Abstract
The genre of landscape photography known as “Liminal Space” (hereafter referred to as “LS”) has been developing as a genre of Internet aesthetics since around 2019. LS is characterized by the absence of people in places where their presence is expected, and the images have a unique texture. This category is defined in various formats and content in academic formats and on social networking sites, and the most common definitions include the main characteristic of “no people” and the transitional nature indicated by the word “liminal”. However, when images follow these definitions, there has not been sufficient consideration of whether they evoke an LS feeling, and it is possible that simply dark spaces or images with low contrast can generate an LS feeling. Furthermore, although there have been evaluations of the pleasantness or unpleasantness of the impression caused by the presence or absence of people, which is the most distinctive feature of LS, and of the unique texture that is said to contain the eeriness, silence and nostalgia of LS, or in other words, the LS feeling, the examination of what this is has been carried out mainly from an aesthetic perspective, and quantitative analysis is still insufficient.
The aim of this study was to conduct an image evaluation experiment, use the results to clarify the conditions that evoke a sense of LS through regression analysis, and then interpret what LS is through factor analysis. Specifically, in Study 1, we conducted a subjective evaluation of LS for 16 images that manipulated the presence or absence of people in the image (Hypothesis 1: LS is stronger when there are no people), the presence or absence of people passing through the location (Hypothesis 2: LS is stronger when people are expected to pass through), and image features such as brightness, contrast, and saturation (Hypothesis 3: LS is stronger when brightness is low, contrast is low, and saturation is high ) were manipulated, and a subjective evaluation of the LS feeling was conducted, and a regression analysis of the main effects was performed. In Study 2, a subjective evaluation was conducted using 25 evaluation words selected from a preliminary experiment using the images from Study 1, and factor analysis was conducted using 15 words that correlated with the LS feeling based on the evaluation.
From the analysis of Study 1, Hypothesis 1 was supported by a regression coefficient of -0.41 ($p<.001$). As for Hypothesis 2, a slight effect opposite to the hypothesis was observed with a regression coefficient of -0.05 ($p=0.02$). Regarding hypothesis 3, brightness was found to have a slight effect with a regression coefficient of -0.06 ($p=0.001$), while the regression coefficients for both saturation and contrast were not significant. As a result, it was shown that the absence of people is a necessary condition for evoking a sense of LS, and that expectations of people passing by and the brightness of the image have a slight effect on the sense of LS. In addition, the analysis of Study 2 identified three common factors (cumulative contribution rate of 50.83%) - “lonely and unnatural depictions and their eeriness”, “artificial sense of confinement and its discomfort”, and “quietness and lack of a sense of life” - and it became clear that, in addition to eeriness, quietness and a sense of confinement are important factors in the sense of loneliness.
This study is the first attempt to provide a quantitative interpretation of LS from an aesthetic perspective, and it will become a new method for analyzing aesthetic category perception analytically while drawing on aesthetic background.
The aim of this study was to conduct an image evaluation experiment, use the results to clarify the conditions that evoke a sense of LS through regression analysis, and then interpret what LS is through factor analysis. Specifically, in Study 1, we conducted a subjective evaluation of LS for 16 images that manipulated the presence or absence of people in the image (Hypothesis 1: LS is stronger when there are no people), the presence or absence of people passing through the location (Hypothesis 2: LS is stronger when people are expected to pass through), and image features such as brightness, contrast, and saturation (Hypothesis 3: LS is stronger when brightness is low, contrast is low, and saturation is high ) were manipulated, and a subjective evaluation of the LS feeling was conducted, and a regression analysis of the main effects was performed. In Study 2, a subjective evaluation was conducted using 25 evaluation words selected from a preliminary experiment using the images from Study 1, and factor analysis was conducted using 15 words that correlated with the LS feeling based on the evaluation.
From the analysis of Study 1, Hypothesis 1 was supported by a regression coefficient of -0.41 ($p<.001$). As for Hypothesis 2, a slight effect opposite to the hypothesis was observed with a regression coefficient of -0.05 ($p=0.02$). Regarding hypothesis 3, brightness was found to have a slight effect with a regression coefficient of -0.06 ($p=0.001$), while the regression coefficients for both saturation and contrast were not significant. As a result, it was shown that the absence of people is a necessary condition for evoking a sense of LS, and that expectations of people passing by and the brightness of the image have a slight effect on the sense of LS. In addition, the analysis of Study 2 identified three common factors (cumulative contribution rate of 50.83%) - “lonely and unnatural depictions and their eeriness”, “artificial sense of confinement and its discomfort”, and “quietness and lack of a sense of life” - and it became clear that, in addition to eeriness, quietness and a sense of confinement are important factors in the sense of loneliness.
This study is the first attempt to provide a quantitative interpretation of LS from an aesthetic perspective, and it will become a new method for analyzing aesthetic category perception analytically while drawing on aesthetic background.
“Liminal Space”（以下”LS”）というジャンルの景観写真は、2019年ごろからインターネット美学の一ジャンルとして発展してきた。LSは人の存在が期待される場所に人がいないという特徴を持つとされており、画像は独特な質感を有する。このカテゴリの定義は学術的フォーマットやSNS上などで様々な形式、内容で行われており、「人がいない」という最大の特徴や"Liminal"という言葉にが示す過渡的性質などが代表的な定義として挙げられる。しかし、画像がこれらの定義に従ったとき、LS感を喚起するのかについては十分な検討がされておらず、単純に暗い空間やコントラストの浅い画像がLS感を発生させる可能性がある。また、LSの最大特徴である人の在不在による印象の快不快、不気味さの評価はなされているものの、LSの不気味感・静けさ・ノスタルジーを含有するとされる独特な質感、つまりLS感とは何かについての検討は主に美学的な視座から行われており、定量的な分析は未だ不十分である。
本研究は、画像評価実験を行いその結果からLS感を喚起する条件を回帰分析によって明らかにし、次にLS感とは何かを因子分析によって解釈することを目的とした。具体的には、Study1において、画像内の人の有無（仮説1: 人がいない方がLS感が強くなる）、場所の人の往来の有無（仮説2: 人の往来が期待される方がLS感が強くなる）、輝度・コントラスト・彩度の画像特徴量（仮説3: 輝度が低い、コントラストが低い、彩度が高いとLS感が強くなる）を操作した16枚の画像に対して、LS感の主観評価を行い、主効果の回帰分析を実施した。Study2では、Study1の画像を用いて予備実験から選定した25語からなる評価語を用いた主観評価を行い、その評価を基にLS感と相関のある15語で因子分析を行った。
Study1の解析から、仮説1は回帰係数-0.41($p<.001$)で支持された。仮説2に関しては、回帰係数-0.05($p=0.02$)でわずかながら仮説と逆の効果が認められた。仮説3に関しては、輝度は回帰係数-0.06($p=0.001$)でわずかに影響が認められ、彩度とコントラストの回帰係数はともに有意ではなかった。結果、人がいないことがLS感を喚起する必要条件であり、人の往来に対する期待と画像の明るさはわずかにLS感に影響することが示された。また、Study2の解析から、「寂くて不自然な描写と、その不気味感」「人工的な閉塞感と、その居心地の悪さ」「静けさと、生活感のなさ」の３つの共通因子（累積寄与率50.83\%）が特定され、LS感は不気味さに加えて静けさや閉塞感が重要であることが明らかになった。
本研究は、美学的視座から行われてきたLSの解釈に対して、定量的な解釈を行う初の試みであり、美的カテゴリー知覚の分析を、美学的な背景を援用しつつ解析的に行う新たな手法となるだろう。
本研究は、画像評価実験を行いその結果からLS感を喚起する条件を回帰分析によって明らかにし、次にLS感とは何かを因子分析によって解釈することを目的とした。具体的には、Study1において、画像内の人の有無（仮説1: 人がいない方がLS感が強くなる）、場所の人の往来の有無（仮説2: 人の往来が期待される方がLS感が強くなる）、輝度・コントラスト・彩度の画像特徴量（仮説3: 輝度が低い、コントラストが低い、彩度が高いとLS感が強くなる）を操作した16枚の画像に対して、LS感の主観評価を行い、主効果の回帰分析を実施した。Study2では、Study1の画像を用いて予備実験から選定した25語からなる評価語を用いた主観評価を行い、その評価を基にLS感と相関のある15語で因子分析を行った。
Study1の解析から、仮説1は回帰係数-0.41($p<.001$)で支持された。仮説2に関しては、回帰係数-0.05($p=0.02$)でわずかながら仮説と逆の効果が認められた。仮説3に関しては、輝度は回帰係数-0.06($p=0.001$)でわずかに影響が認められ、彩度とコントラストの回帰係数はともに有意ではなかった。結果、人がいないことがLS感を喚起する必要条件であり、人の往来に対する期待と画像の明るさはわずかにLS感に影響することが示された。また、Study2の解析から、「寂くて不自然な描写と、その不気味感」「人工的な閉塞感と、その居心地の悪さ」「静けさと、生活感のなさ」の３つの共通因子（累積寄与率50.83\%）が特定され、LS感は不気味さに加えて静けさや閉塞感が重要であることが明らかになった。
本研究は、美学的視座から行われてきたLSの解釈に対して、定量的な解釈を行う初の試みであり、美的カテゴリー知覚の分析を、美学的な背景を援用しつつ解析的に行う新たな手法となるだろう。
Information
Award / Exhibition / Conference
・[Award] 慶應義塾大学SFC 優秀卒業プロジェクト 2024年度春学期
Asuka Ishii, Masaya Shirai, Hiroyoshi Murata