2.2. Framework
The framework of a building comprises the following elements: columns (Japan. 柱径 hasirakei), sleepers (Japan. 地覆 jifuku), hip-penetrating tie beams (Japan. 腰貫 koshi-nuki), neck-penetrating tie beams (Japan. 飛貫 hi-nuki), and head-penetrating tie beams (Japan. 頭貫 kashira-nuki).
Table 2 and
Table 3 present the dimensions of these elements measured in the basic unit (fen).
First, the author focused on the standard diameter of a column (
Table 2). According to the Yingzhao Fashi, columns should have a diameter of 42 to 45 units (fen). From these standard column diameters, the author derived four integer multiples: 42, 43, 44, and 45. The author applied these multiples in the author’s analysis of the extant buildings. The author’s analysis of the Korean buildings revealed the following results. When the multiple was 42, 10 of the Korean buildings had at least a 90% concordance, 12 had a concordance rate of 80% to 89%, and 8 had a concordance rate of 70% to 79%. When the multiple was 43, the results were nine buildings, 11 buildings, and 8 buildings; when it was 44, the results were 9 buildings, 10 buildings, and 7 buildings; when it was 45, the results were 12 buildings, 6 buildings, and 7 buildings. Crucially, the concordance rate was the highest when the column diameter was scaled at 42 units; 30 of the 43 Korean buildings the author analyzed had a concordance rate of at least 70% at this scale. This observation implies that Korean buildings are closely aligned with the standard dimensions regardless of architectural style, time period, and region. This result (30 out of 43) is equivalent to 70%.
Analysis of the Japanese buildings revealed the following results. When the multiple was 42, 14 of the Japanese buildings had at least a 90% concordance, 8 had a concordance rate of 80% to 89%, and seven had a concordance rate of 70% to 79%. When the multiple was 43, the results were 13, 8, and 7; when it was 44, the results were 13, 5, and 6; when it was 45, the results were 12, 6, and 4. When the multiple was 42, the Japanese buildings as a whole generally concorded, with 29 buildings having a close concordance rate.
A temporal variation can be observed: Buildings dated between the late 12th century and late 14th century are out of proportion to all the multiples (42 to 45). Japanese buildings tended to feature smaller sizes than the standard diameters for columns (42 units to 45 units). Buildings built between the late 12th century account for 14 of the 52 Japanese buildings the author analyzed. If these 14 buildings are excluded, then of the remaining 38 buildings, 29 (76%) have a high concordance rate. As for regional variation, discrepancies can be found among buildings in Nara, Nagano, Hyogo, Shiga, Tochigi, and Gifu, suggesting that regional variation occurs across Japan, with no concentration in any particular region.
Subsequently, the author examined the height (vertical length) of a sleeper. The standard length is 17 units or 18 units. The author tested both multiples (17 and 18) on the Korean and Japanese buildings. Analysis of the Korean buildings revealed the following results. When the multiple was 17, 7 of the Korean buildings had at least a 90% concordance, 5 had a concordance rate of 80% to 89%, and 12 had a concordance rate of 70% to 79%. When the multiple was 18, 10 of the Korean buildings had at least a 90% concordance, eight had a concordance rate of 80% to 89%, and eight had a concordance rate of 70% to 79%. Thus, when a sleeper’s height was scaled at 18 units, 26 (74%) of the 35 Korean buildings had a concordance rate of at least 70%; this excludes four buildings for which the sleeper length was unclear and buildings that had a column-centered bracket set (柱心包 Republic of Korea. jusimpo Japan. chushinho) that was inconsistent with the scale. Analysis of buildings with a column-centered bracket set revealed that the components were larger than the standard dimensions. Analysis of the Japanese buildings revealed the following results. When the multiple was 17, 10 of the Japanese buildings had at least a 90% concordance, 12 had a concordance rate of 80% to 89%, and 1 had a concordance rate of 70% to 79%. When the multiple was 18, 10 of the Japanese buildings had at least a 90% concordance, nine had a concordance rate of 80% to 89%, and four had a concordance rate of 70% to 79%. Thus, the concordance rate is higher when the sleeper length is scaled at 18 units. A high concordance rate can be observed among the buildings built in the 12th to 14th centuries, in contrast to the case of the column diameter. Overall, the concordance rate is lower than that of the Korean buildings.
The author then analyzed the standard sleeper width, which is typically 12 units. The author tested this multiple on the Korean and Japanese buildings. Analysis of the Korean buildings revealed the following results. In total, 6 of the Korean buildings had at least a 90% concordance, 8 had a concordance rate of 80% to 89%, and 13 had a concordance rate of 70% to 79%. The author observed the variation by time period and architectural style: The concordance rate was higher among buildings built in the 15th, 16th, and 17th centuries. The sleepers were wider than standard among buildings that used a column-centered bracket set. Analysis of the Japanese buildings revealed the following results. Compared to the Korean buildings, fewer Japanese buildings overall had a high concordance: 11 of the Japanese buildings had at least a 90% concordance, 10 had a concordance rate of 80% to 89%, and 7 had a concordance rate of 70% to 79%. As with Korean buildings, the author observed that the concordance rate was higher among buildings built in the 15th, 16th, and 17th centuries, suggesting that those built before the 15th century were likely to be out of proportion. In both Korean and Japanese buildings, the dimensions tended to be larger than the standard dimensions stipulated in the Yingzao Fashi.
The next dimension the author investigated was the standard height of neck-penetrating and hip-penetrating tie beams. According to the Yingzao Fashi, the standard height of the former is 27 units and that of the latter is 28 units. Analysis of the Korean buildings revealed the following results. When the multiple was 27, three of the Korean buildings had at least a 90% concordance, eight had a concordance rate of 80% to 89%, and nine had a concordance rate of 70% to 79%. When the multiple was 28, three of the Korean buildings had at least a 90% concordance, seven had a concordance rate of 80% to 89%, and eight had a concordance rate of 70% to 79%. In both cases, the concordance rates were low. However, the concordance rate was high among the buildings with a column-centered bracket set (柱心包 Republic of Korea. jusimpo Japan. Chushinpo). Analysis of the Japanese buildings revealed the following results. When the multiple was 27, 12 of the Japanese buildings had at least a 90% concordance, 11 had a concordance rate of 80% to 89%, and 6 had a concordance rate of 70% to 79%. When the multiple was 28, 14 of the Japanese buildings had at least a 90% concordance, 10 had a concordance rate of 80% to 89%, and 3 had a concordance rate of 70% to 79%.
Following this, the author assessed the standard width of neck-penetrating and hip-penetrating tie beams. According to the Yingzao Fashi, both widths are 18 units. Analysis of the Korean buildings revealed that two of the Korean buildings had at least a 90% concordance, four had a concordance rate of 80% to 89%, and two had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that three of the Japanese buildings had at least a 90% concordance, six had a concordance rate of 80% to 89%, and 10 had a concordance rate of 70% to 79%. Thus, in both the Korean and Japanese buildings, the dimensions were inconsistent with the Yingzao Fashi’s specifications; they tended to be smaller than the standard dimensions stipulated in the Yingzao Fashi. However, as was the case with the height of these elements, the concordance rate was high among the Korean buildings with a column-centered bracket set (柱心包 Republic of Korea. jusimpo Japan. chushinpo).
Next, the author focused on the standard height of a head-penetrating tie beam (
Table 3). The standard height is 30 units. Analysis of the Korean buildings revealed that six of the Korean buildings had at least a 90% concordance, 15 had a concordance rate of 80% to 89%, and 2 had a concordance rate of 70% to 79%.
Analysis of the Japanese buildings revealed that 13 of the Japanese buildings had at least a 90% concordance, 12 had a concordance rate of 80% to 89%, and 5 had a concordance rate of 70% to 79%. Among the Japanese buildings, those built in the 12th, 13th, and 14th centuries had poor concordance rates, while the concordance rate was better for buildings built before the 14th century. Among the Korean buildings, those built in the 19th century had poor concordance rates, while the older buildings had good concordance rates. In the Korean and Japanese buildings with non-standard dimensions, the dimensions were smaller than standard, implying a lack of region-specific patterns.
Lastly, the author analyzed the standard width of a head-penetrating tie beam (
Table 3). The standard length is 20 units. Analysis of the Korean buildings revealed that nine of them had at least a 90% concordance, five had a concordance rate of 80% to 89%, and nine had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that 13 of the Japanese buildings had at least a 90% concordance, nine had a concordance rate of 80% to 89%, and six had a concordance rate of 70% to 79%. As in the case of length, Japanese buildings constructed in the 12th, 13th, and 14th centuries had poorer concordance rates. In the Korean and Japanese buildings with non-standard dimensions, the dimensions are smaller than standard.
The analysis of framework elements can be summarized as follows. The greatest concordance rate was seen in column diameter. Given that Japan’s kiwari manuals stipulate column diameter as the basic unit for determining the dimensions, it is unsurprising that, when the three countries are compared, the author find a relatively high concordance rate with the column diameters of the buildings described in Japanese kiwari manuals. Temporal variation was observed: In Japan, buildings constructed in the 12th, 13th, or 14th century were either more likely or less likely (depending on the element in question) to be in concordance with the standard dimensions set out in the Yingzao Fashi. In the century after this time range, Japan entered the Warring States period (it began in 1467 or 1493). After the Warring States period, handcrafted architectural elements became larger [
8]. However, how this change in element size is related to the change in architectural technology remains unclear. The dimensions of the buildings the author analyzed were smaller than the standard dimensions set out in the Yingzao Fashi, with the exception of sleeper dimensions.
2.3. Bracket Set
The bracket set comprises the following elements: bracket arms (Japan. 肘木 hijiki), large bearing blocks (Japan. 大斗 daito), and small bearing blocks (Japan. 巻斗 makito).
Table 4 presents the dimensions of these elements measured in the basic unit (fen). First, the author focused on the standard height of a large bearing block. The standard height is 20 fen. Analysis of the Korean buildings revealed that 11 of the Korean buildings had at least a 90% concordance, 9 had a concordance rate of 80% to 89%, and 4 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that 25 of the Japanese buildings had at least a 90% concordance, 11 had a concordance rate of 80% to 89%, and 6 had a concordance rate of 70% to 79%. Notably, as many as 42 of these 44 Japanese buildings with large bearing blocks had a concordance rate of at least 70%. Moreover, 25 had a concordance rate of at least 90%. As for the Korean buildings, in those with a column-centered bracket set (柱心包 Republic of Korea. jusimpo Japan. chushinpo), the blocks were at least twice as long as the standard length.
The author then turned to the standard width of a large bearing block. The Yingzao Fashi stipulates a standard width of 32 units. Analysis of the Korean buildings revealed that 13 of the Korean buildings had at least a 90% concordance, 13 had a concordance rate of 80% to 89%, and 11 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that 15 of the Japanese buildings had at least a 90% concordance, 17 had a concordance rate of 80% to 89%, and five had a concordance rate of 70% to 79%. Of the Korean buildings, 37 (86%) had a concordance rate of at least 70%. Among the non-concordant cases, the blocks were wider than the standard width. Large bearing blocks were present in 44 of the Japanese buildings. Of these 44 buildings, 37 (84%) had a concordance rate of at least 70%. Among the non-concordant cases, the blocks were wider than the standard width, as in the case of the Korean non-concordant cases.
Following this, the author examined the standard length of a large bearing block. The standard depth is 32 units. Analysis of the Korean buildings revealed that 18 of the Korean buildings had at least a 90% concordance, seven had a concordance rate of 80% to 89%, and 12 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that 18 of the Japanese buildings had at least a 90% concordance, 17 had a concordance rate of 80% to 89%, and 4 had a concordance rate of 70% to 79%. Of the Korean buildings, 37 had a concordance of at least 70%, the same number as in the case of width, but the number of buildings with a concordance of at least 90% was higher than that in the case of width. The Japanese buildings, likewise, had a higher concordance rate than they did in the case of width.
Table 5 presents the standard height of a small bearing block. The standard height is 10 units. Analysis of the Korean buildings revealed that 7 of them had at least a 90% concordance, 9 had a concordance rate of 80% to 89%, and 11 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that one of them had at least a 90% concordance, four had a concordance rate of 80% to 89%, and three had a concordance rate of 70% to 79%.
The standard width of a small bearing block. The standard width is 16 units. Analysis of the Korean buildings revealed that 15 of them had at least a 90% concordance, 10 had a concordance rate of 80% to 89%, and 8 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that four of them had at least a 90% concordance, seven had a concordance rate of 80% to 89%, and five had a concordance rate of 70% to 79%. Among the Korean buildings, the concordance rate was at least 70% in 32 of the 36 buildings that had a multi-bracket set (多包 Republic of Korea. Dapo, Japan. Taho).
Lastly, the standard length of a small bearing block. The standard width is 16 units. Analysis of the Korean buildings revealed that 16 of them had at least a 90% concordance, 10 had a concordance rate of 80% to 89%, and 7 had a concordance rate of 70% to 79%. Analysis of the Japanese buildings revealed that four of them had at least a 90% concordance, eight had a concordance rate of 80% to 89%, and nine had a concordance rate of 70% to 79%.
The bracket set also includes bracket arms, but the author already analyzed this element in
Section 2.1.
The above analysis can be summarized as follows. Of the elements analyzed, large bearing blocks were associated with the highest concordance rates. Notably, as many as 42 (95%) of the 44 Japanese buildings had a concordance rate of at least 70% for the height of a large bearing block. Similarly, as many as 86% of the buildings in Korea had a concordance rate of at least 70% for the width and height of a large bearing block.
Regarding dimensions for a small bearing block, among the Korean buildings that featured a multi-bracket set (among which the author had confirmed discrepancies based on architectural style. 多包 Republic of Korea. Dapo, Japan. Taho), the small bearing blocks had smaller-than-standard dimensions, while buildings that featured a column-centered set had small bearing blocks with larger-than-standard dimensions. One possible reason for this observation is that multi-bracket sets use more small bearing blocks than column-centered bracket sets (柱心包 Republic of Korea. jusimpo Japan. Chushinpo). Among the Japanese buildings, the author found many instances where the dimensions were larger than the standard dimensions prescribed in the Yingzao Fashi.