JAPANESE MINKA XXVII - BEAM FRAMING

After examining the foundations of the minka, then the floor framing, then the wall and post framing, we are now ready to move onwards and upwards to consider easily the most complex, arguably the most interesting, and easily the most beautiful part of the minka structure: the roof structure, or koya-gumi (小屋組). Literally translated, koya means ‘hut’ or ‘small house’ and kumi 組 means ‘set’ or ‘assembly’ and the name is apt: because most minka roofs are pitched very steeply, they are usually the most dominant presence in any minka, both structurally from within (assuming there is no ceiling) and visually from without, making up a far larger proportion of the elevational area of the building than the almost unnoticed walls sheltered beneath their deep eaves. In a way, the roof of the minka is the house.

The thatched roof dominates the exterior elevation of this relatively modest minka.

The minka roof structure consists of various sub-elements: the roof covering, the principal rafters (sasu 扠首) and/or common rafters (taruki 垂木 or 棰), the underpurlins (moya 母屋 or yanaka 屋中), the penetrating ties (nuki 貫), the various short posts (tsuka 束) that transfer loads from the underpurlins to the roof beams (koya-bari 小屋梁), and the roof beams themselves. We begin this subject by looking at the beam framing (hari-gumi 梁組), a huge topic in its own right.

The main role of the beam framing is to resist horizontal forces (from wind and earthquakes) by tying together the heads of the perimeter and internal posts, and so preventing any twisting, warping or leaning of these elements.

In simpler, smaller-span roofs, the beams do not receive vertical roof loads directly; they act purely as tie beams to prevent these opposing walls from spreading apart under the thrusting action of the principal rafters (sasu 扠首). The transverse tie beams and the sasu together form a strong triangulated or ‘trussed’ structure. This type of roof is known as the sasu-gumi (扠首組).

In larger roofs, or where the dead or live roof loads are large, the beams also receive point loads directly from crown posts and other short posts that stand on them, and transfer these loads to the posts below. This type of framing is known as wa-goya (和小屋, lit. ‘Japanese roof’).

Sectional diagrams illustrating the differences between sasu-gumi (top) and wagoya-gumi (bottom) roof framing. In the former, the principal rafters (sasu 扠首) bear on the beams as closely as possible to their ends at the wall posts and wall beams (keta 桁). The sasu carry the underpurlins (yanaka 屋中), which in turn carry the common rafters (taruki 垂木 or 棰). In the latter, the beams support various posts (tsuka 束) at intermediate points; on these posts are carried the underpurlins (moya 母屋) and ridgepole (munagi 棟木) which carry the common rafters.

Japanese carpenters of the past were of what might be called the ‘paperweight school’: the thinking was that by making the roof structure as heavy as possible, and thereby pushing the whole house down from above, a more robust structure could be achieved. Remember that minka were not fixed to their foundations, as today’s houses are with steel anchor bolts and such; rather the posts simply sat on their foundation stones, with nothing to stop the building from lifting or shifting in a typhoon or earthquake, other than its own weight. The general custom was to make beams larger than necessary, and to stack them up in double or triple layers, even where the expected stresses were not particularly great from the point of view of preventing twisting and racking in the walls, especially where there were few, small openings in the perimeter.

Beams were typically pine (matsu 松), left more or less as ‘trunks’ in the round, either only with the bark removed (oni-gawa-mugi 鬼皮剝ぎ, lit. ‘ogre skin peeling’), or worked with an adze to form roughly flat surfaces between rounded corners (tsuma-kawa-mugi 爪皮剝ぎ) . The tsuma-kawa 爪皮 is the toe covering of geta, the traditional wooden sandals of Japan. More elaborately, the sides of the beam might be worked back to achieve a consistent width (taiko-otoshi 太鼓落とし, lit. ‘drum reduction’), or the beam might be shaped into an octagonal profile (hachi-men-dori 八面取り, lit. ‘eight side taking’).

When using bent or twisted timbers as beams, these timbers would be oriented to take advantage of the bend to form an arch, with the ‘rise’ (mukuri 起り) or ‘back’ of the beam oriented upwards, which is both structurally stronger and visually ‘correct’.

A magnificent example of wa-goya roof framing, showing beams running in both directions, arched beams, stacked beams, supporting posts, penetrating ties, principal rafters, underpurlins, and rafters.

VERNACULAR PICTURES 17: CORRUGATED LANESCAPES

The central and iconic role played by corrugated iron (of course, what is commonly referred to as ‘corrugated iron’ or even ‘tin’ has for a long time been corrugated steel, and has never been tin) in the vernacular building of Australia is well recognised; I have even seen Australia referred to as ‘the spiritual home of corrugated iron’. But what comes to most people’s minds when they think of this material is rural farm buildings isolated in paddocks, or the painted and galvanised iron roofs of inner-urban terraces and worker’s cottages, or the modern elevation of corrugated iron into an ‘architectural’ material by Glenn Murcutt et. al. Less well recognised is the use of corrugated iron in the sheds and fences lining the back alleys and laneways of older Australian cities and towns.

Back lanes present an entirely different picture to the well-manicured gardens and orderly weatherboard and brick houses that front the long, narrow blocks of old Australian neighbourhoods. The ‘lanescape’ is a jumble of weathered colours and textures, jungle-like growths of trees, plants and weeds, gravel and cobble paving, bits of machinery, dumped appliances, and building detritus. Often the iron used has been taken from the roof of the main house when it was no longer fit for that purpose, so there are mismatched sheets, flaking paint, lap-lines, rust holes, etc. Graffiti and stickers are common modern additions.

These lanes in a sense represent the purest kind of vernacular architecture, since they are completely unselfconscious, completely artless, and their ‘design effects’ are completely unintentional.

Sadly, due to the immense development pressure exerted on traditional residential neighbourhoods by Australia’s insanely high rate of migration and attendant population growth, backyards are being relentlessly infilled with bland unit developments, with their back fences and sheds replaced by treeless, charmless straight lines of colorbond. As the facades and streetscapes of old neighbourhoods are often heritage-protected, these serene frontages are truly a ‘facade’ in the sense that they conceal the destruction and loss of the ‘inner’ character going on behind.

Rear of a modern unit development with ubiquitous ‘fence profile’ colorbond fence.

JAPANESE MINKA XXVI - DAIKOKU POSTS 2

As the largest and most important ‘principal posts’ (yaku-bashira 役柱) in a minka, daikoku-bashira (大黒柱) will naturally be located fairly centrally within the structure, which could hardly be otherwise when you consider that the reason they are regarded as ‘principal’ in the first place is because they gather roof loads over a wide area. There may be more than one ‘principal post’ in a single dwelling, and whether or not these posts are considered to be daikoku-bashira, or some dialect variation thereof (as we will see there is a bewildering variety of alternative names), or some other type of ‘principal post’ entirely, seems to be a somewhat arbitrary or nominal matter, or at least to a large degree a matter of regional convention. Nor is there any specific structural role or configuration that one can point to as definitive of a daikoku-bashira, or as distinguishing or not distinguishing a daikoku-bashira from another type of yaku-bashira, other than that the size and structural importance of both set them apart from the ‘regular’ posts in the dwelling.

In a regular four-room (seigata-yon-madori 制型４間取り) layout, with the partitions between the raised-floor rooms arranged in a cross, there will usually be a daikoku-bashira at the intersection between the two rooms adjacent to the earth-floored doma and the doma itself; there may also be one at the centre of the cross where the four raised-floor rooms meet, and one in the doma. These principal posts are usually mortised to receive tenoned beams or lintels on all four of their sides; the roof loads are transferred via these beams into the posts. Because the beams usually all enter the post at the same height, the cross-sectional area of the post at this height is greatly reduced. To compensate, the post must be ‘super-sized’ well above that of a regular post so that even after mortising it is still strong enough to carry the loads expected of it. These posts are typically 20 to 30cm square, but posts of 50cm or more are not uncommon. Furthermore, whereas small-section posts are usually of conifer species such as Japanese cedar (sugi 杉, Cryptomeria japonica), Japanese cypress (hinoki 桧, Chamacyparis obtusa), or various species of pine (matsu 松 Pinus sp.), daikoku-bashira are usually made from stronger hardwoods such as Japanese chestnut (kuri 栗, Castanea crenata), Japanese zelkova (keyaki 欅, Zelkova serrata) or oak (kashi 樫, Quercus sp.), although in older houses large pine posts left in their natural trunk-like forms can also be found.

When placed in a line of smaller, normal-sized posts, such large posts will of course be much wider than the sills and headers tenoned into them. On the tatami-floored zashiki (the raised-floor ‘living’ area of the minka) side of the post, the corners of the oversized post will intrude into the corners of the tatami mats abutting it. One early solution to this was to make the tatami mats in a special shape, with a corner notched out to accommodate the post. This solution seems to have been disfavoured, giving rise to the inverse practice of instead taking a notch out of the post itself at the level of the tatami, and slotting a normally-shaped tatami into it. Alternatively, where the daikoku-bashira bordered the doma, a later solution was to shift the post off-axis towards the doma, so that the face of the post on the zashiki side was aligned with the sill and rail, though this could mean that the post might be loaded somewhat eccentrically. In any case, the method by which this problem was solved is one of the clues available to the architectural historian in trying to determine the age of a minka, or at least in establishing the general era in which it was built.

The daikoku-bashira is much wider than the sills and beams tenoned into it. In this case, the solution adopted is to shift the daikoku-bashira off its axes towards the doma, so that the faces of the post on the zashiki side are aligned with the inner edges of the sills, allowing the tatami mats to remain regular rectangles in shape. This method suggests that the minka is of relatively recent construction.

The plans below illustrate some of the different positions daikoku-bashira can take within different types of minka.

Three examples of the different positions daikoku-bashira can take in different minka types. The first plan is a regular four-room layout (seigata-yon-madori 制型４間取り) with three principal posts: one at the intersection of the four raised-floor rooms, one at the edge or the raised floor and the earthen-floored doma (どま), and one in the doma at the corner of the stable (umaya 厩). The second is a ‘divided ridge’ (buntо̄-gata 分棟型) layout, with a single daikoku-bashira towards the rear of the earth-floored ‘cookhouse’ (kamaya 釜屋) near the oven/stove (kamado 釜土), indicated as two circles inside a rectangle. The dashed parallel vertical lines indicate the box gutter between the two roofs. The third is a ‘Yamato ridge construction’ (yamato-mune-zukuri 大和棟造り) layout.

As mentioned, daikoku-bashira are most often, but not always, located at the boundary of the doma and the raised-floor ‘living’ rooms (‘A’ in the first and third plans above). There are many dialect variants for the name daikoku-bashira, including daigoku-bashira (大極柱 lit. ‘great-most post’), tatezome-bashira (建初柱 lit. stand-first post, ichiban-bashira (一番柱 lit. ‘number one post’), ichino-daikoku (一の大黒 lit. ‘number one daikoku’, and teishu-bashira (亭主柱 lit. head/master/husband post’. In the Chūgoku region, the daikoku-bashira enshrines the oven/stove (kamado 釜土) deity (kami 神) Dokujin (土公人), so it is called the rokkū bashira (ロックウ柱).

The ‘B’ posts that stand in the centre of the raised-floor section of the minka are variously called uwa-daikoku or ue-daikoku (上大黒 lit. ‘upper daikoku’), yokoza-daikoku (横座大黒 lit. ‘side sit daikoku’), ushiro-daikoku (後大黒 lit. ‘behind daikoku’), chо̄ja-bashira (長者柱 lit. ‘rich man post’), miyako-bashira (都柱 lit. ‘capital post’), ko-daikoku (小大黒 lit. ‘small daikoku’) ebisu-bashira (エビス柱 lit. ‘Ebisu post’; Ebisu is another of the Seven Gods of Fortune), naka-bashira (中柱 lit. ‘middle post’), etc.

The posts marked ‘C’ in the plans stand in the doma and ‘look across’ at the ‘A’ posts. They are often called niwa-daikoku (庭大黒 lit. ‘garden daikoku’; niwa is an alternate name for doma), shita-daikoku (下大黒 lit. ‘lower daikoku’, or umaya-daikoku (廐大黒 lit. ‘stable daikoku’; other regional variations are mukau-daikoku (向かう大黒 lit. ‘facing daikoku’), nirami-daikoku (睨み大黒 lit. ‘glaring daikoku’), ko-daikoku (小大黒 lit. ‘small daikoku), ebisu-bashira (蛭子柱 lit. ‘Ebisu post’), ushimochi-bashira (牛持ち柱 lit. ‘cow holding post’), etc. When the doma is deep front-to-back, there may be a second principal post in it, in the vicinity of the ‘fireplace’ (kajiba 火事場) near the back doorway (sekoguchi 背戸口); this post is variously called the mizu-daikoku (水大黒 lit. ‘water daikoku), kama-bashira (釜柱 lit. ‘oven/stove post’, or kо̄jin-bashira (荒神柱 lit. ‘storm god post’).　

A daikoku-bashira (here called a niwa-daikoku 庭大黒) standing in the middle of the doma (here called the niwa 庭). The beams tenoned into it support at least two mezzanine levels.

View of the cooking area of a doma. In the foreground is the niwa-daikoku, towards the rear is the mizu-daikoku or kama-daikoku. These posts support a mezzanine floor (chū-ni-kai 中二階) between them for storage.

In the ‘divided-ridge’ type (buntо̄-gata 分棟型) minka (where the living and cooking functions of the dwelling are separated into two structurally distinct entities, each with its own roof ridge) of the Tokai (東海) region, there is no daikoku-bashira in the living structure or ‘main house’ (omoya 主屋) part of the minka; that title is reserved for the principal post in the rear part of the ‘cookhouse’ or ‘stove house’ (kamaya 釜屋). In the minka of Yamato Kawauchi (大和河内), the ‘C’ post, called hanakami-bashira (はなかみ柱) is larger and more esteemed than the ‘A’ post. Some say hanakami-bashira is an alternate pronunciation of kamakami-bashira (釜神柱), but the name is more commonly thought to be a holdover from a time when living and cooking quarters were separated, as in the the aforementioned example of the buntо̄-gata minka, where the daikoku-bashira is in the cooking structure, and the post was in the rearward or ‘upper’ (kami 上) part of the ‘edge’ or ‘peripheral’ (hana 端) house, that is to say the cookhouse; therefore hanakami-bashira can be written 端上柱.

Image of a ‘divided-ridge’ type (buntо̄-gata 分棟型) minka showing the ‘passage’ area of the doma where the main house to the left and cookhouse to the right meet; above is a box gutter linking the eaves of the two roofs, supported by a short stump on a beam tenoned into the daikoku-bashira standing next to the oven/stove (kama or kamado 釜 or 釜) in the ‘cookhouse’ (kamaya 釜屋). Here the daikoku-bashira is called an ue-bashira or uwa-bashira (上柱). Note that there is no principal post in the main house. This layout corresponds to the third plan in the examples depicted above.

JAPANESE MINKA XXV - DAIKOKU POSTS

In the Japanese pantheon, Daikoku-ten (大黒天), or simply Daikoku (大黒), is the god of wealth and guardian of farmers, and one of the ‘Seven Gods of Luck’ (shichi-fuku-jin 七福神). He is associated with the continental deity Mahākāla (who is both the Tantric Buddhist protector of the Dharma and the manifestation of the Hindu god Shiva), and also with the native Shintо̄ kami Ōkuninushi (大国主). Daikoku also gives his name to the daikoku-bashira (大黒柱) or ‘daikoku post’, a term which originally denoted a pillar or pole that enshrined the deity, but at some point took on the more purely architectural definition it has today: the most important and largest post in a traditional Japanese house.

A massive daikoku-bashira at the boundary of the doma and the raised floor, with deep beams tenoned into it on all four sides.

In minka, the daikoku-bashira is usually located at the boundary or point of intersection between the earth-floored doma and the front and rear raised-floor rooms adjacent to the doma; these are all ‘social spaces’ and thus relatively large, so the daikoku-bashira must bear the load from a large area of the roof, transferred to it via the various roof beams spanning these spaces. The daikoku-bashira is thus considered to be the heart of the house, structurally, symbolically and even spiritually, as an atavistic reminder of the timber poles used in ancient Japanese religious ceremonies, of a time when the ‘material subject’ was pre-eminent and central in Japanese architecture and was served by the space around it, before the later evolution of a pure ‘architecture of space’ (a progression traced by Inoue Mitsue in his book Space in Japanese Architecture, perhaps a good topic for a future post).

A man squatting next to a daikoku-bashira, giving a sense of its dimensions. Various beams with their long, deep tenons are also shown.

As discussed in a previous entry, in early minka posts were spaced about one ken (approx. 1.8m) apart and roof loads were evenly distributed over these posts. ‘Special’ posts of unusually large dimensions were rarely required, and even where they appear, they were not normally given any special status or name. Later, when modes of occupying the minka became more sophisticated and there arose the need for conveniences such as being able to comfortably pass between two rooms or use them as a single space, these closely-pitched posts came to been seen as a nuisance, and the desire to remove them motivated important advances in joinery (sashimono 指物), notably the invention of the long, deep lintel beam (sashigamoi 差鴨居), tenoned at either end into posts and with grooves or tracks cut into its soffit (underside) to receive sliding partitions, whether of the paper-covered lattice (shо̄ji 障子) or opaque variety (fusuma 襖). The sashigamoi combines the functions of the shallow, grooved ‘header’ or ‘head jamb’ known as the kamoi (鴨居) and the beam (梁) into one member; put differently, the kamoi was greatly deepened, transforming it into a beam capable of carrying the load of the roof structure. In this way, clear spans (and thus clear openings) of two ken (3.6m) or more were achievable. Incidentally, some degree of deflection or ‘creep’ (sag) in the sashigamoi did not cause the sliding partitions to bind in their tracks, as the grooves in the sashigamoi are cut much deeper than those in the corresponding tracked sill (shikii 敷居) in the floor. This extra head space is necessary to be able to remove the partitions, which is done by lifting the partition up into the ‘pocket’ so that its lower edge clears the tracks in the shikii and can be swung outwards; the partition can then be freed by lowering it out of the sashigamoi.

A deep sashigamoi spanning two ken between rooms, tenoned into a daikoku-bashira on the left. The tracks cut into the sashigamoi to receive sliding partitions (removed) are visible in its soffit.