JAPANESE MINKA X - TWO RIDGE MINKA

After covering several varieties of magari-ya in last week’s post, today I will look at another general minka type, the futamune-zukuri (二棟造り), or ‘two ridge’ minka. As the name suggests, these minka are basically two structurally independent buildings; they may be either entirely separate, requiring one to go outside when passing between them, or joined to some extent. They are most commonly found in the southern regions of Japan, from Okinawa to southern and central Kyushu. The two buildings of the whole are the hon-ya (本屋), the ‘main house’ with raised floor for living, sleeping, etc., and the earthen-floored oto-ya or ‘cookhouse’, (lit. ‘pot/cauldron house’ 釜屋); they are roughly the same size, and are arranged either with their ridges parallel (the formal name for this arrangement is 平行二棟式, heikou-futamune shiki ‘parallel two ridge type’) or at right angles to one another.

Example of a ie-nakae secchaku minka with the ridges of each building arranged at right angles to each other.

It’s not difficult to understand why this type of minka is found predominantly in southern Japan. In subtropical climates like Okinawa and southern Kyushu, separating the ‘cooking’ and ‘living/sleeping’ functions of a house into two separate buildings means that heat from the stove can be kept out of the main dwelling; another advantage of isolating the hon-ya from the oto-ya is that in the event that the oto-ya is destroyed by fire, the hon-ya will stand some chance of surviving (depending on the wind direction of course!). The scale of these dwellings tends to be quite modest; another advantage of the two-ridge type might have been that the two buildings can be completed in two stages, as funds become available. And of course, as with any vernacular building form, we shouldn’t overlook the cultural factors behind the evolution and persistence of any particular typology or design element, i.e. “that’s the way it’s always been done.”

The ie-nakae secchaku (イエ・ナカエ接着) is one name for a variant of the two ridge minka in which the two building volumes (here termed the ie and the nakae) touch (secchaku) at their eaves, and are joined below by walls to form a relatively unified interior space. The ie, or sometimes omote, are regional variant names for the hon-ya; the nakae is the oto-ya.

Map showing the distribution of various configurations of ie-nakae secchaku minka on the Satsuma Peninsular, southern Kyushu

Where the two roofs meet, a box gutter is provided. In our era when long lengths of sheet metal are cheap and readily available, box gutters are taken for granted; in pre-industrial times constructing a waterproof box gutter was a much more impressive technical feat, using only split bamboo, or perhaps a hollowed-out half-log.

Ie-nakae secchaku minka, showing the box gutter and infill wall where the two buildings meet. The box gutter is a combination of old and new: the (presumably original) technique of split bamboo cleverly lashed together into a kind of Spanish tile arrangement, with the upper convex ‘capping’ sections of bamboo directing water into the lower concave ‘gutter’ sections; and the modern addition of cheap and timesaving sheet metal to make the old bamboo waterproof rather than replace it.

JAPANESE MINKA IX - L-PLAN MINKA

Last week’s post presented a map showing the general distribution of different types of minka across Japan. Today I would like to look more closely at one of the types included in that map that I haven’t yet covered in previous posts on the subject: the magari-ya (曲り屋).

Magari-ya literally means ‘bent house’; in other words, a minka with an L-shaped floor plan. But what really characterises the magari-ya is its particular mode of occupation, of which the plan is merely the spatial outcome: the cohabitation of humans and animals (typically horses) under one roof. The main volume, the omoya or moya (母屋), is for humans; the ‘stable wing’, umaya or maya (馬屋), is for the animals, and the two volumes are arranged at right angles to one another, with each forming one leg of the ‘L’. The umaya usually extends out southwards from the south facade of the omoya, because this position enjoys the best access to sunlight: an indication of just how valuable horses were to the occupants of these houses.

Magari-ya are most commonly found in the Touhoku region of northern Honshu (the main island of Japan).

A map of the Touhoku region, showing modern prefectures by colour and principal cities.

Perhaps the magari-ya variant most synonymous with the form is the Nanbu magari-ya, found in the Touno district of central Iwate (the area was once the domain of the Nanbu clan, hence the name). Taking the former Fujiwara residence as representative, we see that the main entrance to the building is in the umaya volume, placed centrally and very practically at the inner corner where the two volumes meet, and leading directly into the niwa, the earthen-floored utility area. From this central position one can go right to the umaya (which has its own independent entrance, next to the main entrance), left to the daidoko (kitchen) at the rear/north, or around the corner to the partitioned, raised floor section of the omoya, which contains the ‘living’, ‘sleeping’ and formal areas of the home.

A grand example of a nanbu magari-ya, the former Fujiwara residence, Iwate Prefecture. The photo is taken from the west, showing the umaya wing to the rear right, and the omoya in the foreground. The two dark openings in the umaya volume are the main entrance to the niwa (left) and the umaya entrance (right).

Floor plan of the former Fujiwara residence, showing the stable wing (umaya) at the bottom (south).

Another variant of the magari-ya is the chuumon-zukuri 中門造り, which like the Nanbu magari-ya is closely associated with the Tohoku region, but this time on the western, Sea of Japan side. Characteristic of this variant is that the entry (with horse) to the dwelling is via a door in the gable end of the umaya, into the chuumon entry/passage adjacent to the stable, where the horse is left before proceeding through to the earthen-floored niwa in the main interior space of the omoya.

A thatch-walled chuumon-zukuri magari-ya, the former Yamada residence.

Floor plan of the former Yamada residence.

Chuumon-zukuri with a sculpted thatch umaya: the former Satou residence in Niigata Prefecture.

Floor plan of the former Satou residence

It’s striking how functionally similar the plan of the chuumon-zukuri is to that of the modern Australian garage-fronted house, where you enter in your car, ‘stable’ it in the garage, and then go on through into the house proper via a door between it and the garage. You don’t need to be an architectural historian from the future to point out that the size and position of garages in our homes says the same about the status and centrality of the car in our lives as the position of the umaya does about the importance of the horse to the occupants of the magari-ya.

JAPANESE MINKA VIII - DISTRIBUTION

Previous posts in this series presented the basic categorisation system for minka floor plan layouts put forward by Kawashima Chuuji. While this system is satisfactory in a general sense, and is useful in reconstructing the evolution of the minka from its earliest forms, Kawashima himself emphasises that such a neat taxonomy can never really capture the great diversity of minka seen throughout Japan, according to function, occupation, custom, climate, topography, socioeconomic status, and other variables. And what holds true of the country as a whole also holds true when it comes to attempting to identify distributions and patterns of minka by region, whether that be the regional distribution of any particular typology of minka, or the typology of any particular region. While it is understandable that the most typical or common minka type in any area will be held up as representative of that area, it should also be remembered that even within a particular region there will be many variations on the representative type, as well as other types, oddities and anomalous forms that resist classification. Thus it is impossible to create a really accurate fine-scale map of minka types according to region; such a map will be unavoidably low-resolution. Nevertheless, the exercise isn’t meaningless or futile, because the patterns are there, however messy they might be, and the alternative would be an analysis so granular and microscopic that any sense of them is lost. In any case, here is Kawashima’s own map:

Kawashima’s map showing the distribution of minka types across Japan (excluding Hokkaido)

The map’s legend enlarged

The legend reads:

• 広間型および広間的間取り　Hiroma-gata and ‘hiroma-like’ layouts

• ４間取り系間取り（田字型）Yon-madori kei madori (ta-ji-gata) four room layouts

• 曲り屋　Magari-ya literally ‘bent house’ i.e. L-shaped plans.

• 中門造り Chuumon-zukuri A sub-category of magari-ya.

• 妻入り（前土間、片側住居、本棟造り）Tsuma-iri (maedoma, katagawa juukyo, honmune zukuri) Gable-entry minka.

• 二棟造り（主屋無土間）Futamune zukuri (omoya mudoma) ‘Two ridge’ i.e. two building minka (without doma)

• イエ・ナカエ接着 (Ie-nakae secchaku) A form of two ridge minka where the two building volumes are joined to form a unified interior.

• 踏込み土間型 (Fumikomi doma-gata) ‘Step-in’ doma type minka

The hiroma-gata and yon-madori kei madori have been covered in previous posts. The other typologies shown on the map fit less neatly into Kawashima’s classification system; in the next few posts in this series I will look at them in more detail.

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.

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.