University of Virginia Library

V. 1

PREVIOUS INTERPRETATIONS

V.1.1.

THE CLASSICAL SCHOOL

FERDINAND KELLER, 1844

The first to speculate about the design of the guest and
service buildings of the Plan of St. Gall was Ferdinand
Keller. In his "facsimile" edition of the Plan, published
in 1844, Keller described them as structures "of oriental
style in which the living quarters are arranged around an
open central court toward which the roof slopes down from
all four sides of the building."[1] Keller explained the squares
—which are designated by the term testu[2] in the explanatory
titles—as "courtyard houses" or "garden huts." He
did not illustrate his views with graphic reconstructions,
but the kind of structures he had in mind must have
resembled the houses sketched in figure 264 A-C. Keller's
interpretation is perplexing because it contradicts the
explanatory titles of the Plan itself, which he had so carefully
transcribed, translated, and annotated. In seven out
of the nine times this house-type appears on the Plan of
St. Gall, the area that Keller interprets as an open inner
court is explicitly designated as the "common" or "principal"
room of the house by inscriptions such as domus
communis (House of the Fowlkeeper), domus ipsa (House for
Sheep and Shepherds), domus familiae (House for Visiting
Servants), domus pauperum et peregrinorum (House for
Pilgrims and Paupers). Keller, however, was not truly
consistent in this matter; for in the case of the House for
Distinguished Guests, where the function of the center
area as an indoor space is visually elaborated by the insertion
of pieces of indoor furniture such as tables (mensae),
benches, and a fireplace (locus foci), he interprets the center
space of the house correctly—and in full accord with its
identifying inscription as "dining room" (domus ad prandendum).
The incompatibility of this interpretation with
that of the other guest and service structures as open courtyard
houses does not appear to have raised any doubts in
his mind of the validity of the latter.

ALBERT LENOIR, 1852

We do not know what specific prototypes Keller had in
mind when he explained the St. Gall house as the descendant
of an oriental courtyard house, but once this idea was
suggested it was inevitable that the design of the St. Gall
house should also be connected with that of the Roman
atrium house. This idea was pursued in 1852 by Albert
Lenoir.[3] Lenoir derived the St. Gall house from a subvariety
of the Roman atrium house that Vitruvius had
called "Tuscan" (tuscanum)—a house with an open inner
court which was partially roofed over (fig. 265), but which
retained in its center a large rectangular "rainhole"
(compluvium) and on the ground below it, the classical
Roman rain catch basin (impluvium).

To the difficulties of Keller's reconstruction, Lenoir
thus added a further one, since testu[do] can be translated
as neither "rainhole" nor "catch basin." Whatever the
specific implications of this term may be, its basic meaning,
"tortoise" or "turtle shell," points in the opposite direction,
namely, to that of a protective shield or cover.[4]

J. R. RAHN, 1876

Probably aware of these inconsistencies in Keller's and
Lenoir's interpretations, the Swiss art historian J. R. Rahn
presented a new solution in 1876, which was incorporated
into a graphical reconstruction of the entire settlement in
a bird's-eye view drawn up for him by Georg Lasius
(fig. 266).[5] Without explicitly refuting or even discussing
Keller's and Lenoir's views, Rahn reconstructed the St.
Gall house as a masonry structure of basilican type with a

large rectangular center room that rose above the outer
rooms and received its light through clerestory windows.
The squares that are inscribed in the Plan alternately as
locus foci and testu[do] Rahn interprets as open fireplaces
surmounted on the level of the main roof by a lantern with
openings for the escape of smoke. This solution was suggested
to him by similar architectural contraptions "still
nowadays in use in certain rural houses of northern
Germany and also occasionally found in England."[7] Rahn's
reconstruction has subsequently found the widest circulation
by being reproduced in Cabrol-Leclercq's Dictionnaire
d'archéologie chrétienne.[8] It also formed the basis for a
masterful three-dimensional model, executed in 1877 by
Julius Lehmann, which found a permanent home in the
Historisches Museum of the city of St. Gall (fig. 267).[9]

The advantages of Rahn's reconstruction over those of
Keller and Lenoir are obvious at first sight. It establishes
correctly, and in accordance with the legends of the Plan,
the large rectangular center space of the house as a covered
room. Second, it associates the term testu[do] with a
device that is compatible with its etymology (protective
shield, or cover). Third, it offers a constructive solution to
the interchangeability of the terms testu[do] and locus foci,
since "hearth" and "lantern," if arranged in the manner
Rahn suggested, would merely be two complementary
aspects of the same device—namely, an open fire with a
smoke hole in the roof surmounted by a lantern.

JULIUS VON SCHLOSSER, 1889

Rahn's reconstruction of the St. Gall house as a basilican
masonry structure with a lantern-surmounted central
hearth had been a purely theoretical venture. He could not
prove—and did not even attempt to prove—that houses of
this description actually existed.

Misconceptions about the "displuviate" and
"testudinate" Roman courtyard house

Julius von Schlosser[10] tried to overcome this weakness by
demonstrating that Rahn's St. Gall house was historically
the descendant of a once widespread Roman house type to
which the ancients referred with the terms "displuviate"
(displuviatum) and "testudinate" (testudinatum). Schlosser's
theory, unfortunately, was based on two erroneous assumptions
that were current in his day, which imparted to the
discussion of the design of the guest and service structures
of the Plan of St. Gall an element of further confusion.
The first of these misconceptions pertained to the precise
meaning of the terms "displuviate" and "testudinate";
the second concerned the origins and structural evolution
of the Roman atrium house.

To begin with the former: what Vitruvius and Varro
designated by the terms "displuviate" and "testudinate"
can under no circumstances be interpreted as structures
of basilican design. They were atrium houses in the full
constructional sense of the term, i.e., houses in which the
living quarters were ranged around an originally open
center space. The terms "Tuscan", "Corinthian", as well
as "tetrastyle", "displuviate", and "testudinate" (tuscanicum,
corinthium, tetrastylon, displuviatum, and testudinatum)
merely referred to the different degree or manner in which
these inner courtyards were roofed over.[11] In the Tuscan
atrium house, for instance, the courtyard roof sloped down
toward the center (fig. 265); in the displuviate house it
sloped upward. But in both cases the courtyard roof
encompassed in its center a rainhole (compluvium) that had
under it not a hearth, but a catch basin (impluvium).
Schlosser did not realize that in connecting the St. Gall
house with the displuviate Roman atrium house, he had
actually retrogressed to Lenoir's views (fig. 265), whose
weakness Rahn's reconstruction (fig. 266) had already
successfully overcome.

The same applied to Schlosser's attempt to connect the
St. Gall house with the courtyard house referred to by
Vitruvius and Varro by the term "testudinate." It is an
atrium house like all the others, as must be inferred not
only from the language of the opening sentence with which

[ILLUSTRATION]

ROMAN ATRIUM HOUSE WITH RAIN
CATCH-BASIN

265.B PERSPECTIVE redrawn after Kähler, 1960, suppl. 53, fig. 31

265.C

265.A PLAN, based on Luckenbach, Kunst and Geschichte, I: Altertum,
Munich, 1910, 94

SECTION, authors' interpretation

The plot, 60 feet wide, is 1/2 ACTUS (120 feet)

See remarks on Roman land surveyor's measure, page III. 140

In the form here shown, the Roman atrium house has its open inner
court partially covered by an inward-sloping roof with its center
open to the rain (COMPLUVIUM); beneath this opening in the center
of the court is a collecting basin (IMPLUVIUM). The dining room
(TABLINUM) lies to the rear of the house and opens onto the garden
(HORTUS). The kitchen stove might have been located in any of the
cubicles adjacent to it.

Vitruvius introduces his subject ("Cava aedium quinque
generibus sunt distincta . . ."; "the inner courts of houses
are of five different styles," etc.), but also from the detailed
descriptions that follow ("Testudinata vero ibi fiunt, ubi non
sunt impetus magni et in contignationibus supra spatiosae
redduntur habitationes . . ."; "Testudinate courtyards are
employed when the span is not great, and they furnish
roomy apartments in the story above").[12] In contradistinction
to the other four types in which the courtyard was only
partially roofed over, the testudinate atrium house was a
house in which the inner court was entirely covered. It was
an atrium house in which the courtyard had lost the character
of an open space by being covered over with a second
story, but it was still an atrium house.[13] The Romans used
this type of construction in houses of relatively small
dimensions, as Vitruvius himself suggests, and probably in
response to restricted land conditions prevailing in the
crowded Roman cities.

The Roman atrium:
an open yard developing into a covered court

Schlosser's misinterpretation of Vitruvius' and Varro's
definitions of the displuviate and testudinate Roman atrium
house was in itself conditioned by the faulty historical
assumption held by many leading classical archaeologists
at that time, that the Roman atrium was originally not a
court but the principal living room of the house which
gradually developed into an open yard.[14] This theory was
taken up and widely propagated by one of the greatest
connoisseurs of Roman house construction, August Mau.[15]
But, curiously enough, it had not originated from any
archaeological evidence, which in fact seemed to contradict
it, but from a questionable etymological speculation by
certain Roman authors who believed that atrium came from
ater ("black") and referred to the blackening of the atrium

[ILLUSTRATION]

266. PLAN OF ST. GALL. BIRD'S-EYE VIEW OF A RECONSTRUCTION OF THE PLAN OF THE
MONASTERY

MADE FOR J. R. RAHN BY GEORG LASIUS (1876, fig. 12, 91)

This is the first, and for its period, truly outstanding attempt to show in an accurately constructed bird's-eye view, what the monastery might
have looked like had it actually been built. It formed the basis of the three-dimensional model reconstruction shown in figure 267. Rahn's
interpretation of the guest and service buildings as covered basilican structures with central hearths and openings in the roofs above, serving as
smoke escape and light inlet, was a great improvement over Keller's (fig. 264) and Lenoir's interpretation, but like theirs, suffers from being
modeled after Classical prototypes rather than those historically and archaeologically related to those of the era and location of the Plan of
St. Gall.

[ILLUSTRATION]

267. ST. GALL. HISTORICAL MUSEUM OF ST. GALL

ARCHITECTURAL MODEL. A RECONSTRUCTION OF THE BUILDINGS OF THE PLAN

The model was made by the sculptor Jules Leemann of Geneva, in 1877, on the basis of drawings furnished by Georg Lasius and has ever since
been on display in the Historical Museum of the city of St. Gall. It is a masterpiece of its kind, built to scale, and executed with supreme
craftsmanship. Length of base: 70¼ inches (1.78m). Width: 49½ inches (1.25m). The roofs of the houses, as well as the Church, can be lifted,
exposing the furniture on the ground floor levels. The second stories, the Refectory and the Cellar can be lifted out in their entirety. The
reconstruction of Church, Cloister, and Novitiate are essentially correct. The height of the Church (a little over twice the width of the nave) is
excessive for the period. The reconstruction makes no distinction between masonry and timber. Entirely unconvincing is the design of the majority
of the guest and service structures (see caption, figure 266).

walls by the smoke from its hearth.[16] Since the beginning
of this century, this view has been increasingly challenged
by a trend of thought that holds that, on the contrary,
the Roman atrium was originally an open yard, which
gradually developed into a covered court. The main exponents
of this theory are Antonio Sogliano,[17] Giovanni
Patroni,[18] and Axel Boethius,[19] who believe the Roman
atrium house to be the product of a gradual transformation
of an early Italic farmstead, whose individual buildings
had been scattered loosely around a central open yard, into
an organized architectural system under the hands of the
Etruscan conquerors.

They assume that the principal building of this Italic
farmyard was a prostyle farmhouse with hearth and bedstead.
Through a gradual process of axial co-ordination
of this main house with the subsidiary structures and the
yard enclosure, the Etruscans, according to this theory,
developed the irregular Italic farmstead into the aggregate
depicted in figure 268. Two further developments, in their
opinion, led from this hypothetical prototype form to the
emergence of the classical Roman atrium house, the
"Pompeian primehouse"; the coalescence, namely, of the
roofs of the subsidiary structures with that of the main
house on one hand, and the roofing-over of the courtyard
on the other. As this process unfolds itself, the hearth is
shifted from the original farmhouse (now tablinum) into
one of the adjacent smaller rooms.[20]

Whatever the merits of this theory may be, this much
appears to be certain: we do not know of a single Roman
atrium house, excavated or otherwise attested, that shows
in the center of its covered court either the traces of a
hearth[21] or any evidence in the roof above it for the existence
of a protective lantern (testudo). In the Roman atrium
house this spot is the traditional place for the catch basin
(impluvium) and directly above it, in the roof, for a rainhole
(compluvium), which also served as air or light source
(fig. 265). The hearth lay, as a rule, in one of the smaller
chambers to the side of the tablinum, or in one of the other
peripheral cubicles, but in any case entirely outside the
atrium space. The testudo of the Roman atrium house, then,
is an altogether different architectural entity from the
device that carries this name on the Plan of St. Gall. The
latter device called testu on the Plan of St. Gall is coextensive
with the hearth site (and could very well be interpreted,
as Rahn suggested, as a protective shield or lantern
that covers an opening in the roof above the hearth); the
testudo of the Roman atrium house by contrast is the designation
for a shielding roof which covers the Roman atrium,
either as a peripheral shed (as in the atrium Tuscanum) or
as a continuous roof (as in the smaller and rarer atrium
testudinatum).

The ash-urn house of Poggio Gaiella

The same objections have to be raised with regard to Schlosser's
comparison of the St. Gall house with an Etruscan
ash urn from Poggio Gaiella (fig. 269) and other imitations

[ILLUSTRATION]

268. ARCHAIC ETRUSCO-ROMAN HOUSE. RECONSTRUCTION

REDRAWN FROM PATRONI, 1941, 294

Patroni's ideal conception shows the form that the early Italic farmstead had attained after individual buildings, formerly scattered loosely
around an open yard, were axially aligned into an organized architectural scheme by the Etruscan conquerors of the Italian peninsula about the
12th century B.C. The peak of Etruscan culture was achieved during the 6th century B.C.

of the displuviate atrium as they are found in the architecture
of some Etruscan tombs. All of these specimens belong
to the courtyard type. They have an opening at the very
spot where the St. Gall house calls for a protective cover,
and there is no suggestion whatsoever that their hearths lay
under this opening or had any functional or developmental
relation to this opening. If the St. Gall house were reconstructed
analogous to the house from Poggio Gaiella, it
would have its hearth on the very spot where every squall
of rain or sleet would kill the fire and drench the occupants
of the adjacent benches and tables. As well as it may have
been adapted to the temperate conditions of a southern
climate, the layout of the house from Poggio Gaiella would
hardly meet the housing requirements of a climate where
heavy downpours and freezing temperatures are matters of
course for periods of considerable duration.

FRANZ OELMANN, 1923-1924

A shaky premise

Whatever the merits of Schlosser's theories might have
been—and even if he had been correct in his assumption
that the Etruscan and Roman atrium houses that he discussed
were of truly basilican type—a problem of major
magnitude was still presented by the formidable gap—
chronological, topographical, and cultural—that separated
Rahn's St. Gall house from its presumptive Etruscan and
Early Roman prototypes. To bridge this gap Franz Oelmann,
in 1923/24, attempted to demonstrate that houses
of the Poggio Gaiella type (fig. 269) were common in Roman
imperial times and continued to be in use in the provincial
territories of Germany and Gaul even after they had been
conquered by the Franks.[22]

A faulty interpretation of the
Gallo-Roman courtyard house

Oelmann subscribed to Rahn's idea about the St. Gall
house and felt convinced that Schlosser was right in
defining it as a descendant of the Etruscan urn house of
Poggio Gaiella, which both he and Schlosser, however,
interpreted wrongly as a structure of basilican type. Oelmann
conceded that the city of Pompeii, with all its wealth
of architectural information, "does not furnish any convincing
parallels,"[23] then added, "but in the secular architecture,
not so much of Italy as of the Gallo-Germanic
provinces of the North, we can find analogies for practically
each and every subvariety of the houses of the Plan of St.
Gall."[24] He undertook to support this assertion by assembling
the plans of a considerable number of Gallo-Roman
houses and juxtaposing them, type by type, with what he
believed to be their constructional counterparts on the
Plan of St. Gall. In establishing these parallels, he adopted
a procedure, as unorthodox as it is startling, by simply
reversing the views of the archaeologists by whom these
houses had been excavated. The latter were convinced that
what they had unearthed were the foundations of typical
Roman courtyard houses, i.e., houses in which the rooms
were ranged peripherally around an open central court as
in the Roman atrium house. In two of them, a farmhouse
in the vicinity of the village of Nendeln, Liechtenstein
(fig. 270),[25] and a Roman villa in Bilsdorf, Luxembourg
(fig. 271),[26] they had found the remains of a large impluvium,
tangible evidence of the correctness of this interpretation.
Oelmann did not conceal these facts,[27] but simply
brushed them aside with the contention that what the
excavators declared to be an inner court was in reality a
covered hall, and that the rectangular basins found in the
center of these structures had to be interpreted not as catch
basins—as the excavators thought—but as hearths!

It is difficult for me to see how an experienced excavator
would confuse the straight and careful lining of a Roman
impluvium that had never been exposed to fire with the
scorched and blackened remains of a hearth whose rims
were never as regularly set; but what makes Oelmann's
categorical reversal of the thinking of his predecessors even
more perplexing is the fact that in one case at least, namely
that of the villa at Bilsdorf in Luxembourg, the excavator
had unearthed not only the remains of the catch basin
itself, but also a good portion of its drainage ditch. In his
account of the villa of Bilsdorf, Oelmann is guilty both of
factual distortion and of suppression of vital archaeological
evidence. He does not tell us that from the presence of heat
ducts found in the walls of chambers A and J (fig. 272) the
excavators had concluded that at least the avant-corps of
the villa must have been a double-storied structure. He
leaves us in ignorance about the fact that, while many of
the peripheral rooms were carefully paved with tiles (Rooms
B, C, and F) or opus signinum (Rooms A, M, L, K, and J),
the floor of the court consisted of nothing but stamped
clay. And least to be excused, he furnishes us with a plan

[ILLUSTRATION]

271. BILSDORF, LUXEMBOURG [after Oelmann, 1928, 127]

[ILLUSTRATION]

270. NENDELN, LIECHTENSTEIN [after Oelmann, 1928, 127]

Oelmann's rendering of the villa at Bilsdorf is a highly simplified version of the very detailed and exemplary plan published in 1910 by its excavators, E. and R.
Malget (fig. 272). It suppresses information (given in our caption to figs. 272-273) that makes incontrovertible the fact that the inner part of the house was an open
court surrounded by a covered walk, with a center opening (IMPLUVIUM) beneath which rain collected in a catch basin (COMPLUVIUM). Thus Oelmann's interpretation
of this villa as a house of basilican design is untenable.

The Bilsdorf villa was destroyed by fire, possibly set by invading Franks, toward the end of the 3rd century A.D. A welter of skeletal remains suggests the house burned
in the course of a violent battle. Both Bilsdorf and Nendeln are typical Roman standard houses with living quarters ranging around an open inner court, of the type
commonplace throughout the length and breadth of the Roman empire. For the original excavation report on Nendeln see S. Jenny, Mitteilungen der K.K.
Central Commission, XXIII, 1897, 121ff.

(fig. 271) that wholly suppresses[28] all of the four heating
units (two furnaces, one hypocaust, and one brazier) that
the excavator found in the peripheral chambers (Rooms A,
C, F, and J) and carefully recorded in his own original plan
(fig. 272). All this evidence taken into account suggests
precisely what its excavator thought it to suggest—namely,
that the villa of Bilsdorf was a classical example of the
tetrastyle Roman atrium house, i.e., a house in which a
peripheral suite of rooms, ranged all around a central open
court, was surrounded by a covered walk that had a large
rectangular opening in the middle of the roof through which
the rain drained off into a central basin in the floor beneath
it. The rooms could be heated by classical Roman heating
devices (hypocaust, furnaces, brazier), either in pairs or
individually. In the tetrastyle Roman atrium house, Vitruvius
tells us, the roof of the surrounding gallery of the
court "was supported at the angles by columns."[29] In the
villa of Bilsdorf all of the base blocks of these posts were
found still in their original emplacement. In its vertical
elevation, then, the villa of Bilsdorf bore not the slightest
resemblance to Rahn's St. Gall house, but rather might be
imagined to have looked like the house shown in figure
273.[30]

What I have tried to demonstrate with regard to the villa
of Bilsdorf holds true for all of Oelmann's other comparisons.
In not a single case could he actually demonstrate
on the basis of controllable evidence that his houses looked
as he claimed them to look; and in whatever cases I have
been able to check, his own interpretation of the facts
either contradicted that of the men by whom these houses
had been excavated or were open to at least one other
explanation.

In claiming that Oelmann's attempt to trace the missing
Gallo-Roman prototypes of Rahn's St. Gall house was a
failure, I do not mean to imply that houses of the type that
Oelmann had in mind might not have existed. But as long
as the proof of their existence rests on authoritative assertion
rather than on archaeological demonstration, I cannot
see how such a house type could be used as a prototype
form for the reconstruction of the guest and service structures
of the Plan of St. Gall.

A faulty interpretation of "testu"

While Oelmann agreed with Rahn's interpretation of the
St. Gall house as a structure of basilican type, he took
exception to the latter's explanation of the testu[do] square
as a lantern surmounting a smoke hole in the roof above the
fireplace. Such a device, he claims, is attested to neither by
classical nor by medieval house construction. He suggests,
instead, that what the drafter of the Plan of St. Gall had
in mind is more likely to have been a huge freestanding
chimney stack on pillars or arches, which rose from the
ground to the ridge of the roof, protruding through the
latter, and ejected its smoke into the open air (fig. 274).[31]
But here again the reader is not furnished with any corroborating
historical evidence. That Rahn's testudines have no
equivalents in classical Greek and Roman architecture may
well be the case, but the assertion that they are not attested

to in the Middle Ages is easily contested.[32] And as far as
Oelmann's own suggestion is concerned, it must be pointed
out that all the presumptive medieval parallels that he
adduces turn out upon inspection to pertain not to houses
but to kitchens.[33] The reader will recall that the squares on
the Plan of St. Gall which are alternately designated as
locus foci and as testu[do] are by no means confined to the
houses for distinguished persons. They are an integral part
of even the humblest among the stables. It is difficult to
imagine that an extremely tall and costly masonry stack
such as Oelmann had in mind should have adorned the
houses of swineherds, shepherds, and goatherds at a period
when such devices were a novel rarity even in the dwellings
of the nobles. Most perplexing of all, however, is
Oelmann's identification of the term testu[do] with "chimney
stack"—an equation that finds no support on any
grounds—since it is neither possible to demonstrate that
the term was ever used in this sense in classical or medieval
Latin, nor reasonable to presume that it might ever have
been used in this manner. Its basic meanings (protective
shield, covering lid, tortoise, turtle shell[34] ) are in outright
conflict with the idea of a hollow flue or duct which underlies
the concept of a smokestack.

V.1.2

THE NORTHERN SCHOOL

All the theories heretofore reviewed have in common the
fact that they attempted to explain the guest and service
structures of the Plan of St. Gall in the light of house types
presumed to have existed in Etruscan, Roman, and Gallo-Roman
times. The most ardent exponent of this school,
Franz Oelmann, expressed himself in no uncertain terms
when he summarized his views with the phrase, "The Plan
of St. Gall, then, must be derived in its entirety from the
Classical tradition, i.e., from Roman architecture, and of
Northern influences . . . there can be no question whatsoever."[35]
The uncompromising fervor of this assertion is
clear evidence that at the time these lines were written, the
issue had already entered a highly controversial phase.
And indeed, as early as the last two decades of the nineteenth
century, the views of the classicists had begun to be
progressively challenged by the speculation of an opposing
school that proposed to reconstruct the guest and service
structures of the Plan of St. Gall in the light of northern
rather than classical building traditions. The seeds of this
theory may actually be discovered in the writings of some
of the exponents of the classical school. When Rahn, in
1876, interpreted the testu[do] squares as symbols for a
lantern-surmounted opening in the roof above the hearth,
he breached the thinking of the classicists, since this was a
solution suggested by analogy with northern rather than
with southern building types. Yet, apart from this "intrusive"
detail, Rahn's reconstruction was essentially a product
of the classical school. A square attack on the theories
of the latter, however, was launched in 1882 by Rudolf
Henning.

RUDOLF HENNING, 1882

In a study entitled "Das Deutsche Haus in seiner historischen
Entwickelung,"[36] Rudolf Henning stressed the resemblance
of the plan of the St. Gall house to certain
house types still used in Upper Germany and in Switzerland,
in territories once occupied by Frankish, Alamannic,
and Bajuvarian tribes. In dwellings of this type (fig. 275)
such as is exemplified by a house from the Engadin in
Switzerland, the hearth is, as a rule, located in a common
center room (Eren) from which access is gained to all the
subsidiary outer rooms. It is surmounted by a wooden
smoke flue of pyramidal shape which projects beyond the
roof like a chimney and can be closed and opened by an
adjustable lid. A similar arrangement is found even today
in old farmhouses of Denmark (fig. 276). Henning did not
propose that the St. Gall house was equipped with such a
smoke flue. He believed, on the contrary, that it had an
open hearth and, in the roof above the hearth, a lantern-surmounted
opening that served as a smoke outlet and as a
light source. He imagined the St. Gall house to have been a
spacious, steep-roofed structure with inner wall partitions
that did not obstruct the view of its enclosing walls and
rafters. He felt supported in this assumption by a passage
in the Lex Alamannorum which makes the paternal right
of inheritance dependent on the ability of the newborn
child to encompass the roof and the four corners of the
house as he opens his eyes.[37] The existence of a house of
this description, Henning felt, must be postulated as the
medieval prototype form of the modern Swiss and Upper

[ILLUSTRATION]

BILSDORF (HAUTE SURE), LUXEMBOURG. PLAN OF A ROMAN VILLA

272.A

272.B

[redrawn after Malget, 1909, 354]

Malget describes features of the villa's rooms:

A. Heatable room with brick-column supported, raised floor of limestone-bedded,
crushed rubble. Within it: socle for an altar (a); brick-paved area for the
brazier probably used to heat the room (m); tile smoke flues (d).

B. Room with limestone-bedded crushed rubble floor.

C. Tile-paved room with hypocaust, heated by furnace (e).

D. Room with crushed brick floor.

E. Lodging for slaves (floor material not identified).

F. Room paved with rectangular brick.

G. Unpaved room, probably for slaves assigned to heat F.

H. Room paved with stamped clay; I, room paved with stamped clay, probably,
for slaves heating J.

J. Warming room over hypocaust heated by furnace in I, with raised floor
(unidentified substance) supported on brick columns.

German farmhouses with central hearth and central accessibility
of the subsidiary outer rooms; and, in the houses of
the Plan of St. Gall, he believed to have discovered the
first pictorial evidence of this prototype form. The development
that leads from this archetype to its modern derivative,
Henning assumed, was characterized by the gradual
substitution of a stone-built stove with smoke flue for the
originally open fireplace, and by the removal of the light
source from the ridge of the roof to the walls, which became
necessary when the opening in the ridge was obstructed by
the installation of a central smoke stack.

The beginnings of this displacement of the open fireplace
by stone-built stoves, Henning suggested, may already be
observed in some of the more distinguished structures of
the Plan of St. Gall, such as the House for Distinguished
Guests, where the common central fireplace is already
supplemented by stone-built corner fireplaces.

KARL GUSTAV STEPHANI, 1902-1903, AND
CHRISTIAN RANCK, 1907

Henning refrained from embodying his ideas about the
St. Gall house in a visual form, and in the fifty years that
followed, this theory found neither support nor acceptance.
The views expressed in Karl Gustav Stephani's encyclopedic
work on the early German dwelling and its furnishings,[38]
as well as those in Christian Ranck's percursory but
widely read cultural history of the German farmhouse,[39]
are literal repetitions of Julius von Schlosser and show that

[ILLUSTRATION]

273. BILSDORF (HAUTE SURE), LUXEMBOURG. PERSPECTIVE VIEW

AUTHORS' RECONSTRUCTION

K. Vestibule with damaged crushed rubble floor.

L. Water basin reached from K by steps paved in square brick.

M. Entrance hall with concrete floor of limestone, broken brick, and sand on a
dry-stone bed.

N. Rain catch-basin (impluvium) set in a square of opus signinum; the rest of
the atrium was paved with red clay. (Judging from the many pieces of
slate found scattered through the atrium, we may assume the house was
roofed in slate.)

Figure 273:

Of square plan but for the generous setback of the entrance wing, the villa's
size (over 80 × 80 feet on the ground floor) and its double-storied corner wings
made it a structure of imposing presence. The reconstruction above is based on
Malget's description, which holds a wealth of specific detail.

the work of even those who specialized in the history of the
German house was still entirely under the spell of the
thinking of the classicist. But in the third decade of this
century, the method that Henning had initiated, namely,
that of attempting to reconstruct the St. Gall house in the
light of its modern derivatives rather than of its historical
prototypes, found a sudden revival in a number of visual
reconstructions that marked a complete departure from
the thinking of the classical school. These reconstructions
(figs. 277-281) came from the hands of men who were not
primarily historians but professional architects, and they
were the product of intuitive speculation rather than of
documentative historical study. The first of these was made
by H. Fiechter-Zollikofer in 1936.

H. FIECHTER-ZOLLIKOFER, 1936

Mr. Fiechter-Zollikofer, a Swiss engineer, wrote an article
entitled "Etwas vom St. Galler Klosterplan aus der Zeit
um 820," which was published in the Schweizerische Technische
Zeitschrift,[40] a journal not normally read by the
architectural historian of the Middle Ages. In this article
Fiechter-Zollikofer reproduced not only an over-all reconstruction
of the entire monastery shown on the Plan of St.
Gall, in bird's-eye view (fig. 277), but also a separate reconstruction
of the exterior of the Outer School (fig. 278), the

[ILLUSTRATION]

PLAN OF ST. GALL. OUTER SCHOOL. INTERPRETATION OF OELMANN (1923-4). AUTHORS' DRAWING

274.C WITH ROOFS

Like Rahn (see figs. 266-267) Oelmann interprets the St. Gall
house as a structure of basilican elevation with a large central hall
rising above the roofs of the perpheral rooms and receiving its light
through windows in the clerestory walls.

274.B WITHOUT ROOFS AND WALLS PARTLY REMOVED

In contrast to Rahn (figs. 266-267) Oelmann interprets the squares,
which on the Plan are alternatingly referred to as TESTU (i.e.
lantern) and LOCUS FOCI (i.e. hearth) as freestanding chimneys
rising all the way up to and through the ridge of the house. This
interpretation of TESTU is philologically not convincing.

274.A WITHOUT ROOFS, AND FULL HEIGHT WALLS

The real weakness of Oelmann's interpretation lies in its bases on
purely theoretical considerations, giving no heed to the vernacular
building tradition of the north (not well understood and known in
Oelmann's days) which offers better and more convincing parallels
for the interpretation of the guest and service buildings of the Plan
(see below, pp. 88ff).

Until challenged by student who felt that the guest and service buildings of the Plan should be interpreted in light of the vernacular building
tradition of the north, this interpretation prevailed. But its proponents could not prove that houses of this type existed in Carolingian times.

exterior of the Abbot's House (fig. 254), as well as a number
of perspectives and cuts of the church and the claustral
structures.

Fiechter-Zollikofer was convinced that the traditional
concept of the St. Gall house as a dwelling that received
its light in the Italian manner through windows in its
clerestory walls was incompatible with the climatical conditions
prevailing in transalpine Europe, and that a solution
infinitely better adapted to the rain and snowswept foothills
of the Alps could be found if the St. Gall house were
reconstructed in the light of certain rural timber dwellings
still used in many districts of Switzerland.

Accordingly, he reconstructs the St. Gall house as a
low-roofed, low-walled gable house of logs with corner-timbered
protruding beams (fig. 278). The center room
of this house receives its light through a large tapering
shaft mounted upon the ridge of the roof which could be
opened and closed through an adjustable lid (fig. 279); the
outer rooms were lighted through windows in the peripheral
log walls. Fiechter-Zollikofer's reconstruction is the first
attempt to interpret the guest and service structures of
the Plan of St. Gall in the light of an actually existing
vernacular house type. It is a handsome reconstruction,
but the prototype after which it is modeled, the Alpine
log house, is too closely associated with local conditions to
have been adopted in a master plan that was drawn up for
the whole of the Frankish empire. Log construction depends
on abundant stands of fir trees, such as are available in the
Alps, the Black Forest, and the mountain ranges of Scandinavia;
but in the lowlands this material was lacking.

Moreover, Fiechter-Zollikofer did not enter into any
detailed analysis of the internal layout of these houses. He
did not support his reconstructions with any specific parallels
with comparable structures still in existence, or attempt
to trace this house type to its historical past.

OTTO VÖLCKERS, 1937

Fiechter-Zollikofer's article had barely been published when
the German architect, Otto Völckers, touched upon the
problem of the St. Gall house in a small, handsomely
illustrated book in which he reviewed the history of the
European house from the Stone Age to the present.[41]
Völckers exemplified his views with a reconstruction of
St. Gall's House for Distinguished Guests (fig. 280). This
he imagines to have been a steep-roofed structure, hipped-over
on the narrow ends of the building. The walls are low
and masonry-built with windows giving light to the external
rooms. The center room is lighted by an opening in the
ridge above the hearth site, which also serves as a smoke
outlet and is surmounted by a small protective roof that
shields the opening against any downpour. The heating
units in the bedrooms of the distinguished guests are
interpreted as corner fireplaces with masonry stacks protruding
through the roof above them. Völckers did not discuss

[ILLUSTRATION]

276. DANISH FARM HOUSE

[after Steensberg, 1943, 20, fig. 7]

A north European variant of the house type shown in figure 275.
Chimney-surmounted hearths of this type work well in relatively
small houses, but would involve constructional hazards of frightening
magnitude in most of the larger houses of the Plan of St. Gall.

A BAKING OVEN

B MALT KILN

C KETTLE

D FIRE RECESS

E HEATER

F MANTLE

[ILLUSTRATION]

277. PLAN OF ST. GALL. BIRD'S-EYE VIEW OF THE MONASTERY

RECONSTRUCTION BY FIECHTER-ZOLLIKOFER [1936, 405, fig. 2]

A great advance over the reconstructions offered by the classicists (figs. 266, 267 and 274) and the first attempt to interpret the guest and
service structures of the Plan of St. Gall in the light of an existing building tradition—wholly workable in structural terms, yet too dependent
on local alpine conditions (log construction) to be applicable to a document worked out in the heart of the Frankish Empire and conceived to
reflect more general conditions.

the structure in any further detail, but judging from an
interior view of the dining room (fig. 281) which he published
in a subsequent book,[42] he appears to think of the
inner wall partitions as likewise being built as solid masonry
walls.

KARL GRUBER, 1937

The same year that Völckers published his pictorial review
of the history of the German house, and probably independent
of both Völckers' and Fiechter-Zollikofer's proposal,
Karl Gruber published still another reconstruction of the
Plan, in bird's-eye view, in a superbly illustrated book,
entitled Die Gestalt der deutschen Stadt (fig. 282).[43] Like
Fiechter-Zollikofer, Gruber reconstructs the St. Gall house
as a house with low pitched roof and straight gable walls on
the narrow sides. It receives its light from windows in the
supporting walls and emits the smoke of its hearth through
a louver in the ridge of the roof. The latter, as in Völckers'
reconstruction, is rendered as a miniature roof, raised above
the level of the main roof to protect the opening over the
hearth site. Gruber is not specific about the material used
in the construction of his houses. The uniform mode of the
rendering of the walls suggests that he thought of them as
being built in masonry.

V.1.3

A RENASCENCE
OF THE CLASSICAL SCHOOL

ALAN SORRELL, 1966

The interpretations of Fiechter-Zollikofer (1936), Völckers
(1937), and Gruber (1937) represented a radical departure

[ILLUSTRATION]

278. PLAN OF ST. GALL. OUTER SCHOOL

[Interpretation by Fiechter-Zollikofer, 1936, 407, fig. 6]

The St. Gall house is here reconstructed as a low-roofed, low-walled
gable house, of log construction, the center room receiving its light
through a tapering wooden shaft mounted upon the ridge of the
roof (see fig. 279).

from the views of the classicists by discarding all reference
to Roman or Gallo-Roman house construction. In 1965
this trend was reversed by a dramatic reconstruction painting
published in a beautifully illustrated book, The Dark
Ages, under the general editorship of a distinguished
Byzantinist, David Talbot Rice.[44] The painting (fig. 283)
carries the signature of Alan Sorrell and is executed in the
flamboyant chiaroscuro that characterizes the hand of this
great interpretive draftsman to whom we owe so many
other impressive reconstructions of medieval and Anglo-Roman
buildings now in ruin.

Unluckily, the scholarship that accompanied this drawing
is not commensurate with the skill of its draftsmanship.
The Church with its steep proportions and its elaborate
blind relief of pilasters and arches looks more like a
Romanesque cathedral than a Carolingian monastery

church. The design of the houses for the animals and their
keepers which occupy the tract to the west of the Church,
conversely, is a return to the superannuated concept of the
courtyard house, proposed by Keller in 1844 and by Lenoir
in 1852. The majority of the guest and service structures
are rendered as buildings of basilican design, in conformity
with the views expressed by Rahn in 1876, Schlosser in
1889, and Oelmann in 1923-4. Sorrell was obviously not
familiar with any of the reconstruction drawings published
by the opposing school (Fiechter-Zollikofer, Völckers, and
Gruber), nor for that matter with a reconstruction of my
own published by Poeschel in 1957 and by myself in 1958.[45]

[ILLUSTRATION]

280. PLAN OF ST. GALL. EXTERIOR PERSPECTIVE
HOUSE FOR DISTINGUISHED GUESTS

[as interpreted by Völckers, 1949, 34]

Völkers reconstructs the St. Gall house—correctly in our opinion—as a
steep-roofed structure hipped at the narrow ends, its center room lighted
by an opening in the ridge that is surmounted by a small protective roof.

[ILLUSTRATION]

281. PLAN OF ST. GALL. INTERIOR PERSPECTIVE
HOUSE FOR DISTINGUISHED GUESTS

[as interpreted by Völckers, 1949, 18]

The large center hearth on the Plan is termed "locus foci." Along the
walls are tables and benches where visitors and their servants take meals.
The rooms under the hipped portions of the roof are heated by their own
fireplaces.

His reconstruction gives the impression of being based on
the description of the Plan published in 1848 by Robert
Willis, who recognized that Keller's interpretation of the
St. Gall house as a courtyard house conflicted with the
inscriptions of the Plan, but did not relinquish this interpretation
in the case of houses for animals and their keepers.

V.1.4

SUMMARY AND CONCLUSIONS

Thus, by the sixth decade of this century the discussion
of the guest and service structures of the Plan of St. Gall
still remains stalemated by two opposing schools. The
classicists had tried to explain the St. Gall house in the
light of a presumptive Roman or Romano-Etruscan type
of house that antedated the Plan of St. Gall by over a
millennium, and whose existence they could not really
prove. On the other hand, the house types on which the
reconstructions of the opposing school were based reached
no further back than the sixteenth, or at best, the end of
the fifteenth century. The proponents of neither of these
two contending approaches were able to demonstrate that
the type of building they had in mind was actually in use
at the time the Plan was drawn. And the great cultural
alternative that Henning had raised in 1882—the suggestion
that the guest and service structures of the Plan of
St. Gall should be interpreted in the light of northern
building tradition rather than in the light of classical
Roman architecture—still loomed as an unsolved problem
over the entire controversy.

Oelmann had summarized this condition correctly in
1923-4, by stating, "The entire quibble about the Northern-Germanic
or Southern-Roman derivation will only be
decided once the existence of layouts that correspond
exactly can be proven in either one or the other area."[46]
However, he added to this statement a passage of questionable
validity when he amplified it with the remark, "The
North is totally excluded, for neither are any contemporary
house remains preserved which would be worthy of mention,
nor is it possible to infer from later specimens earlier forms
of an identical type."

Doubtful in 1923-4, Oelmann's latter idea had become
untenable by 1971 (when this chapter was written).
It is true that at the time of Oelmann's writing research
into the problem of the Northern house was still in its
infancy. Nevertheless, some significant discoveries about
transalpine house construction in the Middle Ages had
already been made in Sweden and on the islands of Gotland
and Iceland.[47] To be sure these were few and scattered;

[ILLUSTRATION]

282. PLAN OF ST. GALL. BIRD'S-EYE VIEW OF THE MONASTERY

RECONSTRUCTION BY KARL GRUBER (1937, 25, fig. 15)

This handsome reconstruction was published in the same year that Otto Völckers published his reconstruction of the St. Gall house (figs. 280-81).
In structural terms Gruber's concept is as workable as that of Völckers and Fiechter-Zollikofer (figs. 277-278), but the conjecture that the guest
and service buildings would all have been constructed in masonry in a part of the world where houses were by tradition built in timber (see
pp. 23ff) is unconvincing.

Very interesting and historically defensible, but not supported by the Plan itself (see I, 163ff) is Gruber's conjecture of a tower over the
intersection of nave and transept. Abbot Haito's church at Reichenau (I, fig. 117) and the abbey church of St. Riquier (I, fig. 196) had such
towers.

[ILLUSTRATION]

283. PLAN OF ST. GALL. BIRD'S-EYE VIEW OF THE MONASTERY

RECONSTRUCTION. PERSPECTIVE BY ALAN SORRELL (Rice, 1965, 279-280)

The design of the houses for animals and their keepers (foreground) is based on the superannuated concept of the courtyard house, proposed by
Keller in 1844 (fig. 264). Most of the other houses are interpreted as basilican structures, in conformity with suggestions made by Rahn in
1876 (fig. 266), Schlosser in 1889, and Oelmann in 1923-24 (fig. 274). Sorrell was not familiar with the writing and reconstructions of scholars
who interpreted the St. Gall house in the light of northern building traditions (fig. 277, 278, 280).

but the numerical lack of strong, convincing archaeological
material was made good to a considerable extent by the
availability of a substantial body of literary and textual
references to house construction which had been touched
upon as early as 1882 by Rudolf Henning[48] and was discussed
at length in 1902-3 by Karl Gustav Stephani's
comprehensive treatise on the German dwelling.[49]

During the last three decades this material has been
enriched by a veritable flood of archaeological discoveries
bearing upon the problem. As I propose to deal with this
material at length in a separate study, I shall review it here
only to the extent necessary for the typological identification
of the guest and service structures of the Plan.