Structure vs. the Expression of Structure

Jonathan Ochshorn

contact | homepage | index of selected writings


In a recent "Reader Poll" on design preferences published in Progressive Architecture, a majority of respondents agreed that "structural systems should be expressed" in works of architecture. But what does it mean to "express" structure? This question is not merely a matter of defining terminology, but goes to the heart of the relationship between the art of architecture and its technological basis. In examining how structure differs from expression of structure, the following propositions are developed: 1) exposed structure does not "express" structure but simply is structure; 2) what structure expresses is not "structure," but a set of culturally-determined beliefs about structure; 3) the expression of structure is different from the actual behavior of structure; 4) the behavior of structural elements and systems is not self-evident; and 5) form is not determined by structure.


Architecture has often concerned itself with questions of expression, a concept first articulated in Classical Greece with the invention of rhetorical devices for speech and writing. Gombrich (4) describes an application of expression to architectural works in the writings of Vitruvius: the Doric order was to be used in temples dedicated to strong-willed gods, leaving the other orders for gods with more peaceful attributes. Considering the entire range of emotions that might be expressed, Vitruvius placed the Doric order at the more "severe" end of the scale. But there is nothing in the Doric column itself that is "severe." It is the context, both the culture in general and the range of architectural choices in particular, which makes such an interpretation possible. Transposed to a different context, the same column could express a completely different idea (Fig. 1).

doric column compared to Loos' Tribune competition

Figure 1. The Doric column expresses "severity" in its Classical context (left); in a different context—Loos' Chicago Tribune competition entry—it can easily express a completely different idea (right). Tzonis and Lefaivre, Classical Architecture, MIT Press, 1986; Gravagnuolo, Adolf Loos, Rizzoli, 1982.

Structural engineering, on the other hand, concerns itself with structure: not expression, metaphor or poetry, but beams, columns and bending moments. Looked at in this way, the architect and the engineer seem to be concerned with quite different things. Of course, the architect must design a building that contains structure; and the engineer must design a structure that supports a particular work of architecture. But aside from that area of overlapping interest, the two disciplines often seem to operate on different ground.

It was therefore surprising to come across a recent "Reader Poll" in Progressive Architecture (8) in which a majority of respondents agreed that "structural systems should be expressed" in works of architecture. Here we have, linked in together in the same sentence, the proprietary concerns of the architect and the engineer: "expression" and "structure." But what does it actually mean to express structure?


The first point to be made in examining this question is that exposed structure does not express structure merely by being exposed: it simply is structure (Fig. 2). What makes a work of architecture expressive, what gives it its meaning, cannot be discovered by examining the physical properties of its beams, columns, walls or slabs. As Juan Bonta (2) has written: "The materials of ...architecture are not stones, any more than the material of literature is ink. The materials of these arts are not inert matter but the creation of man, charged with the cultural heritage of a community—no more, but certainly no less than language."

doric column compared to Loos' Tribune competition

Figure 2. Exposed structure does not express structure merely by being exposed: it simply is structure. Condit, American Building, University of Chicago Press, 1982.

doric column compared to Loos' Tribune competition

Figure 3. In Sullivan's Guarantee Building (Buffalo, NY), the true structural bay is revealed only at the ground floor levels. Scully, Modern Architecture, Braziller, 1965.


The second point to be made is that what we express when we express structure is not the actual structure, but a set of culturally-determined beliefs about structure. For example, Louis Sullivan (10), writing about the potential of the newly-invented steel frame, described its appeal and inspiration as lying in "the suggestion of slenderness and aspiration, the soaring quality as of a thing rising from the earth as a unitary utterance, Dionysian in beauty."

Claude Bragdon, in his foreword to Sullivan's autobiography, credits Sullivan with being "the first squarely to face the expressional problem of the steel-framed skyscraper." A comparison with the actual steel frame shows how Sullivan's expression of structure not only differs from the structure itself, but is a product of a particular mind working within a specific cultural context. As the architect Paul Phillipe Cret (3) wrote in 1925, it is a "conception of steel construction which assumes that in a [structural] system of a girder resting on two posts, the posts have a metaphysical nobility which entitles them to a special magnification."

The subsequent development and expression of steel-framed office buildings shows that Cret's sarcasm was certainly justified. What was characteristic of Sullivan's day gave way to new forms of expression, in spite of the fact that the actual structural frame behind the expressive facade looked more or less the same. Thus, the cornice, a necessity in Sullivan's expressive system, disappeared in later curtain wall construction. Curt Siegel (9), typical of architectural critics writing in the early 1960s, remarked that the simple flat roof with no cornice was "in keeping with the nature of skeleton construction." He added: "The unadorned cubic form appears to be taking hold. At any rate, hardly any of the more prominent buildings erected in the last ten years have prominent cornices." Siegel also objected to the expression of false intermediate columns such as occur in Sullivan's Guarantee and Wainwright buildings, where the true structural bay is revealed only at the ground floor levels (Fig. 3).


When Sullivan described the expressive possibilities of the steel frame, he did so not in the language of statics and strength of materials, but with metaphor and poetry. This brings up a third point concerning structure: the expression of structure is different from the actual behavior of structure. This is because the real behavior of a structural system depends to a large extent on loads applied, types of connections between structural members, material strengths and relative stiffnesses, etc., factors which are commonly hidden from view and hard to evaluate by external observation. Expression of structure, on the other hand, deals with notions about structure that may have only a superficial correspondence to the actual structure and its behavior.

doric column compared to Loos' Tribune competition

Figure 4. S.O.M.'s Hancock Building in Chicago (left) and Stubbins' Citicorp Center in New York (right) both contain trusswork on their outer surfaces. S.O.M. "expresses" this structural system; Stubbins' trusses are concealed by the building's skin. Architectural Record, Oct. 1972 (top left); Ludman, Hugh Stubbins and his Associates, 1986 (top and bottom right); Architecture of Skidmore, Owings & Merrill, Architectural Book Publishing Co., 1974 (Bottom left)

As Mies van der Rohe is said to have remarked in justifying the application of non-structural I-beams to the columns and mullions of his apartment houses at 860 Lake Shore Drive, the real reason was that it "looked right." Even when the expressed structure more faithfully corresponds to the actual structural system, as in S.O.M.'s Hancock Building in Chicago, it is still important to distinguish between the structure and its expression, since they retain their independence and their separate purposes. The articulation of trusswork on the building facade may tell an educated observer something about how lateral forces can be optimally resisted in a high-rise building; but its expressive intent is not so literal nor so limited in its appeal. Rather, the expression is metaphorical, applying a form familiar from other contexts to convey an image of strength, utility, or of technology in general. Presenting this trusswork as a key element of the building's architectural design therefore has little to do with the utility of the real trusses in performing their structural function. Where a different expressive intention is at work, such as in the Citicorp tower in New York, the chevron trusses on the face of the building are simply covered by the building's skin (Fig. 4).


The fact that the expression of structure is different from structural behavior brings up a fourth point: the actual behavior of structural elements and systems is not at all self-evident. Even Galileo, to whom we can attribute more than common intelligence, was unable to visualize correctly the distribution of stresses in a cantilever beam. Without the benefit of Hooke's law, formulated 50 years later, and without the insight of exactly how tension and compression were distributed within the beam's cross-section to equilibrate the external moment, Galileo misrepresented the strength of the beam by a foctor of three (11).

Even understanding the behavior of a simply-supported beam or slab is far from simple. As Rowland Mainstone (7) wrote in his book on structural form: "Their primary structural actions, considered in terms of the principal internal tensions and compressions, are ... among the most complex. The beam is capable of spanning between supports without exerting any horizontal thrusts only by virtue of containing within itself the actions of both arches and catenaries. But all these actions take place beneath the surface and are belied by the simple exteriors."

In fact, all structural action takes place "beneath the surface;" our view of structure is, literally, superficial. We do not see tension in a suspension bridge cable nor compression in a stone column. In beams, the stress trajectories described by Mainstone never appear on the beam's surface (Fig. 5). Even the deformation of structural elements under load is, in general, invisible to our senses. If Siegel (9) complains that the "curtain wall clothes the [structural] skeleton from the outside, obscuring or even totally concealing the structural principle," it would be more to the point to suggest that no structure, however exposed, reveals its "principles" merely by being visible. Galileo's encounter with the cantilever beam should be evidence enough of that fact.

Galileo and bending stress

Figure 5. Galileo's bending test, from Due Nove Scienze, 1638 (left). Because he could not visualize correctly the distribution of stresses, he overestimated the beam's strength by a factor of 3. The actual distribution of combined bending and shear stresses (right) is not at all self-evident. Cowan, The Master Builders, John Wiley & Sons, 1977; Timoshenko, Strength of Materials, Part I, Van Nostrand Reinhold, 1958.

Maillart and moment diagram

Figure 6. Maillart's Salginatobel bridge in Switzerland can be seen as a physical embodiment of its moment diagram; yet even here, structural considerations alone did not determine form. Bill, Robert Maillart, Praeger, 1969.


The necessity for all works of architecture to contain structure brings up a final point: even though it contains structure, architectural form is not determined by structure. Structural considerations may constrain or may inspire the invention of architectural form, but in most cases, the primary decisions about form arise from other considerations.

Of course there are notable structures whose form seems to be actually derived from structural analysis. Both the Eiffel Tower in France and the Salginatobel bridge of Maillart in Switzerland can be seen as physical embodiments of their respective moment diagrams: the tower as a cantilever subjected to lateral forces; the bridge as a 3-hinged arch subjected to gravity loads. Yet even here, structural considerations alone did not determine form (Fig. 6).

First, the most basic decisions concerning the number and types of supports, the materials to be used, the degree of indeterminacy, etc., are not at all fixed. Second, the decisions resulting in a given moment diagram hardly exhaust the formal possibilities for the finished structure. As Billington (1) has written, both Maillart and Eiffel brought an artistic sensibility to the task of creating form that set their work apart from other engineering structures. Third, other non-structural issues play a large part in determining form, even where structure seems to control the design. In the case of Maillart's bridge, for example, the need to accommodate vehicular traffic certainly influenced the overall shape to a considerable degree.

Finally, though a structural solution may seem uniquely appropriate to a given form, it is often the expressive possibilities of the structure rather than any unique merit it may possess as structure, that underlies its appeal. Saarinin, far from being forced into the design of his Dulles International Airport Terminal Building because of structural considerations, arrived at his structural system out of formal considerations. His concern was that a "strong form that seemed to rise from the plane and hover over it would look best...." In fact, as Howard (5) has shown, the forms have little to do with structural efficiency, but are designed from the point of view of architectural and structural expression (Fig. 7).

Having discussed the difference between structure and its expression, we can conclude by returning to the question posed by the P/A reader poll: should structural systems be expressed? The affirmative response of its readers was taken by the editors of P/A as a validation of modernist dogma. That being the case, it seems fair to close with a quote from Le Corbusier (6), whose position on this issue is typical of his modernist peers: while admitting that the revolution in structural materials and techniques required the invention of new forms, they viewed the creation of architectural form as a problem in its own right, separate from, though related to, the techniques of construction. Le Corbusier wrote: "The conclusion has often been drawn that architecture is construction. It may be that the effort put forth by architects has been mainly concentrated on the constructional problems of the time; that is not a reason for mixing different things....Stones are dead things sleeping in the quarries but the apses of St. Peter's are a drama."


  1. Billington, David P., The Tower and the Bridge, Basic Books, 1983.

  2. Bonta, Juan P., Architecture and its Interpretation, Rizzoli, 1979.

  3. Cret, Paul P., letter to Fiske Kimball, 6 May 1925, The Architectural Record, Vol. 65, 5 May 1929, quoted in Bonta (2).

  4. Gombrich, E. H., Art and Illusion, Pantheon Books, 1960.

  5. Howard, H. Seymour, Jr., Structure, An Architect's Approach, McGraw-Hill, 1966.

  6. Le Corbusier, Towards a New Architecture, Holt, Rinehart & Winston, 1960.

  7. Mainstone, Rowland, Developments in Structural Form, MIT Press, 1975.

  8. "Reader Poll—Design Preferences," Progressive Architecture, Oct. 1988.

  9. Siegel, Curt, Structure and Form in Modern Architecture, Van Nostrand Reinhold, 1962.

  10. Sullivan, Louis H., The Autobiography of an Idea, Dover Publications, 1956.

  11. Timoshenko, Stephen P., History of Strength of Materials, McGraw-Hill, 1953.

Saarinin's Dulles International Airport

Figure 7. Saarinin's Dulles International Airport Terminal Building: the forms have little to do with structural efficiency, but are designed from the point of view of architectural and structural expression. Howard (5).