Further Explanations for Difficulties in Reading Recent Japanese Names Correctly

A previous study analyzed the names of babies born in Japan between 2004 and 2018 and demonstrated that common writings have many variations in readings, which makes it difficult to choose the correct reading at first glance (Ogihara, 2021c). The study provides empirical evidence of the difficulties in reading recent Japanese names correctly. The current article answers three questions on Ogihara (2021c). Specifically, I explain 1) it is still difficult to read recent Japanese names correctly even if one tries to remember the most frequent readings of names 2) specific numbers of reading variations are not important, and 3) how we should deal with these difficulties. This new information would certainly help to further comprehend the difficulties in reading recent Japanese names correctly, which contributes to a better understanding of names and naming practices not only in Japan but also across the entire Sinosphere.


INTRODUCTION
analyzed the names of babies born in Japan between 2004 and 2018 and demonstrated that common writings have many variations in readings, which makes it difficult (or almost impossible) to choose the correct reading. For instance, one of the common writings for boys, 大翔 had 18 variations in reading, and for girls, 結愛 had 14 variations in reading. These variations differed remarkably in pronunciation, length, and meaning. This shows that it is difficult to read recent Japanese names correctly at first glance. Prior research had mentioned these difficulties (e.g., Sakata, 2006;Sato, 2007;Kobayashi, 2009;Ohto, 2012), but they had not been empirically examined. Thus, the study provides empirical evidence of the difficulties in reading recent Japanese names correctly at first glance, contributing to a better understanding of names and naming practices not only in Japan but also across the entire Sinosphere (the vast regions where Chinese characters are/were used).
I found questions on this study that are not discussed in Ogihara (2021c) and are helpful to better understand Japanese names and naming practices. Thus, in this article, I answer some of them and add further explanations for the difficulties in reading recent Japanese names correctly. These explanations would contribute to a further understanding of names and naming practices not only in Japan but also across the entire Sinosphere. 1

IS IT TRULY DIFFICULT TO READ RECENT JAPANESE NAMES CORRECTLY? Question
One saw the distribution of readings of the name 結愛 [ Figure 1; Figure 5 of Ogihara (2021c)] and insisted that the name is easy to read because the frequency of the most common reading, Yua, was approximately 60%. In other words, even if one does not know the correct reading of this name, reading it as Yua will be correct with a 60% probability. This questions whether it is truly difficult to read recent Japanese names correctly.

Response
This is incorrect for at least four reasons. It is still difficult to read recent Japanese names correctly, even if one tries to remember the most frequent readings of names.
First, frequencies of the most common readings remarkably differ in names, and it is not always the case that the most common readings have high frequencies. It is important to understand that distributions of name readings in Japan vary widely for names. Indeed, the distributions of reading variations of the eight names that were examined in Ogihara (2021c) are different from each other ( Figure 1). For example, the frequencies of the most common readings range from 50.00% (愛: Ai) to 83.20% (杏: An). Moreover, it is not always true that the frequency of the most common reading is much higher than those of the remaining readings. The distribution of the readings of 愛 is one example of this [ Figure 1; Figure 7 of Ogihara (2021c)]. The most frequent reading was Ai (50.00%), but the second most frequent reading was Mana (38.89%), which was also a frequent reading and its frequency was close to that of the most frequent reading.
Second, distributions of reading variations of names other than those analyzed in Ogihara (2021c) are still unclear. As stated above, distributions of reading variations differ in names. Ogihara (2021c) demonstrated the most common readings for the eight popular writings, but other writings have not been investigated and clarified. Thus, it is still difficult to read other names. The distributions of the readings became clear "after" the study was conducted. Before the study, the distributions had been unknown. In this sense, one of the purposes of Ogihara (2021c) was partially fulfilled, which was to help people read Japanese names correctly.
Third, even if all the most common readings of writings are clear (though this has not been achieved), it is difficult (or almost impossible) to remember all the most common readings of every name. Because there are too many writings, this is not realistic.
Fourth, even if one remembers all the most common readings of every name (though this is impractical), they can change in the future. It is also important to know that distributions of name readings are changeable over time. Thus, the most common readings are not fixed, which leaves it difficult to read Japanese names correctly.
Yet, it should be noted that there may be individual differences in the difficulties of reading recent Japanese names correctly. Not every individual has the difficulties to the same extent. For example, people who engage in educational or medical services may be somewhat better at reading recent Japanese names correctly at first sight. Teachers, instructors, school staff, doctors, nurses, and hospital staff are some examples. This is because they see many recent names with both their writings and readings, store them, and update them regularly. In other words, they have a larger amount Frontiers in Education | www.frontiersin.org February 2022 | Volume 6 | Article 799119 2

Ogihara
Difficulties in Reading Japanese Names of data on baby names and the distributions of their readings stored in their own heads. Thus, they may have better predictions for readings of names. Yet, it is still difficult even for them to read recent names correctly at first sight.

ARE SPECIFIC NUMBERS OF READING VARIATIONS IMPORTANT?
Question Ogihara (2021c) analyzed and revealed the distributions of reading variations of eight popular writings. For example, the study showed 18 variations of readings of the writing 大翔 and 14 variations of readings of the writing 結愛. Some have emphasized these specific numbers of reading variations. For instance, one argued whether a name with seven reading variations is more difficult to read correctly than a name with five reading variations. This questions whether such specific numbers regarding reading variations matter.

Response
Specific numbers of reading variations are not very important. As written in Ogihara (2021c), the study indicated specific numbers because it was necessary to empirically demonstrate the difficulties of reading recent Japanese names and provide readers a clearer understanding of the difficulties. Ogihara (2021c) did not collect all names that were given to newborn babies in Japan between 2004 and 2018. Thus, the numbers of reading variations are the results of data (Ogihara, 2020a; Meiji Yasuda Life Insurance Company, 2020) examined in Ogihara (2021c), and they are not necessarily consistent with those of populations in interest (all names that were given to newborn babies in Japan between 2004 and 2018).
Thus, if more names are investigated, more variations of readings will likely be found. Indeed, there are other readings than those listed in Ogihara (2021c). For example, 大翔 was read Takeru and 結愛 was read Yuria for the same period, both of which were not found in Ogihara (2021c).
Further, numbers of reading variations can change over time. As stated above, distributions of name readings are changeable over time. New readings can appear in the future, which increases the number of reading variations. Similarly, some readings may be less frequently given to Chinese characters, which decreases the number of reading variations. Thus, specific numbers of reading variations are not very important.

SO, HOW SHOULD WE DEAL WITH THE DIFFICULTIES? Question
One agreed with the difficulties of reading recent Japanese names correctly and then asked about how we should deal with these difficulties. Reading names correctly is important in daily communication. Misreading names should be avoided because it is generally considered to be rude in Japan. 2 This question is also important for non-native Japanese speakers who learn Japanese names, language, and culture.

Response
I recommend four strategies to face these difficulties.
First, accepting the difficulties is necessary. The difficulties are somewhat shared in Japan, but there seem to be variations in some layers such as generation and region. Accepting the difficulties and understanding that the difficulties are shared should decrease the pressure to read names without mistakes. It would be particularly effective for non-native Japanese speakers to know that even native Japanese speakers cannot read recent Japanese names correctly.
Second, explicitly asking how a name is read is recommended. When there are many possibilities in readings and no hints as to choosing the correct one, asking for the correct reading is the most straightforward solution. As stated above, at times even native Japanese speakers cannot read names correctly at first glance and misreading names is generally considered to be rude in Japan. Thus, asking for a correct reading before making mistakes is basically not impolite in Japan.
Third, it is effective to learn patterns of reading names in Japan (for a summary, see Ogihara, 2015;2021b). For example, 海 (meaning marine) is read as Marin and 月 (meaning moon) is read as Runa. These are based on a strategy to read Chinese characters as foreign words (e.g., English, Latin, and French) that correspond to the semantic meaning. The percentages of unique names have increased in Japan (Ogihara, 2021a;Ogihara, in press;Ogihara et al., 2015). Thus, capturing how to read uncommon names would be beneficial. This increases the possibility of reading Japanese names correctly. Related to the second point above, it is also effective to ask about the origins and meanings of names, which helps one to understand and remember the names. Yet, it is still difficult to read recent Japanese names correctly at first sight because predicting which pattern is applied to a name is difficult (or almost impossible).
Fourth, knowing more names and updating distributions of reading variations would increase the probability of reading Japanese names correctly. In particular, it is effective to learn how to read common writings as it was conducted for the eight writings in Ogihara (2021c). However, this is difficult for ordinary people who do not have many opportunities to see recent Japanese names.
As stated above, professionals in education and medical services may be better at reading recent names correctly. This is because they have their own databases of names in their heads, update them, and communicate with people (e.g., students, patients) using them in their workplaces, which in turn causes them to implement the four strategies explained above (accepting the difficulties, asking how a name is read explicitly, learning patterns of reading names in Japan, knowing more names, and updating distributions of reading variations).

DISCUSSION
This article answers three questions on Ogihara (2021c). Specifically, I explain 1) it is still difficult to read recent Japanese names correctly even if one tries to remember the most frequent readings of names 2) specific numbers of reading variations are not important, and 3) how we should deal with these difficulties. This new information would certainly be helpful for further comprehending the difficulties in reading recent Japanese names correctly, which contributes to a better understanding of names and naming practices not only in Japan but also across the entire Sinosphere.

AUTHOR CONTRIBUTIONS
The author confirms being the sole contributor of this work and has approved it for publication.