.
The outer portico on columns in round temples, which at first glance
might look similar, is really entirely different. Its openness, the lack of a vault
and especially its height which normally equals the height of the cella make
the dissimilarity quite clear. Indeed the combination of a vaulted centre-room
with either an inner or a closed and relatively low, outer ambulatory seems
only to be found in one group of Roman buildings: that is, in sepulchral
architecture. Roman mausolea of the 3rd and 4th centuries use all the
different patterns which occur in the baptisteries, from the simple round or
octagonal plan with or without niches to the complicated forms with inner or
outer ambulatories. Round mausolea with niches in the thickness of the walls
or enclosed in a surrounding ring of masonry are found for instance in the
Tomb in the Villa de' Gordiani (P1. 5a, b), the so-called Tor de' Schiavi, in the
Mausoleum of St. Helena near Rome, in that of the West Roman dynasty
near St. Peter's, in St. George at Salonica and perhaps in the Mausoleum of
Constantine at Constantinople.2 In the monument of the Turcia family the
niches in the four main axes protrude so as to merge circular and cross plans,
with a result similar to that of the Baptistery of Canosa.3 In the Mausoleum
of Diocletian at Spalato columns are arranged in a double order between the
wall niches and though the interior is round, the exterior is octagonal. Another
octagonal tomb with a round inner chamber with eight supports along its
walls was situated near Frascati.4 Similar patterns must have been quite
frequent. Often the tomb chamber is situated below the ground and surrounded
by an outer ambulatory which may be square or semicircular or
round in plan; it communicates with the inner chamber through only one or
two doors. Such ambulatories are found, for instance, in the tomb of the Furia
family near Tivoli, in the Mausoleum of the Vigna Cavalieri (P1 5c), in that
of the Servilii on the Via Appia and in one of the tombs of the Via Latina.5
As late as the 4th century exactly the same type occurs in the necropolis of
1 On plans of Roman Thermae what sometimes
looks like such an outer corridor, is
really nothing but a furnace passage.
2 H. Koethe, "Zum Mausoleum der westr6mischen
Dynastie," Rimische Mitteilungen,
46, 1931, p. 9 ff.; idem, "Das Konstantinsmausoleum
und verwandte Denkmailer,"
Jahrbuch des Archdologischen Instituts, 48, 1934,
p. I85 ff.
3 L. Canina, Gli edifizi di Roma ... antica e
sua Campagna, Rome, 1848-56, VI, 2, pl.
CXXIII; the monument which now goes by
the name of Sta. Maria della Tosse, seems to
have been erected in the middle of the 4th
century.
4 Canina, op. cit., VI, I, pl. LXXXII.
Similar patterns, sometimes of a more complicated
type, are frequent among the Roman
monuments drawn in the I6th century, for
instance, by Bramantino, Le Rovine di Roma
S. . , ed. Mongeri, Turin 1879, or by G. B.
Montano, Le Cinque Libri di Architettura, Rome
1621. Yet these drawings seem to be frequently
"variations on Roman themes"
rather than actual surveys. Thus it seems inadvisable
to depend on them too much.
r B. de Montfaucon, L' Antiquiti Expliqude,
V, I, Paris, 1719, pls. I8, Io8, III, II8. The
history and the significance of these subterranean
tomb corridors in pre-Roman and
Roman times have been discussed by G.
Welter, "Zwei vorr6mische Grabbauten in
Nordafrika," Riimische Mitteilungen, 42, 1927,
p. 84 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
26 RICHARD KRAUTHEIMER
Tarragona.' The octagonal mausoleu
even later; its round inner tomb ch
by an annular corridor.2
This originally subterranean barr
slightly transformed, as late as the
Christian mausolea, Sta. Costanza
from numerous i5th and I6th centu
building was surrounded above grou
on columns. The motive of an encir
by the circular peristyles of round
peristyles were used for mausolea, fo
at Spalato. But the outer portico of
motive with that of the Roman tom
to the subterranean origin of the
ambulatory covered by a barrel-vau
room by twelve pairs of columns m
this 4th century mausoleum unites al
the group of central baptisteries fr
Sta. Costanza was in no way uniqu
sepulchral architecture. Another Ch
tory supported by eight columns,
Obviously there is no way of tellin
'Tombs' with inner or outer ambul
in I6th century drawings,5 were of t
memorial churches closely akin to
device of the inner ambulatory fro
ticularly in the Near East.6 The com
and outer ambulatory characterizes
churches: the Anastasis at Jerusalem
Thus it seems that the baptisterie
these Roman mausolea and partic
features which they do not share
third Baptistery of the Lateran can b
a derivative of the mausoleum of S
glance, the links between the baptis
in content as well as in pattern. Un
symbolically to cleanse the catechu
only aspect: other connotations a
1 E. Junyent, "I monumenti Cristiani d
Spagna," Atti del Illo Congresso, op. cit.,
255 ff.
2 Gsell, op. cit., II, p. 421 ft.
3 C. Cecchelli, S. Agnese fuori le mura e S.
Costanza (Le chiese illustrate di Roma, Io),
Rome, n.d.
4 Gsell, op. cit., II, 410 f.
5 Ligorio, Cod. Vat. lat. 3439 f. 70" Templum
Platonis et Proserpinae;" f. 25 " Templum Isidis et
Serapis." See, however, above p. 25, n. 4.
6 H. Koethe, op. cit., Jahrbuch des Archaeologischen
Instituts, 48 (I934), pp. 85 ff-.,
especially 198 ff.
7 P. Styger, "Nymphiaen, Mausoleen,
Baptisterien," Architectura, I, 1933, p. 50 ff.
has taken a strong stand against the thesis
that baptisteries had any connection with
thermal rooms or with nymphaea. He suggests
the possibility that a great number of
baptisteries were in reality originally
mausolea-for example the Baptisteries at
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 27
Paul's letter to the Romans started Christian thought on one
mental tenets of baptismal mysticism: "An ignoratis quia quicu
ati sumus in Christo Jesu, in morte ipsius baptizati sumus? Con
sumus cum illo per baptismum in mortem: ut quomodo Christ
a mortuis per gloriam Patris, ita et nos in novitate vitae am
mystical equation seems to be established between baptism
resurrection, death meaning the dying of the old Adam and at
a mystical imitation of the death of Christ. This two-fold equ
perfectly clear by St. Basil in his Book on the Spirit, when discuss
letters: "How then are we made in the likeness of His death? In that we were
buried with Him by baptism. What then is the manner of the burial? And
what is the advantage resulting from the imitation? First of all, it is necessary
that the continuity of the old life is cut. And this is impossible unless a man
be born again, according to the Lord's word; for the regeneration, . . .is
a beginning of a second life. So before beginning the second . . . it seemed
necessary for death to come as mediator between the two .... How then do
we achieve the descent into hell? By imitating through baptism the burial of
Christ. For the bodies of the baptized are, as it were, buried in water ....
For there the death on behalf of the world is one, and one the resurrection
of the dead, whereof baptism is a type."2
Obviously this mystical death in baptism holds out the hope of future
resurrection and is at the same time regeneration and resurrection in itself.
It is also a symbol of the resurrection of the Lord through whose death
Christian resurrection in and through baptism becomes possible. The same
idea recurs in other Patristic writers. To St. Augustine baptism is nothing
but a "similitudo" of the death of the Lord and at the same time a death of the
old Adam "since we have been baptized in the death of Christ."3 According
to Hilarius of Poitiers baptism is a sacrament of regeneration and of resurrection;4
it is a symbol of the future resurrection of man as well as of the resurrection
of the Lord, a promise of rebirth in eternity, of eternal life.5 Consequently
the symbolum fidei, professed at the occasion of baptism, strongly
emphasized Christ's resurrection, His return to judge over the quick and the
dead and their entering into eternal life. Leo the Great discusses the same
idea in similar terms; baptism is a mystical imitation of Christ's death, of his
burial and of his resurrection, "ut ... per similitudinem formamque mysterii
Naples, at Ravenna and at Agliate. No proof
is given to support these suggestions. On the
other hand he vehemently objects to the possibility
that the type of the baptistery as such
should be derived from mausolea types. The
resemblance of the niche over the baptismal
font at Dura with Eastern Roman tomb types
has been pointed out by Hopkins, op. cit.,
P. 249.
1 Paulus, Epistola ad Romanos, VI, 3, 4; cf.
also Epistola ad Colossenses, II, 12 and Petrus,
Epistola Prima, III, 21.
2 The book of Saint Basil on the Spirit, cap.
XV, 35, Nicene and Post-Nicene Fathers, VIII,
Oxford and New York, 1895, p. 21 f. Cf.
R. Reitzenstein, "Heilige Handlung," Vortrdge
d. Bibl. Warburg, 1928-29, Leipzig 1930,
p. 21 ff.
3 Augustine, Enchiridion, cap. LII, Migne,
Pat. Lat. XL, c. 256 f.; idem, Contra Julianum
Pelagium, cap. V, 14, Migne, Pat. Lat. XLV,
c. 683.
4 Hilarius of Poitiers, De Trinitate, lib. IX,
cap. 9, Migne, Pat. Lat. X, c. 288(265).
5 Pseudo-Augustine (possibly Faustus),
sermo clxviii, 2, Migne, Pat. Lat., XXXIX,
c. 2070.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
28 RICHARD KRAUTHEIMER
ea quae geruntur in membris his quae
dum in baptismatis regula et mors int
turam triduanam imitatur trina dem
instar est de sepulcro."l As late as the
repeats literally St. Paul's classical for
cuiusdam mortis et sepulturae."'2 From
the Middle Ages, Easter, the day of the
day when baptism could be lawfully
was Pentecost, the day when the Holy
the first time the people were baptiz
Christ's baptism (and originally of His
Leo the Great.3
Evidently baptism and resurrection and therefore symbolical death and
burial were closely linked in the minds of Early Christians. Thus it is not
surprising to find baptisteries and sepulchral architecture related to one
another. Occasionally baptismal fonts were placed in catacombs; the
existence of a baptistery in the catacombs of Sta. Priscilla is suggested by
remains of walls and conduit pipes and on the testimony of a possibly related
late 4th century inscription.4 Although the connection of these remains with
baptism has been contested,5 a baptistery certainly existed in the catacombs
of S. Pontianus. It was evidently laid out in the 5th or 6th century inside an
older tomb chamber; the piscina and the steps leading down to it are preserved,
as well as a 6th century fresco which represents the baptism of Christ
and the stag drinking from the fountain of life.6 A baptistery connected with
a cemetery basilica was erected possibly as early as the 4th century in a
necropolis near Tarragona.7
Occasionally tombs were placed in the baptistery proper. In one of the
hymns of Prudentius, a baptistery is mentioned which commemorated the
death of two martyrs who were either buried or slain on the site.8 Tombs are
known to have existed in the Baptistery of the Arians at Ravenna and at least
one of them was contemporary with the structure.9 The prohibition of burials
in baptisteries issued in 578 by the Council of Auxerre only proves the
existence of the habit, reveals at least that burials in baptisteries were not
uncommon.'0 Still another link in this chain should be mentioned. It has
been pointed out that the pictorial decorations of catacombs time and again
depict baptism as a symbol of resurrection; on the other hand the mosaic
decorations of 5th century baptisteries allude frequently to death and
1 Leo Magnus, Epist. xvi, cap. 3, 3, Migne,
Pat. Lat., LIV, c. 698 (719).
2 Anselm of Canterbury, De azymo et fermentato,
cap. IV, Migne, Pat. Lat., CLVIII,
c. 544 (136).
3 Leo Magnus, ibid., c. 696.
40. Marucchi, "La basilica papale del
Cimiterio di S. Priscilla," Nuovo Bull. di Arch.
Crist., XIV, 1908, p. 5 ff., especially p. 48 ff.
5 G. P. Kirsch, Le Catacombe Romane, Rome,
1933, P-. 93 ff.
6 Kirsch, op. cit., p. 229.
7Junyent, op. cit., Atti del Illo Congresso,
p. 283 ff-
8 Prudentius, Peristephanion, VIII, v. I ff.,
Migne, Pat. Lat., LX, c. 430 ff.
9 Gerola, op. cit., passim.
10 F. W. Unger, "Uber die christlichen
Rund- und Octogon-Bauten," Bonner Jahrbiicher
41, 1866, p. 52, n. 2. As late as the 9th
century, the Baptistery of Santa Severina was
crowded with tombs; see above, p. 24, n. 9.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 29
resurrection, for example at S. Giovanni in fonte at Napl
Baptistery of the Orthodox at Ravenna.
These links may help to support the thesis that the central
baptisteries as they appear from the late 4th century onward
one of their roots, and quite an important one, in sepulchral a
must have seemed perfectly natural to any Early Christian beli
pattern of a mausoleum for an edifice in which his old sinful Adam
and where he was to be buried with Christ so that he might
with Him. In the mausolea he would find a type similar enoug
rooms to be merged with their pattern and thus to carry over
cleansing from the thermae into the round baptisteries; on the ot
mausoleum type would transfer to the baptistery all the implicati
and resurrection which Early Christianity connected with bap
Roman mausolea would contain an element which in connection with a
sepulchral monument was bound to hint specifically at resurrectio
octagonal pattern which was in itself a symbol of resurrection and rege
tion. An inscription attributed to St. Ambrose which decorated the Bapti
of Milan cathedral and which has been conserved in the Sylloge Laures
mensis III carefully elaborated this symbolism. In eight distichs it is po
out that the edifice was octagonal and that its shape and that of the octa
piscina corresponded to the significance of the number; for eight is the num
of salvation and regeneration of the death of the old Adam and of the b
ning of new life. As Doelger2 has pointed out, these verses conform comple
to the symbolism of numbers in the writings of St. Ambrose and of other
Christian authors. Time and again they emphasize the character of bap
as a spiritual regeneration which is symbolized by the number eight. Bap
is a "creation from the womb of the water" a rebirth into the "spir
octave."3 This explains the predilection for octagonal patterns which pr
throughout Early Christian baptisteries.4
These patterns which connect the baptistery with the mausoleum and
with the idea of resurrection continued throughout the Middle Ages. Examp
are numerous, particularly in Northern Italy and they occur even Nor
the Alps, although there baptisteries were rare after the I2th century
circular shape with eight engaged columns (Agrate Conturbia), the struc
1 F. J. Doelger, op. cit., Antike und Christentum,
IV, 1934, P. I53 ff.; idem, "Die Inschrift
im Baptisterium S. Giovanni in fonte .. ."
Antike und Christentum, II, 1932, p. 252 ff.
While Doelger strongly emphasizes the symbolical
link between baptism and resurrection
he derives the plan of the baptistery from
thermal rooms only; see also Sauer, op. cit.,
p. 78.
2 Doelger, op. cit., Antike und Christentum, IV,
1934, P. I53 ff.
3 Clemens of Alexandria, Stromata IV, 25,
I6o, see Doelger, op. cit., 1934, P. 179 f.
4 Doelger, op. cit., 1934, p. 182 ff. On p.
187 Doelger points out that sometimes a
hexagonal piscina is arranged inside an
octagonal baptistery; he interprets the hexagon
as symbolizing Mother Church. The
pattern seems to be particularly frequent
along the Dalmatian Coast and in North
Africa. I should be inclined to interpret it
rather as symbolizing the 'Old Adam.'
Adam, since he was created on the sixth day,
is represented by the number six; see Hrabanus
Maurus, De Universo, Lib. II, cap. I,
Migne, Pat. Lat. cxi, cap. 31, and A. Goldschmidt,
"Friihmittelalterliche illustrierte Enzykloptidien,"
Bibliothek Warburg, Vortrdge
1923-1924, Leipzig-Berlin 1926, p. 218 f.
3
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
30 RICHARD KRAUTHEIMER
with octagonal inside and round outsid
the octagon with alternating semicirc
Giovanni in Atrio), the plain octagon
columns and with octagonal ambulato
these do not depart essentially from
octagonal shape or the design of eight
regeneration and resurrection remain p
holds good, of course, for shrines which
necessarily containing a baptismal fon
in cemeteries, thus emphasizing the
burial and resurrection. As early as
cemetery of the convent ofAniane was d
in the i th century a chapel "in hon
aliorum sanctorum .. ." was erected in th
somewhat earlier, a new baptistery in fr
tombs was built at Aquileia. Shortly
erected in the early mediaeval cemetery
on the site of an older baptistery. Also t
was constructed on an early mediaeval ne
the present I I th century structure had
shape as whether the site around the c
Middle Ages. Yet at least three person
inside the baptistery: Bishop Raineriu
the main part of the structure, Bishop
John XXIII (d. 1419).4
The question arises whether mediaeva
early patterns whose origin can be link
At times it seems they went further
Anastasis in Jerusalem, where Christ h
prototype of resurrection and symboli
such a connection, not for baptisteries pr
ings at Constantinople, of which at lea
Unfortunately the identification of th
reconstruction from older descriptions
influence of the Anastasis on baptisteries
1 Porter, op. cit., passim; Reggiori, op. cit.
2 Schlosser, Karol. Kunst, no. 578.
3 Casus monasterii Petrishusensis, lib. II, cap.
x6, Lehmann-Brockhaus, op. cit., no. I095.
4 The tomb of Guccio de' Medici which is
now inside the Baptistery was transferred
there from the Piazza. The burials inside the
Baptistery of Florence prove, by the way,
that the prohibition of the Council of Auxerre
was not of much effect.
5 Unger, op. cit., 25 ff. has strongly emphasized
the possibility of such a link between
the mediaeval baptisteries and the Anastasis.
In his opinion all baptisteries were laid out
on an octagonal plan from the 7th century
onwards, because they all were derived from
the Holy Sepulchre; this, in his reconstruction,
combined from 614 on an octagonal
exterior with a round interior. Unger's reconstruction
is, of course, mistaken, and consequently
also his conclusions are erroneous.
Yet his fundamental assumption seems to be
quite correct; he is wrong in his reasoning
but right in the ultimate reasons for his
reasoning.
6 G. Millet, "L'6glise ronde de Preslav,"
Comptes Rendus de l'Acadimie des Inscriptions te
Belles Lettres, 1933, p. 169 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 31
more natural since as early as the end of the 4th century the Anast
an important r6le in the baptismal rites at Easter in Jerusalem
describes the rites in detail : the catechumens had been catechized for seven
weeks in the basilica and had been instructed in the literal meaning of the
Scriptures. On Palm Sunday they professed the Creed and thus became
Neophytes. Then during Easter week they were led every day into the
Anastasis to hear "the teachings of the deeper mystery, that is of Baptism
itself. . . . There the bishop stands, leaning against the inner rails which are
in the cave ... and explains all things that are done in Baptism. In that hour
no catechumen approaches the Anastasis, but only the neophytes and the
faithful, who wish to hear concerning the mysteries, enter there, and the doors
are shut lest any catechumen should draw near .... And truly the mysteries
are so unfolded that there is no one unmoved at the things that he hears to
be so explained." Nothing could stress more strongly the link between
Resurrection and Baptism than this scene: the bishop's voice coming out of
the Tomb whence the Lord had risen and explaining to the neophytes the
mystical meaning of Baptism, the mystical death and the spiritual resurrection
which they were to undergo during the last hours of the week. At Jerusalem
at least, the significance of the Anastasis within the baptismal ceremonies is
quite evident. By the later Middle Ages this connection of the Anastasis with
Baptism had evidently become so close that at least in popular usage St. John
the Baptist was sometimes linked as co-patron to churches dedicated to the
Holy Sepulchre; a 13th century chronicler reports that after the first Crusade
a church had been built at Huy "in hon. s. sepulchri Domini et b. Johannis
Baptiste ob venerationem et recordationem ecclesie lerosolimitane, que
ecclesia in hon. predictorum patronorum dicitur esse fundata."2 The same
combination occurs about the middle of the 12th century in the dedication of
S. Giovanni del Sepolcro in Brindisi. In the same way the i ith century
church at the Krukenburg, copied from the Holy Sepulchre at Paderborn,
was dedicated to the Baptist.3 In each case the combined name may be influenced
by the importance of the Order of St. John of Jerusalem to which
was entrusted the care of the pilgrims who came to visit the Holy Sepulchre;
indeed, the churches of this order were traditionally round and 'copied' from
the Anastasis.
All these examples are somewhat tenuous proof for the connection between
baptisteries and the Anastasis in Jerusalem. Yet there exists at least one
mediaeval baptistery which is an actual copy after the Rotunda of the
Anastasis: the Baptistery of Pisa (P1. 3d). The structure begun in 1534 is
characterized by a number of elements which clearly are 'copied' from the
Anastasis. It is circular in shape and its centre-room is surrounded by an
ambulatory and by an upper gallery. Four cross-shaped piers alternate with
1 S. Silviae . . . peregrinatio ad loca sancta,
Geyer, op. cit., p. 35 ff., particularly p. 98 f.
We are quoting from the English translation
by M. L. McClure and C. L. Feltoe, The
Pilgrimage ofEtheria (Translations of Christian
Literature, ser. III, Liturgical Texts), London
and New York, n.d., p. 90go f., particularly
p. 94.
2 Chronica Albrici monachi Trium Fontium,
M.G.H., SS. XXIII, p. 815, Lehmann-Brockhaus,
op. cit., no. I787.
3 See above, p. 6.
4 Rohault de Fleury, op. cit., p. 56 ff., pls.
XVIII-XXI; Salmi, op. cit.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
32 RICHARD KRAUTHEIMER
eight columns so that twelve supports carry
on the upper floor, four stronger alternating
exactly the rhythm of the lower arcade. The
room, a steep conical roof, originally truncated
served beneath the I4th century dome. As
pointed out, this roof in itself clearly proves t
ing the Rotunda at Jerusalem; so does the
groin-vaulted ambulatory and gallery, and its
departure from the number eight which had
and the replacement by twelve supports poin
new prototype. Even the alternation of piers
repeated though it is accomplished by chang
piers and three columns for each quarter cir
pier and two columns. As in any mediaeval c
up into its single elements; a selection of them
parts have been re-arranged, possibly under t
structures. The rhythm of supports at Pisa mig
more refined pattern of pilasters and colum
Baptistery of Florence.
There cannot be any doubt that the Baptister
copy of the Anastasis at Jerusalem. The questio
of the Anastasis in a baptistery is an isola
impossible to give a definite answer to this que
element to be found in a great number of I I th
which looks suspiciously as though it had bee
from the I I th century onwards galleries make
starting with the small village Baptistery of
I I th century. Half a century later at S. Gio
of all the buildings of this type, narrow dwarf
rhythmical orders of pilasters and columns o
seem to have been inspired by the Pantheon
in the Baptisteries of Cremona ( 1176) and of
obviously a more distant derivative of Fl
galleries appear late in the I I th century at St.
out behind the cathedral, possibly in place of
I 130 in the Baptistery of Arsago. Obviously
ing and could be of no practical use in bapti
galleries becomes general throughout ecclesia
as well as in the Rhineland during the late i i
well have been introduced into baptisteries
influence from the Anastasis need not be eli
appearance of galleries in baptisteries occurs
introduced into copies of the Holy Sepulchr
Sto. Stefano at Bologna and the Holy Sepul
baptisteries take up their 'unfunctional' gall
1 W. Horn, "Das Florentiner Baptisteriurn,"
Mitteilungen des Kunsthistorischen Instituts
in Florenz, V, 1938, p. 99 ffThis
content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms
"ICONOGRAPHY OF MEDIAEVAL ARCHITECTURE" 33
very time when the Anastasis exerted more and more of an i
imagination of laymen and architects throughout the Occiden
intend to give any definite answer to this question. But it ma
remember that many of the 'approved' copies of the Anastasi
erected "instar dominici sepulchri Ierosolimitani" resembled th
no more than did the Baptistery at Florence.
Ij. Hubert, op. cit., Bull. mon. 90o, 1931, p. 91 ff.
This content downloaded from 147.251.6.77 on Sun, 04 Jun 2017 18:47:39 UTC
All use subject to http://about.jstor.org/terms