The Pantheon: Unanswered Questions

by Lothar Haselberger, appearing in Volume 40

Author’s Note: Thomas—you made us see classical architecture with new eyes. May your vigorous inspiration stay with us!

The Pantheon, as we know (or don’t know) it, has been the object of puzzlement for several centuries. And it continues to haunt us with the uncomfortably nagging questions it still poses. This is a most welcome situation—as long as we admit the predicament, and do not gloss over it. Matters are complicated, and much is unresolved.

Even with its unanimously accepted name, Pantheon, we are living on borrowed terms by drawing on a strange hybrid from the Middle Ages. The ancient name was Pantheum in Latin, and Pantheion (Πάνθειον) in Greek. Principally, we do know this; we just have difficulties accepting it. And there are other pieces of oft-repeated wisdom that we should not accept, such as continually assuring ourselves that the Pantheon was erected by Hadrian during the period AD 118-125, because we do not know this with precision. In fact it remains perfectly unclear that it was Hadrian at all who initiated the building.

In a way, our refusal to admit to the open questions surrounding the Pantheon maintains the willfully blind and destructive mindset of the eighteenth century that could not bear any longer the “problematic” interior order. Frustration boiled over in view of a nagging question still unresolved today: the blatant misalignment of almost all interior axis lines, along with a puzzlingly flat and spindly upper order where the pilasters were ostensively unable to live up to their loadbearing task, thus leaving the rotunda’s entire dome without a plausible base. As well-known and in plain sight today, the problem was “resolved” in 1756 by hacking off the offending pilasters and revetment and replacing an inconvenient ancient reality with neutral stucco panels.

What follows is a list of—still—unresolved issues: my wish list of questions that I believe we must keep open for further investigation, hoping, of course, that new documentation may bring us closer toward solutions.

Curvature in the Façade?

First, I would like to know whether the “show side” of the Pantheon, its pronaos façade, features an intentional curvature, that upward bending of all seemingly horizontal lines from the stepped platform to the entablature. The issue is almost a century old.

In 1912, William Goodyear claimed such a curvature for the front entablature of the pronaos (a vestibule at the front of a classical temple, enclosed by a portico and sidewalls), based on modern visual technology: a photograph, taken at a flat angle across the façade at the height of the entablature, shows a distinct regular upward curve under the pediment’s horizontal cornice as it exists today. However, subsequent scholarly discussion of the complexities of curvature demonstrated that a deformation of this sort is not a reliable indication for the intentionality—and thus the originality—of ancient curvature.

Much rather, technical indicators had to be found: measurable deviations from right angles and straight lines cut by the ancient stone masons in order to create and accommodate those curves. For example, as Nikolaos Balanos pointed out in 1925 with regard to the Parthenon, the existence of a “donkey’s back” on top of the capitals delivers tangible proof that the architraves described a polygonal line closely approximating the building’s curvature and are, in fact, themselves cut with oblique angles so as to follow that curve.

Yet, while G.P. Stevens measured the Pantheon columns with painstaking accuracy in order to document their entasis curves in 1924, neither he nor anyone else has ever put the question of the Pantheon’s curvature to the test. We urgently need specific documentation here—one that does not merely rely on a documentation of the status quo, but draws on a detailed examination of the stone-cutting features in the architectural elements of the pronaos.

Misfitting Elements

Next comes a hugely consequential question that concerns the technical fit of the Pantheon’s dressed stone elements or, more precisely, the sometimes blatant absence of their appropriate adjoinment. Quite a few “misfits” have been observed in the pronaos architecture—where they are interpreted as evidence for a change of plan in the pronaos design.

Yet we can also, and almost regularly, find such misfits within the rotunda. There too, the misconceived alignment and width of the elaborate pilaster capitals and their fluted shafts—well-crafted in themselves but ill-suited to the structure as a whole—is striking. Thus, this quality (or lack thereof) is not restricted to the pronaos, and instead it appears to be an overarching characteristic of the Pantheon’s dressed stone architecture: carefully carved individual pieces, often badly fit into the context of other pieces. Does this seeming madness have any method? Could it be the result of a rigorous and professional division of labor applied by the imperial building yards?

In the case of Augustus’ temple of Mars Ultor, Joachim Ganzert compellingly reached just this result. Hardly any drum of the still--standing columns aligns its flutes properly with those of the adjoining drum. Unlike Greek standard practice, the drums were all completely prefabricated, flutes included, which permitted a vastly increased speed of building, though with compromises in detail. The possibility of deliberate, acceptable “compromises” in quality has not been taken into account for the Pantheon, while it appears to explain best what looks so strange otherwise.

In fact, such misfits emerge as a broader, still unappreciated characteristic of Roman-imperial building yards, in particular when working on grand-scale projects. For the Pantheon, the list of technical shortcomings can easily be confirmed.

In the pronaos itself, the column capitals, deemed to be “the finest Roman capitals of the Corinthian order we know of,” sit on monolithic shafts whose entasis curve is by no means as perfect as Stevens’ “best preserved” shaft might indicate. Even sighting along the column shafts with the bare eye reveals this without measurement.

Markedly deficient appears the entasis curve of the façade column right next to the one Stevens selected; it is one of two front-row columns with an elaborate repair at the top of the shaft, where it shows a strangely flat (if not actually concave) contour. Furthermore, the perfection of the capitals contrasts stunningly with the irregular spacing of the modillions of the entablature, with random variations amounting to about half the width of a modillion (or about 12 cm over an axial average distance of 81-82 cm). And as long observed, there is no vertical concordance in the modillions of the pediment, both in themselves and in relation to the column axes.

Why the Misfit

Imperial haste, a temporary shortage of material, or the lack of skilled manpower might explain this evidence. But one should also consider whether the various shortcomings in the Pantheon’s cut stone architecture were part of a characteristic picture—one in which “messy” solutions in detail form an intrinsic ingredient in the planning and building process.

Despite masterly stone-carving, a lack of coordination, especially in joining operations, can be observed. No clear seams separate the perfect from the less-than-perfect results. Rather, this phenomenon emerges as a recurrent feature throughout the structure and, while not confined to certain locations, affects lines of juncture in particular.

It is, for instance, unlikely that the same subtle-minded designers and stone-masons responsible for the capitals also produced the cornice blocks with their irregularly spaced modillions. Also, neither the workmen carving the pilasters (in both the pronaos and the rotunda) nor the site-supervising architect seem to have been aware that the pilasters were too wide for their capitals.

Despite the isolated perfection of the finely sculpted capitals and masterfully fluted pilasters, their combination is awkwardly mismatched. This is not a matter of hasty workmanship since the fine quality achieved in the capitals alone amounts to a multiple of the work required to fix the lack of alignments and flawed connections—had they been known, or anticipated, at the right time and place!

Division of labor, prefabrication, standardization, stock-piling, surplus production, long-distance trade in Roman imperial cut-stone architecture as an innovative building strategy: it is along these lines, I think, that we first have to understand the puzzling contradictions in workmanship and design of the Pantheon (or the Temple of Mars Ultor, for that matter). In the wake of J.B. Ward--Perkins’ seminal work, this trait in Roman imperial building has recently received new attention within Rome’s metropolitan perimeter through the marble yard studies of Patrizio Pensabene and Martin Maischberger.

It has been recognized that the prefabrication of cut-stone elements critically increased in the periods after Augustus toward ever more specialized trade and processing patterns ranging, in the course of the second century AD, from major changes in the operating systems of imperial quarries to a notable circulation of semifinished products and an “internationalization” of local stylistic forms. Yet we still have to consider the emergence of new, disconcertingly “un-Greek” building strategies, such as the use of pre-fluted column drums in a major Augustan temple, in conjunction with the Pantheon.

A Calculated Design

Even then, our understanding of the Pantheon’s misfits may still not be complete. The lack of concordance in the pronaos modillions (between horizontal and raking cornice as well as to the column axes) may well lie in the conscious artistic decision to abandon a vertical alignment in favor of a more fluid treatment, a development acutely observed by Torsten Mattern for the post-Augustan and Flavian periods in Rome.

Most important in this context is the stunning avoidance of alignment in the Pantheon’s interior orders, as documented in detail by Marco Pelletti and M. Fosci. Adding to the disjointed organization is the rotunda’s square-based pavement design, which leads to drastically truncated floor patterns at its perimeter. So far, no one has interpreted these mismatches as results of a change of plan, with an embarrassed Hadrian accepting yet another flawed design.

Rather, attempts must be made to explain such “disconnects” as a calculated design strategy. Analyzing the numerical relationships in the divisions of the interior orders and the cupola, Gottfried Gruben arrived at a system of rationally staggered proportions based on the number twenty-eight of the cupola division as leit-motif.

In view of the consistent discordances of the interior zones from floor to cupola, Mark Wilson Jones has sensibly described the rotunda’s composition as weaving “a magical dance around a syncopated, almost jazzy, rhythm.”

In this context, we will have to finally see the flat and thin attic pilasters. The problem is not their visual “denial” of a load-bearing role, but our foisting such a role upon them. The cupola was not supposed to appear to rest on them; instead, with the actual flow of forces consciously veiled, the dome defies the impression of weight.

Traditional perceptions, as reinforced in the façade of the Pantheon, were fundamentally reversed in its interior. The weightless “syncopation” of the domed space reveals itself as a momentous step toward metaphysical architecture. By literally plastering over the most offensive, and at once most telling, part of this endeavor for lack of scholarly explanation, one missed a culminating point in the development of classical architecture.

The Tympanum Lacuna

My next wish leads us to the Pantheon’s pronaos pediment, as carried out, and more specifically to the huge blank pedimental field it presents today. There exists no specific and sufficiently detailed study of the holes and cuttings in the tympanum, many of them no doubt ancient, but perhaps not all of them. All the more confidently repeated is the reconstruction of the pediment decoration with eagle, wreath, and ribbon—which, as is well known, essentially rests on the Trajanic relief built into the portico of Santi Apostoli church.

The best published documentation of the pediment, along with the tympanum cuttings, remains Achille Leclère’s elevational drawing of 1813, published as a needle-sharp heliography by Hector d’Espouy in 1910. Yet even this fine documentation omits, spectacularly, the massive ancient stone projection readily seen in the center of the tympanum (and deemed by some to have served as support for the hypothesized central eagle).

The admirably drawn traditional visualization by Sheila Gibson, authoritatively presented in Ward-Perkins’ Roman lmperial Architecture and widely accepted, has now even more suggestively become “petrified” in John Burge’s computerized reconstruction of the Pantheon. Yet the only correct description of the state of affairs has been spelled out by William Lloyd MacDonald with laconic precision: “The fastigium sculpture is unknown.”

We urgently need a thorough documentation and study of the Pantheon’s pronaos tympanum, in order to fill this monumental, and blatant, lacuna in our knowledge of Roman art. A visual documentation alone, even if superbly precise, will not do. It must be accompanied by a direct, eye-to-hand examination in the established tradition of Bauforschung (building research).

The Alleged Portico Fiasco

Finally in my list of unresolved questions comes the much-debated “compromise” in the design of the existing Pantheon façade. According to Wilson Jones’ argumentation, the existing portico is “second best” compared to what was really intended to be built. And this is the reason for the flawed connections between pronaos, transitional block, and rotunda, which are tangible in a whole list of architectural misfits, most visibly in the two stacked, interfering pediments of the pronaos and transitional block.

The claimed original design of the Pantheon façade is supposed to have featured a taller front architecture utilizing fifty-foot column shafts, so that the height of the pronaos pediment coincided with the second pediment in the transitional block. The existing Pantheon façade becomes the result of a remarkable, in fact regrettable, imperial compromise. Since the emperor could not receive the fifty-foot monolithic column shafts in time, he had no choice but to give in and settle for the second best with all the (alleged or real) infelicities that came with this monumental “portico fiasco.”

Did the emperor in the end, by having Agrippa’s name inscribed on the portico, “not want to take the credit for a flawed outcome?” Only by “mentally restor[ing] a taller portico,” we are told, will we truly gain “a Pantheon consonant with the thunderous climax of Roman architecture that is Hadrian’s legacy.”

Yet despite the almost universal applause this hypothesis has received, I wonder whether accepting this dramatic solution and “redesigning” the Pantheon as it should have been might not prevent us from a subtler understanding of what is. Many of the structural-architectural misfits used to buttress the argument of a design fiasco can, and to my mind should, be explained in view of the now-emerging “non--classical” traits of Roman-imperial building mentioned above.

Further Questions

Not trying hard enough to understand these might obscure, rather than clarify, our view. Does the accessory pediment of the Pantheon’s transitional block have no parallel in ancient architecture? Do we really have to accept the point that “no ancient building copies this arrangement?”

In the case of classical propylaia on the Athenian Akropolis, the double pediment is considered an integral element of Mnesikles’ pioneering creation. The unprecedented “break” in the axis of this building in line with its five-doorway wall defines the momentous division between the exterior and interior, between the profane and the sacred. Athens’ paradigmatic gateway with its double pediment could well have stood as the model for the gateway to the Pantheon.

Of more direct importance is the younger sibling of the Pantheon in the Greek east that has, strangely enough, been completely excluded from the discussion: the Temple of Zeus Asklepios in Pergamon, built during the AD 120s-130s in direct response to Rome’s Pantheon.

Enough pieces of its frontal architecture survive to prove beyond doubt the existence of two staggered pediments. It remains unclear only whether two pediments were indeed, as shown in the published reconstruction, so neatly separated or whether they might not have intersected, just as in the lamented Pantheon compromise.

The immediate impact of the Pantheon in Rome on the design of the Pergamene “Pantheon” is undisputed.  Should we then assume that Hadrian promulgated the imperial building defeat in the capital and imposed a similar deficient solution on a prominent Pergamene sanctuary—declaring beautiful, as it were, the emperor’s new clothes?

Still, we have to ask why later ancient reinterpretations of the Pantheon in the vicinity of Rome, such as the Tor de’ Schiavi and the Mausoleum of Maxentius (both from around AD 310), do not show the “compromise solution” of the existing Pantheon façade.

Was it, from the distance of two centuries, finally acceptable to realize an uncompromised design—one that echoes the postulated original design of the Pantheon façade and appears, in our eyes, as the clearly preferable solution?

For the comparatively small Tor de’ Schiavi mausoleum such a façade design is indeed documented (while not preserved anymore), yet still deemed gravely deficient. As for the Mausoleum of Maxentius, reaching about half the size of the Pantheon and thus a closer comparison, there is a different explanation.

The relevant parts of the Maxentian Mausoleum are a free reconstruction, with no specific elements of the superstructure found at all, and we prefer this solution for the same reason that a classical-minded Thomas Jefferson wished to see it at his design for the rotunda of the University of Virginia. In the spirit of an age that plastered over the Pantheon’s disconcerting interior order, Jefferson improved on the ancient model, thus avoiding the puzzling double pediment and its alleged “disastrous appearance.”

The Essential Question

The essential question that arises from this discussion is whether, by today’s subtler level of research, we do ourselves—and the realities of history—a favor by evading aspects of the Pantheon that we cannot find acceptable. The easy acceptance of the “portico fiasco” narrative might prematurely close a file that should, in my mind, remain open. We risk plastering over the uncomfortable realities of the Pantheon yet again.