We have finally reached the end of this long series on minka layouts, which began with single-space minka and progressed through one-room, two-room, three-room, four-room, and multi-room layouts, with many digressions, tangents and reiterations along the way.
Here is a good time to remind the reader that this series, like the one before it on minka structure, is a loose translation of Internal Layouts, Structure, and Interiors (Madori・Kо̄zо̄・Naibu 間取り・構造・内部), the second volume of Kawashima Chūji’s (川島宙次) three-volume work on minka titled Horobiyuku Minka (滅びゆく民家, lit. ‘Disappearing Minka’). The first and third volumes are Roofs and Exteriors (Yane・Gaikan 屋根・外観) and Sites/Auxiliary Structures and Typologies (Yashiki-mawari・Keishiki 屋敷まわり・形式) respectively.
So by way of conclusion, and for the last post of the year, this one is strictly for the nerds: an explanation of the various symbolic notation systems, including Kawashima’s own, developed to represent minka layouts non-pictorially.
Kawashima points out that such a system had already been pioneered by Ishihara Kenji (石原憲治) before World War II. Ishihara’s work in organising, classifying, and statistically analysing a great number of minka layouts has been extremely useful to later researchers, and in Kawashima’s view he deserves great respect for bringing the study of layouts into focus earlier than would otherwise have been the case.
Ishihara’s notation system, expressed with numerical and arithmetical symbols, is able to capture the basic gist of minka layouts, but it is highly reductive and has two major shortcomings: first, though it is possible to visualise the number and general arrangement of rooms from the notation, the names or uses of the rooms are not indicated: the system does not convey whether a particular room is allocated for use as, for instance, a formal zashiki or as a bedroom (nando), so it is impossible to know the mode of habitation of the rooms, or of the dwelling as a whole.
The second drawback of Ishihara’s system is that it is not even always possible to determine the layout type from the notation. For example, a front-zashiki type three-room layout (mae-zashiki-gata san-madori 前座敷型三間取り) and a parallel lineup three-room layout (jūretsu-gata san-madori 縦列型三間取り), two very different layouts, are both simply represented as ‘3’.
This latter deficiency can be addressed by adopting a symbolic system that indicates not only the number of rooms but also the layout type or style. For this purpose, Kawashima offers the following symbols and definitions:
H Hiroma-type (Hiroma-gata 広間型) layout
S Regular (Seikei 整形) layout
K Staggered (Kui-chigai 食違い) layout
Y Perpendicular lineup type (heiretsu-gata 併列型) layout. Since ‘H’ is already taken, ‘Y’ is used for the equivalent term Yoko-narabi (横ならび, lit. ‘horizontal lineup’). The lineup of rooms runs perpendicular to the room-doma boundary.
T Parallel lineup type (jūretsu-gata 縦列型) layout. Since ‘j’ is already taken, ‘T’ is used for the equivalent term Tate-narabi (縦ならび, lit. ‘vertical lineup’). The lineup of rooms runs parallel to the room-doma boundary.
Throughout this series on layouts, we have used the ‘X-type N-room’ (or alternatively ‘N-room X-type’) format, as in ‘regular four-room layout’ (seikei yon-madori 整形四間取り) or ‘three-room layout hiroma-type’ ( 三間取り広間型). To convert a minka layout written in this way into Kawashima’s ‘layout type + number of rooms’ format, we simply take the letter that represents the layout type, and follow it with the number that represents the number of rooms, with the two separated by a dash. So the ‘regular four-room layout’ (seikei yon-madori) is written ‘S - 4’ and the ‘three-room layout hiroma-type’ (san-madori hiroma-gata) is written ‘H - 3’. For a staggered four-room layout (kui-chigai yon-madori 食違い四間取り), we write ‘K - 3’.
Kawashima’s system also remedies the first shortcoming of Ishihara’s system, by using alphabetical characters to represent room names (and thus uses), a method which he notes is adopted as-is from that used by Nishiyama Uzо̄ (西山夘三) in the farmhouse volume (Nо̄ka-hen 農家編) of his work Japanese Dwellings (Nihon no Sumai 日本の住まい). Use of an upper-case letter indicates that the room runs the full width (harima 梁間) of the dwelling; lower-case letters stand for the names of rooms that are not full width, i.e. are divided from other rooms by partitions perpendicular to the room-doma boundary.
The room symbols and definitions employed by Kawashima are as follows:
n Sleeping space (shinshitsu kūkan 寝室空間). Common names: nema, heya, nando
z Reception space (sekkyaku kūkan 接客空間). Common names: zashiki, genkan
d Dining/ ‘housework’ space (shokuji・kaji kūkan 食事・家事空間). Common names: daidoko, katte
j Family ‘public’ space (kazoku no kо̄kūkan 公空間). Common names: jо̄i, chanoma, iroma, nakanoma
o Public space admitting guests/visitors (kyaku o fukumeta kūkan 客をふくめた空間). Common names: omote, dei
W Interior (generally earth-floored and full-width, thus capitalised) work/utility space (yanai sagyо̄ kūkan 屋内作業空間). Common names: niwa, daidokoro, doji
m Livestock space (kachiku no kūkan 家畜の空間). Common names: maya, daya, umaya
s Storage (shūnо̄ 収納) space such as a closet (oshi-ire 押入, mono-iri 物入), cupboard (todana 戸棚); or a decorative alcove (toko 床, toko-no-ma 床の間), Buddhist alcove (butsuma 仏間), or the like.
e Liminal or ‘edge’ (en 縁) space such as a ‘verandah’ (engawa 縁側) or entry vestibule (genkan 玄関).
As an example, using this ‘room name + spatial arrangement’ system to describe regular four-room layout, consisting of a nando, zashiki, daidoko, omote, and niwa, gives us:
n d
—— W
z o
If we render each of the major layout types and styles into the four descriptive formats (the conventional word-based format and the three symbolic notation systems) discussed above, we obtain the following table of 18 plan diagrams, where each plan is accompanied by its description in each of the four formats, usually to the right of the plan.
Plan diagrams of the various minka layout types. To the right of each plan diagram is a ‘quadrant’ of four ways of describing the layout, in either words or symbolically. The quadrant contains the following: top left, the layout description in words; bottom left, a symbolic representation consisting of the layout type followed by the number of rooms, and that followed optionally by the stagger type, e.g. ‘K - 4T’; top right, a symbolic representation of the function and position of each room in the layout, with the arrangement of letters conveying the location of each room and spatial relationships between them; bottom right, in parentheses, Ishihara Kenji’s system of notation, using numbers, the arithmetical symbols ‘+’ and ‘x’, and occasionally a character suffix such as ‘併’ (hei) to indicate what I call the ‘perpendicular’ lineup’ layout, or ‘全’ (zen) to indicate that a room is ‘full (width)’.
By way of example, the first and simplest layout is shown below.
Plan diagram of a one-room layout (ichi-madori １間取り) and its various descriptions.
At the top left is the layout description in words: ‘prototype one-room layout’ (genkei hito-madori or genkei ichi-madori 原型１間取り).
Below that, at bottom left, is the reduction of this description into its alphanumeric layout symbol or ‘code’: in this case, simply ‘1’. There is no letter here to indicate a layout type, because in a one-room layout there is no possible variation in this regard.
At top right is the symbolic spatial description of the room layout, which adds room and spatial information to the alphanumeric code: here, ‘J W’, meaning an earth-floored utility space (W) and an adjacent family ‘public’ room (J), both capitalised since both run the full width of the dwelling.
The number is parentheses, here (1), is the layout expressed in Ishihara Kenji’s notation system.
Now let’s take a more complex example, shown below: a perpendicular stagger four-room layout.
Plan diagram of a a perpendicular stagger four-room layout (tate kui-chigai yon-madori 縦食違い４間取り) and its various descriptions.
Its description in words is tate kui-chigai yon-madori (縦職違い４間取り).
Expressed alphanumerically, this becomes ‘K - 4T’. The ‘T’ here comes after the number, so indicates that the stagger of the four-room layout is of the perpendicular type (tate kui-chigai 縦食違い), and is not to be confused with a ‘T’ before the dash, which indicates a parallel lineup type (jūretsu-gata 縦列型) layout. Likewise, a ‘Y’ after the number would indicate that the stagger of the layout is of the parallel type (yoko kui-chigai 横食違い), whereas a ‘Y’ before the dash indicates a perpendicular lineup type (heiretsu-gata 併列型) layout.
The spatial description of the room layout is:
n d
/ \/ W
z o
The horizontal three-bar zigzag indicates the perpendicular stagger; further, the fact that the peak of the zigzag is on the left and the valley on the right indicates that the partition between the nando and the zashiki is rearward of the partition between the daidoko and the omote.
Finally, Ishihara’s notation for this layout, shown in parentheses, is ‘2+2’. The plus sign indicates this as a staggered four-room layout, distinguishing it from a regular four-room layout, which is written ‘2x2’.
If we want to be able to reconstruct plan diagrams from our symbolic notation systems with more detail and greater accuracy, we can introduce the elements of floor area and linear dimensions. For example, to the letter indicating the ‘layout type’, we can add numerals that indicate the overall building dimensions (length x width) in ken (間, 1 ken = 1.818m). For example, a hiroma-type layout minka that is 12 ken long and 5 ken wide would be written as:
H 12 x 5
Note that in adding the dimensions to this ‘layout type + number of rooms’ notation system we have lost the ‘number of rooms’, so for this ‘layout type + building dimensions’ notation to be of any use, it must be complimented by the ‘room name + spatial arrangement’ notation system; or, to put it another way, when accompanied by the ‘room name + spatial arrangement’ notation system, the numeral indicating the number of rooms is redundant. Combining both the ‘number of rooms’ and the ‘building dimensions’ in the ‘layout type’ notation would probably require putting parentheses around the ‘building dimensions’ to avoid the formula becoming messy and ambiguous. If the above ‘H 12 x 5’ minka were a four-room layout (‘H - 4’), for example, we could write the whole as:
H - 4 (12 x 5)
We can add the element of area to the ‘room name + spatial arrangement’ notation system too, by affixing numerals indicating the area of each room, measured in tatami mats, to the letters representing the room names. A tatami mat is one ken in length and half a ken wide, i.e. 1.818m x 0.919m, so the area of a single mat is around 1.67m2. The counter suffix for tatami is jо̄ (帖); an eight-mat zashiki, expressed symbolically here as ‘z8’, would be read as hachi-jо̄ no zashiki (８帖の座敷). To obtain the area of a room in tatami mats or jо̄, simply multiply the length of the room by the width, both measured in ken, and then multiply the result by two. So a room measuring 3 ken long by 2 ken wide has an area of 2 (3 x 2) = 12 jо̄.
Let’s now apply all this to a real-world example: the former residence of the Sakuta/Sakuda/Tsukuda (作田) family, originally of Sanbu County (Sanbu-gun 山武郡), Chiba Prefecture, but now relocated to the Japan Open-Air Folk House Museum (Nihon Minka-en 日本民家園) in Kanagawa Prefecture, and designated an important cultural property.
The plan of the house is fairly complex: a six-room wrapped-hiroma type (tori-maki hiroma-gata 取巻き広間型) layout, with the added complication of being a ‘separate ridge style’ (buntо̄-shiki 分棟式) construction, meaning that the dwelling is comprised of two buildings that are structurally independent and have separate roofs, but (in this example at least) are internally continuous.
Plan of the former Sakuta/Sakuda/Tsukuda (作田) family house, a wrapped-hiroma type (tori-maki hiroma-gata 取巻き広間型) ‘separate ridge style’ (buntо̄-shiki 分棟式) minka. Labelled are the earth-floored utility area (doma どま), which occupies its own separate building; the large, board (ita 板)-floored hiroma or omote (here the kami かみ) with firepit (irori, marked ro 炉), Buddhist alcove (butsuma, marked manji 卍) and shallow decorative alcove (oshi-ita 押板); the ‘dining room’ (daidoko, here cha-no-ma ちゃのま) with firepit; the bedroom (nando なんど); the rear formal room (zashiki, here oku おく) with decorative alcove (toko とこ); the ‘middle room’, also formal (naka-no-ma なかのま); and the formal entry ‘anteroom’ or ‘vestibule’, the liminal (en 縁) genkan げんかん). Part of the doma is taken up with the stalls of a stable (maya まや, unlabelled).
Expressing the layout in both the ‘layout type + building dimensions’ and ‘room names + room areas + spatial arrangement’ notation systems gives us the following:
Two complementary symbolic descriptions of the Sakuta house. On the left, the ‘layout type + building dimensions’ description. On the right, the ‘room names + room areas + spatial arrangement’ description.
On the left we have the ‘layout type + building dimensions’ description. ‘TH’ stands for for ‘wrapped-hiroma type (torimaki hiromagata 取巻き広間型), and this is followed by the total dimensions of the main building: 8 ken long by 5.5 ken wide, or around 14.5m x 10m. ‘W’ indicates the earth-floored work or utility area (the doma 土間 or niwa にわ); the ‘+’ separating the ‘TH’ and the ‘W’ indicates that these are two separate buildings; in other words, the minka is a buntо̄ (分棟 ‘separate ridge’) construction. The ‘doma building’ is 3 ken long and 6 ken wide, or around 5.5m x 11m.
Using only the ‘layout type + building dimensions’ formula of ‘TH 8 x 5.5 + W 3 x 6’, it is possible obtain a preliminary outline plan like that shown below. The dashed lines representing room divisions are ‘anticipatory’ and cannot be derived from the ‘layout type + building dimensions’ formula alone.
A preliminary reconstruction of the external walls and overall length (maguchi 間口) and width (okuyuki 奥行) dimensions (measured in ken 間) of the two buildings of the Sakuta house, obtained from the description ‘TH 8 × 5.5 = W 3 × 6’. Note that the half-ken difference in depth between the doma building and the main building’ is resolved here by showing the doma building extending out at the rear, with the two buildings drawn flush on the facade side, but this is an assumption that cannot be derived from the symbolic description alone. Likewise, without knowing the customary style of minka in the area, one cannot know from the symbolic description alone that the two buildings are internally continuous and not fully separate.
The intimidating-looking formula on the right is the ‘room name + spatial arrangement’ description, with the addition of numerals that the floor area of each room.
For the sake of convenience, let us assume that the building is oriented so that the facade (the hiroma or omote side) faces south, though this information is not carried in the symbolic descriptions, as not relevant.
Starting on the left, we have first a capital ‘E’, for an engawa running the full width of the building; the width of this space is not described.
From the next part of the formula we can determine that there are six rooms: a rear zashiki with associated storage or alcoves, another zashiki, a facade-side engawa or other liminal space, a bedroom (nando), a dining-kitchen (daidoko), and the main room (omote).
Next, the ‘W’ before the plus sign indicates that there is a full-width earth-floored utility area (doma or niwa) at the opposite end of the main building to the full-width engawa (‘E’).
The two horizontal lines represent partition lines: the top line terminates at the ‘W’ and the bottom line terminates at the ‘O’ (which should be lower case); from the positions of the room symbols relative to these lines we can determine that there is a ‘column’ of three rooms non-adjacent to the niwa (the rear zashiki with rear storage/alcove, the ‘middle’ zashiki, and front en space), that the omote runs from this column to the niwa, that the combined length of nando and daidoko is equal to that of the omote, and that the daidoko is adjacent to the niwa but the nando is not, being between the daidoko and the rear zashiki.
The floor area of each room, measured in tatami mats, is indicated by the number suffixed to it. So the rear zashiki ‘z’ is ten tatami mats in area (jū-jо̄ １０帖), and so on.
Now for the difficult part: reconstructing the dimensions (length and width) of each room from the given floor areas. Let’s start from the column of three rooms on the left, the combined width of which must equal the width of the building, 5.5 ken.
We know that each zashiki is 10 tatami mats (10 jо̄) in floor area, and that the en space is 5 jо̄ in area. Keep in mind that a tatami mat is 1 ken (1.818m) long and 0.5 ken (0.909m) wide. We know that the rear zashiki has a storage closet (oshi-ire 押入), decorative alcove (toko 床), or the like at its rear, as indicated by the small ‘s’ above the ‘z10’. The depth (the width or transverse dimension) of this ‘s’ space is not given, but let’s assume that it is the standard 0.5 ken (0.909m). This leaves us with 5 ken of building width (oku-yuki 奥行, lit. ‘rear going’) to distribute across the two zashiki and the en space. Assuming that the layout is regular (seikei 整形) and therefore the transverse (vertical) partitions of these three rooms are aligned, it makes sense to also assume that each room is 2.5 ken (2.5 tatami lengths or 5 tatami widths) long; by dividing this length into the area of each room and then dividing the result by 2, we can obtain the widths of the rooms with reasonable confidence that they are correct. For each zashiki:
(10 jо̄ / 2.5 ken) / 2 = 2 ken
For the en space:
(5 jо̄ / 2.5 ken) / 2 = 1 ken
So the dimensions of the two 10-mat zashiki are 5 tatami widths by 2 tatami lengths, i.e. 2.5 x 2 ken, and the en space is 5 tatami widths by 1 tatami length, i.e. 2.5 x 1 ken.
Next, we know that the bedroom (nando) and kitchen (daidoko) are both 10 tatami mats in area, but again, assuming the layout is regular (seikei) and therefore that the longitudinal (horizontal) partition boundary between the nando/daidokoro and the omote is aligned with that between the two zashiki, the width of these rooms must be the same as that of the rear zashiki (2 ken) and its ‘storage’ space (0.5 ken) combined, i.e. 2.5 ken. Knowing the width (2.5 ken) and the areas (10 tatami mats or 10 jо̄ each) of the nando and daidoko, we can calculate their length by dividing the area (10 jо̄) by the width (2.5 ken, i.e. 5 tatami widths), which gives us 2 ken, or two tatami lengths. Therefore the nando and daidoko are both 2 ken long by 2.5 ken wide.
We know that the omote is 24 jо̄ in area. Its length must equal that of the combined length of the nando and daidoko, i.e. 2 + 2 = 4 ken, and its width must equal that of the combined zashiki and engawa space, i.e. 2 + 1 = 3 ken. Doing the area calculation
2(4 ken x 3 ken) = 24 jо̄
confirms that that these dimensions are correct.
As mentioned, there is a full-width engawa (‘E’) at one end of the main building, and an earth-floored utility area (niwa, ‘W’) at the other end. We know that, as the overall width of the building is 5.5 ken, these spaces must also be 5.5 ken in width, but their length is not given. We do know, however, the overall length of the main building (8 ken), and from our calculations of the room dimensions and orientations we also know that the total length of the rear zashiki, nando, and daidoko, and the total length of the front zashiki/en and omote: both sum to 6.5 ken. By subtracting this length from the overall length of the main building, we obtain a leftover length of 8 ken - 6.5 ken = 1.5 ken. It is impossible to know what share of this 1.5 ken is apportioned to the engawa and what to the niwa, but it is reasonable to assume a standard 0.5 ken for the engawa, leaving 1 ken for the niwa.
Next is the separate niwa building, the ‘W’ in ‘TH 8 x 5.5 + W 3 x 6’. We know from the ‘3 x 6’ that this building is 3 ken long and 6 ken wide, for a total area of 2(3 x 6) = 36 jо̄ (or equivalent, since this area is earth-floored so the tatami figure here is ‘virtual’). Comparing the width of the niwa building with that of the main building, we see that the former is 0.5 ken wider than the latter; whether because the distance between the external structural wall/posts lines themselves are more widely spaced, or because a 0.5 ken under-eave space to either the front or rear of the niwa is enclosed, we cannot tell.
Finally the, ‘m8’ after the ‘W’ indicates that within the niwa building there is a stable (maya) or similar (‘m’) and that it is 8 jо̄ in area, though its exact position within the niwa is not captured by the symbolic description.
The layout of the Sakuta house, as reconstructed from the formulas given.
In general, though it is possible with effort and practice to arrive at the correct reconstruction using only the formulas given, there remain ambiguity and the possibility of error in translating them into the correct length and width dimensions of the various rooms, the correct room orientations, and the correct partition locations.
In terms of legibility and detail, Kawashima’s system is a significant improvement on Ishihara’s, though by Kawashima’s own admission it still isn’t perfectly comprehensive or able to capture every particularity of minka layouts in all their messy variety. Because only the area of the rooms is given, not their length and width, the interpreter is left to puzzle out the ‘aspect ratio’ and orientation of each room: an 18 jо̄ space, for example, might be 1 ken by 9 ken, or 2 x 4.5, or 2.25 x 4, or 3 x 3, or 4 x 2.25, or 4.5 x 2, or 9 x 1. Though the correct arrangement can usually be arrived at with common sense and some thought, this effort could be eliminated by including the length and width dimensions for each room, just as they are given for the structure as a whole.
In the end, a notation system capable of capturing the full range of layout types and all nuances of their sub-variations to a degree of detail that guaranteed accurate reconstruction of every possible minka plan would probably be so complicated and unwieldy that one might as well use plan diagrams, which at least have the advantages of being comprehensible at a glance and of instantly conveying the ‘feel’ of the minka interior in a way that symbolic notation cannot.
But, as Kawashima points out, there are benefits to using these symbol systems as a kind of shorthand in certain situations, such as when surveying multiple minka of a particular style in a particular region: one can undertake a detailed plan-drawing of a single representative example, and for all the others record only symbolic data, which can interpreted and reconstructed at leisure ‘back in the office’ by reference to the plan, thereby greatly reducing the labour and time (money) required for fieldwork. Kawashima also suggests the possibility of using computers to digitise and automate the process of reconstructing and generating plan diagrams from symbolic inputs.
This may seem quaint to us, but it is a reminder, and important to recognise, that Kawashima, Ishihara and others were surveying minka at a time either entirely before computers, or in the era of punch-cards and room-sized machines, when digital memory was scarce and expensive (Disappearing Minka was first published in 1973), to say nothing of digital cameras and all the other conveniences of our own time, including modern roads and transportation networks; in pre-war and immediate post-war Japan, just reaching some of the remoter mountain villages and islands would have been a challenge in itself.
