DOUBLE GLAZING DEBUNKED, PART ONE

Insulated glazing unit (IGU) is the industry name for any glazing product that consists of two or more panes of glass separated by a metal or polymer spacer, with the whole assembly forming a thin sealed chamber that contains an insulating layer of air or other gas (typically argon). Insulated glazing was first patented as far back as the 1860s, and IGUs have been commercially available since the 1940s. Though triple, quadruple and even sextuple glazing is available for use in colder climates, double glazing is by far the most common type of IGU seen in Australia, where it has steadily gained market share to the point that it is now arguably seen as the standard choice (at least outside the tropics) in new houses, particularly since achieving first a five-star, then a six-star energy efficiency rating became mandatory in most states in the 2000s. IGU’s themselves are not mandated in the building code, but they are one of the easiest ways to ‘tick the boxes’ in the formal and largely meaningless exercises known as thermal energy assessments (which is a whole other subject in itself). Indeed, double glazing has become somewhat emblematic of ‘green’, ‘eco’ or ‘sustainable’ architecture - feel-good, nebulous and largely sham concepts that generally indicate the uncritical application of energy-intensive, high-tech solutions to perceived ‘problems’ in building design and construction.

But does double glazing work? Well, that depends what you mean by ‘work’. IGUs perform as advertised out of the box, but will they work for the lifespan of your house? Almost certainly not. Lifespans (and warrantees) given for IGUs range from around 10 to 25 years; the failure mode is almost always the failure of the seal, and an IGU is only an IGU as long as the seal retains its integrity. If you look closely at the strip of metal or plastic separating the panes of glass in an IGU, you will see two rows of tiny holes. Under these holes is a layer of desiccant. Once the seal fails, moist air enters the gap, the desiccant eventually becomes saturated, and all you have at that point is two expensive and very closely spaced single-glazed windows prone to internal condensation. If being ‘green’ is your concern, bear in mind that the whole IGU must now be replaced, with all the additional embodied energy that implies.

A sectioned timber-framed IGU showing the desiccant layer (white) under a perforated metal strip

Older, low-tech alternative to IGUs exist that provide much of the insulative benefits of IGUs without the limited lifespan. One very old solution is the use of external storm shutters, but these have the disadvantage of not being able to be used during the day. A more modern solution, common in cold climates from the early 20th century until the advent of IGUs, is just to use two single-glazed openable units in a single frame, separated by ten centimetres or so. While the large gap does mean that there will be some convection of air which will reduce the insulative performance, it also allows the internal faces of the panes to be easily cleaned, and the fact that the cavity is not sealed means that there is no seal to fail - the inevitable fate of all IGU’s in the end.

But perhaps the most fundamental ‘solution’ to this ‘problem’ of heat transfer across windows doesn’t require the application of technology at all, either high or low. Rather it simply requires a change of attitude, which is perhaps why it is almost never mentioned. It requires us to go right back to basics and challenge one of the assumptions that underlies the adoption and perceived necessity of double glazing in the first place: the idea that larger windows are always better and more desirable than smaller.

In next week’s post, we will demonstrate how this ‘no tech’ approach works, by first reviewing the physics of heat transfer and looking at how the insulative properties of materials and building elements are measured and calculated, and then applying this knowledge via a practical example to highlight the influence of window size on heat loss from a room or building.

JAPANESE MINKA VII - FOUR ROOM LAYOUTS

The four room type (yon-madori gata) represents something of a point of completion or fulfilment in the evolution of the minka, having first appeared in the relatively advanced and affluent Kinki region at the beginning of the Edo Period (1603 - 1867), and from there spreading around the country.

In this type, as the name suggests, the raised floor portion of the minka is divided into four rooms; in the paradigm example below, the divisions are in the form of a cross, known in Japanese as the ta-no-ji-gata-madori, ta being the Japanese character for rice paddy, ‘田’. In this example the four rooms are the ‘everyday’ room, here called the dei; behind it the katte for eating; the formal zashiki; and behind it, the heya for sleeping.

In the following examples the rooms have different names, but the functions are the same. In them we can see how the ta form can be easily adapted to meet the ‘weighting’ requirements of the various rooms, simply by shifting one of the lines of partition off centre.

Any later development of the minka beyond the four room type, such as minka with five, six, or more rooms, or minka with multiple wings or other complex plan-forms, is limited to a relatively small number of examples of upper class dwellings rather than types per se, and are thus difficult to fit into any generalising classification system.

LOCAL HEATING

Heating the entire volume of a house or room with a fan-forced convection device such as a split-system air conditioner is a very recent luxury. Before gas and electricity, heating was far more ‘local’ to the body, and was usually achieved with a radiant heat source, be that an open fire, stove, or brazier. Then as now, conductive heating was also employed, and at the most local level possible: by using the heat of the body itself to warm the layer of air trapped between it and clothing or blankets.

In the unsealed and uninsulated traditional Japanese house, there were three main ‘stations’ of heat that the inhabitants used to keep warm throughout the day and night: the kotatsu, the bath (heat by conduction), and bed.

The kotatsu is an excellent example of the kind of evolved emergence and holistic integration of parts that is so often found in vernacular ‘design’. It is a low table with a top that sits loose on the frame; between the frame and top is sandwiched a padded futon (here meaning a blanket or quilt rather than ‘mattress’) which drapes down on each side to the floor and is placed over the laps of those sitting at the table, so enveloping their legs in the heated space created between the floor and the futon.

A modern Japanese kotatsu

In the modern version, the heat source is a small electric space heater attached to the underside of the frame. In the traditional version, the hori-gotatsu or ‘sunken’ kotatsu (presumably evolved from the irori, the hearth sunk into the floor of Japanese ‘living rooms’ in farmhouses and elsewhere), there is a pit sunk into the floor that contains a small charcoal brazier and is covered by a grate flush with the floor to protect the legs. In some cases, there is a pit for the legs roughly the size of the table itself and the depth of the lower legs, so users can sit as if in a chair rather than cross-legged; the brazier is contained in a smaller pit within this pit.

Extended family gathered around a farmhouse irori.

The modern kotatsu (top) and the more traditional hori-gotatsu (bottom).

The key to the effectiveness of the kotatsu is in the clothing of those using it: traditional Japanese clothing such as the kimono are open at the bottom, allowing the heat from the kotatsu to rise up into the space between the clothing and the body; the clothing can also be drawn closed or open at the neck to prevent or allow the heated air from escaping as necessary. The kotatsu also forms the locus of the social activity the Japanese call kazoku-danran: sitting together in a family ‘circle’ to eat, talk, play games, and so on. So the kotatsu can be seen as part of a system, a highly satisfying vernacular solution that integrates not only the function of heating with the furniture and the architecture, but also with the clothing, and even with the manner of social interaction.

A birds-eye view of kazoku-danran around the kotatsu

Similar solutions can be found in the west, though perhaps not so sophisticated as the kotatsu. The high-backed, winged armchair, for example, achieved its form for functional reasons in the days before central heating. When faced towards an open fire, the cupping shape of the chair collects the radiated heat; the high back and wings block cold draughts to the head, and the the arms allow a blanket to be more securely draped over the legs.

ANIMAL ARCHITECTURE

Animal Architecture is a great book by the German ethologist Karl von Frisch, on the subject of (you guessed it) animal architecture. Von Frisch is probably best known for deciphering the dance of the honey-bee; this book is not one of his academic works, but is intended for the general reader. I highly recommend it to architects and designers- not, mind you, as a collection of forms to be turned verbatim into buildings (the world does not need any more spiral shell floor plans or treelike columns, thanks) but as a source of analogues and guiding principles. I don’t have a copy, but have always remembered one line from it on the topic of scale, which was brought to mind today when driving past the edgy new government building that has recently gone up in my town: a monolithic, undetailed monstrosity that completely dwarfs not only the people below it but also the existing buildings around it. The line goes something like: “The hummingbird does not build his nest out of branches, nor the eagle his of gossamer.”

WHAT HAPPENED TO COLOUR?

Is our era the most monochrome in Australian architectural history? Light gray-dark gray-white, and other equally drab exterior colour schemes, have held sway here for years, and show no signs of going away any time soon.

Most people know by now that ancient Egyptian, Greek, and Roman buildings were a riot of colour:

Egptian columns

Reconstruction of a polychrome Greek temple

As were Gothic cathedrals:

Gothic clustered columns

Even Victorian and Federation vernacular buildings, though their builders had only a limited range of relatively subdued natural (and a few synthetic) pigments to work with, seem positively joyous compared to our desaturated modern streetscapes (but good luck finding a house from those periods that hasn’t been ‘refreshed’ to look ‘contemporary’).

Period Federation colour scheme

Probably a big part of the motive here, for both developers and home owners, is the same as that behind the fact that the vast majority of vehicles are white, silver-grey, or black: the desire for ease of resale. Houses are now painted not to present the individuality and taste of their long-term owner to the street, but to be as bland and inoffensive as possible, with one eye to flipping them for a profit a few years down the track.

This is a great pity, especially in the emphatically not-grey country of Australia, where a short walk in the bush will provide you with endless colour ideas, and where you could spend an entire career working only with the palette found on a single parrot or eucalyptus tree.

IN DEFENSE OF (SOME) MODERNISM

As a proponent of traditional design and architecture, I sometimes find myself in the position of wanting to defend the work of certain ‘modernist’ architects against the more strident ‘traditionalists’ on twitter and elsewhere who are as reflexively dismissive of all ‘modernist’ architecture as architectural progressives are of traditional forms. This blanket dismissal suggests to me that these critics haven’t really understood that what makes a building ‘traditional’ in part or whole is the degree to which it displays the underlying principles that constitute the ‘traditional’ in design, and are instead relying on superficial attributes or associations, such as era or style, in passing judgement. I always emphasise that traditional design has nothing to do with historicism or classicism, and that it is perfectly possible to do traditional architecture that is neither.

Traditional architectural principles are broadly hierarchical, and died in stages: first to go was ornament, but lack of ornament isn’t necessarily fatal to a building. Most of the architects of the period of early or ‘high’ modernism, though their work may be shorn of ornament, nevertheless preserved many of the other, arguably more foundational, principles of traditional design that were progressively lost over the following decades: natural materials, a degree of fractal scaling, local symmetries, a careful sense of proportion, plumb walls, rectilinear windows, and so on. Were you to bring them back, most of these architects would be appalled by the sterile, anti-human, parametric horrors of the architect-priests of our own time.

The modern cult of individual creative genius may have been disastrous for architecture as a whole, but that doesn’t mean that such figures don’t exist. And these architects certainly had their failures- the problem with free-floating, intuitive inspiration, as opposed to vernacular or classical design anchored in the communal rules of tradition and so almost infinitely forgiving of mediocrity, is that if the muse deserts you you aren’t left with much. But the best of the work of the best is, to me at least, undeniably beautiful, and represents a self-conscious but successful high-architectural invocation of the spirit of vernacular architecture. You might even, with some justification, call it ‘traditional modernism.’

Alvar Aalto

Alvar Aalto

Alvar Aalto

Gunnar Asplund

Gunnar Asplund

Luis Barragan

Luis Barragan

Jorn Utzon

Jorn Utzon

DESIGN CONDESCENSION

From time to time I come across articles on interior design blogs or in other places where the writer traces the development of a particular aspect of architectural or interior design through its history. In these articles, there is often a faint undercurrent of condescension or superiority, as if to say, ‘haha look at those silly premoderns, luckily we moderns know better.’ This attitude is driven by an underlying assumption of inevitable and endless progress, be it social, material or technological, that confers redundancy on everything that came before the present.

A good example of this is kitchen design. The author will sketch out the history of kitchens, comparing the separated and poky little lean-to kitchens of the nineteenth century unfavourably to the modern ‘open plan’ that is ubiquitous today, and imply bafflement that anybody would have chosen to do it any way other than we do. As an aside, it is stating the obvious to point out that between the two ends of this kitchen design spectrum there are all kinds of in-between ‘semi-open’ design possibilities that allow the best of both worlds, but for whatever reason these possibilities are rarely explored; nor in any case are the eminently rational motives behind the design decisions buried in these old and ‘primitive’ kitchens.

Before electricity and even gas, all cooking was done with wood or coal, and the risk of fire was very real. By separating the kitchen off the back of the house, the risk of a kitchen fire taking out the entire house was reduced, particularly in the case of a brick house where the lean-to kitchen was effectively fire-separated from the main dwelling. Cooking fires also generate a lot of heat, which isn’t necessarily wanted in the rest of the house, especially in an Australian summer.

No electricity also means no mechanical extraction fans, so a separate kitchen was the only way of preventing smoke, soot, oil, cooking smells, and water vapour from permeating the walls and furnishings of living areas.

These are only some of the ‘technical’ reasons for kitchens being the way they were; there are also social factors that I won’t go into here. The point is that the design decisions of past buildings shouldn’t be dismissed as historical or superannuated, but rather taken seriously and even learnt from.

Design, like evolution, has no telos; design features, like the features of biological organisms, simply represent the fittest or best responses to the prevailing conditions of the environment in which they exist. If, as I believe, we are leaving our historically anomalous environment of extreme energy and resource abundance, and re-entering an environment of energy and resource scarcity that is almost beyond living memory in the first world, then we will also witness a reversal of the design ‘progress’ seen by techno-progressives as irreversible, and the re-emergence of many of the design elements, and much of the design wisdom, contained in old kitchens and other spaces.

JAPANESE MINKA VI - THREE ROOM LAYOUTS 2

Last week we examined the three room layouts that evolved within the tatebunwari pattern, where the basic principle of room division is that of transverse ‘columns’ across the dwelling - the room adjacent to the doma (typically called the hiroma) bounds the doma for its full width, and the rooms further ‘in’ are generally parallel to the hiroma and also span the full width of the dwelling. This week we will look at the other subgroup of three room layouts: those that developed from the yokobunwari pattern, where room divisions are longitudinal, and more than one room bounds the doma.

The first subtype of the yokobunwari pattern is called the mae-zashiki-gata 前座敷型or ‘front zashiki’ type. In the example of this type shown below, we have the front zashiki of the title, where more formal or public-facing activities would take place, and also possibly more utilitarian activities in the area of the zashiki bordering the doma. To the rear of the zashiki are two rooms: the doma-bordering daidoko 台所, where eating of meals and other household activities were undertaken. The daidoko might also be used for sleeping. At the most ‘interior’ part of the dwelling is the nema 寝間, used mainly for sleeping.

The maezashiki type, yokobunwari pattern.

The second type is called the tatenarabi sanma-dori 竪ならび三間取り which I will call the ‘row type’ in contrast to the ‘column type’ discussed in the last post. Here the three rooms are arranged parallel to one another so that each borders the doma on their short side. The example below is typical, with again the front zashiki, the middle daidoko, and the rear heya for sleeping.

Tatenarabi sanma-dori type of the yokobunwari pattern.

Analysing these patterns and layouts and contemplating the possibilities inherent to each pattern and type can be a productive exercise for any architect or designer. Without corridors or other distracting auxiliary spaces, they have the purity of architects’ schematic bubble diagrams, but made real; there is an appealing directness and clarity to the functional and spatial relationships they contain.

JAPANESE MINKA V - THREE ROOM LAYOUTS

Further to last week’s post on two room layouts and the two ways in which these rooms can be arranged - the tatebunwari and yokobunwari patterns - I would now like to examine the sub-variations that emerge from these two patterns when they are developed into three room layouts, beginning this week with tatebunwari layouts.

The tatebunwari pattern can be further broken down into two sub-types: the heiretsugata, or what I will call the ‘column type’ layout, and the hiromagata or ‘hiroma type’ layout.

In the heiretsugata type, the rooms are arranged in transverse ‘columns,’ with the ‘outermost’ room fully and exclusively bordering the doma. In the example shown below, this room is called the gozen, typically where meals, family ‘together time’ and handwork would take place; further in comes the omote, for sleeping and other activities, and then the innermost tsubone, for receiving guests and other more ceremonial or formal activities.

A typical tatebunwari pattern minka of the subtype heiretsugata or ‘column’ type.

In the hiromagata type, the ‘everyday’ space (in the example below called the hiroma) again fully borders the doma. Hiroma in general usage simply means a wide or large room; in the context of rural minka it is the ‘general’ room for eating and other everyday activities. The inner portion of the raised floor area is here divided not transversely but longitudinally, into the rear heya (literally ‘room’) for sleeping, and the front zashiki, a formal space for the entertaining of guests, etc.

A typical tatebunwari pattern minka of the subtype hiromagata or ‘hiroma type’.

JAPANESE MINKA IV - TWO ROOM LAYOUTS

In its simplest and probably most common form, the minka is rectilinear in plan, and so a useful way of thinking about the internal partitioning and functional organisation of the minka is in terms of two axes: the longitudinal and the transverse. The transverse axis might be thought of as the ‘front-back’ axis, with the front as the public side, the ‘face’ of the house, ideally the south or sun side, and the back the private, ‘dark’ side; the longitudinal axis might be thought of as the ‘in-out’ axis, with the doma at the public, ‘out’ end and the most private or formal areas at the ‘in’ end. This can be illustrated by the following example of the hito-ma or ‘single room’ minka discussed in last week’s post.

Two room minka are a natural evolution from the single room typology and represent a greater need for functional differentiation and/or a greater level of affluence. Two room minka were still typically found amongst the lower and poorer strata of society, however, and as such they were only required to fulfill the most essential functions of everyday life, with relatively little ‘specialisation’ of spaces, and little need for exclusively formal rooms for activities such as entertaining guests or conducting ceremonies.

The single room layout can be developed into a two room configuration in one of two ways, depending on which axis the ‘room’ in the above plan is divided. In the tatebunwari (竪分割) or ‘transverse partition’ type, the room is divided transversely, so that the doma and the two rooms are arranged in series along the ‘in-out’ axis. In the example shown below, the hiroma 広間 is roughly equivalent to a living room, an every day space for eating, handwork, etc. and also used for sleeping. The zashiki 座敷 is a more formal space than the hiroma, for the use of the master of the house and his guests.

A two room minka of the ‘vertical division’ type.

In the yokobunwari (横分割) or ‘longitudinal partition’ type, the room is divided longitudinally, so that the two rooms are on the ‘front-back’ axis, and each borders onto the doma. In the below example, the nema (寝間) is a sleeping space, but also used for other activities. The omote (表 or おもて) is the more formal ‘front room,’ but not typically as reserved in its use as the zashiki.

A two room minka of the ‘horizontal division’ type.

JAPANESE MINKA III - SINGLE ROOM LAYOUTS

After looking at the ancient antecedents of the minka in the previous two posts - the tateana pit dwelling and the takayuka raised floor dwelling - in this post we will examine the first step in the evolution of the minka proper- the combination of these two archetypes.

Note also that here we will be considering only the subcategory of minka known as nou-minka, the rural farmhouse, and not the better-known machiya, the urban townhouses so characteristic of cities like Kyoto.

To anyone with both romantic and ascetic inclinations, the purity of minka interiors is compelling. Without internal corridors and often without permanent internal partitions, even many ‘multiple room’ minka are still in a sense one-room dwellings, or at least ‘one space’ dwellings, united under a single ceilingless roof.

Interior view showing the roof structure of the Hakogi sennenya, the oldest extant minka in Japan, dated to the late Muromachi era (1336 - 1573)

The vast majority of minka consist of both a doma, the earthen-floored area which contains the dwelling entrance and is used for cooking and ‘utility’ work, and where footwear remains on; and the timber-framed raised floor takayuka, which is generally accessed by ‘going up’ via the doma after footwear is removed. Since the doma is universally present, it can be omitted in analyses of the interior layout, and any count of the number of rooms in the minka does not include the doma. So a ‘one-room’ minka contains two areas that are functionally differentiated, but in most cases not physically divided and so constituting a single space: the doma, which is not considered a ‘room’, and and the takayuka, which is.

As the humblest and simplest of minka, one-room layouts were found all over Japan, often for the use of religious or other ‘retirees.’ This single space happily accommodated all the activities of pre-modern daily life, and, as is the case with single-space dwellings of other cultures around the world, the apparent simplicity of the plan belies the unspoken but well-evolved and sometimes severe conventions that dictate the use of the space, conventions that display what you might call ‘folk rationality.’

In the example from Shiga Prefecture shown below, the doma (here called niwa) is used for cooking, indoor farm-work, and the storage of food and agricultural implements; the threshold area of the takayuka heya (literally ‘room’) adjacent to the niwa is used for taking meals and ‘handwork’; the narrow nure-en or ‘verandah’ along the facade is used for conversing with neighbours and as the entry point for guests, who were received in the area in front of the butsudan or Buddhist altar; the ‘back’ corner of the heya in front of the tokonoma or alcove was used for sleeping. In this informal division of the space by function, we can see the germ of later multi-room minka in which the single room has been partitioned off into three, but the functional relationships nevertheless remain intact.

A hito-ma or one room dwelling showing the functional division of the space.

JAPANESE MINKA II - THE RAISED-FLOOR DWELLING

As discussed briefly at the end of the last post, the tateana-juukyo pit dwellings of the Jomon period gradually gave way from around 2,300 years ago to a new building typology brought from the Asian mainland by the Yayoi people: the takayuka-shiki juukyo or ‘high/raised floor style dwelling’, often shortened to takayuka juukyo. Where the Jomon were hunter-gatherers, the Yayoi were rice agriculturalists; it is likely that the original impetus behind the development of raised-floor structures was the need to preserve the rice harvest from both damp and vermin, and that the earliest of these structures were grain stores rather than residential buildings. But it couldn’t have been long before people realised that the raised floor confers the same advantages to humans as it does to grain.

The typical takayuka juukyo consisted of four or more posts sunk deep into the ground, on top of which was built the elevated floor structure, walls of plank, reed, or clay, and a gable roof of log underpurlins and rafters, topped with thatch.

A raised-floor granary standing next to a pit dwelling, presumably a common sight (though probably with greater fire separation!) in the transitional period before pit dwellings gave way to raised floor dwellings. A much higher level of sophistication is evident in the raised floor structure, in both the structural system and in the dressing and joining of timbers. Note the disc-shaped caps on the posts to prevent rats and other vermin from entering.

A highly refined example of the raised floor typology, with finely worked timbers and close-fitting plank walls.

In the pit dwelling and the raised floor dwelling we have the two antecedents to, and near-universal elements of, Japanese residential architecture up to the present day. Until relatively recently, all Japanese houses consisted of a both a raised floor ‘interior’ of planks or tatami mats, where sleeping, relaxing, eating, receiving guests, praying and the like took place; and a doma: an earth-floored area where all the dirty ‘utility’ activity of the household, including cooking, happened.

An expansive earthen floored doma in a traditional farmhouse, with a raised floor of thick planks beyond.

Modern Japanese houses are almost all raised floor; there are few houses with significant doma and almost none of the ‘slab on grade’ floors that predominate in Australia, as the Japanese building code requires in principle that the finished floor level of habitable rooms be at least 450mm above ground level. The pit dwelling survives only atavistically as the genkan, the ‘sunken’ entrance area to the Japanese home, found even in the tiniest apartments, that functions as a transitional space between outside and the raised floor of the interior, where shoes are taken off before ‘going up’ into the house.

A tiny apartment genkan demonstrating one of its functions: stopping leaves and other debris from going further into the house.

This genkan in a traditional building, with granite paving stones set into a beaten-earth floor, is evocative of its Jomon ancestry.

JAPANESE MINKA I - INTRODUCTION AND PIT DWELLINGS

Feeling ambitious, I have decided to do a series of posts on minka, the traditional vernacular residential architecture of Japan. Minka (民家), literally ‘people’s house’ or ‘folk house,’ is the Japanese word for any ‘common’ or vernacular dwelling, traditional or contemporary, as opposed to both the refined and self-consciously ‘classical’ historical tradition, represented by the villas and tea houses of the aristocracy, residences attached to temples and shrines, and the like, and to modern ‘architectural’ design. In practice, minka is often used more narrowly to refer to traditional residential structures built until the middle of the 20th century.

My main reference in this series will be Kawashima Chūji’s comprehensive three-volume survey and study titled Horobiyuku Minka (滅びゆく民家) or ‘Disappearing Minka’. There is already an excellent English translation of this work, albeit abridged into one volume, by Lynne E. Riggs, but neither the original nor the translation are widely known or cheaply available. Totalling almost 900 pages and heavily illustrated with photographs, sketches, diagrams, plans and sections, the work and its subject certainly deserve a wider audience, so I hope to be able to present at least some of its contents here to anyone who is interested. The three volumes are subtitled: Roofs and Exteriors (Yane・Gaikan 屋根・外観); Internal Layouts, Structure, and Interiors (Madori・Kо̄zо̄・Naibu 間取り・構造・内部) ; and Sites/Auxiliary Structures and Typologies (Yashiki-mawari・Keishiki 屋敷まわり・形式) respectively. I will probably be focusing mostly on the second volume.

Before diving in, however, it would probably be a good idea to lay the groundwork by looking briefly at the archaeological and historical origins of the minka.

The history, or rather pre-history, of minka begins with tateana-shiki jūkyo (竪穴式住居) or simply tateana jūkyo 竪穴住居), the pit (tateana 竪穴) dwellings thought to have first appeared in the late palaeolithic, but more closely associated with the Jо̄mon period (roughly 14th to 1st millenium BC) and surviving into the subsequent Yayoi period (roughly 3rd century BC to 3rd century AD). These structures consist of a pit, round or later rectilinear, with the excavated material often used to form a low wall or berm around the perimeter. The depth of the pit floor to the top of the berm varied by period and region; in cold areas it could be two metres or more. Posts (typically four but sometimes two, three, five, or more) were set into the floor of the pit and tied together with beams, which supported a roof consisting of rafters running from ground to ridge, purlins, and a covering of earth/turf or later thatch.

Modern reconstruction of a tateana-jūkyo (gable entry).

Modern reconstruction of a tateana-jūkyo (side entry).

Reconstruction of the interior of a tateana jūkyo

Cutaway showing the structure of a tateana jūkyo.

An excavated tateana jūkyo pit showing post holes and fireplace, the structural framework, and a reconstruction of the external appearance showing thatching, entrance opening, and smoke openings in the gables.

A series of sections showing the evolution of the tateana jūkyo roof structure, from a simple earth-covered A-frame to a thatched structure with differentiated wall and roof, essentially identical to a modern house but for the sunken floor.

Later, relatively sophisticated examples of the form, showing square plan, hipped-and-gabled roof, ‘chimneys’, and perimeter wall posts and beams which allow the roof structure to be raised clear of the ground.

The Japanese climate is classified as ‘temperate’ over most of its range, which might surprise anyone who has been there in August or February. Builders of houses in most parts of the archipelago have always been faced with the challenge of balancing the competing requirements of hot, humid summers and cold, humid winters, often with significant snowfall. The tateana jūkyo sucessfully addressed many of these challenges. The insulative thatch and the thermal mass of the earth surrounding the pit acted to keep the interior within a comfortable temperature range, around 23 degrees celsius in summer and 20 degrees in winter. The smoke holes in the gables provided effective cross ventilation and exhausted the bulk of the smoke from the fire; at the same time the permeable thatching, while preventing rain from entering the house, also allowed smoke to diffuse from the inside out, which both fumigated the thatch against insects and rodents and preserved it against rot.

Diagram showing the environmental performance of the tateana jūkyo.

Despite these advantages, the tateana jūkyo gradually gave way to the takayuka jūkyo (高床住居) or ‘raised floor’ dwelling, introduced to Japan by the Yayoi people in their migration from the continent beginning in the 3rd century BC, just as the rice agriculture of the Yayoi eventually displaced the hunter-gathering of the Jо̄mon. The takayuka jūkyo will be the subject of the next post.

CEILING HEIGHTS

If you’ve lived all your life in newer buildings, you’re probably familiar with the sense of expansiveness and ease you feel on entering a Victorian or Edwardian house, then noting how high the ceilings are compared to those in your own home. What happened?

Regulation of ceiling heights in Australia goes all the way back to 1810, when, under the Governorship of Lachlan Macquarie, an order was issued to the effect that “no Dwelling-house is to be less than nine Feet high” (this figure probably refers to the ‘pitching height’ of the rafters, which is de facto roughly the ceiling height). Presumably the order was felt necessary because builders and developers were trying to skimp on material costs by building low, and nine feet (2.7m) was settled on as the minimum required to provide amenity to occupants. In the Australian climate, tall rooms have the advantage of being cooler in summer, because warm air will pool near the ceiling, leaving cooler air near the inhabited zone at floor level- the difference can be 5° C or more. In the short mild winters, high ceilings presented less of a disadvantage than they do today, because heating then was radiant- open fireplaces heat surfaces and bodies directly, rather than heating the air of the entire space, as is the case with modern air conditioning systems. Taller rooms also allow for taller windows, allowing light to penetrate more deeply into rooms.

As the prosperity of the colonies grew, so did ceiling heights. In particular, a fall in material costs in the 1860s saw ceiling heights of twelve or even fourteen feet (3.6 or 4.2m) becoming relatively common in the homes of the affluent.

The 20th century saw ceiling heights swing back in the other direction. Following World War 2 in particular, austerity conditions and materials shortages put pressure on building regulations to reflect new economic realities, and the minimum ceiling height for habitable rooms was reduced from 9 to 8 feet (2.7 to 2.4m), where it remains today. This represents a reduction of just over 10% in required wall materials. Taller ceilings may also require taller cornices, skirting boards, doors and windows if they are to remain in proportion. Ceiling lights need to be more powerful or more numerous the further they are from the floor. In two storey houses, increasing the floor-to-floor height means more space and material required for the stairs. So dropping the ceiling can mean substantial savings, and for the ‘marginal’ prospective buyer whose ability to afford a house is borderline, the difference might mean being able to scrape together the deposit for a mortgage on an off-the-plan volume-built house.

While ceilings have dropped over time, the size of the average Australian house has more than doubled since 1950.  One way of interpreting this is that we've sacrificed vertical space for horizontal, and not because families have grown (they’ve shrunk), but to accommodate all our extra stuff. Vertical space is seen as not as useful for this purpose; its value is more intangible, more difficult to articulate, and harder to defend against the material advantages of ‘building out.’

A MIX OF COMPATIBLE MATERIALS

Over recent years, the idea has taken hold among architects and planning authorities alike that building facades need to display ‘modulation,’ ‘articulation,’ and ‘a variety of materials and colours’. The state of New South Wales seems particularly intrusive in this, going so far as to promote and even mandate these notions in its planning schemes. The Hornsby Development Control Plan 2013, for example, contains the following provisions for the street facades of medium density housing developments (basically townhouses and the like):

• Articulation should be achieved by dividing all facades into vertical panels. Wall planes should not exceed 6 metres in length without an offset of at least 1 metre and a corresponding change in roof form.

• Buildings should include structural elements such as sunshades, balconies and verandahs that provide variety in the built form.

• Facades should incorporate a mix of compatible materials such as face or rendered brickwork and contrasting areas of light weight cladding.

• Sunscreens and awnings comprised of timber battens or metal frames are encouraged.

It seems that what is being attempted, albeit in a crude and inchoate way, is the reintroduction of some degree of fractal scaling into the streetscape, although it is highly unlikely that the authors of the planning scheme would have described it in these terms. Rather, town planners probably perceived a need to respond to a creeping featurelessness or blandness of the modern developer-driven ‘builder’s vernacular’ without at the same time going to the other extreme of giving free reign to local architects with pretensions of ‘genius’ along the lines of a Gehry or Hadid. The problem they are faced with, probably insurmountable, is how to reconcile these two aims - the avoidance of ‘monotony’ on the one hand, and the imposition of ‘order’ on the other - within the framework of a modern architectural orthodoxy that regards them as contradictory and antithetical.

Traditional design, which is essentially self-regulating, had solved the uniformity/monotony - variety/chaos problem before it even arose. Traditionally, streetscapes displayed a stylistic and material uniformity and harmony within each building, and a degree of variety across different buildings, but each still bound by the constraints of traditional design and materials; today, on the other hand, we have chaotic variety within each building, and a kind of monotonous but equally chaotic sameness across buildings.

The traditional architect had no problem with a long, ‘flat’ facade plane completely lacking in ‘offsets,’ because he knew he could easily avoid a monotonous or oppressive appearance by effectively articulating it on both a wider and a finer range of scales than is typically seen today - that is, by the use of pilasters, string courses, cornices, window sashes with small panes set in muntins and deep window reveals in thick walls, a variety of brick bonds, material textures, and so on. Steps in and out in the facade are typically on the order of centimetres, not metres. Today we start with a flat facade, consisting of flat window frames set close to flush in a flat wall surface made up of flat panels of flat industrial metal and flat industrial brick, and grossly overcompensate by insisting that this facade be arbitrarily stepped in and out by metres, and that materials, colours, roof pitches, etc. be varied equally arbitrarily and randomly, with the aim of somehow providing ‘interest’. Predictably, the result this is that every duplex or townhouse development is essentially indistinguishable from any other: dutiful use of ‘a variety of materials and colours,’ thick square or three-sided ‘picture frames’ of alucobond or fibre cement around balcony openings and garage doors, upper levels cantilevered out over brick lower levels, glass balustrades, and random skillions.

A particular modern favourite is to use two or more different brick colours in large ‘panels’ in a facade. Look at any old brick building, in contrast, and you will rarely find more than one brick colour used; where you do, there is a clear dominant or ‘ground’ brick colour, and the other is the ‘figure’ employed to pick out highlights at corners, around windows, and so on. The variety is in the service of expressing a structural or functional differentiation. It was understood that buildings need to project a sense of visual unity.

19th century townhouses in Millers Point, Sydney. With no variety of materials or offsets in the facade plane, presumably this ‘design’ would not be permitted today.

Hotel in The Rocks, Sydney. The facade displays fine-grained ‘offsets’, ornament, and a subtle variety of colours and finishes, differentiated rationally and functionally.

TRADITIONAL DESIGN II: UNIVERSAL DISTRIBUTION

This is the second in a series of posts exploring the ideas of the mathematician and design theorist Nikos Salingaros, and by extension his collaborator, the architect Christopher Alexander.

The previous post in this series examined the idea of a universal scaling hierarchy and how it could assist designers in deciding how many scales should be employed in buildings, and what the ratios between them should be.  This post and the next will consider the concept of universal distribution, as a way of answering another important question: how ‘full' should each scale be? or how many elements should each scale in the hierarchy contain? 

Imagine you set out to design a tree. While universal scaling is concerned with how big the 1st order branches should be in relation to the trunk, how big the 2nd order branches should be in relation to the 1st order branches, and so on, universal distribution is concerned with how many branches should be on the tree, and how many twigs, and how many leaves.

A good place to begin exploring universal distribution is by looking at the structure of fractals.  For our purposes, a fractal is any pattern that is generated recursively, and has the property of scaling symmetry: it is self-similar at various scales, meaning that you can zoom in on any part of the pattern and it will look identical or near-identical to any other level of magnification, and will contain the same amount of detail.  Fractals also have the interesting property of fractal dimensionality, but this property is less relevant to our purposes. 

One of the simplest fractals is the Koch snowflake.  Starting with an equilateral triangle, add to it three triangles with sides 1/3 the length of the original (meaning the scaling factor is 3); to the sides of these triangles add nine triangles with sides 1/3 their length; and so on.

The Koch snowflake

Another simple triangle-based fractal is the Sierpinski gasket. In a sense it is the inverse of the Koch snowflake: it is subtractive (‘perforated') where the snowflake is additive (‘accretive'), and' ‘ingrown' where the snowflake is ‘outgrown.'  Also, the scaling factor here is 2, not 3.  Neither of these factors are very close to 2.72, which was proposed in the last post as a good approximation of the universal scaling hierarchy, but this doesn’t matter here: we are using fractals not to prove a point about universal scaling (we could easily create a fractal with a scaling factor of 2.72 if we wanted), but to introduce the concept of universal distribution.

The Sierpinski gasket

The distribution of elements in the Sierpinski gasket is as follows:

0th order scale:  1 element

1st order scale: 3 elements

2nd order scale: 9 elements

3rd order scale: 27 elements

4th order scale: 81 elements

The distribution factor in this case is 3, i.e. each scale contains three times the number of elements of the previous scale.

As with the Fibonacci sequence (also recursive), fractals are everywhere in nature: trees and river systems are familiar examples.  People find these recursive, scale-symmetrical structures inherently pleasing, because they generate just the right amount of information compression in the brain: they are neither monotonous, like a grid of triangles all of the same size, nor chaotic, like a field of triangles of random size, position, and rotation.  The former possesses the necessary quality of order, but has no sense of life; it doesn't contain enough complexity to hold the mind's attention.  The latter produces a sense of anxiety, because the mind cannot derive any rules or patterns from it to reduce the computational load.  The objects and environments that give us the greatest satisfaction occupy a ‘Goldilocks zone' between these two extremes.  They exhibit scaling coherence: they have structures at different scales, a scaling hierarchy, and a high degree of self-similarity at different ‘magnifications.' Neither simplistic nor chaotic, they stimulate the mind without overwhelming it.

How does this apply to buildings?  Consider the baroque facade of Santa Maria in Vallicella, Rome.

Santa Maria in Vallicella, Rome

There are bare areas devoid of detail, and within these areas there are centres of focus where the detail is concentrated, particularly around the parts of the facade where one's attention is naturally directed, such as doors and windows (incidentally, but not coincidentally, the human face displays the same kind of detail distribution: bare areas such as the cheeks and forehead, and smaller areas of concentrated, expressive detail, such as the mouth and eyes).

The facade of the gothic/romanesque Siena Cathedral shows a much denser distribution of detail typical of much gothic architecture - its distribution factor is higher than that of Santa Maria in Vallicella - but even here there is a hierarchical arrangement of blank areas and areas of greater detail.

Siena Cathedral

Now let’s look at some pathological examples of universal distribution, or rather its lack.  Take a textbook example of ‘high modernism,' the Villa Savoye by Le Corbusier.

Villa Savoye

The facade contains only a few large scales; it is almost completely devoid of small-scale details, and the mid-range scales are absent.  The result is flat, barren, and unnatural; the building lacks the characteristics of universal scaling and distribution found in nature.

The lifeless character of ‘white box modernism’ was recognised as a problem, at least implicitly, by the post-modernists and deconstructivists, but in rejecting high modernism without understanding the root cause of its shortcomings, they only fell into another set of problems.  The facades below are representative. They contain only one scale, or perhaps not even that. Are they all detail, or all blank?  There is no hierarchy, only monotony.  In the end, the effect is every bit as dead as the modernist deadness these contemporary architects were presumably seeking to avoid.

21 DESIGN RULES FROM 1855

The following is taken from The Register of Rural Affairs, published in America in 1855.  I think it holds up pretty well :)

1.  Always compare the cost with the means, before deciding on the plan.  It is much better to build within means, than to have a large, fine house, hard to keep in order, and encumbering the owner with a heavy and annoying debt.  A great error with many is an attempt to build finely.  Attend to real wants and substantial conveniences, and avoid imaginary and manufactured desires.

2.  Study a convenient location rather than a showy one: a house on a lofty hill may make a fine appearance, but the annoyance of ascending to it will become greater on each successive day.

3. Build of such good materials as are near at hand.  An interesting index is thus afforded to the resources and materials of that particular region, with the addition of great economy over the use of such as are “far brought and dear bought."

4. Prefer lasting to perishable materials, even if more costly.  A small well built erection, is better than a large decaying shell.

5. Discard all gingerbread work, and adopt a plain, neat, and tasteful appearance in every part.  Far more true taste is evinced by proper forms and just proportions than by any amount of tinsel and peacock decorations.  A marble statue bedizened with feathers and ribbons, would not be a very pleasing object.

6. Proportion may be shown in the smallest cottage as well as in the most magnificent palace - and the former should be carefully designed as well as the latter.  However small a building may be, let it never show an awkward conception, when a good form is more easily made than a bad one. 

7. Where convenient or practicable, let the plan be so devised that additions may be subsequently made, without distorting the whole.

8. More attention should be given to the convenient arrangement and disposition of rooms in constant daily use, that those employed but a few times in the course of a year.  Hence the kitchen and living-room should receive special attention.

9. In all country houses, from the cottage to the palace, let the kitchen (the most important apartment,) always be on a level with the main floor.  It requires more force to raise a hundred pounds ten feet upwards, whether it be the human frame or an assortment of eatables, than the same weight one hundred feet on a level.  To do it fifty times a day is a serious task.  If the mistress superintends her own kitchen, it should be of easy access.  For strong light and free ventilation, it should have, if possible, windows on opposite or nearly opposite sides.

10. There should be a set of easy stairs from the kitchen to the cellar.  Every cellar should have, besides the stairs within, an outside entrance, for the passage of barrels and other heavy articles.

11. The pantry, and more especially the china closet, should be between the kitchen and dining room for easy access from both.

12. The bathroom should be between the kitchen and nursery, for convenience to warm water.

13. Let the entry or hall be near the center of the house, so that ready and convenient access may be had from it to the different rooms; and to prevent the too common evil of passing through one room to enter another.

14. Place the stairs so that the landing shall be as near the center as may be practicable, for the reasons given in the preceding rule.

15. Every entrance from without, except to the kitchen, should open into some entry, lobby, or hall, to prevent the direct ingress of cold air into rooms, and to secure sufficient privacy.

16. Let the partitions of the second floor stand over those of the lower, as nearly as possible, to secure firmness and stability.

17. The first floor of any house, however small, should be at least one foot above ground, to guard against dampness.

18. Flat roofs should be adopted only with metallic covering.  Shingles need a steeper inclination to prevent the accumulation of snow, leakage and decay - more so than is frequently adopted.  A steep roof is, additionally, cheaper, by admitting the use of a less perfect material for an equally perfect roof, and giving more garret room.

19. The coolest rooms in summer, and the warmest in winter, are those remote from the direction of the prevailing winds and from the afternoon sun.  Hence parlors, nurseries, and other apartments where personal comfort is important, should be placed on this side of the house where practicable.

20. Always reserve ten per cent. of cost for improvement and planting.  Remember that a hundred dollars in trees and shrubbery produce a greater ornamental and pleasing effect than a thousand in architecture.

21. Lastly, never build in a hurry; mature plans thoroughly; procure the best materials, and have joiner-work done at the cheaper season of winter, and the erection will be completed in the most perfect manner, and with the greatest practicable degree of economy.

TRADITIONAL DESIGN I: UNIVERSAL SCALING

This is the first in a series of posts exploring the ideas of the mathematician and design theorist Nikos Salingaros, and by extension those of his mentor and collaborator, the architect Christopher Alexander.

introduction

Buildings, like biological organisms, are organized at a number of different scales, from the largest (the overall dimensions of the building) to the smallest (the texture of the sand in a render coat). This fact presents the architect or designer with a choice: he may either explicitly address it, and attempt to answer the questions and challenges it raises in the design process; or he may choose to ignore it and evade its challenges entirely; either way, his choice will be evident in the results.

How many scales should a building contain? What should the ratios between them be? How many elements should each scale contain? And what should the ratios between the number of elements in each scale be? All of these questions point to a deeper issue: why are some buildings so beautiful and seem so healthy, while others are so ugly and pathological?

This and subsequent posts will try to answer the above questions via an exploration of the ideas of Nikos Salingaros, with the aim of outlining a simple and practical methodology that can might be useful to anyone whose interest is in designing buildings that are beautiful and healthy, rather than ugly and sick.

the universal scaling sequence

You are probably familiar with the Fibonacci sequence, where each number in the series is the sum of the previous two numbers, beginning with 0 and 1:

0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377 . . .

From this we can derive another sequence, more relevant to our purposes, called the universal scaling sequence, which is obtained by removing alternate terms from the Fibonacci sequence:

1, 3, 8, 21, 55, 144, 377 . . .

The universal scaling sequence can be applied to architectural design in the following way: assign the arbitrary size of 1 to the largest scale in a building, then to the next scale down in size (in the same linear dimension) assign the value 1/3, then 1/8, then 1/21, and so on.  Or, start at the other end and assign to the smallest scale in a building the arbitrary size of 1, then the next scale up in size should be 3, then 8, then 21, and so on. 

To see how this works, try the following: start with a blank building facade with an overall height of 10m, which is your ‘first-order’ or ‘1’ scale.  From this, the sequence might suggest a floor-to-floor height of 10m/3 = 3.33m, a window height of 10m/8 = 1.25m, an ornamental cornice height of 10m/21 = 475mm, a sill or window frame height of 10m/55 = 180mm, and finer ornamental details at 10m/144 = 70mm and 10m/377 = 25mm.  Now do the same for the horizontal dimensions.  Say the facade is 6m wide: 6m/3 = 2m, 6m/8 = 750mm, 6m/21 = 285mm, 6m/55 = 110mm, and 6m/144 = 40mm.  Using these figures, I drew the facade below in about 15 minutes, letting the scales determine the design without much input from my ‘individual creativity.’

It’s nothing special, but that’s the point- designing in this way is highly forgiving.

Interestingly, if you take any term from the Fibonacci sequence and divide it by the previous term (e.g. 233/144), the number obtained approaches the famous golden mean, 1.618, as the two terms get larger.  Likewise, if you divide any term from the universal scaling sequence by the previous term, the answer approaches 2.618, which is the square of 1.618. However, because the Fibonacci sequence is not a true geometric sequence (where each term is the nth power of the previous term), it is impractical to use it in most design situations.  Nikos Salingaros proposes the natural logarithm e, 2.718, as an acceptable geometric substitute.

The Fibonacci sequence in nature

A logarithmic spiral in nature

These constants are not just arbitrary abstractions: they are found throughout the natural world (hence the name universal), from spiral galaxies to molluscs to the number of petals on flowers.  Their great aesthetic and mathematical appeal sometimes tempts architects into designing golden rectangles into their buildings, citing the (possibly apocryphal) example of the Parthenon, or designing buildings that look like seashells and other organic forms.  These kinds of applications are over-literal and fundamentally misconstrued.  The real significance of the universal scaling sequence is that it provides a useful tool for checking that a building's various scales (the dimensions of building elements as measured along the same axis) are a reasonable approximation to the ‘natural' hierarchy of the universal scaling sequence, i.e.:

1. Few scales of the sequence are missing;

2. There are no significant scales that fall between the terms of the sequence; and

3. The ratios of adjoining scales are close enough to 2.618, or 2.718. 

Of course, real-world considerations mean that the scales in actual buildings will rarely conform to the mathematical ideal. In practice, the ratios are rounded off into rules of thumb, like the vernacular builder's ‘rule of three’: each scale in a building should be roughly three times the size of the next smallest scale, and 1/3 the size of the next largest.  At any rate, the important thing is not strict adherence to the numbers, but understanding the concept of the universal scaling hierarchy as an ideal to aim for.

Adoption of the universal scaling hierarchy has several benefits: it imposes non-arbitrary limitations on design (limitations are good); it guides the designer in making more effective design decisions; and it aids in the diagnosis of design flaws.  If a building or facade feels too busy, for example, it may be because it contains too many scales that fall between those on the universal scaling sequence.  Conversely, omitting scales from the sequence results in a collapse of the scaling hierarchy and a barren, lifeless appearance. 

Inappropriate spacing of scales and collapse of the hierarchy: only a few large and seemingly random scales are present in this facade.

This classical facade presents a full hierarchy of scales, from large (the ‘bay’ or distance between columns) to small (the width of the dentils in the cornice).

Consider another example: a door and its architrave.  If the door is a standard 820mm wide door, universal scaling would suggest an architrave with a width of 820 / 2.718 = 300mm or so.  This might sound excessive, but if the architrave itself is further subdivided (by mouldings, painted or carved patterns, or other ornament) into successively smaller scales in the hierarchy (say 110mm, 40mm, and 15mm), the result is a door with great presence.  Economic realities generally meant that such opulent doors were reserved for classical or civic architecture; in humbler vernacular buildings, architraves were typically around 100mm wide, which skips a scale but is still far more effective in expressing the idea of ‘doorness’ than a modern ‘architectural’ door, which might have an ‘architrave’ as thin as 10mm.  In this case there are three scales missing between the scale of the door width and the scale of the frame width, which is a bit like a tree consisting of a single, massive trunk covered in tiny twigs: our brains cannot ‘span the gap' between the two scales to form a coherent connection between them, and the hierarchy collapses. 

Classical door with ornamented surround

Modern door with "frame"

Given that the human perceptual system evolved in the natural world, where these scaling sequences and ratios are the literal rule, it should not be controversial to suggest that people have an instinctive affinity for correct scaling ratios, and that buildings designed around a universal scaling hierarchy hold an innate aesthetic and emotional appeal for us, as evidenced by the fact that such buildings are found across all ages and cultures, in both classical architecture and vernacular building traditions.  In fact, as Salingaros points out, there are only two significant exceptions to this universality: one being the ‘death architecture’ of Egyptian Pyramids and defensive fortifications, both of which are deliberately designed to be repellent; and the other being modern architecture.

SKILLION ROOFS

A skillion roof or sometimes shed roof is a single-pitch or mono-pitch roof, in contrast to the traditional dual pitch gabled roof, where the two ‘pitches’ slope down symetrically from a central ridge to the longer walls of the building, producing the triangular gables on the shorter walls.

A skillion roof (a) and gable roof (b)

The skillion roof is generally defined as having a pitch (or gradient or fall) of at least 3 degrees or so; roofs shallower than that are usually referred to as flat roofs.

The use of skillion roofs in residential buildings seems to have originated in Australia, with architects such as Robin Boyd employing them as early as the 1950s, but the skillion roof remained largely confined to ‘magazine architecture’ for many years.

Date House by Robin Boyd, 1955

Much of its present popularity, and many of the ‘architectural’ examples of the form from the 1990s on, can be traced to the influence of a single figure: Australia’s defacto architect laureate, Glenn Murcutt, though his influence seems rarely acknowledged (Nemo propheta in patria?)

Murcutt’s skillion roofs are typically clad in corrugated iron, with unlined eaves supported on tapered steel or timber rafters and purlins and sometimes struts, a clerestory of sashless glazing running around the perimeter, and a clear datum separating the clerestory from the walls or glazing below.  The roof runs up to the north (southern hemisphere), the ceiling follows, and the depth of the eaves overhang on that side is carefully designed to exclude summer sun but allow deep penetration of winter light. Shading of the glazing below the datum is accomplished with external adjustable louvres.

Simpson-Lee House by Glenn Murcutt, 1993

But where Murcutt’s skillion roofs - influenced by his love of high modernism, fastidiously detailed, and genuinely functional - bought the form to a higher degree of refinement than earlier examples, the skillion’s later diffusion, first across the architectural world and then ‘down’ into the ‘builder’s vernacular’ to the point that it is now an established element and a common sight in volume-built subdivisions (though it has never come close to supplanting the gabled or hipped roof in popularity), has seen it often reduced to the status of empty stylistic gesture, a lazy, shorthand way of bringing together those shadowy twin concepts of contemporary and sustainable.

Examples at the ‘architectural’ end of the spectrum are often shamelessly plagiarised from Murcutt, but rarely executed with either his aesthetic subtlety or his fine-boned structural clarity. 

This is a winery, not a house, but a good example of what can happen when all you have is “ecologically sustainable outcomes”.

The ‘builder’s vernacular’ skillion has boxed eaves and fascia boards right around the roof: a stumpy, graceless profile.  It may be oriented any which way, and eaves depth is often arbitrary or insufficient.  There might be two or more skillions pitched in different directions on the same building.  There may be no clerestory and the ceiling underneath may be flat.  The skillion here is purely in the service of fashion or style, not function. 

A good example of the dog’s breakfast that is the skillion roof in the ‘builder’s vernacular’.

Though a well-designed and detailed skillion roof can be an effective solution to various environmental or other design considerations, one might still object to the form on a deeper level - call it psychological, or aesthetic, depending on your preference. That is, where the ceiling follows the pitch of the roof, the enclosed space, though dynamic in its asymmetrical upwards ‘loft’, lacks the stillness and serenity desired in a residential space.  The space of the room ‘drains out’ through the clerestory, as opposed to the way it ‘pools’ in the cathedral ceiling, with its obvious metaphors of the inverted hull or cupped hands, or in the flat ceiling, which forms a kind of shoebox lid on the room.  There is something settling about the traditional dual-pitch, symmetrical roof, with each side coming down from a central ridge to ‘cap’ the walls beneath, and in many cases eaves that project out over the walls, protecting them from weather, and if visible from within, serving as a comforting ‘cap-brim’ to the view. 

STEEP AND LOW ROOFS

One of the most characteristic elements of the 19th century Australian worker’s cottage is its roof. Steeply pitched with short spans and therefore low and compact in form, it is perfectly in keeping with the modest volumes it shelters. There are two basic types: either a parallel series of hipped or gabled units, themselves parallel to the street and separated by box gutters; or a U-shaped hipped roof, whose form is not immediately apparent when viewed from the front and sides, but becomes clear when viewed from the back: a box gutter, perpendicular to the street, runs down the middle of the house, separating the two hipped (or occasionally gabled) roofs that form the uprights of the ‘U’.

One explanation given for the emergence of these forms is that the unsophisticated colonial builders had a poor understanding of structural principles: the ceiling joists weren’t tied to the rafters to form a primitive triangulated truss and prevent the rafters from spreading the walls, and so the thrust exerted on the walls by the roof could only be controlled by keeping the span of the roof, and thus its mass, to a minimum. Low roofs with simple rise:run ratios such as 1:1 (45 degrees) or 1:1.3 (a 3-4-5 triangle, 37 degrees) were also easier to construct and required only short rafters.

Aside from these practical and material factors, early builders also no doubt had their aesthetic motivations, and understood very well that low, steep roofs suit these humble cottages perfectly and give them their unique appeal.

On the left: parallel gable roofs separated by a box gutter. On the right: a hipped ‘U’ form roof with an extremely long central box gutter (hidden).

On the left: a ‘U’ form roof with a lean-to off the back. On the right: a parallel series of three hipped roofs separated by box gutters.

A ‘U’ form roof shown from the back, with twin hipped roofs separated by a box gutter