The Relationship Between Personalities and Faces of Manga Characters

Mangaka (manga artists) are limited to portraying the personalities of characters through static images and limited changes in facial expression. It is imperative that the personality and appearance of a character complement each other. In this study, we found that participants rely heavily on facial features to identify the personalities of characters. However, we also found that their first impressions are inclined towards two out of a total of sixteen specific personality types. This preference for specific personality types may be a unique feature of manga. Perhaps because readers are entirely reliant on the visual presentation of manga, facial features provide a relatively rich source of information when determining the personality of a character. The physical appearance of characters is somewhat limited by the stylization typical of manga. Possibly the facial features of characters have become restricted to a few select types, enforcing stereotypical images in the minds of readers. This encapsulation of personalities is important to the development of manga. Simplifying personality types helps readers to identify with characters and understand the plot. However, over-simplification can make characters dull or indistinguishable from each other


Introduction
We would like to suggest the soul of a manga character lies within its role. Readers decide whether they want to read a manga book based mainly on their attraction to the roles of its characters. Psychologists have long known that the first items presented in a new series of information are remembered better than those presented later, and that a person's memory of this information is difficult to change in the following days. This is called the primacy effect (e.g., Waugh and Norman 1965;Glanzer and Cunitz 1966). As a result, readers form their first impression of a character based on its appearance when they first see it on the book cover or in the story. Although they gradually gain a more in-depth understanding of the character as the story progresses, their first impression is difficult to change. The design of characters is the first point of contact for manga readers as well as the first step in getting to know their roles. Therefore, the physical appearance of characters is crucial.
Personality is the first building block of role development and mangaka assign personalities to their characters through the script, which they write before designing any graphics (e.g., Hirohisa 2011;Hikaru 2007;2012;Hikaru and Tubura 2010). It is imperative that the personality and appearance of a character complement each other or the reader may feel confused, misinterpret the story, or even give up reading it altogether. As expressed by Yomota (1994), manga can only be seen, not touched or heard. Therefore, the graphic presentation of the appearances, poses, movements, and facial expressions of its characters all we have to understand them. These physical characteristics represent not only a character's form, but also highlight its personality and unique role in the story. Successful differentiation enables the reader to easily remember and identify each character's role and thereby become absorbed in the story. However, an initially well-presented character is not only easy to differentiate but also allows a reader to accurately deduce the personality of a character prior to reading.
Li (2012) indicated that the publishers of manga (e.g., Kodansha Ltd.; Shogakukan Inc.; Shueisha Inc.) segment the market based on age and gender. There are yonen (kodomomuke) manga for children in the lower grades of elementary school, shonen manga for boys, shojo manga for girls, seinen manga for men, and josei manga for women. Among these manga categories, Shuffle Alliance (2013) has argued that shonen manga is mainstream in today's market. According to the Japanese Ministry of Justice (2008), the term 'juvenile' in Japan refers to a person under twenty years of age. Shueisha Inc. (2014) Mangaka (manga artists) are limited to portraying the personalities of characters through static images and limited changes in facial expression. It is imperative that the personality and appearance of a character complement each other.
In this study, we found that participants rely heavily on facial features to identify the personalities of characters. However, we also found that their first impressions are inclined towards two out of a total of sixteen specific personality types. This preference for specific personality types may be a unique feature of manga. Perhaps because readers are entirely reliant on the visual presentation of manga, facial features provide a relatively rich source of information when determining the personality of a character. The physical appearance of characters is somewhat limited by the stylization typical of manga. Possibly the facial features of characters have become restricted to a few select types, enforcing stereotypical images in the minds of readers.
This encapsulation of personalities is important to the development of manga. Simplifying personality types helps readers to identify with characters and understand the plot. However, over-simplification can make characters dull or indistinguishable from each other.
clearly specified that shonen manga is aimed at male readers of this age; the age of the main character and scope of the content in many works are focused on high school students (e.g., Assassination Classroom created by Matsui 2012; Slam Dunk created by Inoue 1990;2012;YuYu Hakusho created by Togashi 1990;1992). However, Kodansha Ltd. (2014) has stated that a significant number of shonen manga is also suitable for readers of both genders of all ages. This results in the large readership that shonen manga boasts today. We therefore adopted shonen manga as the focus of this study.
This study seeks to determine what physical clues readers rely on to identify the personality of a character, and whether current shonen manga works are successful at enabling readers to recognize personality based on appearance. We also explore how many different personality types can be construed from readers' first impressions of a character.

Study 1 1 Procedure
To understand what physical clues readers use to distinguish the personalities of shonen manga characters and whether these clues enable them to successfully deduce personalities, we recruited 41 undergraduate and graduate students (25 female and 16 male, with an average age of 23.3 years). Each participant was asked to describe the clues he/she used to appraise manga characters. The appearance-based clues used to identify personality were rated in ascending order of importance (least important = 1). By summing up these scores, we discovered which physical characteristics were used to infer personality and the order of their importance.
To find out whether readers could accurately judge personality from appearance, participants were asked to indicate whether they felt they could guess a character's personality based on its appearance before reading the story and how accurate they felt their guesses were upon completion of the story.

Results
We found that the most important physical clues used to identify personality were facial features (i.e., hair, chin, ears, eyebrows, eyes, nose, and mouth), accounting for 54.7 per cent. Of the 41 respondents, 34 (82.9 per cent) indicated that before reading a story they speculate about the personality of a character based on its facial features; 31 (91.2 per cent) believed their guesses to be accurate after having read the story, feeling that they were able to correctly deduce personality from facial features. Seven of the 41 respondents (17.1 per cent) indicated that they do not try to guess the personality of a character based on its facial features before reading the story. However, six of these (85.7 per cent) reported that once they had come to understand a character's personality through reading the story, they felt the facial features of the character seemed to match its disposition.

Study 2 1 Procedure
After identifying the important role that facial features play in helping readers to identify the personalities of shonen manga characters, we wanted to know how many different personality types readers could identify from first impressions of a character's facial features.
From the previous sample group, we recruited 34 undergraduate and graduate students (19 female, 15 male; average age 22.1 years old) from those who indicated they tend to try to guess the personalities of manga characters based on their facial features before reading the story. The stimuli were 88 characters selected from 44 well-designed, current shonen manga works (one male and one female character were selected from each work). We ensured that the selected works were new to the participants, thereby eliminating bias that might arise from prior knowledge of the characters' roles. All the works were drawn by different mangaka so that participants would not be unduly influenced by any particular style. For each character, we presented eight common facial expressions to prevent any specific expression from inordinately influencing the judgment of participants. The characters had fairly stable personalities and their facial features did not change too dramatically. All images were extracted from the manga and efforts were made to select full-frontal depictions that displayed complete, consistent facial features. Textual elements were eliminated from the images. Figure 1 shows the eight facial expressions used by participants to guess the personality type of a particular character. We adapted the Myers-Briggs Type Indicator (MBTI) developed by Myers (1962) to define personality types. According to The Myers and Briggs Foundation (2014), the MBTI comprises four pairs of psychological preferences. The first pair is Extroversion (E) and Introversion (I). It refers to where and how people direct their attention and energy. People who are extroverted put their attention on people and things in the outer world, and draw energy from actions. Conversely, those who are introverted draw energy from dealing with ideas and images in their own inner world. The second pair of psychological preferences is Sensing (S) and Intuition (I). It refers to how people prefer to deal with information. The former refers to people who tend to trust information they receive through their five senses. They pay attention to physical reality, and prefer to look for details and facts. Those who are described by the latter label place more value on impressions and symbolic meanings. They may be more interested in the future than the past. They also like to work with symbols and abstract theories. The third preference pair is Thinking (T) and Feeling (F). It refers to modes of decisionmaking. People who prefer thinking tend to be causal, logical, and consistent in their choices. They try to be objective. Conversely, those who prefer feeling put more weight on personal concerns and the people involved and are concerned with harmony. The last pair of psychological preferences is Judging (J) and Perceiving (P). They refer to how people interact with the outer world. People who prefer judging like to have things settled and organized. They favor a more structured and decided lifestyle. Those who prefer perceiving like to understand and adapt to the world rather than organize it. They like to keep options open, and prefer a more flexible and spontaneous lifestyle. These pairs are used to identify 16 distinctive personality types (MBTI types).
After reading the stories to gain a full understanding of their roles and personalities, we used the MBTI method to assign personality types to manga characters. From a sample of 452 characters, we selected 88 evenly distributed among the 16 MBTI types (5-6 characters for each type). The participants also rated the characters using the MBTI method, although their judgments were made based purely on facial features.

Results
We found the participants' scoring of the personality indicators to be highly reliable. Kendall's W for the four pairs of psychological preferences were .489, .383, .450 and .385, respectively, with p values below .001, indicating significant correlation at a moderate level. This indicates that participants were moderately consistent in their personality ratings of the same character. In other words, there is evidence that readers usually rely on generally accepted cues when guessing the personality types of characters prior to reading.
However, we also found that the MBTI types were unevenly distributed when rated according to first impressions. Table 1 shows the MBTI types of the 88 characters as rated by participants and researchers. This table can be used to compare the differences between personality type as determined from facial features or as determined from reading the story. Although we distributed the characters evenly among the 16 MBTI types, participant ratings were concentrated on two MBTI types: (a) ISTJ: serious and generally quiet; practical and able to singularly focus efforts and abilities, giving the impression of stability and dependability; realistic and logical, decisive and relatively pragmatic; focusing only on existing rules and plans, resulting in very orderly and organized lives; and (b) ENFP: energetic and enthusiastic; lively and friendly, emphasizing personal relationships; imaginative, making quick connections between information; waiting for the support and affirmation of others before making decisions; spontaneous and flexible, frequently making use of their ability to improvise.
Four MBTI types did not feature in the participant ratings at all: (a) ISTP: relaxed and easygoing, quiet observers, preferring objective analysis and taking an interest in exploring causal relationships, leading to an ability to quickly determine the root of a problem and identify solutions; (b) ENTJ: enthusiastic, knowledgeable and decisive-born leaders, able to identify potential problems and make decisions through rational analysis, preferring to make long-term plans and set long-term objectives, and readily assuming leadership of others; (c) ENTP: smart, quick, confident and articulate; creative and imaginative; analytical and often objective in treatment of others; resourceful in trying new methods, enjoying new challenges, possessing a wide range of interests and tending to frequently switch their attention from one thing to another; (d) INFJ: attentive and compassionate, happy to discreetly serve others; wanting to understand relationships and their meaningfulness; able to develop a clear visions of how best to serve the common good and implement these visions step by step, leading to them being respected and emulated by others.
Participants gave little consideration to three other MBTI types: (a) ESTP: fun and good at making friends; analytical, solving problems rationally; energetic, focusing on  the here and now rather than bothering with abstract concepts or theories; enjoying a challenging environment and rebelling against existing rules and procedures; (b) ISFP: quiet and kind; avoiding conflict and seeking to get along with others; not forcing their opinions or values on others; enjoying the present moment, living freely and enjoying their own space and working within their own time frames; (c) INTP: quiet, with little interest in social interaction; logical thinkers who enjoy analyzing and solving problems, value flexibility and do not concern themselves with trivia.

Discussion
In Study 1, we found that facial features were the most important clues in identifying the personalities of shonen manga characters. Perhaps because readers are entirely reliant on the visual presentation of manga, facial features provide a relatively rich source of information when determining the personality of a character and what emotions he/she is conveying. Mangaka often zoom in on the faces of characters who are in dialogue or otherwise interacting, highlighting their facial features, which then naturally become important clues to identifying personality types. We also found that participants often make judgements about the personalities of characters based on their facial features, before reading the story. It is possible that this allows for readers to understand the characters more quickly and logically judge their behaviour the first time they appear, enabling them to better understand the story. In choosing whether to read a specific manga book, readers assess the appeal of its characters. At this point, readers consider not only the appearance of characters but whether their personalities are attractive. Therefore, a well-designed face helps readers to understand the personality of a character and better identify with him/her in a short period of time, making it easier to immerse themselves in the world created by the mangaka.
After reading the story, most participants felt that their judgments were accurate, despite the narrow range of personality types allocated. Possibly the readers believe this narrow range is an appropriate representation of the personalities present in manga, given the limitations of the genre. Readers are somewhat vague in their definition of personality types and therefore tend to identify many personality traits as a single type. As the plot develops and they gain a better understanding of the characters, they continue to expand their definition to support their initial judgment.
Gender identification was found to affect judgments about personality type. For example, some participants believed a character to be female based on its facial features, but when advised after the experiment that the character was male, they changed their initial guess of personality type. This is perhaps because the participants made stereotypical assessments of gender roles. A softfeatured appearance was considered appropriate for a female role; however, when informed that the character was male, participants felt he was 'unmanly' and changed their judgement of personality type. Readers are often confused by gender when reading manga, and this is increasingly being used as a deliberate tactic by mangaka (such as females dressing as males). The question of how gender influences the personality judgments of readers is a topic worthy of further investigation.
In Study 2, we found that participants tend towards the identification of specific personality types, particularly ISTJ and ENFP types. To explain this tendency we turn to Field (1979;2005), who purported that 'all drama is conflict'. He further explained that without conflict, there is no action; without action, there is no character; without character, there is no story. Therefore a good story contains conflict and its resolution, prompting mangaka to often present opposite or conflicting characters -ISTJ and ENFP are an appropriate example. Characters with such significant differences in personality make for a great deal of conflict, adding dramatic tension and strengthening the personality traits of each. This serves to further animate the characters and draw readers into the story. It may be the case that manga editors have found that ISTJ and ENFP characters are the most popular with readers, and they consequently mandate the inclusion of these roles in the works of the mangaka. It is also possible that ISTJ and ENFP types are easier to present graphically, or lead to more vivid and attractive visual effects. Readers of manga may then become more familiar with these personality types, leading to a tendency to place characters within these developed stereotypes at the start of the story.
This tendency may be exacerbated by the preference of shonen manga readers for characters with simple, easyto-understand personalities. Although the shonen manga audience is fairly broad and covers different gender and age brackets, the readers of this sub-genre are generally male high school students. According to Li (2012), in shonen manga, the elements of the story and the personalities of the characters should match the interest and the world view of the reader to be attractive and inviting. Mangaka therefore include a higher proportion of characters of this gender and age group in order to draw these readers into the story. In addition, as is typical of young adult fiction, the personalities of the characters are somewhat simplified. The impulsive, spontaneous nature of the ENFP personality type makes it popular among adolescent readers and it lends itself well to themes of friendship, struggle and victory, which are common threads in manga stories (e.g., Son Goku in Toriyama 1984;2004;Hanamichi Sakuragi in Inoue 1990;2012;Uzumaki Naruto in Kishimoto 2000;Eren Jaeger in Isayama 2010;2011). The ISTJ type is a clear contrast to ENFP, highlighting the characteristics of the latter. The ISTJ type is therefore also common in manga, often in an important supporting role to the protagonist as a companion or teacher, or even as the ultimate rival (e.g., Piccolo in Toriyama 1984;2004;Kaede Rukawa in Inoue 1990;2012;Uchiha Sasuke in Kishimoto 2000;Levi in Isayama 2010;2011) (Figure 2). A pair of conflicting characters and indeed multiple pairs of these make a deep impression on the reader, so much so that readers might develop a tendency to categorize characters into these conflicting types before reading the story. Although this interaction is shaped by the relationship between the mangaka and the target audience, this predisposition will affect all readers of the text.
Shonen manga -with its short frequent publication cycle-grants mangaka only 10-20 pages to attract the attention of readers. If the plot is too complex, and demanding of lengthy character development, readers may lose interest in the story. This might explain why more complex personality types like ISTP, ENTJ, ENTP and INFJ, as well as ESTP, ISFP and INTP, were not initially considered by readers. The subtleties inherent in these types are difficult to represent. Therefore even though mangaka may work hard to develop a wider range of character types, readers might tend to neglect their more complex aspects because they do not personally prefer these personality types. This preference for specific personality types may also be a unique feature of manga. Although dialogue, movement and development of specific events are used to enrich the personalities of characters, their physical appearance is somewhat limited by the stylization typical of manga. Perhaps as a result, the facial features and expressions of characters have become restricted to a few select types, enforcing stereotypical images in the minds of readers. Thus when readers guess the personality type based on facial features, they may consider only a limited range of personalities, having come to make the judgement with the previously formed belief that a certain combination of facial features automatically places a character within a certain category. The question of which combinations are associated with which personality type is an issue for further research.

Conclusion
We found that in general readers rely heavily on facial features to identify the personalities of shonen manga characters and are very confident in their judgments. However, we also found that their first impressions are inclined towards specific personality types. This encapsulation of personalities is important to the development of manga. Simplifying personality types helps readers to identify with characters and understand the plot. However, oversimplification can make characters dull or indistinguishable from each other.
Mangaka are limited to portraying the personalities of characters through brief dialogue, static images and limited changes in facial expression, in contrast to film productions, which provide viewers with a great deal of information on their characters through audio effects, continuous action, and rich, subtle changes in demeanour. Novelists are also privy to a wider range of resources in terms of character development, having unlimited textual space to detail the innermost selves of their characters. Mangaka, however, must use exaggerated motions or brief dialogue, as well as interaction with other characters, to portray complex human personalities. This makes it difficult for readers to gain a complete picture of a character's personality within a short period of time. Initially, readers rely on facial features to identify personality; however, the stylization inherent to manga limits the differentiation of facial features, causing readers to identify characters with different personalities as a single type, which leads to various reading difficulties. If mangaka did more to highlight the dissimilarity of facial features, this could help readers to correctly identify the intended personality types at the beginning of the story, making the plot more understandable and potentially more enjoyable.