University of Virginia Library

II.2.2

THE CHURCH AS MODIFIED IN THE
LIGHT OF ITS CORRECTIVE TITLES

To reconstruct what the Church of the Plan would have
looked like had it been modified in the light of the corrective
titles is an intriguing historical task and has produced
a variety of different proposals. It is not surprising that no
agreement has ever been reached in this matter. The
attempt involves some delicate changes on which even
Carolingian architects might not easily have come to terms
with one another.

GEORG DEHIO (1887) & JOSEPH HECHT (1924)

One of the distinctive and historically most fascinating
features of the Church drawing is that it is constructed
according to a system of squares (fig. 61), exhibiting a
principle of spatial organization that became a guiding
feature in certain schools of the Romanesque, two centuries
later.[191] To reduce the Church to the requested length of
200 feet implies the abandonment of the square schematism;
anyone who attempts to redraw the Church using
the measurements listed in the explanatory titles has made
this distressing discovery. Not wishing to totally relinquish
this feature Georg Dehio, who belonged to a generation of
architectural historians profoundly interested in the problem
of modular geometricity in medieval architecture,
retained the squares in the transept and in the fore choir.
By diminishing the interstices of the arcades of the nave
to the stipulated twelve feet, he then arrived at the compromise
length of 218 feet (fig. 130).[192] Joseph Hecht, pursuing
similar lines of thought, arrived at a length of 224
feet.[193]

[ILLUSTRATION]

124. PARIS. ST. GERMAIN-DES-PRÈS

[after Deneux, 1927, 50, fig. 70]

[ILLUSTRATION]

125. PARIS. ST. PIERRE DE MONTMARTRE

[after Deneux, 1927, 50, fig. 71]

Two typical examples of the SPARRENDACH, a relatively steep-pitched
roof with supporting frames of narrowly spaced rafters in continuous
sequence, with light scantling, and braced internally but without purlins
or ridge beams.

[ILLUSTRATION]

126. ROME. ROOF OF OLD ST. PETER'S RECORDED IN 1694 IN
CARLO FONTANA'S TEMPLUM VIATICANUM ET IPSIUS ORIGO. Detail of engraving same size as original

[after FONTANA, 1694, 99]

In publishing this design of what he refers to as "the trusses which sustained the
roof over the nave of Old St. Peter's" (LE INCAUALLATURE, CHE SOSTUEUANO LI
TETTI DELLA NAUE MAGGIORE . . . DEL ANTICA BASILICA VATICANA) Carlo Fontana
(1634-1714), disciple and collaborator of Lorenzo Bernini and architect in charge
of the Pontifical Office of Architects and Engineers, informs his readers that his
engraving was made after an "accurate drawing" (UN GIUSTO DISEGNO) tendered
him by an "informed person" (UNA PERSONA DILETTEUOLE); and that it was
because of the extraordinary constructional "sophistication" (INTELLIGENZA) as
well as the soundness of the timbers employed in these trusses that the roof of the
Constantinian basilica survived intact for so many centuries—to the extent that
when finally taken down, it was found to be in such good condition that its timbers
could be reassembled to sustain the roof of the Palazzo Farnese.

If timbers from the roof of Old St. Peter's were re-used in the Palazzo Farnese
(1546-1589) they must have come from the western half of the nave dismantled
by Bramante (1502) to make room for the construction of New St. Peter's. The
eastern half of the nave (closed off from the construction site by a provisional wall
under Paul III, 1534-1549) was demolished only in 1606 to make room for Carlo
Maderna's westward elongation of New St. Peter's. The roof timbers of this
portion were also re-used, this time for the Palazzo Borghese (1605-1621).

The names of the component members of the truss shown in Fontana's engraving
are enumerated on two scrolls which form part of the drawing. On these the tie
beams are referred to as CORDE MAGGIORI (B), the collar beams as CORDE
MINORI (C), the rafters as PARADOSSI (D), the center post suspended from the
apex of the truss as traue pendente adVso di monaco (E).

During the 12th to 13th centuries of its existence, the roof of the Constantinian
basilica of Old St. Peter's was, not surprisingly, in need of numerous repairs. A
complete account of them, including what in the sources is referred to as a "renovation"
by Pope Benedict XII (fecit fieri de novo tecta huius Basilicae sub anno
1341) is given by Michele Cerrati in Tiberii Alpharini, "De Basilica Vaticana
Antiquissima et Nova Structura" (STUDI E TESTI vol. 26 Rome, 1914, 13 note 2;
brought to my attention by my colleague Loren Partridge). There is no compelling
reason to presume that Benedict's renovation involved any basic changes in the
roof's design.

Fontana's rendering of the trusses of Old St. Peter's is in complete accord with
that which Vitruvius recommends for broad spans, except that all of the principal
members of the truss are doubled, and that the tie beams are fashioned in two
pieces, joined midway by an overlapping scarf joint. Owing to the extraordinary
width of the nave of Old St. Peter's (23.6 3m.), two-piece tie beams were necessary
since it would have been hard to find trees of sufficient height to yield single timbers
to span the whole space. Doubling all of the principal members was an extremely
wise constructional feature—probably the primary contribution to the longevity
of the Constantinian trusses—which evolved from the strategic function made of
the TRAVE PENDENTE. The scheme is a laminative one; a kind of truss-sandwich
is formed in which the structural components are assembled and joined in a function
that yields, in effect, a "pair" of trusses, but which is really a single homogeneous
creation of remarkable simplicity and purity of concept—revealing a mastery of static
mechanics that transcends Vitruvius and commands admiration today. Yet the design
does not seem to have found general acceptance. On the contrary, a medieval carpentry
truss, when it is impressive, gains our attention rather by its quaintness, its
intricacy of joinery and the complexity of its members.

The construction is ingenious. Transmitting the entire roof load to the two outer ends
of the tie beams, the principal rafters D-D are in compression and thus act as
columns as well as beams. Column action augmented by the deflection of beam
action is resisted by the horizontal strut C-C (collar beam) which functions in
compression. These minor chords support the rafter pairs midway in their span, a
construction that reduces the effective length of the rafters to approximately one-fourth
the nave span. Strut C-C is supported at mid-point by the vertical member
E-E (MONACO) which concurrently serves as a tension member to prevent sag in
the great lower chords. The scarf joint of these tie beams, tabled, locked, and
girdled by iron bands, prevents them from separating in the horizontal plane. In
contemplating the brilliance and simplicity of the design, remember that the wall-to-wall
span was above 84 feet—reflecting a state of theory of mechanics and
knowledge of structure existing in the 4th century!

[ILLUSTRATION]

127. TWO BASIC ROMAN ROOF TYPES

[after Vitruvius, Fourth Book on Architecture; interpreted by Barbaro in his translation of 1556]

HANS REINHARDT (1937 and 1952)

Hans Reinhardt, who tends to under-evaluate the square
schematism of the Church of the Plan of St. Gall,[194]
attempted to resolve these discrepancies by developing a
drawing in which the Church was shown reduced to 200
feet. But to attain this goal he found himself compelled to
reduce the fore choir and the space of the crypt beneath it
to one-fourth of their original dimensions, and thus he
arrived at a modified plan which appears to retain no spiritual
kinship to the original concept of the drawing (fig.
131).[195]

Reinhardt contracts the Church most severely where
contraction hurts most: in the all-important area around
the high altar and the tomb of St. Gall, where the entire
body of the monks assembled daily for a total of four hours
or more, in common chant and the celebration of the divine
services. He placed the high altar against the very edge of
the raised choir, where it drops vertically down to the floor
level of the transept leaving no space for the officiating
priest and his attendants (fig. 132). A step on the eastern
side of the altar suggests that Reinhardt imagines the priest
to stand behind the altar facing west. This not only is
incompatible with what is known to have been a general
custom in Carolingian liturgy,[196] but also in open conflict
with fourteen other altars in the Church of the Plan (figs.
84, 93 and 99). Their layout leaves no doubt that the
officiating priest stood west of the altar, facing east; the

[ILLUSTRATION]

128. SYRIA, BATUTA CHAPEL. PORCH, SOUTH SIDE

[courtesy Princeton University Archaeological Expedition to Syria]

location of the altar barriers and the position of the crosses
leaves no doubt on this score.

One feels equally puzzled about Reinhardt's modification
of the crypt. The drafter of the Plan provided the
monastery with two crypts with different but complementary
functions. One is an outer corridor crypt in the
shape of a crank, which takes the pilgrims and the other
secular visitors to the tomb of St. Gall. The other is an
inner crypt which lies beneath the high altar and is reached
from the crossing through a passage marked accessus ad
confessionem, between the two flights of steps that lead up
to the fore choir (fig. 99). Being accessible from an area

[ILLUSTRATION]

130. PLAN OF ST. GALL. CHURCH

[after Dehio, 1887, pl. 42 and fig. 2]

Dehio reconciles the conflict between the "corrective" titles of the plan
of the Church and the manner in which it is drawn, by reducing the
arcade spans to 12 feet. Leaving Transept and Presbytery untouched,
he retains the full measure of space (and incidentally its square
schematism in the liturgically most vital part of the Church, where
monks assembled daily for no less than four hours of religious services.
This solution is acceptable—even perfect—if one excludes the western
apse, as Dehio seems to have done, from the 200-foot length prescribed
for the Church. Dehio's church measures 218 feet from apex to apex of
its apses.

[ILLUSTRATION]

131. PLAN OF ST. GALL. CHURCH

[after Reinhardt, 1952, 20]

Reinhardt proceeded by assuming that the 200-foot prescribed length
of the Church must be understood to include both apses. He adopted
Dehio's solution for the nave of the Church and accomplished a further
reduction of the overall length to exactly 200 feet by eliminating the fore
choir, moving the high altar into the apse, dispensing with the altar of
St. Paul, placing the tomb of St. Gall beneath the altar, and the high
altar itself, into a position where it had to be serviced from the east
rather than the west as was customary in this period.

[ILLUSTRATION]

129. SYRIA. BRAD CONVENT. PEDIMENT OF PORCH

[after Butler, 1929, 199, fig. 201]

Had Forsyth not discovered the 6th-century roof of St. Catherine's on Mt. Sinai, Brad and Batuta (fig. 128) would be sole evidence for
reconstructing the design of timbered roofs in Early Christian Syria. St. Catherine's roof is still functioning, in perfect condition. See The
Monastery of St. Catherine at Mt. Sinai, 1973.

reserved for the exclusive use of the monks, this can only
have been a hall crypt providing the monks with prayer
space around the tomb of St. Gall.[197] Reinhardt eliminates
this confessio altogether and thus creates a spatial vacuum
in one of the most spiritually vital spots of the Church.

From a liturgical and functional point of view the
removal of the fore choir is fatal. Moreover, it is devastating
in its effect on the subsidiary spaces of the Sacristy and the
Scriptorium, which are built against the fore choir and, like
the latter, each cover a surface area of 40 × 40 feet. What
does Reinhardt propose to do with them? To reduce them
proportionately would render them unusable;[198] to retain
them as originally planned would amount to an aesthetic
degradation of the apse which seems incompatible with its
liturgical and architectural function.

Reinhardt's proposal also seems unsuitable in general
historical terms. The interposition of a separate spatial unit
between apse and transept is one of the new and original
features of Carolingian architecture. It appeared in
Neustadt-on-the-Main shortly after 768-769 (figs. 116, 133);
in the abbey church of St. Riquier (Centula) between 790799
(fig. 135); in the church of Vreden around 800 (fig. 136);
in the cathedral of St. Mary and St. Peter of Cologne, prior
to the death of its founder, Archbishop Hildebold, d. 819
(fig. 139); in St. Mary at Mittelzell on Reichenau, as rebuilt
by Abbot Haito between 806 and 816 (fig. 134); and in the
abbey church of Hersfeld, if Groszmann's reconstruction is
correct, between 831 and 851.[199] The primary motivation
for this new spatial entity was, as Thümmler has correctly
pointed out, the desire to isolate and strengthen the importance
of the high altar, at which the choral services were
held, and to provide more space for the officiating clergy.[200]
The increasing dimensions of the crypt, and the latter's
division into an outer corridor crypt for the pilgrims and an
inner confessionary for the monks, is directly related to this
development. Both of these innovations were responses to
pressing liturgical needs.

ARGUMENTS IN FAVOR OF
DEHIO'S INTERPRETATION

It is easy to understand why Dehio was reluctant to
undertake any changes in the eastern parts of the Church
and took the step, for which he was subsequently so
severely criticized, of making the Church a little larger (218
feet) than the stipulated 200 feet. However, there remains
the question whether Dehio is really guilty of such a compromise.
His reconstruction may in fact be based upon a
more accurate interpretation of the title which prescribes
the reduction. Dehio's critics interpret the propositional
phrase AB ORIENTE AD OCCIDENTē to mean "From the
apex of the eastern apse to the apex of the western apse."
There is no assurance whatsoever that this is in fact what
the title meant to convey. The first five letters of the phrase,
AB ORI, are inscribed into the eastern apse, which means
that this apse was a component part of the designated
length. But the inscription does not run into the round of
the western apse; it stops in the westernmost bay of the
nave with the numeral .cc. Literally interpreted this would
mean that the western apse was not meant to be included
in the designated length of 200 feet. If it was not, then
Dehio's reconstruction (fig. 130) would run only 8 feet

[ILLUSTRATION]

132. PLAN OF ST. GALL

CRYPT & ALTAR SPACE, Reinhardt's Interpretation

[after Reinhardt, 1952, 20]

Eliminating the fore choir and moving the high altar into the apse
would have reduced space occupied by monks during divine services to
less than half that foreseen in the Plan. In addition to incongruities
described in fig. 131, this would have led to congestion of disastrous
proportions in this most heavily used part of the Church.

beyond the stipulated length (nine arcades of a span of
12 feet = 108 feet; crossing unit = 40 feet; fore choir =
40 feet; apse = 20 feet. Total = 208 feet)—close enough
to be acceptable; and acceptable without any shadow of
doubt, if the radius of the eastern apse were shortened from
20 feet to 12 feet.[201]

It is imperative, in this context, to draw attention to the
fact (entirely disregarded in previous discussions of this
subject) that the reconstruction proposed by Georg Dehio
appears to conform, indeed, with the manner in which
Abbot Gozbert and his builders interpreted the Plan when
they rebuilt the church in 830-836, as August Hardegger
inferred from the measured architectural drawings made of
the church by Pater Gabriel Hecht, in 1725/26, when
much of the Carolingian fabric of the church was still
identifiable.[202]

WOLFGANG SCHÖNE (1960)

By far the most radical attempt to reconcile the drawing
of the church of the Plan of St. Gall with that of its corrective
explanatory titles was that which Wolfgang Schöne
published in 1960.[203] Schöne not only shortened the church
to the desired 200 feet, but applied the same reduction to
all the other buildings of the Plan. However, in advancing
this theory he either overlooked or disregarded the fact that
the same proposition had already been discussed and convincingly

rejected twenty-two years earlier by Fritz Viktor
Arens, who pointed out that if one were to redraw the Plan
according to the measurements given for the length of the
Church (i.e., 200 feet), the Cloister and all of the service
structures of the Plan would be too small to perform their
designated functions.[204]

My own analysis of the scale used in designing the Plan
confirmed this view. Were the Plan redrawn in this manner
not only would the monks, including the abbot and the
visiting noblemen (i.e., Monks' Dormitory, Abbot's House,
and House for Distinguished Guests) no longer fit into
their beds, but the Refectory of the Monks would be too
small to seat the full contingent of monks, the horses would
lack the required floor space to stand in their stables, and
the workmen could not carry out their respective crafts and
labors.[205]

The most decisive counter argument, however, to
Schöne's interpretation of the Plan is to be found in a
statement made by a man who lived at the time when the
Plan was drawn. In his commentary on the Rule of St.
Benedict, written around 845 in the monastery of Civate,
Hildemar, a monk from Corbie, declared that in his days
"It was generally held that the cloister should be 100 feet
square and no less because that would make it too small."[206]

Schöne reduces it to a little less than 67 × 67 feet. It is
historically incongruous to assume that a scheme of paradigmatic

[ILLUSTRATION]

134. REICHENAU-MITTELZELL

HAITO'S CHURCH (CONSECRATED 816)

[after Reisser, 1960, fig. 285]

Built by the "author" of the Plan of St. Gall, between 806-817 (but
the nave perhaps not before 811), this church continues the tradition of
St. Riquier with its extended altar space and tower-surmounted
crossing. (Also see figs. 117 and 171 for comment on its alternating
supports, and its underlying modular concepts.)

significance should propose a cloister whose dimensions
fall by one-third below what at the time was
considered to be the lowest suitable limit.

ADOLF REINLE (1963-64)

The reconstruction of Adolf Reinle (fig. 137), because of
his radically different interpretation of the axial explanatory
title of the Church, occupies a position entirely apart from
those of any of the previous students of the Plan. Translating
the axial title of the Church "AB ORIENTE AD OCCIDENTE[M]
PED .CC." in the sense of "THIS PLAN IS DRAWN
AT THE SCALE OF 1:200," he felt himself under no compulsion
to reduce the Church to a length of 200 feet, as so
many others had tried to do. He rather endows it with its
full length of 300 feet. However, in adjustment to the title
which designates the intercolumnar interstices of the
arcades of the nave to be 12 feet, Reinle consequently
increased the number of arcades from nine to fifteen.
Reinle draws support for this interpretation from the
observation that arcades of a span of 20 feet (6.8 m.) are
not known to have existed in any of the large colonnaded
basilicas of the first millennium.[207] This being as it is, he
concludes "we must assume that the columnar order of the
Church of the Plan of St. Gall is rendered in a schematic
manner in logical explication of the system of squares which
controls the Plan of the Church."[208] He categorically rejects

[ILLUSTRATION]

136. VREDEN. PLAN

CHURCH OF SS FELICISSIMUS, AGAPITUS, & FELIGITAS

[after Thummler, 1953, 306]

Vreden is a three-aisled cruciform basilica with westwork and
extended choir, plus an annular crypt, built ca. 800 (W. Winkelmann,
1953) or ca. 839 (H. Claussen). See Claussen-Winkelmann,
"Archäologische Untersuchungen unter der Pfarrkirche zu Vreden
(Vorbericht)," Westfalen XXXI, 1953, 304ff.

[ILLUSTRATION]

137. PLAN OF ST. GALL. CHURCH

REINLE'S INTERPRETATION OF THE CHURCH OF THE PLAN AS MODIFIED
BY ITS CORRECTIVE TITLES

[after Reinle, 1962/3, 100]

SCALE 1/64 INCH = ONE FOOT [1:768]

any connection of this geometricity of the Plan of the
Church with the square schematism of the Romanesque.[209]

This is too simple a way, in my opinion, to explain a
complex historical phenomenon. Columnar interstices of
20 feet, it is true, are not attested for the period in which
the Plan was drawn. But this does not mean that such a
solution was not within the grasp of an imaginative
Carolingian architect. Our analysis of the scale and
construction method used in designing the Plan[210] has
shown that the author of this scheme proceeded with an
acute awareness of the dimensional realities involved in
whatever he drew. It is inconceivable, in my opinion, that
an architect whose punctilious observance of spatial needs
is reflected in the dimensioning of even the smallest detail
throughout the entire width and length of the Plan, should
have reverted to a radically different method of rendering
when he drew the Church of the Monastery and should
have spaced the columns at a distance of 20 feet when in
fact he meant them to be placed at intervals of 12 feet. A
consistent interpretation of the dimensional layout of the
Plan permits no other conclusion than that the draftsman
meant what he drew. Nor is there evidence to presume
that the instruction to make the columnar interstices 12 feet
wide stemmed from fears that arcades spanning 20 feet
would be a constructional hazard. Our reconstruction (figs.
107-110) demonstrates this point clearly enough. The
shortening of the arcade spans was simply an inevitable
consequence of the reduction of the overall length of the
Church from 300 to 200 feet. It dealt a deadly blow to the
square schematism as applied to the nave of the Church—
one of the draftsman's favorite and most original ideas—
but it was the most reasonable way out of the dilemma
caused by the overall reduction of the length of the Church.
By reducing the spatial depth of each bay, the corrective
title permitted the retention of the original number of altar
stations, while at the same time it safeguarded the original
concept in those parts of the Church where a reduction
would have impaired the primary function of the sanctuary,
the conduct of the sacred services in transept and choir.

[ILLUSTRATION]

PLAN OF ST. GALL

137.X CENTULA
(ST. RIQUIER)

see also page 185 and figs. 135, 168, and 196

141. ROME. OLD ST. PETER'S

[after Jongkees, 1966, 34, Pl. I and II]

The church's inner length was 112m (368′), its height 84m (276′), and its width
58m (190′). Probably begun after 324 and finished by Constantine's death in 377,
its precise dates are unknown.

138. FULDA

[after Groszmann, 1962, 351, fig. 5]

Ratger's church of 802-817. Precise measurements unknown

139. COLOGNE

Hildebold's church of SS Peter and Mary

[after Weyres, 1966, 408, fig. 10]

140.