‘Vertebrate buildings: the architecture of structured space’, by W.G. Howell. Part III

Part three of Bill Howell’s 1970 talk, originally published in the RIBA Journal (vol. 77, no. 3, March 1970) and reproduced with the kind permission of RIBAJ. The second part can be found here. GF

*                                *                                 *

I want now to consider for a few moments what is involved architecturally in the choice of a structure. Most buildings consist of a set of flat floors, held up in space and forming the ceilings of the rooms below, which need to be supported at certain intervals by walls or columns or both. These supports can hold up a flat slab or one differentiated into deeper and shallower elements, this being usually a more economic solution. The elements that go to make up the slab will be either linear or planar. True planar elements can only be made up out of an isotropic material (one with the same properties in all directions): for instance, concrete, plywood or steel plate. In many structural systems there will be a hierarchy of differentiated members – boards on joists on beams, say, each successive group being laid counter to the one below.

The choice of a system will be influenced by a number of factors – eg, the requirements of the cellular subdivision of the building, and the impingement, if any, of the partitioning on’ the structure, and of the services. Should the architect wish to display the hierarchy of parts of which the structure consists, then he must decide whether or not the building is one where a linear, directional arrangement is appropriate or not. It may be so – say, in the case of offices ranged either side of a linear internal route. If, however, the architect decides that the spaces in the building, or the relationship of the building to its site, make a directional structure inappropriate, then it will be necessary to devise a system with equal, or at any rate similar, characteristics in two directions.

Curiously enough, we do not have, as far as I know, convenient accepted terms for describing these types of structural anatomy. I have battered my brains and my dictionary, also several dons and one classicist vicar, in an attempt to dig up an appropriate set of terms. The vicar jumped straight in and said ‘Quite simple: monothreptic and diathreptic – growing in one direction and growing in two directions’. Now I know engineers are well educated chaps, but I can just imagine saying down the telephone, ‘Well, actually I thought the beam arrangement ought to be more, well, kind of diathreptic, you know.’ So let us avoid being too fancy, and stick to Anglo Saxon three–letter words, and talk about one–way and two–way structures – and, I suppose, three–way and many way structures.

One-way structure
One-way structure

A one–way structure is one whose basic set of parts has certain characteristics in one direction and different ones in the other, and which therefore adds together with its sideways neighbours in a different way from a neighbour fore and aft. The two–way structure has a basic set whose characteristics are the same along the x and the y axis , though it need not actually be square; it can be distorted into a rectangle and still have the same characteristics, though not the same dimensions, along the two axes. It will abut similar neighbours in the same way on all sides.

‘Well, actually I thought the beam arrangement ought to be more, well, kind of diathreptic, you know.’

Two-way structure
Two-way structure

This will have little architectural significance if the structure will in due course disappear from view. Indeed, a one–way structure may be clad with a suspended ceiling of square (ie, two-way) acoustic tiles, which may relate to a two–way grid of moveable partitions. If the structural form is going to remain on show, however, it becomes the architectural form, and the choice of a one or two–way structure is an important decision. For instance, the building may be sited on a corner. A two–way anatomy goes round a corner, one might say, automatically, but when a one–way structure comes to a corner, all it can do is stop. You can then start next to it, running along the other street another one–way structure. They become neighbouring one-way structures, with or without a gap between them.

If the external anatomy of the building bears some relationship to the structural happenings within, it is likely that the external form of the building will bear a similar relationship to the two streets if there is a two–way structure: but probably a one-way structure will present, literally, a different face to the two streets. This may be entirely appropriate, since one street may be much less important than the other. But often it is entirely unjustified, and results in all those thoughtless blank–ended matchboxes we see at the junction of two equivalent thoroughfares. This corner problem illustrates how in our view the choice of a structural system relates even to the town–planning context of a building. While thinking about structural economy, the relationship of internal partitioning to downstanding beams, the relationship of cladding to the structure, and so on, you are taking decisions which affect the relationship of the anatomy of the building to its site and to its neighbours.

As well as choosing between a one-way and a two–way structural system, the architect must decide on the location of his columns in relation to the enclosing skin of the building. The columns can be outside the skin or in the line of the wall, or the whole column can be inside the building. Columns in the line of the wall cause all those well–known problems of cold bridges and tolerances, creating a sort of hiccough in the planning grid, and perhaps of ending up rather lost , if not entirely so, both from inside and out, because the thickness of a column and a wall these days are often not dissimilar. We have always had a sneaking regard for columns set within the space (ie, inside the enclosing skin), provided this can be done without sterilising useful space, because if the intention is to create an architecture whose form consists of structural enclosures, these become even more apparent and communicative if you are living with your support system; they create that room within a room effect, of which John Soane was the pastmaster.

‘Vertebrate buildings: the architecture of structured space’, by W.G. Howell. Part II

Here’s the second installment of Bill Howell’s 1970 talk, originally published in the RIBA Journal (vol. 77, no. 3, March 1970) and reproduced with the kind permission of RIBAJ. The first part can be found here. GF

*                                *                                 *             

Before glancing at the Renaissance, let me cease to be insular and look back at Greece and Rome. Greek temples when in use, tended to have somewhat perfunctory and ill–lit interiors where the kind of sophisticated architectural use of structural forms we see on the outside played little part. As we see them now, of course, exploded (literally in one case) and eroded by the elements, the local inhabitants and acquisitive peers, they have become stark structural enclosures of space, Stonehenge–fashion, and as such are the most moving and satisfying artefacts – but in a way never envisaged by their designers.

Roman engineering is, of course, all Structure, and where it enclosed space, it did so in a highly vertebrate manner. A lot of surviving Roman interiors have had their architecture removed by generations of popes and other predators. and to us are all the better for it. They too have become structural enclosure – of a kind not envisaged by their designers, but very splendid to those of us who like being enclosed by engineering. To paraphrase the immortal bard, Roman building’s much improved once it’s had its tiles removed.

“Roman building’s much improved once it’s had its tiles removed”.

Interior renaissance architecture, at any scale other than the small domestic, consisted of space articulated by structural forms. They were, very often, fictive structural forms: ie, the column, the pilaster and the entablature were used as a set of images to define the enclosure of the space, but were not always – indeed not often – its actual structural system. Using the word ‘fictive’, rather than ‘false’ or ‘phoney’ as the architectural moralists did , enables us to see why we can enjoy and admire this sort of architecture even though one may be a vertebrate to the backbone. One can enjoy fiction as well as enjoying real life. The baroque is not just fiction – it is poetic drama: if you don’t like this, you won’t like opera. Its forms are not only non–structural but often anti–structural – and deliberately so. In its religious application. it sought to create an other–worldly enclosure that did not just conquer gravity but looked as if it had actually freed itself from gravity. (Ronchamp, still to me the one really magical modern building, is anti–structural in just this way.)

Baroque space is articulated by fictive structure — actual structure is not part of the message.
Baroque space is articulated by fictive structure — actual structure is not part of the message.

When Gothic forms returned to supplement late renaissance ones, they appeared first as fictive Gothic – a system of articulation of interior space pretty remote from Gothic engineering. It was of course, the ‘hard’ Gothic boys who wanted to do what we like doing – enclosing the space by actual structural engineering (I assume they chose the title, thereby leaving their adversaries no alternative but to be ‘soft’).

And even more than the hard Victorian architects, the cast–iron men enclosed space with structural devices, as demonstrated at the Palm House at Kew, at Paddington Station , at the Oxford Museum, or at the late lamented Coal Exchange, and in the great mills and tiny pissoirs and bandstands. Japanese architecture, ancient and modern, appeals to us largely because it consists of an integration of a system of structural members with the planning system of the whole interior.

Transverse section of the Crystal Palace at Kew
Transverse section of the Crystal Palace at Kew
Japanese architecture
Japanese architecture

In the early days of the modern movement, there was a lot of talk about ‘structural honesty’ and ‘expressing the structure’, but curiously enough this seldom took place in the interior. For example, consider Corb’s Maison Domino, the most perfect ideogram of structure as a basis for the new architecture. But turn it from an ideogram into a building, from a structure into an enclosure, and what do you get? Indeed, what did result in the houses of the period? Usually an interior enclosed by plain walls, uncommunicative of whether they were support elements or not, a plain flat lid, and here and there a white column (or could it be a pipe duct? – which, as we all know, in one celebrated case it was).


When my partners and I were students after the Second World War, the first buildings to fire our imagination were the Hertfordshire schools; they had all the beauty of a simple but sophisticated child’s toy which built up into a series of habitable cubes defined by square columns and lattice beams – out of the Japanese house by Meccano Ltd. Their interiors were gay and friendly and appropriate to the junior citizens who inhabited them, like an assembly of space–age Wendy houses. Of course there were unsolved problems. The grid was too large to be flexible enough. The fibrous plaster column casings looked (as indeed they were) soft. There wasn’t anywhere to put the services. The lattice beams led to problems where partitions came up to the ceiling.

In the early Hertfordshire schools, space is articulated by the actual structural system
In the early Hertfordshire schools, space is articulated by the actual structural system

The outcome was the 3ft 4in grid system pioneered by Bruce Martin, which has proved to be the ancestor of so many subsequent schools systems – the deep floor with suspended ceilings to house services, and the randomly placed columns avoiding partitioning and cladding: the Maison Domino came to life. Dry–built, clean, taut, springy. Even when clad and filled with its clip–together shiny partitions, it retained this exciting feeling of brittle tension and elegance. But something had gone. We thought back to the beautiful silver frame sitting in the field, and longed for it not to have been tidied away. It anticipated the interiors of so many buildings today a world enclosed by acoustic tiles above us and wallboard around us. The wonders of technology have brought us to a world of wallpaper. The principal lesson of Hunstanton was that here was a building that looked marvellous as an unclad frame and also looked marvellous when clad and enclosed – indeed, it looked much the same, because there, both inside and out, was the frame. We are all too familiar with that moment when, with most buildings, the structural engineering is finished and the architecture starts going on, and you know it’s over the hump and will never look good again.

‘Vertebrate buildings: the architecture of structured space’, by W.G. Howell. Part I

Bill Howell (1922-74) didn’t write much — HKPA said that they preferred making buildings to making propositions — but the few texts he left bear close examination. The following article, based on a talk he gave on 20 January 1970 at the Royal Institute of British Architects, is something like a credo. It was published in the RIBA Journal (vol. 77, no. 3, March 1970) and is reproduced here with the kind permission of RIBAJ. I’ve broken up the text into six instalments — posthumous guest posts if you will. Howell’s illustrations are included with their original captions where possible, and the original spelling and punctuation have been maintained. GF

*                                *                                 *                                

When I was asked to take part in this series, I discussed with my partners whether or not we should attempt to devise a sort of group sermon – chanted in unison, perhaps, in order to demonstrate that in so far as we have a philosophy, it is a group one. However, since on the last occasion when we exposed ourselves in public (at the AA four years ago*) we did more or less that, we felt that to make such a meal of it again might prove indigestible as well as repetitious; so though I may seem to appear without visible means of support, remember that anything I say has been quarried from the sedimentary deposits of endless dialogues and conversations, not only among ourselves but with structural engineers; and to some extent quarried not from verbalised attitudes, but by looking back at what we have done and wondering what it is, if anything, that characterises it – and why.

An embryonic suggestion of my theme appeared in our AA talk. On that occasion we confined ourselves to half a dozen issues where we felt we had attempted some contribution. One of them was our attitude to structure, and it is this I want to develop here. My theme is a certain kind of architecture. It represents a minority party in modern architecture, though I’m glad to say it has a most respectable ancestry. Briefly it is architecture in which the interior volume is defined and articulated by actual, visible structure. Perhaps the term ‘Vertebrate Architecture’ comes somewhere near identifying what I mean. The only danger of using such an expression is the suggestion that everything I am not talking about is by definition spineless, and therefore no good. But in fact I like lots of architecture, old and new, that is outside the scope of what I am here defining.

Reyner Banham, in his Architecture of the well–tempered environment, has identified the contribution of the services engineer to 20th century environment and architectural form. In so doing he has given the M&E boys a deserved niche in architectural history. I can’t presume in a short paper to do anything so grand for the structural engineers; and in any case, compared to a hi–fi historian at full volume, I am but a tweeter. All I can hope to do is to raise my hat in passing to the structural engineer, and to try to show that when his imagination and skill and those of the architect combine to create a visible (as opposed to concealed) result, a specific sort of architecture happens.

Let me start with a brief glance at history. This is not my selection of the in–buildings of the past, implying that all the others are wicked. I’m trying to define without preaching ie, to avoid doing a Ruskin. The vertebrate buildings are, of course, the ones he would have approved of, and the rest he would certainly have thought sinful. I suppose that standing on the threshold of the second permissive decade, we may now admit that sin can be fun.

“I suppose that standing on the threshold of the second permissive decade, we may now admit that sin can be fun”.

Let us start near home, with Stonehenge – space defined (not yet fully enclosed) by actual, palpable structure. The simplest and most dramatic ideogram of man’s defiance of gravity, a heavy object supported in space so that you can walk underneath it. The intentions behind this amazing prototype work, and the ritual and astronomical purposes it served, could have resulted in other solutions. It would have been easier to define the significant place and align the stars with a set of wooden poles, as the same people did elsewhere on the Wiltshire downs. But here they felt impelled to go out of their way (miles out of their way) to conquer gravity in perpetuity, and devised a system of curved geometry made up from linear elements which is both simple and sophisticated. Whoever the designer was, it was quite likely his first (may be his only) structure, which must make this as remarkable a first building as Castle Howard and the Smithsons’ Hunstanton school.

Stonehenge, like most primitive structures, is among the vertebrates...
Stonehenge, like most primitive structures, is among the vertebrates…

The Celtic beehive tombs are the ancestors of the other basic group of enclosing structural forms – the planar as opposed to those differentiated into linear and infill elements. The enclosing form is the structural form. Linear and planar elements come together in Durham Cathedral. an assembly of structural sets of piers, ribs and vault, enclosing and articulating a spatial continuum. There followed the whole Gothic phenomenon, a stone architecture whose impact on us derives mainly from the directness with which a proliferation of manifest structural sets builds up into a flowing spatial whole.

...as is most mediaeval build ing (eg Durham Cathedral).
…as is most mediaeval building (eg Durham

* ‘Attitudes to architecture, 1’, AA Journal, no.82, pp 95-122.


Model of 1960 development plan for St Anne's College, Oxford
Model of 1960 development plan for St Anne’s College, Oxford. Photograph © Geraint Franklin

My new study of Howell Killick Partridge & Amis will be published by Historic England in 2017. It is part of the Twentieth Century Architects series, published in collaboration with the Twentieth Century Society and the Royal Institute of British Architects.

As I research the book, I’ll be adding all sorts of bits and pieces here.

If you’re interested in my project, do get in touch and follow me on twitter and this blog. Better still, help to make publication possible with a donation! Each author in the series is required to seek external funding of £5,000 to make the title commercially viable for the publisher.

All support is gratefully received, but for donations of £100 or over, you’ll receive a signed copy of the book and a special mention in the Acknowledgements section. For further details please get in touch.

Thanks for reading,

Geraint Franklin