Layouts and Layers - SKKUsjeas.skku.edu/upload/200605/05_Sarvimaki Marja.pdf · Layouts and Layers:...

Layouts and Layers: Spatial Arrangements in Japan and Korea

Sarvimaki Marja*

A b s t r a c t : The aim of this paper is to clarify spatial conceptions ofJapanese and Korean aristocratic residences built during a 250-yearperiod since 1600, in order to interpret meanings of these built forms.The first topics of discussion are the materials that surround thespaces and, second, the elements that are not only space dividers, butalso reflections of social order. Finally, the layouts and spatial layers ofKorean and Japanese houses are examined in terms of linguistic and‘non-verbal’ aspects of the ‘language of architecture’ in both cultures.Throughout this interpretation, special attention is paid to the continu-ous space of Japanese and Korean architecture, in which of mostimportance are the distinctions of multipurpose rooms, ‘inside’ and‘outside’, as well as private and public areas, defined by such featuresas behavior and language, besides architectural expressions.

Keywords: banga, buke, gan, ken, ma, module, oku, shoin, sukiya.

8 0

Sungkyun Journal of East Asian Studies, Vol. 3, No. 2, 2003, pp. 80-108.

© 2003 by the Sungkyunkwan University, The Academy of East Asian Studies

* Marja Sarvimaki has specialized on East Asian architecture. She completedher Master’s in Architecture from the Helsinki University of Technology in1985 and received her Doctor Degree in Science in 2000 from the same univer-sity. In addition to her career as an architectural designer, she has done exten-sive studies and fieldwork on China, Japan, and Korea. She has written booksand articles on East Asian Culture and Society. Dr. Sarvimaki is working as an Assistant Professor at the University of Hawaii at Manoa, School of

P r o l o g u e

For some time, scholars have been ‘unwrapping’ Japan anddemystifying phenomena related to the cultural and ethnic ‘unique-ness’ of the Japanese. Personally, I have found the recent works ofa n t h ropologists in general, and those of Joy Hendry in particular,inspiring in ‘re - reading’ the ‘language’ of Japanese arc h i t e c t u re — afield that, according to a widely held view, is especially unique. Thissentiment, however, is to a great extent based on contrasting Japanand ‘the West,’ conducted by Japanese and Westerners alike. But ifwe compare Japan to other East Asian countries, the point of viewchanges, in a manner of speaking.

In many ways, buildings can be considered ‘packages’ or ‘boxes’(as modern arc h i t e c t u re is often re f e r red to by laymen) which, like agift and its wrapping or a locution, express such relations as polite-ness, power, and status between people and/or diff e rent socialg roups, in addition to aesthetic values. In Hendry’s words: “In theJapanese case, architectural style and the layout of domestic and re l i-gious edifices can be seen to use layers of ‘spatial wrapping’ in theway they enclose their inner sanctums, just as people use layers of‘linguistic wrapping’ to express themselves.”1 She also suggests thatthe Japanese ‘wrapping principle,’ which is a mode of indirect com-munication in one particular cultural milieu, could provide a modelfor cross-cultural comparisons, and, I would like to add, for interd i s-ciplinary studies.

T h e re f o re, this paper is an architect’s attempt to shed new lighton the interpretations of Korean and Japanese languages of arc h i t e c-t u re from a ‘Hendryan’ perspective, in spite (or perhaps because) ofmany diff e rences, as there are also similarities, above all linguistica ff i n i t y. Due to limited space, the examination does not include

Sarvimaki Marja 8 1

A r c h itecture, and is at present working on a forthcoming publication on thephilosophy of Korean architecture.1 . Hendry 1995, p. 4. That is why Joy Hendry is the only person quoted here.

China, nor city layout, gardens, palaces, temples, and farmhouses,or geomancy and other related thoughts, which are importantaspects of East Asian arc h i t e c t u re. Instead, we focus on the spatiallayers and their definitions in aristocratic residences, Korean b a n g aand Japanese s h o i n - z u k u r i (below s h o i n for short), built between theearly 1600s and mid-1800s.2 Besides that these mansions of Japaneses a m u r a i elite and Korean y a n g b a n gentry reflect aesthetic, cosmologi-cal and sociological ideologies of their times, they are good exam-ples of ‘spatial wrapping.’ They are also elucidative objects of com-parison being designed for the upper classes during the same periodin neighboring countries.

Spatial Wrapping Materials

Most pro b a b l y, any discourse on space eventually leads us deal-ing with the elements that surround it, in order to grasp somethingas abstract as the concept of space. As Hendry has pointed out, thematerials of Japanese houses are much the same as in gift-wrapping,and this applies to Korea as well. The main stru c t u re is built of tim-ber and is an object d’art in its own right, like the beautiful woodenboxes, in which the most precious Japanese and Korean artifacts are

8 2 Layouts and Layers

2 . The Japanese s h o i n - z u k u r i is a building type of s a m u r a i houses (b u k e), impe-rial residences, and abbot’s quarters in temples. This style started to devel-op in the twelfth century, but was more common from the seventeenthcentury on, when it often included s u k i y a style semiformal rooms and rusticsoa n teahouses or tearooms. To simplify, s h o i n, also meaning an individuals h o i n style formal room, indicates here the whole house analogously tob a n g a that are specifically the residences of Korean y a n g b a n gentry (the b a nof b a n - g a and y a n g - b a n being the same ideogram for ‘nobleman’) to distin-guish them from m i n g a, or ‘peoples’ houses.’ It must also be pointed outthat m i n g a is not exactly the same as Japanese m i n k a which refers to farm-houses, whereas Korean m i n g a indicates houses in general, including thehouses of the ‘middle classes’ (j u n g i n and i g y o) between the commonerclasses (y a n g i n and c h o n i n) and y a n g b a n a r i s t o c r a c y .

packed. However, while almost all Japanese stru c t u res until theMeiji era were built of wood, by the middle of the Korean Joseondynasty (1392-1910) there was shortage of good quality timber onthe peninsula.3 Wooden buildings such as b a n g a w e re hence ‘statussymbols.’ In Japan, where high-grade timber was available in gre a tamounts, wealth was demonstrated, for instance, by s h o i n style thatwas restricted from the lower classes till the late Edo period (1600-1 8 6 8 ) .

A c c o rd i n g l y, there were diff e rences in the appearance of s h o i nand b a n g a, the most obvious being the winding wooden members ofK o rean houses, contrasting to the extremely straight-lined andrefined Japanese carpentry. During a recent field trip, it struck methat one only has to observe Korean forests with their bush-like tre e sto understand why buildings have more organic details than inJapan. Naturalism, on the other hand, is considered one of the mostcharacteristic design principles of both cultures. In this context, it isnot possible to delve into the culturally variable perceptions ofn a t u re, and I only mention two examples of this phenomenon ins h o i n. The first is the main post of a t o k o n o m a alcove, preferably leftin a ‘natural’ state, albeit carefully selected in the repository ofn a t u re and surrounded by a perfectionist framework of other stru c-t u res. Secondly, the rustic style of tearooms (soa n, ‘thatched hut’),which was inspired by humble farmhouses, iro n i c a l l y, became a partof highly elaborated and expensive art of the aristocrats, as ‘linguis-tic wrapping’ culminates in the ritualized language of tea cere m o n y.

Paper is another parallel between wrapping and arc h i t e c t u re aswell as between Japan and Korea. Both in s h o i n and b a n g a, rooms areseparated mainly by papered panels, but Japanese s hoj i doors are

Sarvimaki Marja 8 3

3 . As one result of the timber decrease in Korea, the wooden members weresmaller in size than before and were used primarily in the roof structuresand in the details of doors, windows, and the kind. Even in temples,palaces, and aristocrats’ residences, which had timber frames, the use ofbrick and stone increased and they continued to be the main buildingmaterials of thatch-roofed commoners’ houses. Kim Dong-uk 1994, p. 65.

c o v e red with paper on the outer side of the wooden framework,w h e reas it is on the inner side of Korean j a n g j i doors. (The concepts‘inside’ and ‘outside’ will be discussed below, but here the formerrefers to a room and the latter to anything around it.) This cre a t e sd i v e rgent appearances, accompanied by panels that have paper onboth sides, or are built of wooden planks.4 Generally speaking,K o rean interiors are more dominated by paper, since not only arethe inner surfaces of the door/windows and walls papered, but soa re also the ceiling and the floor of b a n g rooms with o n d o l h y p o c a u s t(fig. 1).5 Because it would be uneconomic to heat wider spaces thannecessary b a n g is a small and low room for sleeping, sojourn, andworking. These rooms are, naturally, used in wintertime, thoughalso in the summer, as o n d o l is not heated then and the doors arehoisted on overhead hinges and hooks, or opened sideways. Whenthey are closed, the atmosphere is almost sealed: Only paper is visi-ble, and one truly feels as though one were in a paper box.

In addition, Korean houses have wood-floored m a r u rooms (seefig. 13). Contrary to b a n g, m a r u is a high space rising all the way tovisible roof rafters, sometimes including a partial wooden ceiling,and the biggest parts of the walls are possible to open; in southernb a n g a, these areas do not always have exterior doors at all. In otherw o rds, m a r u functions in a similar way as Japanese verandahs orrooms that have porous t a t a m i mats and sliding or removable wall/


4 . By Western vocabulary it is rather difficult to define these sliding or pivot-ed elements, as they are either doors, windows, or walls depending ontheir adjustment, though in Korea and Japan there are also permanentwalls and openings that are specifically windows. In general, s hoj i a n dj a n g j i can be regarded as all of them because they do not only serve as pas-sages when opened and partitions when closed, but also provide light andventilation in the room and view to outside, which are the basic functionsof a window.

5 . In o n d o l method, the smoke from a fireplace either in kitchen or outside isconducted into the under-floor flues, where the warm air heats the flooruntil passing out the chimney. The floor is made of clay and stones that arecovered by several layers of oiled paper.

door elements. Together with elevated floors and deep eaves, theyc reate horizontal lines of the facades in both building types, exceptthat Korean rooflines are more curved. Also, unlike the Chinese, theK o reans and Japanese continued living on the floor of these multi-purpose rooms with minimum furniture .

Straw is the material that dominates Japanese interiors due tothe t a t a m i mats (fig. 2). Originally t a t a m i, literally ‘to fold,’ was afoldable carpet, similar to the Korean d o t j a r i, still in use today. Thethick straw mats were first portable daises for the most importantpeople to sit on, and started to cover the whole room areas of s h o i nresidences only at the turn of the fifteenth century. The reason for thepopularity of t a t a m i was, undoubtedly, its suitability for the sultryclimate, but because the mats were (and are) expensive, they did notbecome widely used in commoners’ houses until in the eighteenthc e n t u r y, alongside better living standards of the rising merc h a n t

Sarvimaki Marja 8 5

Figure 1. A b a n g room (s a r a n g b a n g) of Yun Hwang’s house. Early 1600s, movedto its present site in the eighteenth century, Nonsan, South Korea.

class—thus, one way todemonstrate wealth.

As for other wrap-ping materials surro u n d-ing Korean and Japanesespaces, the roofs of b a n g aand s h o i n a re usually cov-e red with tiles, especiallyside buildings also withs t r a w, bark or shingle,while brick and stone areused in Korean fire - re s i s t-ing walls (b a n g h w a j a n g) ;necessary between o n d o lf i replaces and woodens t ru c t u res. The Japanesehearth is also normally inthe dirt-floor kitchen, butin the middle of it, or

sometimes at the center of a t a t a m i room, and walls are almost with-out exceptions made of plastered bamboo net between square pillarsand beams, again, to improve ventilation. It is the method in b a n g aas well, although the appearance is slightly diff e rent because thewall is of close to the same depth, if not thicker, than the frame stru c-t u re, thereby providing insulation needed during the colder winterson the Korean Peninsula. For the same reason, b a n g a’s openingsoften have two layers of sliding or pivoted elements, occasionallyeven a third set of framed mosquito nets, and a papered built-in casefor the extra pairs of panels.

To summarize, the innermost realm of Japanese s h o i n is mainlycomposed of straw floors, wooden ceilings, and walls with paperdoors and plastered areas, both interspersed by wooden framework,c reating the famous contrast of materials and colors. The interior ofK o rean b a n g a, on the other hand, is characterized by the interplay of


Figure 2. Interior of Katsura Imperial Villa. Seventeenth century, Kyoto, Japan.

b a n g and m a r u rooms. The former are dominated solely by paperand the latter by wood with plaster and paper between the woodenmembers, being reminiscent of verandahs, which are equally essen-tial parts of s h o i n and b a n g a.

Social Wrapping Elements

Most of the above phenomena are solutions to practical consid-erations and climatic conditions, yet, arc h i t e c t u re always re f l e c t salso ideological currents. In that respect, Edo Japan and JoseonK o rea departed from each other substantially, even if both were feu-dal societies. As is well known, the Joseon dynasty with its central-ized imperial rule was an ideal Confucian state that strained to ban-ish all other thoughts, whereas Confucianism was not rooted inJapanese soil of equal depth, but Buddhism and Shintoism cohere das influential religions. In the beginning of the peaceful Edo period,when the Tokugawa shogunate established control over the court,the Japanese s a m u r a i warriors became mainly bureaucrats who con-stituted their own elite, comparable to the Korean y a n g b a n g e n t r y.These were local officials, or served the administration in Seoul,while other family members provided for the household b a n g a in thecountryside. Some scholars also built similar houses as ‘hermitages,’w h e re they re t i red (or were sentenced) to compose poems, practicec a l l i g r a p h y, and cultivate their mind by other activities suitable forConfucian gentlemen.

As a result of the ‘alternate attendance’ system, the Japaneses a m u r a i leaders, d a i m yo, had to spend years in the Tokugawa head-quarters in Edo (today’s Tokyo) maintaining residences there, inaddition to the mansions and castles in their own domain, and oftenhouses even in the imperial capital Heiankyo (Kyoto). Since theirlifestyle re q u i red more rooms and partition between them, thebuildings incorporated new elements, as sliding panels with deco-rated paper on both sides and less functional ornamental details that

Sarvimaki Marja 8 7

indicated the prestige of the house. These features became con-stituent parts of the s u k i y a style of s h o i n residences, along thei n c reasing use of earlier ‘luxuries,’ like t a t a m i mats, ceilings, tea-rooms, and t o k o n o m a alcoves with adjoined t s u k e s h o i n desks and t a n a(or c h i g a i d a n a) shelves. This was also the era when most verandahsgot s hoj i doors in the outer colonnade, making these areas parts ofthe indoors—or more accurately, corridor-like intermediate spaces—due to the invention of wooden screens (a m a d o, ‘rain door’) that pro-tected the paper and when unneeded slid into a built-in case in theouter wall.

Besides s hoj i and j a n g j i panels, characteristic of both s h o i n a n db a n g a a re various kinds of adjustable space dividers, as foldeds c reens (Jap. b yob u, Kor. b y o n g p u n g), cloth curtains (Jap. n o re n, Kor.m u n y e o m j a), and bamboo or reed shutters (Jap. s u d a re, Kor. b a l) .Although the sliding elements that have paper on both sides (Jap.f u s u m a, Kor. m a e n g j a n g j i) as well as the latticed areas above the open-ings (Jap. r a m m a, Kor. g o c h a n g) are quite similar, the latticework of alltypes of dividers in a b a n g a is usually more complex, and the doorsthat are pivoted above are a Korean speciality; though existed inJapanese s h i n d e n residences that preceded s h o i n. Also, because ofb a n g a’s high foundations most of the verandahs that do not functionas entrances have railings and appear more as balconies, similar tothe Japanese elevated verandahs. There are diff e rences in the layoutof the houses as well. While the rooms of a s h o i n a re often connectedcornerwise in stepped blocks, a b a n g a is generally composed of morerectangular buildings, but arranged much less symmetrically thanthe extremely axial Chinese courtyard houses (figs. 3 and 4).

In all East Asian countries, the details and dimensions of build-ings were restricted by laws according to the owner’s social rank,based on the Confucian hierarc h y. The Tokugawa government, forinstance, prohibited commoners from using s u k i y a style, but a s a m u -r a i house, no matter how small, always included a t o k o n o m a a n doften a tearoom. At least d a i m yo and s hog u n should also have had ano stage, since tea ceremony and no drama, among many other arts,


w e re considered essential parts of s a m u r a i education. Even orna-mental windows, such as s h i t a j i - m a d o ( ‘ u n d e rg round window’), imi-tating farmhouses’ broken walls that disclosed the bamboo nets t ru c t u re, expressed social hierarchy because they were mainly usedin tearooms, or were attached to t o k o n o m a.

Sarvimaki Marja 8 9

Figure 3. Floor plan of Katsura Imperial Villa’s s u k i y a style main building with olds h o i n (upper right corner), middle s h o i n, and new s h o i n (below left). Early andmiddle seventeenth century, Kyoto, Japan. Nihon kenchiku shizush 1980, p. 84.

Especially gates, the outermost limit between ‘outside’ and‘inside,’ manifested the status of the house. The entrance of a samu-rai mansion had a curtain displaying the household crest, and thecusp-gabled roof (k a r a - h a f u), reflecting the importance of a Japanesebuilding from the twelfth century on, became prominent part of theEdo period s h o i n (fig. 5). A d a i m yo’s gateway with this kind of a ro o fwas reserved for s hog u n during his visits to the re t a i n e r, or for thereception of emperor at shogunal establishments. Needless to say, as t ru c t u re associated with these social connections imparted specialmeaning. In Korea, too, it was the roof of a gate that depicted the


Figure 4. Floor plan of Yun Jeung’s house. Originally built in the late seventeenthcentury, major alterations in the mid-1800s, Nonsan, South Korea. Joo 1999, p. 61.

rank of the house and, not surprisingly, gates with elevated ro o f(s o s e u l d a e m u n) were restricted for the y a n g b a n, while gates of thesame height as the fence (p y e o n g d a e m u n) were used in the re s i d e n c e sof lower classes, or in b a n g a’s inner walls (fig. 6). Of particular signifi-c a n c e was the j e o n g r y e o gate with governmental tablatures award e dto a virtuous wife or an obedient son.

A somewhat less tangible divider between the spatial layers isthe modular grid of the house. From a practical point of view, amodule brought about an order in the building, but both Kore a nand Japanese modular methods, which drew their origins fro mChina, were also means to express social ord e r. The basic unit,depicted by the Chinese ideogram of ‘sun’ within ‘gate’ (j i a n in Chi-nese, g a n in Korean, k e n, m a, etc. in Japanese), was a length, square ,and volumetric measure between two or four columns or theirheight re s p e c t i v e l y, and its dimensions varied on periods and are a s ,

Sarvimaki Marja 9 1

Figure 5. A guardroom of Ikedamon gate with k a r a - h a f u roof. Late seventeenthcentury, Tokyo, Japan.

even in a building.6 M o re o v e r, it distinguished the owner’s socialrank, as not only was the amount of units, but also the size of theunit itself controlled. For example, the length of g a n was diff e rent ford i ff e rent social groups: for commoners 6-7 c h e o k, for y a n g b a n 8 - 9(sometimes up to 11), and for the royal family 11-12 c h e o k.7 (1 c h e o k i sca. 30 cm.) This, in turn, means that the maximum of 10 ‘square g a n’


Figure 6. S o s e u l d a e m u n gate of Dongnakdang house. Mid-Joseon period,Gyeongju, South Korea.

6 . There was also another Chinese module, c a i (literally ‘material’), a ratio of15 by 10 between the longer and shorter side of bracket arm timber. In theNorthern Song dynasty, it had eight, by the Qing dynasty eleven gradeswith different dimensions (though the same ratio) according to the impor-tance of the building and the rank of the owner. As that type of bracket setswere not used in s h o i n and b a n g a, this module is not dealt with here. Moreabout the East Asian modules in my Ph.D. thesis, Sarvimaki 2000, pp. 161-1 8 5 .

7 . Courtesy of Dr. Kim Do-Kyoung (personal discussion).

(about 45 m2) in commoners’ houses was much less than one thirdof the 30 g a n (at least 170 m2) allowed for the lowest y a n g b a n r a n k .

The Japanese modular system, to which it developed during theEdo period and as we know it today, is more reminiscent of the con-temporary method, though it must be pointed out that k e n also hadvarious dimensions, and still there are four main sizes of t a t a m i m a t sthat are intrinsically connected to the column distance. In simplifiedterms, all horizontal and vertical measures depend on a particulark e n that is a multiple of s h a k u units (ca. 30 cm) and on its fractionsinto smaller intervals, as pillar section of 4 s u n (0,4 s h a k u). Graduallythis structural order was used in all residences, mainly due to thei n c reasing popularity of s h o i n type buildings and the rising livings t a n d a rds in the late Edo period, when Japan, indeed, developedinto a more egalitarian society than Joseon Korea. Yet the dimen-sions of t o k o n o m a, t s u k e s h o i n and t a n a w e re, and are, defined by thek i w a r i module, which is predominantly based on aesthetic ratios andwas originally the proportioning method of s a m u r a i m a n s i o n s ,palaces, and temples.

A l ready in the beginning of the Edo period, however, wealthym e rchants (the lowest Confucian social class of ‘citizens’) built larg eresidences, such as the Nakamura House in modern Osaka, withformal t a t a m i rooms and t o k o n o m a alcoves. Also, even if the style ofgates was limited and the guard rooms of lower d a i m yo ranks were tohave single sloping roofs, they tended to ignore the prohibition andhad curved k a r a - h a f u c o n s t ructions in their houses.8 And in Kore a ,w h e re such details as round pillars and d a n c h e o n g multicolor paint-

Sarvimaki Marja 9 3

8 . Particularly after the devastating fire of Edo in 1657, residences like the so-called row houses (n a g a y a) were controlled because of the scant resourcescaused by the rebuilding of the entire city. The compendium of regulationsand etiquette for the members of the s a m u r a i class, issued by the Tokugawashogunate, restricted the styles of houses and gates, among other rules.They also limited the width of the houses of the highest-ranking d a i m yointo three bays, whereas lower s a m u r a i leaders were required to reducetheir n a g a y a into two and a half bays. Coaldrake 1996, pp. 193-207.

ings were permitted only in temples and palaces, some aristocrats,who were not as humble as Confucian scholars were supposed tobe, did show off by having these forbidden elements in their b a n g a.In many cases, their residences also exceeded the permitted size,s i m p l y, by not including the servants’ quarters (h a e n g n a n g c h a e) in thehouse are a .9

F u r t h e r m o re, due to the Neo-Confucian ideology that empha-sized strict separation of genders—as it was interpreted by the y a n g -b a n scholars—the Korean mansions had to be divided into men’sand women’s quarters, s a r a n g c h a e and a n c h a e, by high walls with amiddle gate (j u n g m u n) and some lesser passages in between. In re a l-i t y, many b a n g a have interesting unofficial routes or ‘grey zones,’which connect the enclosures. Take the Yangjindang in Hahoe Vi l-lage, where everything, at first, looks like the rules were followed(figs. 7 and 8). Had I prior to my visit not seen the floor plan, itwould have been impossible to guess that a pair of doors veils anadjunctive m a r u, since the usual places for a visitor to examine thehouse are the courtyard or reception rooms of s a r a n g c h a e. On thesame trip in Andong, I spent four nights in a b a n g a, where I experi-enced an analogous arrangement that observes the moral principles,but maximizes the functional eff i c i e n c y. It was only in the evening,


9 . For example, the Kim residence in Chongup had 87 g a n, but was reportedto the authorities to be 40 g a n in size, which was the maximum amount ofg a n for the rank of the owner of this house and the total of only the mainbuildings. In principle, non-titled citizens were not allowed more than 10g a n, ‘middle classes’ 20 g a n, third class aristocrats 30 g a n, second class 40g a n, princes and princesses of kings’ secondary wifes 50 g a n, and so on,ending in the 99 g a n palaces. Chun et al. 1999, pp. 37-41. (Here, too, it mustbe taken into account that the size of a g a n was bigger the higher the statusof the owner.) By the late Joseon period, when the scrutiny of regulationsbecame more lax, even rich farmers adopted features of b a n g a in their resi-dences and especially the ‘middle class’ citizens had s o s e u l d a e m u n g a t e s .Therefore, some y a n g b a n aristocrats opposed this decline of social order bynot having s o s e u l d a e m u n gates and rebuilt a p y e o n g d a e m u n main gate intheir houses. Courtesy of Dr. Kim Seong-Do (personal discussion).

while getting the bedding out from the cupboard, when I found a‘ s e c ret door’ to a n c h a e in the rear wall of the fixture and realized thatit, in fact, made the space more a corridor than a storage (fig. 9).1 0

Sarvimaki Marja 9 5

1 0 . During my stay in this b a n g a (Jichonjongtaek, Cultural Properties MaterialNo. 44), I also realized that a plan that allows privacy to the family and, atthe same time, has separate area for the guests, is why the house works sowell in its modern function as an inn. Because I was conducting researchon traditional Korean architecture the hosts, Mrs. and Mr. Kim Won-Kil,kindly let me enter the a n c h a e too. On that side, there was a highboy infront of the ‘secret door’ to block the passage that, naturally, is not meant tobe used by the customers of the inn, since the owners live in the a n c h a e.

Figure 7. Floor plan of Yangjindang house with an adjunctive m a r u b e t w e e na n c h a e (lower left square) and s a r a n g c h a e (upper right rectangle). Mid-Joseonperiod, Andong, South Korea. Contemporary Architecture 2000-1, p. 222.

These types of informal pas-sageways were, of course,needed for women or ser-vants to serve the men enter-taining their guests in theevening, when the middlegate was locked.

The widely used transla-tion ‘men’s and women’sa rea,’ nonetheless, is not quiteaccurate, as the Korean termsdo not refer to sexes. A n c h a ecould be better interpreted as an ‘inner house’ ands a r a n g c h a e as an ‘outer house’;in other words, the former isthe private and the latter thepublic part of b a n g a.1 1 T h emale family members wereallowed to enter a n c h a e,though preferably with a per-mit of the household head (orin the night, when the ser-

vants did not see them), but s a r a n g c h a e’s courtyard and re c e p t i o nrooms were the furthest a visitor could get. Often there was a com-manding view from s a r a n g c h a e to the middle gate, which allowedthe control. No wonder, the official quarters for the household heirand his sons took on symbolic significance as the entire family her-


Figure 8. The doors of an adjunctive m a r u of Yangjindang house seen froms a r a n g c h a e’s courtyard. Mid-Joseon period, Andong, South Korea.

1 1 . The ideogram sa means ‘housing’ and rang ‘corridor,’ whereas chae is asuffix indicating an independent building or group of them. It should alsobe noted that a n, meaning ‘inner,’ is a Korean word and is not written bythe Chinese ideogram a n for ‘safe,’ ‘peace,’ etc., depicted by ‘woman’ underthe ‘roof,’ although the etymology and connotation might lead to this c o n c e p t .

itage and authority of the patriarch—ideally receiving his guests bysitting in n u m a r u, a pavilion-type m a r u that is erected even higherthan the other rooms of s a r a n g c h a e (fig. 10).1 2

Sarvimaki Marja 9 7

Figure 9. Layout of Kim Banggeol’s house, in which a ‘secret door’ in the cupboard of s a r a n g b a n g (lower right corner of the main building) connects it witha n c h a e (upper part of the main building). Mid-1600s (the house was moved fromits original location in 1990s because of the Imha Dam project, but the floor plansof s a r a n g c h a e and a n c h a e are original), Andong, South Korea. ContemporaryArchitecture 2000-1, p. 245.

1 2 . Professor Joo Nam-Chull brought to my attention that sometimes thesmoke from o n d o l fireplace is conducted to the holes in the podium, fromwhere it fills the surface of the courtyard. Then, the s a r a n g c h a e looks like itwas ‘floating on a cloud,’ metaphorically indicating the superiority of thefamily patriarch. Professor Kim Dong-Uk, for his part, pointed out thatmany b a n g a are also preserved till our days because the owners are so

In spite of the diff e rences in the layouts and appearances ofb a n g a and s h o i n, there are similar perceptions. Also Japanese housesa re clearly divided into private and public areas in the way peoplec i rculate in them. Just as Neo-Confucian ideology was not as notablein Japan as in Korea, nor was gender segregation. Consequently(along other possible reasons), s h o i n did not part into men’s andwomen’s quarters to the same extent as did b a n g a. Even more obvi-ous is that social hierarchy was indicated by the arc h i t e c t u re of s h o i nand b a n g a, as was the inside/outside distinction, though not alwaysv i s u a l l y. Let us, there f o re, proceed by ‘looking at’ what is less visiblein the light of Hendry’s statement of the tea ceremony: “The actualspeech used is very sparse, extremely restricted, and we would learnvery little if we looked only at the word s . ”1 3


proud of their y a n g b a n ancestry (personal discussions).

Figure 10. N u m a r u in the corner of the Nakseonjae house’s s a r a n g c h a e and thepivoted doors of its d a e c h e o n g hall hoisted on the overhead hooks. 1847, Seoul,South Korea.

Unfolding Spatial Layers

Without being “overly concerned with ‘unwrapping,’ withrevealing the perceived essence of things,”1 4 as Hendry warns us, itis useful to consider the stages of entering a b a n g a or a s h o i n. Inmany respects, it is reminiscent of opening a gift, which includes theelement of surprise—real or pretended. All the way from the gate,the first transactional space between the ‘very outside’ and the firstlayer of ‘inside,’ the houses and routes in them are care f u l l ydesigned by opening new views and revealing others. Between theentrance courtyard and the main buildings, there might be addition-al yards or gardens; in Japan, usually separated by hedges or fencesmade of wood, bamboo or other ‘light’ materials, whereas b a n g a a redominated by split-levelled masonry or stone walls that have tilec o v e r, creating an almost castle-like atmosphere. The buildings alsohave removable ‘covers,’ like bamboo or reed blinds, and theappearance of the facades greatly depends on whether the woodenshutters and other outer panels are closed or open.1 5 A c t u a l l y, theK o rean doors pivoted above, even better than the sliding ones, cor-respond to the ‘unfolding’ metaphor because first they are foldedhorizontally and then hoisted up vertically.

When entering a Japanese house, a split cloth curtain (n o re n)often partly conceals the entrance, and accompanied by formalg reetings, one arrives in the porch (g e n k a n) removing the shoes.While ‘climbing up’ (a g a r u or n o b o r u) into the building, the host usu-ally offers slippers to be worn in the corridors and other wood-f l o o red areas, except for toilets and bathrooms, where special slip-pers are needed. No footwear is used in the following stage of thet a t a m i-matted areas, of which the first is either o z e n, meaning ‘fro n troom,’ or a more elaborate z a s h i k i reception room with t o k o n o m a

Sarvimaki Marja 9 9

1 3 . Hendry 1995, p. 165.1 4 . I b i d., 1995, p. 5.1 5 . For more examples of Japanese dividers and transactional spaces, see

Nitschke 1993, pp. 49-93.

alcove. Seldom is a visitor allowed any further. If one is, the re l a t i o n-ship has become intimate enough that one is not re g a rded as a com-plete outsider anymore. At the same time, many of the linguistichonorifics are dropped, which Hendry points out: “Just as with lan-guage, layers of polite formality conceal (and occasionally reveal) aninner sanctum, and the nearer the outside one finds oneself, them o re formal is the expected behaviour. ”1 6

A classy s h o i n house has several z a s h i k i rooms, and the signifi-cance of a particular space is indicated by the height and tre a t m e n tof the ceiling and the tiers of the floor from the lower level (g e d a n) tothe middle level (c hud a n), ending in the highest dais (jod a n) for themaster of the house, or a more important visitor. If there are no dif-f e rent levels, the most honorable place is the t a t a m i mat in front ofthe t o k o n o m a. From this public area, a guest may get glimpses deeperinto the house through the dividing elements, which ‘wrap’ theinner areas, or o k u, having connotations of ‘heart,’ ‘interior,’ or‘depth.’ Similarly to the phrases used, the movable elements concealo k u, when closed, and momentarily reveal it, when family membersor servants pass the openings. Then, one possibly sees a next set ofsliding panels, behind one more, and so on, creating the impre s s i o nof depth. Besides, even when the panels are closed, the translucentpaper does not completely separate a room from another—especiall ynot acoustically—nor the interior from the exterior, although theamount of layers and thickness of paper allows the interplay of lightand shade and/or transparency and opaqueness.

In the course of entering a Korean b a n g a, one even more literally‘climbs up’ (o re u d a), as not only are the floors elevated, but the build-ings are placed on podiums above the courtyards that themselveshave terraces at several levels. The shoes are left on the podium’sstepping stone, from which one climbs on a verandah, and a visitorends either in the great m a r u hall (d a e c h e o n g), n u m a r u, or outer b a n grooms of s a r a n g c h a e. Like in Japan, through the layers of openings

1 0 0 Layouts and Layers

1 6 . Hendry 1995, p. 100.

one gets an impression ofcontinuous space; inb a n g a, by glimpses of them a s t e r’s bedroom (c h i m -b a n g) and others a r a n g c h a e’s inner are a s ,while an outsider seesonly the roofs of a n c h a ebehind the high walls(fig. 11). Hence, it is per-haps not too far- f e t c h e dto consider a n c h a e, witha n meaning ‘inner,’ theK o rean equivalence too k u, and vice versa. Thesame is suggested by thew o rds for ‘wife,’ whichboth in Korean andJapanese have connota-tions of ‘inside person.’

R e g a rding Japan, theconcept of ‘empty center’has been widely specu-lated, and some scholarshave been contemplatingw h e re the center of aJapanese house is—by

the hearth or the t o k o n o m a1 7—though I do not find it necessary, orpossible, to locate o k u, the ‘center of emptiness.’ In a b a n g a, the ‘veryinside’ might be somewhere around the first lady’s room (a n b a n g) ,

Sarvimaki Marja 1 0 1

1 7 . In my opinion, and corresponding to the phenomena discussed in thispaper, o k u definitely is ‘deeper’ in the house than in the t o k o n o m a of areception room.

Figure 11. A view through the s a r a n g c h a e o fYun Jeung’s house: in front s a r a n g b a n g, thenm a r u b a n g (in between interesting a n g o j i g idoors that are both horizontally pivoted andsliding), and furthest behind the backyard is thewall of a n c h a e (see also the floor plan in fig. 4).Late seventeenth century, Nonsan, South Korea.

her sleeping room (w i t b a n g, ‘upper room’) and the daughter- i n - l a w ’ sroom (g e o n n e o n b a n g, ‘opposite room’) that is on the opposite side ofa n c h a e’s d a e c h e o n g. Rather than enclosed by permanent elements, thelayers surrounding these ‘centers’ are defined by behavior like aban-donment of shoes, the ‘dirt of outside,’ before entering the ‘cleaninside,’ and the further borders are indicated by some sort of a stepand/or change of flooring material when moving from one spatiallevel to another. Similar kind of temporal expressions are the EastAsian terms for module, since the Chinese ideogram for the columndistance also means both ‘space’ and ‘time’—one more arc h i t e c t u r a lp e rception that is usually re g a rded as uniquely Japanese, famous forthe concept m a.

F rom our Hendryan approach, ‘linguistic wrapping’ is an alter-native means in interpreting the spatial layers and their definitions,as discussed above. In that respect, Japanese residences have ani n t e resting detail, the frieze rail (u c h i - n o r i - n a g e s h i) in the so-called‘honest wall’ (s h i n - k a b e), meaning that the construction system isgenuinely exposed (fig. 12). Paradoxically, since the invention ofconcealed horizontal ties (n u k i), it had no tectonic function exceptabove the openings (even there only to strengthen the k a m o i s l i d i n gtrack), but continued also to be used on the solid walls of high-classs h o i n, though disappeared in many houses of commoners and lowerranking s a m u r a i; among the latter, partly due to the appreciation ofthe semiformal s u k i y a style, associated with the ‘way of tea’ (c h a do) ,the ‘way of warriors’ (b u s h i do), and other ‘arts’ (do). This ‘beam’—like a ribbon of a gift—could well be considered corresponding tothe Japanese ‘beautification language’ (k e i g o), in indicating aestheticvalues and status of the house. Similarly, public or ‘formal’ (h a re)a reas are defined by the formal t a t e m a e speech, comparable to ‘fro n t ’(o m o t e) and ‘outside’ (s o t o), as opposing to private or ‘informal’ (k e) ,‘ re a r’ (u r a), ‘inside’ (u c h i), and the real opinion, or h o n n e. In short,calling the wall ‘honest’ is t a t e m a e, while mentioning that the friezerail is not a structural but an ornamental member is h o n n e.1 8

The view from ‘inside’ to ‘outside’ is of equal importance,


though the perspective shifts in the cases of s h o i n and b a n g a. Sincethe latter is ideally located on a hillside, the openings fro ms a r a n g c h a e p rovide a wide command of the entrance courtyard(s a r a n g m a d a n g or h a e n g n a n g m a d a n g) and far-away mountains, whileanchae opens to the inner yard (a n m a d a n g) and the rear gard e n(d w i t m a d a n g). That is probably why the Koreans pay so much atten-


1 8 . With regard to distinguishing spatial layers by movable and translucentelements, particularly interesting is psychiatrist Takeo Doi’s view, to whichHendry refers: “Japanese people very often know precisely what the h o n n eis in any situation, despite the formal expressions of t a t e m a e. [...] so thatwhen people look at o m o t e they see also the u r a through it.” Hendry 1995,p. 163. She discusses these levels of Japanese language in depth; regardingthe usage of different types of k e i g o, translated in various ways (in the con-text of a structural element that is actually an ornament, I prefer the term‘beautification language’), see especially Hendry 1995, pp. 52-69, abouth o n n e / u r a and t a t e m a e / o m o t e, pp. 145, 167-8.

Figure 12. Continuous space of Honmaru Goten, the s h o i n - z u k u r i residence ofthe lords at the Kawagoe Castle, and the u c h i - n o r i - n a g e s h i frieze rail above theopenings as well as on the solid walls. 1848, Kawagoe, Japan.

tion to the windows andtheir sills, on which onecan rest one’s elbow whenadmiring the scenery.Although Korean yard sthat have very little vege-tation are not, as theJapanese gardens, consid-e red to be viewed fro mindoors, I would arg u ethat at least the ‘flowerstairs’ (h w a g y e) in the gar-den definitely are (fig. 13).The general setting of thehouse surrounded by highwalls and a forested hill inthe background, in turn, iseyed from the direction ofa p p ro a c h .

Japanese s h o i n is usu-ally on a flat lot and theview from the rooms tothe ground level appearsto be more significant; an extreme example are the doors that havevertically sliding lower part allowing a scene solely to the surface ofthe garden, preferably recently covered with snow, which the doors’name y u k i m i (‘snow viewing’) s hoj i implies. More o v e r, the deep andstraight eaves greatly block out the upward vista, and Japaneseopenings that are specifically windows most often have latticework,so that even when the papered panels are open they provide moreor less ventilation and illumination, but only glimpses of the sur-roundings. However, exceptions like Katsura Villa’s elevated are a sand s h o i n’s corner rooms for moon viewing and other admiration ofn a t u re function quite similarly to Korean pavilions and n u m a r u


Figure 13. A n c h a e’s great m a r u hall in YunJeung’s house with a view to the rear gardenand its ‘flower stairs’ and the stand for k i m c h ijars. Late seventeenth century, Nonsan, South Korea.

(always in the corner of s a r a n g c h a e), from where one sees a pond that is important element of both b a n g a and s h o i n. The ‘borro w e dlandscape’ technique in Japanese gardens is also reminiscent of theK o rean and Chinese way to use the wider panorama as part of thec o u r t y a rd s .

It is difficult to say where these houses ‘end.’ The Japaneseverandahs that have colonnades on both sides (wood-floored h i ro - e nor t a t a m i-matted i r i k a w a), to some extent also the outer t a t a m i ro o m s ,as well as Korean m a r u rooms and porches (a p t o e or t o e t m a r u) can beseen as outdoors—not outside—when the doors are open, andindoors—literally—when closed. The outer verandahs that are nor-mally narrower and on a lower level (Jap. o c h i - e n, Kor. j j o k m a r u) andvarious kinds of stepping stones are more clearly extensions of theg a rden, but even the ground level spaces under the eaves, in addi-tion to courtyards, are re g a rded as part of the house. To put it other-wise, although not yet in the buildings, one is already either in‘open-air insides’ or ‘roofed outdoors.’ And further in the gard e nt h e re are side buildings, as teahouses that are integral parts of s h o i n,w h e reas b a n g a has many types of annexes, like pavilions and sepa-rate houses (b y e o l d a n g or a n s a r a n g c h a e) for re t i red parents, marriedfirst son’s family, or secondary wife/s (figs. 14 and 15). Since Confu-cian ancestral worship (j a e s a) was constituent part of y a n g b a n l i f e ,t h e re is also a primogenitors’ shrine (s a d a n g) or several of them inthe rear end of the lot backed by pine tre e s .

F rom here we could, and should, proceed to geomancy andother relevant cosmological ideologies, though it is not possible toextend the scope of this paper to these interwoven topics, such ascity layout reflecting the society at large. To just, once again, quoteHendry: “In Japan, the whole country was from 1600 to the middleof the nineteenth century closed off from the outside world so thatthe nation virtually wrapped itself entire l y. ”1 9 The same applies toK o rea, the so-called Hermit Kingdom, even a few decades longer.


1 9 . I b i d., 1995, p. 116.


Figure 14. S hokintei teahouse in the garden of Katsura Imperial Villa. Seventeenthcentury, Kyoto, Japan.

Figure 15. Hwallaejeon garden pavilion by the lotus pond of Seongyojang house.1816, Gangneung, South Korea.

This is one more reason why juxtaposition of Jeoson Korea and EdoJapan is elucidative, as both were closed societies, in which very dis-tinct cultures evolved. Indeed, I am not claiming that Japanese andK o rean architectural expressions would always correlate, nor thatthe phenomena discussed above (especially division of houses intopublic and private areas, or connection between time and space)w e re something unique for Japan and Korea. What I do tend toa rgue is that the divergent appearances of s h o i n and b a n g a a re morea question of diff e rences in architectural idioms than meanings—idiom itself being a linguistic term for a phrase that is typical for adialect or a language, in this context, the language of arc h i t e c t u re ,including its ‘non-verbal’ modes.

R e f e r e n c e s

Ben-Ari, Eyal & Moeran, Brian & Valentine, James (eds.). U n w r a p -ping Japan. Honolulu: University of Hawaii Press, 1994.

Chun, Jin-hee et al. (Choi, Jae-Soon & Hong, Hyung-Ock & Kang,Soon-joo & Kim, Dae-nyun & Min, Chan-hong & Oh, Hye-kyung & Park, Young-soon). Hanoak: Traditional Korean Homes.Seoul: Hollym International, 1999.

Coaldrake, William H. A rc h i t e c t u re and Authority in Japan. London:Routledge, 1996.

Contemporary Arc h i t e c t u re: The Traditional Arc h i t e c t u re of Kore a, No. 29,2 0 0 0 - 1 .

H e n d r y, Joy. Wrapping Culture: Politeness, Presentation, and Power inJapan and Other Societies. Oxford/ New York: Oxford Universi-ty Press, 1995.

Inaji, To s h i ro & Vi rgilio, Pamela (translation and adaptation). The Gar -den as Arc h i t e c t u re: Form and Spirit in the Gardens of Japan, China,and Kore a. Tokyo/New York/London: Kodansha International,1 9 9 8 .

Joo, Nam-chull. Hanguk jutaek geonchuk ( K o rean Housing Arc h i t e c-


t u re). Seoul: Iljisa, 1980._ _ _ _ _ _ _. Hanguk geonchuk uijang ( K o rean Architectural Design).

Seoul: Iljisa, 1997._ _ _ _ _ _ _. Hangukui jeongtong minga ( K o rea’s Traditional Houses).

Seoul: Daeuhaksulchongsa, 1999._ _ _ _ _ _ _. Hangukui mungwa changheu ( K o rean Gate and Wi n d o w s ) .

Seoul: Daewonsa, 2001.Kim, Dong-uk. ‘The City and Arc h i t e c t u re of Seoul During the Late

Choson Period,’ K o rea Journal, Vol. 34, No. 3, Autumn 1994, pp.5 4 - 6 7 .

Nihon kenchiku shizushu (Japanese Arc h i t e c t u re Illustrated). To k y o :Shokokusha, 1980.

Nitschke, Gunter. F rom Shinto to Ando: Studies in Architectural Anthro -pology in Japan. London: Academy Editions, 1993.

Sarvimaki, Marja. S t r u c t u res, Symbols and Meanings: Chinese and Kore a nInfluence on Japanese Arc h i t e c t u re. Espoo: Research series of theHelsinki University of Te c h n o l o g y, Department of Arc h i t e c-t u re, 18/2000.

A c k n o w l e d g e m e n t s

If otherwise not mentioned, this paper is based on my disserta-tion thesis S t r u c t u res, Symbols and Meanings: Chinese and Korean Influ -ence on Japanese Arc h i t e c t u re (Sarvimaki 2000), on which I started towork during my studies at the Tokyo National University of Arts andMusic in 1987-89 on a Japanese government M o n b u s h o s c h o l a r s h i p .The years-long study was further funded by several org a n i z a t i o n s ,the most generous having been the Jenny and Antti Wihuri Founda-tion and the Finnish Cultural Foundation in Finland as well as theN o rdic Institute of Asian Studies in Denmark. Most re c e n t l y, andmost importantly, I had the chance to delve into traditional Kore a na rc h i t e c t u re during my post-doctoral re s e a rch from March to October2002, as a Korea Foundation Fellow, affiliated at the Korea University.


Layouts and Layers - SKKUsjeas.skku.edu/upload/200605/05_Sarvimaki Marja.pdf · Layouts and Layers:...

Documents

Transcript of Layouts and Layers - SKKUsjeas.skku.edu/upload/200605/05_Sarvimaki Marja.pdf · Layouts and Layers:...