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]
(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.