Tiny Use Case 5 Part II – Actual Co-Occurrence of Blood Type and Character Personality Traits

Part one of this blogpost introduced the research questions of our fifth tiny use case, which stem from the circulation of popular beliefs regarding blood type and temperament in Japan, and their potential influence on character design practices. The two questions were as follows:

  • Do these beliefs about the connection between blood type and personality translate into character designs, such as a correlation between blood type and specific personality traits/descriptors?
  • Does its representation in character data reflect popular beliefs circulating in Japan or does it differ?

We took our questions to two datasets, one coming from VNDB , another from ACDB .

Contrary to our initial expectation, examination of both VNDB and ACDB revealed that listing blood type information for characters is not as common as one might infer from looking at popular media franchises. On ACDB, 4.48% of female characters have blood type listed in their entries vis-à-vis 4.53% in male character data entries, with negligible or null distributions in characters tagged with other genders. On VNDB, the percentage is significantly higher: 16% of female characters and 17.12% of male characters have blood type information listed. However, neither of these two cases allows us to determine whether the distributions found within them actually represents the distribution of blood types across characters in Japanese visual media in general or not. Therefore, we cannot answer our first research question definitively.

While we lack information about the population distribution of blood types in Japanese visual media characters in general. Acknowledging these limits of our data we still proceeded to explore the distribution of blood type information according various dimensions. One of these dimensions that stood out as being particularly distinctive was in relation to game setting. Characters in media set in worlds resembling our own (in the form of some version of contemporary Japan) are more likely to feature characters with a listed blood type. Within video games tagged as set in modern day Japan, the percentage of characters with a listed blood type is circa 48.51%. In video games tagged as being set in a fantasy universe, the percentage drops to about 11%.

While it is possible that this too might be a result of data availability and/or community interest, the difference is significant enough to establish that, at least in visual novel games, there might be some kind of usage pattern of blood type in character design based on the setting. Setting the work in a fantasy world appears to lead to a lack of blood type information for characters, because maybe there is no reason for the blood type personality theory to exist in a fantasy world. Nevertheless, it is possible to point to examples in generalist media franchises where blood type information can still be found despite the fantastical setting: Naruto, One Piece and Attack on Titan, which are all set in fictional worlds, but still feature characters with a listed blood type.

While we could surmise the presence of some sort of pattern in the listing of blood types in character data entries based on the above, we still do not know if such usage can be mapped onto existing beliefs linked to blood type personality theory. Of our two datasets, ACDB’s is not suitable for answering our question, because its data model does not describe character personality in a systematic fashion. On the other hand, VNDB’s granular data model allows us to see the precise distribution of traits related to character personality. The tables below provide percentage figures for the number of characters with a given trait among all characters of a specific blood type.

Percentage of characters within a given blood type with particular personality trait for female characters (VNDB dataset, top 20 scoring traits)

Percentage of characters within a given blood type with particular personality trait for male characters (VNDB dataset, top 20 scoring traits)

Note: in the Japanese language, personal pronouns and ‘opposite gender voiced’ can be used to denote aspects of a character’s personality (ex. a male character voiced by a female voice actor can be marked as a timid person).

Based on the above tables, we can see some differences between the distribution of personality traits among blood types, however, such differences are not as pronounced as the descriptions provided by Nomi or other authors would seem to suggest. Keeping in mind that attempting to map VNDB traits to Nomi’s own descriptions is simply not possible without extensive arbitrary interpretation, if we nevertheless compare the distributions with the trait-based description offered by Sato, Miyazaki and Watanabe (see part one) results in non-correspondence. In fact characters consistently featured frequent co-occurrences of the same trio of traits through all blood types, namely Kind, Friendly, and Watashi (polite personal pronoun) in female characters and Ore (usually considered a coarse personal pronoun) in male characters. Even ignoring ‘Kind’, ‘Watashi’, ‘Friendly’ and ‘Ore’, and looking at data distributions once more still does not lead to the emergence of any distinctive profiles.

Thus, based on our analysis blood type appears to be more like a piece of trivia that is attached to characters in a way to either link them to a specific setting, as it is within visual novel games, or as a piece of additional trivia for fans. All in all, blood type personality theory, as envisioned by Furukawa Takeji and popularized by Nomi Masahiko, does not seem to circulate in Japanese visual media production beyond being traces of existing categories.

As we round up our work on tiny use case five, we can summarize our results as follows:

  • Beliefs about the connection between blood type and personality do not translate into identifiable practices of character design.
  • Listing of blood type can serve as a way of anchoring characters in realistic video game settings, especially within visual novel games. We can argue that listing of blood type in Japanese visual media might be more connected to the overall setting of a media work, and in particular, whenever the work is set in a fictionalized version of our own everyday life or not. But this does not communicate anything about blood type and character personality.