Architectural

Institute of Japan

Translated Paper

Research on the design process of Kenzo Tange’s

own residence
Saikaku Toyokawa
Department of Urban Environmental System, Faculty of Engineering, Chiba University, Chiba, Japan

Correspondence Abstract
Saikaku Toyokawa, Department of Urban Environmental
System, Faculty of Engineering, Chiba University, 1-33 This study examines the design process of Tange residence based on sketches
Yayoi-cho, Inage-ku, Chiba-shi, Chiba prefecture (Kantou and photographs kept by Akira Tarashima, who was in charge of designing the
area), Japan. residence. The study classifies them according to area size from Type A to Type F.
Email: toyokawa-s@chiba-u.jp Observing the front of the longer direction, it is observed that 32 shaku, 40
The Japanese version of this paper was published in shaku, 50 shaku, and 60 shaku were under consideration, after which Type F
Volume 86 Number 787, pages 2368–2377, https://doi.org/ became 54 shaku. This study confirms that the various study sketches considered
10.3130/aija.86.2368 of Journal of Architecture and Plan- room allocation based on the basic modules of 4 shaku, 5 shaku, and 6 shaku. It
ning (Transactions of AIJ). The authors have obtained is also confirmed that the plans and elevation were drawn simultaneously and
permission for secondary publication of the English ver-
sion in another journal from the Editor of Journal of clarified the process from a modest-size one-story house (Type A) to the two-
Architecture and Planning (Transactions of AIJ). This story completion draft with an exterior worthy of the residence of a master
paper is based on the translation of the Japanese version architect (Type F).
with some slight modifications.
Keywords
Received December 20, 2022; Accepted January 25, 2023 floor level, module, pilotis, sketches, Tarashima Akira
doi: 10.1002/2475-8876.12339

pre-war housing competition, the continuity between the Kishi

1. Introduction
Memorial Gymnasium and Tange residence, and Tange’s use of
The residence built by architect Kenzo Tange in Seijo (herein- sandwiched beams and pillars. Through an interview with Tara-
after referred to as “Tange residence”) has been featured in var- shima, Fujimori learned that the initial plan of Tange residence
ious media, such as magazines and television, since its was similar to that of Farnsworth House by Mies van der Rohe.
completion in 1953. In recent years, Tange residence has been Further, he says that, although he heard from Tarashima about
featured extensively in large-scale exhibitions at the National the existence of the sketch materials related to the design of
Museum of Modern Art and the Mori Art Museum, and has been Tange residence, he could not confirm them.1 With this discov-
positioned as one of the most important residential works of ery, their whereabouts may have been confirmed.
modern architecture in post-war Japan. However, the design pro- The second study is an interview of Tarashima by the author.
cess of the residence has been only a matter of hearsay because It seems that the person initially in charge of Tange residence
few people were involved in the design of the residence and was Yamanashi Kiyomatsu (a member of Tange Laboratory);
related materials such as drawings, sketches, model photographs, since Tange asked Tarashima to help Yamanashi, he participated
and letters are scarce. However, sketches related to Tange resi- in the design of Tange residence. He said that he does not
dence were discovered when Professor Akira Tarashima (former remember what kind of plan Yamanashi was considering at that
assistant professor at Kagoshima Univ.), who was in-charge of time, but it was far from the final plan, and he claimed that “dur-
designing the residence while at Tange Laboratory, passed away ing the implementation design, it became a two-story building,
and his family were organizing his belongings. and an earthquake-resistant wall became necessary.”2
Shaku is a unit of length, widely used in East Asia. In The third study was conducted by the author when he
Japan, 1 shaku is approximately 0.303 meters. restored the model of Tange residence (1/3) at the Mori Art
Museum. The author organized the characteristics of Tange
residence from four perspectives: the site, piloti, module, and
2. Previous Studies, Status of Materials, and Purpose of
collaboration with the artists.3
this Study
In the fourth study, the author discussed Tange’s renovation
Four studies have been conducted on Tange residence. In the process and extension plan, making full use of Tange’s photo-
first, Terunobu Fujimori discusses Tange’s proposal for a graphs of his residence and sketches related to the extension.4

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided
the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2023 The Author. Japan Architectural Review published by John Wiley & Sons Australia, Ltd on behalf of Architectural Institute of Japan.

Jpn Archit Rev | 2023

TOYOKAWA wileyonlinelibrary.com/journal/jar3

This study discusses previous studies on the topic in Chapter are unknown. There are two types of examination plans for
2, discusses the sketches and photographs discovered by this Type B, one written by hand (See Figure 3 B-1) and one
study in Chapter 3, and then uses them to analyze the design drawn with a ruler (See Figure 4 B-2). In the handwritten plan,
process for Tange residence in Chapter 4. Chapter 5 clarifies the main entrance (1) is drawn at the left end and the service
the relationship between the design intent for the residence entrance (2) is drawn at the center, suggesting that they were
and the sketches, based on an explanation Tange gave after the searching for the optimal position of the kitchen and were con-
completion of the building. sidering it in relation to the position of the dining table.
Further, the positions of the bed and storage are clearly drawn
in the parent room and child room, and the plan is such that
3. List of Issues and Findings from the Urawa House
the living room (3) can be approached directly from the
Among the few houses designed by Tange, those published in entrance.
magazines include his own residence and the Urawa house. In the plan (B-2) drawn with a ruler, the layout of the rooms
Tange’s autobiography indicates that the former was designed is being considered based on a 4 shaku 9 2 shaku grid. The
in 1952, constructed between June 1952 and October 1953, positions of the main entrance and service entrance are
and featured in the January 1955 issue of Shinkenchiku,5 The unclear, and the living room (3) has irregular tatami mats,
design and construction periods of the latter house could not similar to Type A. However, it is more spacious than Type A,
be confirmed, but it was featured in the November 1955 issue giving a feeling that many guests are expected. In addition,
of Shinkenchiku. However, materials held at the Tsuboi since a revision regarding the irregular orientation of the
Laboratory, which was in charge of the Urawa house, indicate tatami mats was made, it can be inferred that the corridor lead-
that the draft structural design document [Note 1] (the stage ing from the core (toilet [7] and bathroom [8]) to the living
prior to the final draft) was dated November 25, 1953. Thus, room (3) and the exterior wall facing it were considered thor-
the Urawa house was likely designed in parallel with or soon oughly. Similar to Type A, the core is arranged slightly to the
after Tange residence. left from the center of the building.
When the Urawa house was featured in the magazine, There are two types of elevations in Type B, B-3 and B-4.
Yamanashi evaluated it from the perspectives of layout, floor The former is almost symmetrical, like Type A. While Type A
plan, sunlight, ventilation, heat insulation, structure, scale, and had an even arrangement of the five pillars, Type B adopted
lighting fixtures.6 These are important clues for understanding an irregular grid of 8 shaku: 4 shaku: 8 shaku: 8 shaku: 4
the points that were taken into consideration for housing shaku: 8 shaku (See Figure 5 B-3). On the contrary, the draw-
design in Tange Laboratory at that time. ing of the latter is marked with “Elevation North.” Inferring
Therefore, this study analyzes the materials related to Tange from the position of the double roof (core position), it is pre-
residence based on the points of discussion obtained at the sumed that the study of the Type B plan was done according
Urawa house. The materials listed in Table 1, except for F1 to to the general drawing method, with the upper part of the tra-
F-3 and P-3 to P-6, were discovered at Tarashima’s house. cing paper as the north (Figure 6 B-4). The latter is com-
pressed by about 14 mm (about 2 shaku in 1/50 conversion) in
the longitudinal direction relative to the former to form an ele-
4. Analysis of the Six Types of Sketches
vation of 236 mm (about 38 shaku in 1/50 conversion). At the
4.1 Type A: 32 shaku 3 15 shaku plan same time, the elevation is drawn according to a grid that
Type A is a one-story house with a total floor area of about 45 divides 38 shaku into 10 parts. Since this reference line holds
m2, and the site shape, layout, and orientation are unknown the outside of the pillars at both ends, they seem to have been
(See Figure 1 A-1). The main entrance (1) and the service examining the proportion of changes that would occur on the
entrance (2) are placed back to back, which is very similar to north elevation as a result of trying to fit the skeleton of the
the construction of the Urawa house. Regarding the layout of house within a certain dimension and compressing the indoor
the house, the living room (3) is placed at the left end, the par- rooms.
ents’ room (10) and children’s room (9) are at the right end,
the core (kitchen [5] + toilet [7] + bathroom [8]) is slightly to 4.3 Type C: 50 shaku 3 18 shaku plan
the left (west), and the space for the housekeeping room (12) Type C has a total floor area of 84 m2, and, like Type A, the site
is quite large. While the housekeeping room in the Urawa shape, layout, and orientation are unknown (See Figure 7 C-1).
house accommodated six tatami mats, the Type A plan had a While the Type B plan is based on a grid of 4 shaku 9 2 shaku,
sewing table (6). Four tatami mats of an irregular size Type C uses a grid in which the longitudinal direction (50 shaku)
(1500 9 600) are laid out in the living room (3). is divided into 10 divisions of 5 shaku each. In addition, the
The elevation (A-2) of Type A is symmetrical, and the large short side of Type C is about 2 shaku longer than the short side
opening (sliding door) of the dining table (4) hangs slightly of Type B, and the grid is broken because the large sliding door
towards the inside of the exterior wall, giving depth and sha- facing the dining room (4) is placed inside the pillar. The posi-
dow to the elevation. An open veranda (deck) can be seen at tions of the main entrance (1) and service entrance (2) are close
the foot, giving a feeling of continuity with the garden (See to those in Type A and the Urawa house.
Figure 2 A-2). Type A can be said to be a design close to the In Type B, tatami mats were laid in the living room (3), but
image of the Farnsworth House that Tarashima communicated in Type C tatami mats are laid to clarify the difference from
to Fujimori. In addition, the roof directly above the core is a other floor finishes. Further, in contrast to Type B, Type C has
double roof, probably for ventilation and odor control pur- a new veranda on the outside of the living room (3); however,
poses. since two pillars are placed there, it is thought that this was
intended to be a semi-outdoor veranda with a roof. The front
4.2 Type B: 40 shaku 3 16 shaku plan of the dining room (4) is marked with a rectangle that appears
Type B is a one-story house with a total floor area of about 60 to be an open veranda, and the outside of the children’s room
m2, and, like Type A, the site shape, layout, and orientation (9) is marked with a note for future extension (13).

Jpn Archit Rev | 2023 | 2

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

TABLE 1. Types of materials in Tarashima Archive & Other Archive

Type 種別 Number番号 Content 内容 Size サイズ Scale Material材質

STUDYSketch A-1 Type A Plan (32shaku 9 15shaku)Type A 平面図 (32尺915尺) 320 9 256 1/50 Tracing
Type Aスケッチタイプ A Paperトレペ
A-2 Type A Elevation (32shaku 9 15shaku)Type A 立面図 (32尺 305 9 212 1/50 〃
915尺)
STUDYSketch B-1 Type B Plan-1 (40shaku 9 16shaku)Type B 平面図 (40尺916 305 9 208 1/50 〃
Type Bスケッチタイプ B 尺)
B-2 Type B Plan-2 (40shaku 9 16shaku)Type B 平面図 (40尺916 305 9 208 1/50 〃
尺)
B-3 Type B Elevation (40shaku 9 16shaku)Type B 立面図 (40尺 305 9 212 1/50 〃
916尺)
B-4 Type B’ Elevation (38shaku 9 16shaku)Type B’ 立面図 (38尺 305 9 212 1/50 〃
916尺)
STUDYSketch C-1 Type C Plan (50shaku 9 18shaku)Type C 平面図 (50尺918尺) 305 9 208 1/100 〃
Type Cスケッチタイプ C C-2 Type C Elevation-1, 2 (50shaku 9 18shaku)Type C 立面図-1, 2 305 9 210 1/100 〃
(50尺918尺)
STUDYSketch D-1 Type D 1st Floor Plan, 2nd Floor plan, Elevation 1 (60shaku 870 9 560 1/100 〃
Type Dスケッチタイプ D 9 20shaku)Type D 1階平面図、2階平面図及び立面図-1(60尺
920尺)
D-2 Type D Elevation-2 (60shaku 9 18shaku)Type D 立面図-2 (60 557 9 290 1/50 〃
尺918尺)
D-3 Type D Section-1 (60shaku 9 18shaku)Type D 断面図-1 (60尺 551 9 431 1/50 〃
918尺)
D-4 Type D Section-2 (60shaku 9 18shaku)Type D 断面図-2 (60尺 551 9 445 1/50 〃
918尺)
D-5 Type D Perspective (60shaku 9 20shaku)Type D-3 パース (60尺 800 9 282 – 〃
920尺)
STUDYSketch E-1 Type E 1st FloorPlan, 2nd Floor Plan (56shaku 9 20shaku) 870 9 560 1/100 〃
Type Eスケッチタイプ E Type E 平面図(56尺920尺)
Type FCompletation F-1 Type F Second Floor Plan (56shaku 9 24shaku)Type F 二階平 – 1/100 Collected
Drawing竣工時の図面 面図(56尺 9 24尺)Tange K. Reality and Creativity. Bijutsu Works作品集掲載
Shuppansha. 1968:100.丹下健三『現実と創造』1968 p.100
F-2 Type F Site & 1st Floor Plan (56shaku 9 24shaku) @Michiko 560 9 479 1/100 Blue Print青図
Uchida Archive. Type F 配置図兼一階平面図(56尺 9 24尺)所
蔵:内田道子アーカイブ
F-3 Ty F Section (56shaku 9 24shaku) Type F 断面図(56尺 9 24 – 1/100 Collected
尺)Tange K. Reality and Creativity. Bijutsu Shuppansha. Works作品集掲載
1968:101.丹下健三『現実と創造』1968 p.101
STUDYSketch G Type G Elevation Tange’s Notebook 1952 (Uchida Archive) – – Notebook手帳
Type Gスケッチタイプ G Type G 立面図 丹下健三の1952年手帳 (所蔵:内田道子アーカイ
ブ)
Photograph写真 P-1 West Wall & Pilotis 01 (1953)photographer: Tarashima西側側 – – Print印刷
壁とピロティ01 撮影者:田良島昭
P-2 West Wall & Pilotis 02 (1953)photographer: Tarashima西側側 – – 〃
壁とピロティ02 撮影者:田良島昭
P-3 South Elevation photographer: Tange, K (Uchida Archive)南 – – 〃
外観 撮影者:丹下健三(所蔵:内田道子アーカイブ)
P-4 Katsura Rikyu photographer: Tange, K (Uchida Archive)桂離 – – 〃
宮 撮影者:丹下健三(所蔵:内田道子アーカイブ)
P-5 Living 01, photographer: Y. Ishimoto @Kochi Prefecture, – – 〃
Ishimoto Yasuhiro Photo Center, 居間 撮影者:石元泰博
P-6 Living 02, photographer: Y. Ishimoto @Kochi Prefecture, – – 〃
Ishimoto Yasuhiro Photo Center, 居間 撮影者:石元泰博

Type C has two types of elevations, C-2 Type C Elevation-1 stepped approach is set to 3 mm (0.3 m in 1/100 conversion,
and Elevation-2 (See Figure 8 C-2). Both are lengthwise eleva- about 1 shaku) from ground level. Type C elevation-2, when
tions and are one-storied, like Type A and Type B. The floor seen from the open veranda, seems to abandon the symmetry,
of Type C elevation-1 is set at a level of 6 mm (0.6 m in 1/ and the pillars are arranged at a pitch of 5 shaku: 10 shaku: 10
100 conversion, about 2 shaku) from ground level, and the shaku: 10 shaku: 5 shaku: 10 shaku.

Jpn Archit Rev | 2023 | 3

TOYOKAWA wileyonlinelibrary.com/journal/jar3

FIGURE 1. A-1 Type A plan. Source: Owned by Akira Tarashima Archive

It seems that the plan was to set up an entrance (1) in the

center of the first floor and take the stairs (13) up to the sec-
ond floor from there, but this is not obvious. It also seems that
a toilet and bathroom were planned for the 1st floor, but they
could not be identified from the rough sketch. Similar to the
Type A to Type C room layouts, the second floor of Type D
has a tatami area on the left side of the plan, the children’s
room (9) and parents’ room (10) are placed on the right end of
the plan, and the kitchen (5) and dining room (4) are placed
between them. It is presumed that the left half of the tatami
area was intended for use by visitors and the right half for
daily life as a housekeeping room (12), which (it is assumed)
would be used openly during events (See Figure 10 D-1).
In the Type D elevation, the floor is raised to a level of
20 mm (2 m in 1/100 conversion, 7 shaku) from ground level.
In the longitudinal direction (60 shaku), pillars are arranged at
12 shaku: 4 shaku: 12 shaku: 4 shaku: 12 shaku: 12 shaku: 4
FIGURE 2. A-2 Type A elevation. Source: Owned by Akira Tarashima shaku. The roof’s slope is also a little gentler than that of the
Archive Type C elevation, and the double roof is eliminated.
Type D elevation-2 (D-2) is 365 mm long (18 250 mm in 1/
50 conversion, about 60 shaku), and the floor is raised to a
4.4 Type D: 60 shaku 3 20 shaku plan, 60 shaku 3 18 shaku level of 28 mm (1.4 m in 1/50 conversion, about 5 shaku)
plan from ground level. (See Figure 11 D-2). The pillars divide the
Type D is a two-story plan (See Figure 9 D-1) with a total length into five equal parts, and the plan has a low correlation
floor area of about 109 m2 on the second floor. The first floor with Type D-1 plan. The roof is given a characteristic design
is about 20 m2 (only rooms 1, 7, 8, and 13 are counted). Type with sharp edges, and it is probable that, since the floor was
C uses 10 divisions in 5-shaku increments in the longitudinal raised, the roof shape was examined at the same time.
direction (50 shaku), while Type D uses 15 divisions in 4- There are two types of sectional detailed drawings for Type
shaku increments in the longitudinal direction (60 shaku). D, D-3 and D-4. In D-3, the floor is lifted to a level of 70 mm

Jpn Archit Rev | 2023 | 4

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 3. B-1 Type B plan-1. Source: Owned by Akira Tarashima Archive

FIGURE 4. B-2 Type B plan-2. Source: Owned by Akira Tarashima Archive

Jpn Archit Rev | 2023 | 5

TOYOKAWA wileyonlinelibrary.com/journal/jar3

Tarashima at the time Tange residence was completed contains

several photographs taken from angles very similar to D-5 (see
Figures 16 P-1 and 17 P-2), there are no photographs taken from
such angles among the multiple photographs taken by Tange of
his residence. The Type F plan is slightly different from Type D
because the gable end is 24 shaku, there is a veranda, and the
position of the pillars is also different. However, since the verti-
cal louvers cover the elevation of the gable end of the second
floor, D-5 provides much information useful for understanding
the design process for the house.
This notebook consists of three sketches: a sketch that
divides the west elevation shown in D-5 above into three equal
parts, a sketch with a pillar placed in the center of the west
elevation as shown in the completed photograph (P-1), and a
sketch with a balcony (See Figure 18 G-1). These sketches
were likely drawn during the transition from Type D/Type E
to Type F (final plan). These sketches are also drawn in the
FIGURE 5. B-3 Type B elevation-1. Source: Owned by Akira Tarashima November/December pages of the 1952 notebook diary, and
Archive they correspond to Tange residence construction period men-
tioned earlier (June 1952 to 1953).

4.5 Type E: 56 shaku 3 20 shaku plan

The total floor area of the second floor of Type E is about 109
m2, which is almost the same size as the final plan (Type F)
when looking only at the interior space. The longer side (56
shaku) is divided into 12 shaku: 4 shaku: 12 shaku: 12 shaku:
4 shaku: 12 shaku; the shorter side (20 shaku) is divided into
8 shaku: 4 shaku: 8 shaku; and the design is carried out on the
premise that the elevation is symmetrical. Looking at the first
floor plan of Type E (see Figure 19 E-1 bottom), it is observed
that an L-shaped wall is provided in the upper side of the plan
(probably on the north side) to secure the approach flow line
from the left (probably a visitor flow line from the west side),
which leads to the stairs (12) to the second floor. The rooms
planned for the first floor in the Type D plan have been elimi-
nated, suggesting the intention to have an open plan with as
few rooms as possible in the piloti area.
FIGURE 6. B-4 Type B elevation-2. Source: Owned by Akira Tarashima In the Type E second floor plan (see Figure 19 top), as seen
Archive in the previous plans, the children’s room (9) and parents’
room (10) are placed at the east end, the living room (3) is
placed at the west end, and the core (such as the kitchen [5]
(1.4 m in 1/20 conversion, about 5 shaku) from ground level., and toilet [7]) are placed at the center. Note that the tatami
which seems to have been drawn in conjunction with D-2. mat notation has been omitted in this drawing.
From these sectional detailed drawings, the relationship In addition, this drawing, unlike other study drawings, seems to
between the beam and the roof structure can be clearly under- focus on the examination of earthquake-resistant walls. For exam-
stood, and it can be seen that, at this stage, the plan was to ple, in the second floor plan, the intention was to place an
affix ceiling panels parallel to the floor at about 2560 mm (See earthquake-resistant wall between the children’s room (9) and the
Figure 12 D-3). parents’ room (10); however, if the beauty of the piloti on the first
In the sectional detailed drawing of D-4, the floor is lifted to floor (see Figure 19 bottom) is prioritized, the wall there cannot
a level of 96 mm (1.9 m in 1/20 conversion, about 6 shaku) be lowered. Thus, this suggests an intention to put in a symmetri-
from ground level, and the distance between the outer pillars cal earthquake-resistant wall, as far as possible. Moreover, unlike
is 270 mm (5.4 m in 1/20 conversion, about 18 shaku). In in Type C’s sectional detailed drawing, the area of the core sec-
addition, a 10-step staircase is attached to the right side of the tion was consumed by a hole in the floor intended to provide a
frame, and details such as the interior perimeter zone, sliding staircase. Although the team succeeded in obtaining a novel ele-
doors, and ceiling of the room are written (See Figure 13 D- vation by lifting the floor to the second floor, the level of difficulty
4). In the drawing, the following roof specifications are writ- increased significantly relative to the conventional plan study, and
ten: “roof slope 3 sun 5 bu (Japanese measurements); roof tile this may have embarrassed Tarashima, the person in charge.
– wooden roof tiles; roofing board thickness – 3 bu; rafters 2
sun 9 1 shaku 5 sun intervals; triangular frame 4 sun 9 7 4.6 Type F: 56 shaku 3 24 shaku plan (at completion)
pieces; main building 3 sun 9 3 pieces.” Type F is the completion plan that was published in a maga-
Sketch D-5 can also be cited as a resource useful for under- zine. The total area of the second floor is about 109 m2. The
standing Type D (See Figures 14 and 15 D-5). This sketch was longer side (56 shaku) is divided into 12 shaku: 4 shaku: 12
probably drawn by Tarashima. The rationale for this is that Tara- shaku: 12 shaku: 4 shaku: 12 shaku, and the shorter side (20
shima himself had the sketch, and although the album taken by shaku) is divided into 4 shaku: 8 shaku: 8 shaku: 4 shaku.

Jpn Archit Rev | 2023 | 6

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 7. C-1 Type C plan. Source: Owned by Akira Tarashima Archive

FIGURE 8. C-2 Type C elevation-1,2. Source: Owned by Akira Tarashima Archive

In the second floor plan (F-1), unlike the previous study east side is used for the study room and living room. In addi-
plans, switching the children’s and parents’ rooms from private tion, on the south side, a veranda is placed in the length direc-
rooms to tatami rooms resulted in a floor plan with a clear tion, and care has been taken to create a closer contact with
core style, and the fluidity of the entire second floor was the garden.
greatly improved (See Figure 20 F-1). In addition, the chil- Looking at the first floor plan, it is found that there are two
dren’s and parents’ rooms are moved to the west side, and the walls between the three pillars (left of the table-tennis table
Jpn Archit Rev | 2023 | 7
TOYOKAWA wileyonlinelibrary.com/journal/jar3

FIGURE 9. D-1 Type D 1st floor plan, 2nd floor plan & elevation-1. Source: Owned by Akira Tarashima Archive

FIGURE 10. D-1 Type D plan & elevation-1 (enlarged figure of Figure 9). Source: Owned by Akira Tarashima Archive

[11]) on the western end (See Figure 21 F-2). However, since Looking at the cross-sectional view (F-3), it is seen that the
slits are cut between them and the beams of the second floor, ceiling parallel to the floor is discarded, and the ceiling surface
they would not be likely to function as earthquake-resistant is constructed along the roof with a gentle slope (See Fig-
walls. Further, according to the completion plan, access is ure 22). Tange himself said, “I didn’t fit any glass in the
from the front road on the south side, and there is no entrance transom at the top of the sliding door unless it was necessary
or exit from the road on the west side, as the lines of visitor because I wanted an impression of unlimited space.”7 A roof
and service flow overlap. with a slightly different slope was placed on top of the large
Jpn Archit Rev | 2023 | 8
wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 11. D-2 Type D elevation-2. Source: Owned by Akira Tarashima Archive

FIGURE 12. D-3 Type C section-1. Source: Owned by Akira Tarashima

Archive FIGURE 13. D-4 Type C section-2. Source: Owned by Akira Tarashima
Archive
roof, and this is thought to have helped reduce the heat load
on the living space on the second floor. However, the back of
the floor was always exposed to outside air due to the adoption (1) to escape the discomfort of humidity; (2) to prevent theft;
of a piloti style; most of the exterior walls on the south side and (3) there was an agreement in Seijo not to install fences
were composed of sliding glass doors; and the walls on the or gates, and the vacant land had been a playground for a long
east, west, and north sides were all glass (with some vertical time, so he wanted to make it open to the community [Note
louvers) up to the ceiling above the waist walls. It is thus pre- (see Footnote 2)]. Although each of these three reasons is suf-
sumed that it was cold in winter, although the sunlight, venti- ficiently convincing, Tange may have adopted piloti for other
lation, and moisture-proof performance were perfect. reasons, including (4) his stance on the land ownership system
Moreover, since there is testimony that the outer walls con- and (5) the influence of Katsura Imperial Villa.
tained asbestos, it is presumed that consideration was given to First, regarding the land ownership system, Tange discussed
the heat insulation performance [Note 2]. the land system of his own residence during a roundtable dis-
cussion (1954) with Kiyoshi Seike, Takamasa Yoshizaka,
Masaru Katsumi, and others, and reported that the children of
5. Tange’s Evaluation of his Residence
the neighborhood used to play freely on the first floor of his
5.1 Piloti: Land liberation and Katsura Imperial Villa house (see Figure 23 P-3). In response, Yoshizaka praised
Tange recalled that, after his house was completed, many peo- Tange’s efforts, but Katsumi pointed out that, while the land
ple asked him why he chose piloti, and he was at a loss for an should ideally belong to society, not to an individual, Tange
answer. He offered the following as examples of his responses: would have to pay the rent and taxes for land that he has

Jpn Archit Rev | 2023 | 9

TOYOKAWA wileyonlinelibrary.com/journal/jar3

FIGURE 14. D-5 Type D perspective. Source: Owned by Akira Tarashima Archive

FIGURE 15. D-5 Type D perspective (enlarged figure of Figure 14). Source: Owned by Akira Tarashima Archive

FIGURE 16. P-1. Source: Owned by Akira Tarashima Archive

Jpn Archit Rev | 2023 | 10

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 17. P-2. Source: Owned by Akira Tarashima Archive

FIGURE 18. G-1 Tange’s Notebook 1952. Source: Owned by Michiko Uchida Archive

Jpn Archit Rev | 2023 | 11

TOYOKAWA wileyonlinelibrary.com/journal/jar3

FIGURE 19. E-1 Type E 1st floor plan, 2nd floor plan. Source: Owned by Akira Tarashima Archive

opened to local residents. Tange replied, “That is because the before the completion of his residence, Tange described the
land belongs to the country.”8 Since it was revealed that the war-damage reconstruction plan as follows: “Many cities have
land of about 300 tsubo on which his Seijo residence was built drawn up ideal plans for the future that look like they were
was purchased with the financial support of novelist Takeo drawn on a blank sheet of paper on the burnt fields. However,
Kato, Tange’s father-in-law,9 it is unclear whether Tange him- those plans have been gradually eroded by the old hidden
self could be considered the landlord in a strict sense. powers, but they are trying to come right back.”11
Tange mentioned the land ownership system of post-war Less than five years after the end of the war, the land natio-
cities in two articles. One was “Problems in Construction,” nalization theory had failed to provide a practical solution to
published in the January 1948 issue of Kenchiku zasshi. In the the land rent problem, and the war-damage reconstruction plan
article, Tange criticized the practice of paying large amounts was not feasible unless the issue of feudal land control was
of key money to the landowner when a small enterprise tried resolved. From Tange’s point of view, the land nationalization
to start a business in an urban area, calling it a feudal land theory was extreme and unfeasible, but the liberalization of the
control relationship. He stated that there is no practical solu- first floor of his residence in the early 1950s may have served
tion even if the theory of land nationalization or of national as a tool that enabled Tange to raise questions against the land
land management is highlighted against the landowners’ pas- ownership system.
sive income-generation as represented by the key money. The previous section explained how the floor of the resi-
Hence, he argued that individual citizens (such as independent dence rose up through the design process. The study discussed
small enterprises and working people) had to fight to secure the documents (C-3 and C-4) wherein Tange examined the
fair contracts from feudal capital by analyzing the social struc- subtle floor height of about 1400, which is too low for a piloti.
ture scientifically and acting accordingly.10 The background to this could be Shingoten and Chushoin of
The other article looked back on Japan’s war damage recon- Katsura Imperial Villa (see Figure 24 P-4, photograph by
struction plan. Although Tange participated in the drafting of Tange), which served as an inspiration for Tange. When
reconstruction plans for Hiroshima and other areas, most of his designing the Hiroshima Peace Memorial Park, Tange men-
proposals were rejected, for reasons such as the issue of land- tioned the influence of Ise Jingu and Katsura Imperial Villa in
owner and leaseholder evictions. In 1949, however, four years his dialogue with Isamu Noguchi,12 and it is easy to imagine

Jpn Archit Rev | 2023 | 12

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 20. F-1 Type F 2nd floor plan. Source: Tange K. Reality and Creativity. Tokyo, Japan: Bijutsu Shuppansha. 1968:100

using the Tange’s version of the modulor (hereinafter referred

to as “Tange Modulor”), imitating Le Corbusier. However,
Tange pointed out that the general standard size of the 3 shaku
by 6 shaku tatami mats was a little too small for the design of
his residence, and he considered 6 shaku for Type A, 4 shaku
for Type B, and 5 shaku for Type C as the basic modules for
the length direction of his residence. He also said about Type
F that “My house is designed using 4 shaku 9 8 shaku as
basic modules.”13 Tange said that 4 shaku 9 8 shaku was an
appropriate scale not only for sitting but also for standing and
moving. Analyses of the sketch materials discovered in this
study indicate that, after various examinations, a final plan was
reached that skillfully combined 4 shaku, 6 shaku, and 8
shaku. In addition, although it is considered appropriate not to
use Tange Modulor used for public buildings, due to the dif-
ference in the scale of his residence, it may have been possible
to consider adopting an industrial module, such as that used
for the Tosho Printing Company Haramachi Factory (published
in the March 1955 issue of Shinkenchiku). However, while
being keenly aware of the 3 shaku 9 6 shaku modules that are
often used in traditional wooden houses, Tange tried to find
FIGURE 21. F-2 Type F 1st floor plan. Source: Owned by Michiko
Uchida Archive the optimal module for his residence by repeatedly studying
plans and elevations using 4 shaku, 5 shaku, and 8 shaku
that he was keenly aware of Katsura Imperial Villa when he scales.
designed his residence, which was completed at about the The second problem is the mixing of chair and floor sitting
same time as the final design of the Hiroshima Peace Memor- styles. Checking photographs of Tange residence in use, it is
ial Park. seen that the family meal was served on a table and chairs of
normal height, set up on the tatami mats. However, when large
5.2 Problems in modules and tatami mats number of guests visited the house, everyone sat down (and
Three issues need to be raised about the use of tatami mats in sometimes lay down) on cushions placed on the tatami mats to
Tange residence. The first concerns the size of the mats. Since eat dinner. About this mixing of chair and floor sitting styles,
completing the design of Hiroshima Peace Park, Tange Tange said, “This is only in an experimental stage, and I do
Laboratory had been conducting collective design activities not think it is something that can be recommended as good for
Jpn Archit Rev | 2023 | 13
TOYOKAWA wileyonlinelibrary.com/journal/jar3

FIGURE 22. F-3 Type F section. Source: Tange K. Reality and Creativity. Tokyo, Japan: Bijutsu Shuppansha. 1968:101

FIGURE 23. P-3. Source: Owned by Michiko Uchida Archive

the general public.” [Note 3]. Tange also placed a low pipe relationship; Figures 25 P-5 and 26 P-6 are owned by ⒸKochi
chair (legless chair, P-5) designed by him in the living room; Prefecture, Ishimoto Yasuhiro Photo Center).
he said, “I decided on the ceiling height, transom glass, and The third is his care in choosing the texture of the tatami
brightness of the ceiling and floor so that when sitting in a low mat (P-6). As was observed in the previous section, after much
position, my eyes would be at the center of gravity of the consideration about his residence, Tange changed most of the
space.” [Note 4]. It can thus be seen that, just as he considered second floor into tatami rooms. Tange explained the reason for
tatami mat sizes in search of the most convenient size, he set using tatami mats, saying, “I wanted the comfort and feel that
the center of gravity of the space low in search of the most tatami has.” [Note 5]. Tange also mentions that, if the interior
comfortable viewpoint position and depth (proportional of the house were fully covered with tatami mats, the

Jpn Archit Rev | 2023 | 14

wileyonlinelibrary.com/journal/jar3 TOYOKAWA

FIGURE 24. P-4. Source: Owned by Michiko Uchida Archive

maintenance cost would be higher than that of a wooden floor

[Note 6].

6. Conclusion
This study has examined the design process for Tange resi-
dence based on sketches and photographs of the residence kept
by Tarashima, who was in charge of designing it. The follow-
ing four points have been clarified.
First, the sketches and the completion plan of the residence
can be classified from Type A to Type F according to the size
of the plan. Looking at the front side of the lengthwise direc-
tion, it is seen that, although 32 shaku, 40 shaku, 50 shaku,
and 60 shaku were examined, 54 shaku was chosen in the final
plan. In each study plan, the allocation of rooms was consid-
ered on the basis of the basic modules of 4 shaku, 5 shaku,
and 6 shaku, and the study confirmed that the plan and eleva-
tion were drawn at the same time. In addition, this study was
able to identify a process proceeding from a modest-scale one-
story house (Type A) to a two-story completion plan (Type F),
which is worthy of being called an “official residence.” Tange
skillfully combined 4 shaku, 6 shaku, and 8 shaku in Type F.
Second, Tange and Tarashima envisioned a one-story house
and an open veranda from Type A to Type C, and the layout
that placed the main entrance and the service entrance back-to-
back bears a close resemblance to the layout of the Urawa FIGURE 25. P-5. Source: Owned by ⒸKochi Prefecture, Ishimoto
Yasuhiro Photo Center
house. These have an appearance close to that of the Farns-
worth House and are consistent with the explanation that Tara-
shima gave Fujimori. In addition, while the children’s room
and parents’ room were private from Type A to Type E, in the completion plan (F-3) by about 2.5 m from the ground and
completion plan (Type F), the walls between the private rooms adopted a piloti style. He cited three reasons for this: (1) to
were removed to create an unrestricted indoor space centered escape the discomfort of humidity; (2) to prevent theft; and (3)
on the core. there was an agreement in Seijo not to install fences or gates.
Third, while Type A and Type C had a floor height of about At that time, Tange also worked on the war damage recon-
60 cm from the ground, in Type D, a plan to raise the floor by struction plan, had strong views on the theory of land nationa-
1.4 m and 1.92 m from the ground was considered, and the lization, and had a keen interest in Katsura Imperial Villa;
study was able to confirm the increase in the floor height dur- these were likely the driving forces behind the adoption of
ing the design process. Tange raised the floor of the piloti.

Jpn Archit Rev | 2023 | 15

TOYOKAWA wileyonlinelibrary.com/journal/jar3

Disclosures
None.

Data Availability Statement

The data that support the findings of this study are available
from the corresponding author upon reasonable request.

Notes
Note 1) Materials from Kawaguchi Kenichi Laboratory, University of
Tokyo.
Note 2) Author. Interview with Michiko Uchida.
Note 3) See Reference [7].
Note 4) See Reference [13].
Note 5) See Reference [7].
Note 6) See Reference [13].

References
1 Fujimori T. Tange’s own residence 1953. In: HOME Editrial Department,
ed. Shouwa Juutaku Memory. X Knowledge; 2005:38-41 (in Japanese).
2 Tarashima A. Interview. In: Toyokawa S, ed. Tange Kenzo and Kenzo
Tange. Ohmsha; 2013:114-117 (in Japanese).
3 Toyokawa S. Tange’s own residence. In: Mori Museam, ed. Catalogue of
exhibition, “Japan in Architecture,” Genealogies of its Transformation.
Echelle-1; 2018:148-149.
4 Toyokawa S. Process and extension sketches for modifying Kenzo Tange’s
own residence. J Archit Plan (Transactions of AIJ). 2020;85(777):2427-
2435. (in Japanese).
5 Fujimori T, Tange K. Chronology of Works, Tange Kenzo. Shin Kenchi-
kusha; 2002:478 (in Japanese).
6 Yamanashi K. Residence in Urawa. Shin Kenchiku. 1955;30(11):60 (in Japa-
nese).
7 Tange K. Tange’s own residence. Shin Kenchiku. 1955;30(1):27 (in Japa-
nese).
8 Tange K, Seike K, Yoshizaka T, Katsumi K. Discussion on contemporary
architecture. Geijutu Shincho. 1954;5(5):174-176. (in Japanese).
FIGURE 26. P-6. Source: Owned by ⒸKochi Prefecture, Ishimoto 9 Kato T. Interview. In: Toyokawa S, ed. Tange Kenzo and Kenzo Tange.
Yasuhiro Photo Center Ohmsha; 2013:832 (in Japanese).
10 Tange K. Problems in construction. Kenchiku zasshi. 1948;737:6-8. (in Japa-
nese).
11 Tange K. Prospects for Tomorrow’s Housing and Cities. Shoukokusha;
Fourth, in Type D, Type E, and Type F, the consideration of 1949:322-323.
a two-story plan can be seen, and the study confirmed that 12 Tange K. Hiroshima plan 1946-1953. Shin Kenchiku. 1954;29(1):13 (in Japa-
they had difficulty in placing the stairs and the earthquake- nese).
resistant wall. In addition, although the study was able to con- 13 Gropius W, Ito S, Ikebe H, et al. Discussion—traditional and modern archi-
tecture. Dr. Gropius’s view of Japanese culture. Shoukokusha; 1956:309-
firm the ingenuity of the roof surface from the cross-sectional 310. (in Japanese).
view, the plan likely led to a house that is cold in winter since
many of the exterior walls are made of glass and the under-
floor is exposed to outside air.
As this study was conducted on the basis of sketches, it did How to cite this article: Toyokawa S. Research on the
not consider the foundation, construction method, ingenuity of design process of Kenzo Tange’s own residence. Jpn
construction, and the detailed dimensional plan (such as the Archit Rev. 2023;6:e12339. https://doi.org/10.1002/2475-
relationship between the column dimensions and tatami splits). 8876.12339
These topics will be taken up in future studies.

Jpn Archit Rev | 2023 | 16