Kyôgi Karuta Overseas: Analysing how Phonetic Variation in the Kimari-Ji of The Hyakunin Isshu affects Gameplay in International Adaptations
Volume 24, Issue 3 (Article 7 in 2024). First published in ejcjs on 13 December 2024.
Abstract
‘Kyōgi Karuta’ is a traditional card game from Japan wherein two players compete to match poetry verses from a set of famous Japanese poems known as the ‘Hyakunin Isshu’. International players can choose to play with the original Japanese poems, translated versions or use an alternative anthology of poems in their native tongue. Due to the game’s strong auditory and literary component, the choice of poems can significantly affect the game’s dynamics—particularly for highly skilled players. Of which, the unique starting sounds of poems (kimari-ji) are the most impactful. This research examines statistical properties of kimari-ji across several international adaptations of the poems and their resultant influence on the game. Our findings show notable phonetic differences within these adaptations, which, while inconsequential for amateur play, may have important implications for professional gameplay and foreign standardisation/tournament play. We provide recommendations to future adaptations on how to mitigate these effects.
Keywords: Karuta, Kyōgi Karuta, Hyakunin Isshu, Chihayafuru, Traditional Card Games, Game Translation, Poetry, Waka
Introduction
The concept of information being ‘lost in translation’ is a classic problem with translating a work from one language to the other. Though the term may have become part of the public zeitgeist with the release of the 2003 film Lost in Translation, the problem it describes has existed since works were first translated. In 1923, Walter Benjamin described the problem at length in his essay, "The Task of the Translator" (Benjamin, 1923).
The unfathomable, the mysterious, the “poetic,” something that a translator can reproduce only if he is also a poet? This, actually, is the cause of another characteristic of inferior translation, which consequently we may define as the inaccurate transmission of an inessential content. This will be true whenever a translation undertakes to serve the reader. However, if it were intended for the reader, the same would have to apply to the original. If the original does not exist for the reader’s sake, how could the translation be understood on the basis of this premise?
A particularly famous victim of this phenomenon is the Bible, of which there are several popular translations, differing collections of texts, and many scholarly critiques thereof. Naude (2004) provides an insightful overview of a selection of these, with particular attention to the influence of culture at the time of translation on the resultant text. Such translations hold a tangible influence over Christians and their beliefs to this day.
As Benjamin mentions above, items of artistic quality are often one of the more difficult categories of media to translate accurately, and all the more important to do so, for much of their intangible qualities are what distinguish them from rote literature.
The conflicts for a translator can be summarised succinctly as striking a balance between fidelity and freedom—i.e., between preserving the literal word-for-word definition of a poem, and of adapting it to better reflect the intended message, whether that be rhythm, references, or something else ‘between the lines’.
With languages with such different construction, cultures, and histories as English and Japanese, translation of Japanese poetry is often exceedingly difficult. The book One Hundred Frogs (Sato, 1995) exemplifies this complexity, containing a hundred unique translations of the famous haiku of a frog jumping into a pond by Matsuo Basho. Whilst some are parodical in nature, the book displays a vast range of possible translations where translators prioritise different aspects of the short poem—whether that be attempting to preserve the signature syllable count of the haiku, the imagery of the scene, a word-by-word translation, or a rewrite to try to translate the intangible feeling behind the poem.
古池や An old pond;
蛙飛び込む A frog jumps in -
水の音 The sound of water
Figure 1. Original poem by Basho (1686) and a translation by Blyth (1942)
In Figure 1 above, one of the more popular translations of Basho’s poems is shown. While it depicts the imagery of the scene adequately, it loses some essential characteristics of the original translation, including:
• The distinctive 5-7-5 cadence of a haiku is not present in the translation.
• The word 「蛙」 (kawazu), which has a strong seasonal connection with Spring, whereas ‘frog’ lacks this association (this feature of haiku is called ‘kigo’).
• The 「や」(ya) in the original serves as a “kireji,” creating a ‘cutting’ of a stream of thought. The semi-colon in the English translation attempts to replicate this, but lacks a true equivalent.
All three of these features are considered a requirement of haiku in the original Japanese (Association of Japanese Classical Haiku, n.d.), but are lost when converted to English. While the translated poem may competently convey the ‘spirit’ of the poem for a broader audience, readers familiar with the traditional form of haiku may prefer translations with stricter regard for form. Like this, for many adaptations, the target audience of a media translation will often inform translators on trade-offs to make. There are several in-depth analyses and commentaries on translating this short poem including a study by Jularbal (2023) which explores three specific translations and their chosen interpretation of diction and explanation.
A hotly contested area of debate between foreign fans of Japanese animation is the choice of subtitles or dubbing. Dubbed anime provides a higher level of immersion in the media but is restricted by convey information in the same amount of time as the original voices, whereas subtitled anime often has the freedom to include more verbose or direct translations, or even provide additional clarifying information. Abdallah (2018) found that study participants had a greater understanding of cultural learnings when watching subtitled shows versus the groups watching dubbed or no anime, showing the subtle impact that such decisions can have on the reception of audiences to translated media.
With the increasing popularity of Japanese culture in the Anglosphere (Allen & Sakamoto, 2014), many more translations have been made of Japanese media, both old and new, by scholars and businesses alike. However, one study by Horton (2018) finds that Western translations of media have increasingly diverged from the original Japanese stylistic intentions. This warrants concern for media where preservation of these key features is important. On the topic of business publications, one could hypothesise that inaccuracies in translation might be exasperated by the financial incentive to translate many pieces of media at a low cost, rather than spend expert time deliberating on subtleties of meaning and intent.
In this paper, we examine how the challenges of accurately translating Japanese poetry might affect the competitive card game ‘Kyōgi Karuta’, sometimes known as ‘Competitive Karuta’. This game, which we describe more comprehensively in the following section, is a one-on-one game where players attempt to capture more cards than the other. Players may only claim cards corresponding with poetry verses read aloud by a reciter, giving the poems an additional auditory dimension that may potentially be lost in translation should it not be thoughtfully considered by translators.
Our analysis compares several different international adaptations of the game and their phonetic distributions. We discuss several key differences in gameplay between these versions and the original Japanese poems including variance in required reaction times and the speed with which to identify matching cards. We find that none of the existing adaptations we studied model the phonetic distributions of the original decks perfectly, and provide suggestions on how to bring adaptations closer to the original.
Concluding the paper is a discussion of the findings and actionable suggestions for translators to attempt to mitigate deviance from the original Japanese. For full transparency in the research, we also list short-comings or gaps where applicable. As this research contains many terms that may be unfamiliar to those not familiar with the game, or Japanese poetry terminology, a Glossary is provided at the end of this paper.
Background Information
Kyōgi Karuta: Competitive Karuta
‘Kyōgi Karuta’ (henceforth referred to as simply ‘Karuta’ for the rest of the paper) is a Japanese card game, which gained mainstream popularity in the Edo period (1603-1868) due to improvements in mass production of woodblock printing. Bull (1996), a woodblock printer himself, writes about how this game is some hybrid of both a sport and a cultural event. The game requires both quick mental and physical ability but also embodies a deep connection to the country’s culture. This dual nature can be seen in the player’s attire: While players often wear easy-to-move-in clothing such as fitness wear or t-shirts during practise, official tournaments often require the wearing of traditional cultural attire such as kimono with hakama (All Japan Karuta Association, 2020).
Two players kneel on tatami mats, facing one another with a set of cards laid out between them. Each player has their own ‘territory’ where they organise twenty-five of their assigned cards as they wish within the regulated area in front of them (See Figure 2 for exact measurements). Players will often place cards tactically at locations where they are easier for them to reach, harder for their opponent, or some other strategic choice. Your own territory is called ‘ji-jin’ while the opponent’s area is called ‘aite-jin’ or ‘teki-jin’.
Figure 2. The initial game setup and measurements of the play area.
The rules of the game have remained mostly unchanged since their formal establishment in 1904 at the first karuta tournament held in Tokyo; After a period of 15 minutes (used by players to memorise the placement of each other’s cards or to practise their movements), the players bow to each other and a designated reciter reads aloud an introductory poem (the ‘Joka’) which signals the beginning of play.
The reciter continues on to begin the reading of each poem in the reciter’s deck of cards. As players hear the words from the reciter, they can begin searching for the partnered phrases printed on the cards between them. If spotted (and it may not be, as players only have fifty out of a possible one hundred poems in front of them.), players will try to take the matching card faster than their opponent. They can achieve this by touching the correct card directly, or by sweeping a set of cards (that contain the target card) out of the field of play. Claimed cards are then removed from play. The goal of the game is to have no cards remaining in your territory before your opponent. If you remove a card from your opponent’s territory, you may move one of your remaining cards from your territory to theirs (thus reducing the card total in your territory).
Hyakunin Isshu: The 100 Poets
The poems used in karuta are a standardised anthology of poems known as the Ogura ‘Hyakunin Isshu’/百人一首 (‘百 (One hundred) 人 (people), 一 (one) 首 (poem)’), compiled by Fujiwara no Teika in the mountain village of Ogura, Kyoto (Mostow, 2023). As the name might suggest, it is a collection of one hundred poems, with each poem being written by a unique poet. The list of poets includes former emperors and empresses, priests, counsellors, and other notable persons of high standing. These poems are examples of waka (和歌 ) poetry, a classical poetic form prevelant from the 6th to 14th centuries .More specifically, the poems in the Hyakunin Isshu are ‘tanka’(短歌) poems which consist of thirty-one syllables in five lines in a 5-7-5-7-7 pattern, a progenitor of the more famous haiku which is 5-7-5.
The collection is split into two card decks; The first consists of full poems and are known as ‘yomifuda’. These cards are read aloud by the reciter during the game. The second deck contains the final two lines of each poem and are known as ‘torifuda’. This set of cards are used by players when matching cards with the poem being recited. For clarity, we show Poem 17, its romanisation and its corresponding yomifuda and torifuda in Figure 3. Whilst all of the yomifuda will be read by the reciter, usually, only half of the one hundred torifuda are in play in one game. This shortens the length of the game, increases the difficulty of card counting and requires players to be careful of pre-emptively claiming cards. Figure 3. Poem 17 (left: Japanese; right: English romaji) [1]
Many translations of these culturally significant poems exist but give differing levels of importance to preserving key elements of the poems. In Figure 4, we show four popular translations of Poem 17. (Another popular source is MacCauley (1917); however, The University of Virginia translation we are referencing is a modern attempt at improvement on MacCauley’s and since the versions of Poem 17 differ little, it was excluded from the figure).
Figure 4. Poem 17 translated by various sources.
The range of approaches to the poem’s translation can be seen well here. Mostow and the University of Virginia’s translations favour more literal translation of the original text (though the resultant translations are still quite distinct), while Honda and McMillan take more creative liberty in attempting to preserve the poetic nature of the scene. Only the University of Virginia’s translation preserves the 5-7-5-7-7 form, and McMillan and Honda both use a different number of lines in their translations. While one can intuit some natural breaking points for torifuda and yomifuda, they will be of different rhythmic lengths to the original.
Ultimately, the question of which is the ‘best’ translation is a subjective one. However, the game of karuta requires the torifuda/yomifuda split and is affected by the speed at which players can uniquely identify the poems; thus, we theorise that while an ordering of ‘worst’ to ‘best’ translation may still be difficult to define, there should be some methodology for determining the level of adherence to these qualities and the rough suitability for the game.
Even greater deviation from the poems is taken when considering localised versions of the game. As of writing, there are two prominent localisation companies—McMillan (2019), who produce the ‘Whack A Waka’ card deck which claims to be ‘optimised for chanting’, and the ‘Ogoola Karuta’ company which produces two types of card decks—Localised collections of poems in English, German, French, Swedish, and Chinese, and decks using the MacCauley (1917) translation. Of notable interest is an expansion pack to this translation which preserves some of the ‘Makurakotoba’, or ‘pillow words’ in Romanised Japanese while the rest of the translation remains in English. This might be an easy, if less accessible way to make the poems be more similar to the originals. An example of Poem 17 in both these translations can be seen in Figure 5.
Figure 5. Poem 17 with preserved Makurakotoba in the Ogoola Karuta expansion (left). Version of Poem 17 in ‘Whack A Waka’ (right)
Kimari-ji: Phonics & Auditory Reaction
Critical to the game is the players’ reaction times when listening to the recitation and finding the matching card in front of them. Depending on the poem, players may be able to attempt to claim a card within fractions of a second of the reading beginning or they may need to wait for a longer amount of time until enough sounds from the poem’s beginning have been read aloud that the poem can be uniquely distinguished from the other ninety-nine poems.
An example is given in Figure 6, where a player has four cards to choose from. The player can begin eliminating non-matching cards as syllables are read aloud:
1. Three of the cards share the same initial sound, ‘ち’ or ‘Chi’.
2. Another syllable is read—‘ぎ’ or ‘gi’, eliminating Poem 17 from contention and narrowing the decision down to 2 cards, each with a 50% chance of being correct.
3. Then, ‘り’ or ‘ri’ is read, but both cards have this sound, so no elimination can be performed.
4. ‘き’ or ‘ki’ is read, finally uniquely identifying the poem.
This logic can become more complex depending on what cards may have already been claimed or may not have been drawn for the game (e.g. if poem #75 had already been read, the poem would have been identifiable at step 2).
Figure 6. An example of quick determination of unique poems.
Professional and skilled amateurs will memorise these early sounds of the poems in order to gain an advantage in their games. These unique starting sounds of each card are known as ‘kimari-ji’; however, some elite players are able to distinguish poems before the kimari-ji are completely read aloud. Matsuda (2020) demonstrates that Poems #5, #60 & #95, which have kimari-ji ‘ooko’, ‘ooe’ and ‘ooke’, can be uniquely distinguished at their shared ‘oo’ part of their kimari-ji.
In phonology, these parts of the kimari-ji are known as ‘mora’—an auditory unit that is equal to a syllable, or part thereof. The morae in these poems correspond directly to their hiragana script. These morae can be further broken down into their constituent uniquely identifying phones. (Cutler & Otake, 1994). To illustrate this, we break a simple word (bread) down into its constituent phonologic parts in Figure 7.
Figure 7. The phonic elements of the word ‘パン’ (bread).
Matsuda’s focus for his paper is on these three poems, but a research gap exists for the remaining poems that also share morae in their kimari-ji. Acknowledging this gap in current research, and the limitations of existing phonology software libraries available to expose these extreme subtleties, the research done in this paper into international karuta decks may be less applicable to this highest tier of players, but our analysis on kimari-ji in these decks should still hold for most classes of player.
Not only are listening speeds in the sport lightning fast, but so too are the reaction speeds to target and claim matching cards. Yamada et al. (2018) developed a measurement system for determining karuta card captures and emphasised the need for millisecond accuracy confidently to determine victors of contested claims.
Thinking about how the kimari-ji affect the gameplay of Karuta, if many cards ‘compete’ over the same starting phones in their kimari-ji, there will be more homogeneity in the game’s play style. Conversely, if decks contain very disparate starting sounds, less time may be needed mentally to eliminate cards from selection and a greater emphasis be placed on immediate reaction for claims. The exclusion of half of the torifuda from the play area may have lesser effect in this scenario, as cards that have already been read may be less important for players to keep track of. The original Japanese poems have a diverse range of ‘compete counts’ and syllable lengths which gives the game a good balance in its gameplay.
International Interest
The game has seen significant popularity growth both within Japan and internationally, with much of the recent growth attributable to the popular manga[1] series ‘Chihayafuru’ by (Suetsugu, 2007-2022), which has been adapted into both an animated series and a set of live action films. The series follows a schoolgirl named Chihaya Ayase as she passionately competes in karuta competitions, climbing the ranks (karuta ranks range from E for beginners, to A for high class players) and vies to become the best player in the world. The compelling story, cultural point of interest, and complementary love-triangle, captured many readers’ attention, and the popularity of the series led to a period of high interest in the game known as the ‘Chihayafuru Boom’ (Huaigo, 2016). Internationally, the animated series was available with English subtitles on the popular streaming service CrunchyRoll, and the manga series distributed for purchase in English from publisher Kodansha. While karuta is overwhelmingly played by domestic individuals, the series does contain scenes of competition with international players and gives admiration of their ability to be competitive despite their language and cultural barriers.
This exposure of the sport to an international audience was more than just fleeting interest in the story, for several official international clubs and tournaments have been established since. A significant milestone was the inaugural international karuta tournament held in 2012 with participants from the U.S.A., New Zealand, Thailand, China, and South Korea (Hiroshi et al., 2017). These developments have been welcomed by Japan, with areas such as Otsu City hoping that the growth of the sport will increase tourism to the region (Otsu, the Karuta Mecca Promotion Office, 2020).
Analysing Derivative Decks
Evaluation Methodology
We discussed earlier about difficulties in translation and localisation of both general elements of foreign media and of specifically the one hundred poems, but the impact of misrepresented elements within these translations are subjective. In this section, we will examine various translations of the one hundred poems by some more objective criteria that can have tangible impact on the game of karuta, namely the immediacy of kimari-ji recognition through their phone length and the competitiveness of phones. In Figure 10, we demonstrate the calculation of these values for the earlier example in Figure 10.
Metrics used in the rest of this section are displayed visually in Figure 8. Kimari-ji length is the number of phones within a poems kimari-ji. In this diagram we use hiragana characters for simplicity, but in practise our analysis uses characters from the International Phonetic Alphabet (IPA) (International Phonetic Association, 2021). Competing card values are calculated per-phone and represent the number of cards which might still match the in-progress poem recital.
Figure 8. Kimari-ji length and Competing cards for the ‘Chi’ poems.
In Table 1, the adaptations used in this analysis and their corresponding shorthand, we list the decks of karuta cards which we analysed along with a short string for identification in the result data. As well as a handful of English translations, we include some other languages such as French and German as well as the localisation decks which contain a different set of poems from the local country.
Table 1. The adaptations used in this analysis and their corresponding shorthand.
In order directly to compare languages that have significantly different origins and linguistic features, we use the IPA for a fair comparison. A Python script was written to convert the poems to the IPA using Doherty’s (2023) software ‘IPA-Dict’ for the original Japanese reference and the Epitran Python library by Mortensen et al. (2018) for the derivative decks (As of the time of research, the Epitran library did not support Japanese to IPA translation).
These IPA characters directly correspond to phones—the basic unit of phonic speech analysis. The IPA allows us to inspect and analyse the sounds of these poems which come from a diverse set of languages in a uniform manner.
Distribution of kimari-ji length
International players using translated decks may appreciate that those decks be similar in playstyle to the originals, particularly if they have ambitions to participate in the international competitions which use the original decks. To do this, we consider the length of the kimari-ji. The original Japanese version of the game has roughly 70% of cards recognisable in the first two to four phones, decreasing as more phones are read, aside from a handful of outliers at 12 phones, meaning that cards can generally be claimed very early in the poem recital.
We compare the distribution of how many phones the kimari-ji have in each deck using both the Pearson Correlation Coefficient and Spearman Rank Correlation as measures of correlation. The results are shown in Figure 9, where moderately strong to strong relationships are highlighted in bold (>0.70), as well as visual representations in Figures 3-5 which are separated into 3 groups—(English Translations, Other Languages’ Translations and Localisations).
Interestingly, the correlations to English translations are notably weaker in comparison to the other languages’ translations. When looking at Figure 11: The number of phones required before a poem can be uniquely identified (Other Language Translations), in comparison to the other figures, there are significantly more instances where phones are identifiable by the third phone. In addition, the English translations have a greater number of late-identifiable phones causing further divergence from the original Japanese.
In the original collection, the largest percentage of cards’ kimari-ji are identifiable at 3 phones, however in most derivative decks more cards are identifiable both earlier and later than this. As a result, when moving to the standard deck for international competition, players may find the game to be more consistently fast-paced than in their home games.
One problem with the analysed localised decks (aside from the English version) is that the number of cards in the deck is lower than one hundred. While beginner players often start their practise using a subset of the deck to ease the burden of memorisation, this makes for a game with less variability and challenge, and subsequently, strategic depth. It also makes these localisations vulnerable to changes in their phonetic distributions should additional poems be added to the collection in the future.
Mitigation
Careful word selection in translation could model the original anthology of poems more closely by consciously considering overlapping sounds in the beginning of poems. A translator could be encouraged or discouraged to use poetic interjections such as ‘Lo!’ or ‘Oh,’ (which the University of Virginia translation uses occasionally)—or deliberately to select words that begin with the same phones e.g., ‘Cleopatra’ (kliəˈpætrə) and ‘Clang’ (klæŋ). It may be easier to do this with native poems than with translation of the originals due to the larger dataset that could be considered or even commissioned. Such operations to translations of the original may be detrimental to other qualities of the poem.
While alignment to the original gameplay could be improved, the divergence of the translations to the original are not so significant that they would affect the game adversely for amateur players, which is likely the main audience for existing products. However, as the interest in karuta and the number of competitive international players increase, the differences may be important to mitigate—especially if standard decks in languages other than Japanese were to be established.
Competing Cards
The distribution of kimari-ji does not provide a complete understanding of how alike the derivative decks are to the original. While it tells us how long the kimari-ji are, it does not inform us of the dynamics before the kimari-ji has concluded. Specifically, it does not tell us much about competing cards, or cards that share the same phones and are yet to be disambiguated. Looking back at the example in Figure 6, we can see that there are three competitors for the starting ‘ち’ or ‘Chi’ sound, which is then reduced to two before unique identification. This is an important facet of the game as players need to keep in mind all cards that could be matches if they are playing carefully and trying not to make fouls by striking incorrect cards. A smaller set of competing cards can be easier to keep track of, whereas a larger pool of competitors would be introducing greater complexity to the game and increasing the likelihood of errors and fouls.
Figure 13. Violin plot of cards that share phones.
In Figure 9, we display a comparison of the number of competing phones in each deck. Naturally, the number of competing cards will decrease as more phones are read (as the Y axis increases). The mean and bounds of each data series are shown, with a dotted red line extending from the Japanese deck’s mean. The majority of cards in each deck have less than 5 competitors each, but generally there are a few starting phones which have a high level of competitiveness (In the original deck, the main outlier is the cards which begin with あ/ ‘a’ which has many competing cards at first but becomes discriminable after a few more phones).
When cards compete in the original deck, the mean number of competitors is 3.4. All but the German translation have a higher average, though they are all within +/- 1 card. Fortunately, this indicates that the complexity of the game using these decks should not be adversely affected.
Generally, the outliers with high number of competing cards are, like the original version, the first phone sounded. While the larger and smaller bounds of data are difficult to attribute to a pattern, the selected non-English translations seem to be markedly lower than the original, with the Russian translation having more values at its extremes than the other decks, which generally have 1 or 2 highly competitive phones.
Mitigation
We hypothesise that the tendency to have a lower amount of extremely competitive cards may be linked to the selected languages having a higher number of phonemes than Japanese (Eupedia, n.d.), and hence fewer possible phonetic combinations and distinct sounds. To mitigate this, translations can be careful about the number of unique sounds they include when adapting the opening sound of the poems.
Conclusion
There is an Italian saying—‘Traduttore, Traditore’, a rough translation being ‘Translator, Traitor’[2]—that by translating something, the message has been altered in a way that does not entirely preserve the original intent. The aspects of possible ‘data loss’ are numerous. A sample of such (in no way exhaustive) include wordplay, imagery, nuance, cultural and historical context, religious or political concepts, slang, proverbs, idioms, societal roles, and symbolism. In this research, we ask about translation’s preservation of auditory components for the Hyakunin Isshu, one of the most famous historical anthologies of poems to come out of Japan.
The auditory element is crucial for the cultural game of Karuta, for it is this by which games are lost and won. Listening to a poem reciter, players compete over matching the read poem to one on the floor in front of them. Skilled players optimise their play by learning just the first few sounds of the poems required uniquely to identify them. These unique sounds (called kimari-ji) shape the game and determine how complex the game can become.
While the game is well established in Japan, it is only in recent years that it has attracted an international audience, and with it, attempts further to spread its influence through more accessible means to local people via translated and localised poem decks. However, there is little-to-no published research on whether these adaptations are fit for purpose if they have a similar play style to the original or their own unique characteristics. This is an important consideration should the sport continue to grow abroad and eventually seek to establish standardised regional playing cards.
In this paper, we performed some statistical analysis on the phonics of several translations of the poems and commercially available adaptations of the game to investigate potential discrepancies in play. We find that, encouragingly, there are no concerning levels of deviation from the original anthology of poems, though there do exist enough differences that creating bespoke translations of the poems specifically for their use in karuta may be worthwhile.
Specifically, our results show that:
• The number of phones that uniquely identify adaptations of poems are generally in the correct statistical range as the original, but the exact distribution of these could be improved. Notably, several adaptations contained poems which were only recognisable after a (comparatively) substantial number of phones had been spoken. This might be avoided by abstaining from the use of long starting words that are shared with other poems, or the careful analysis of an adaption’s usage of poetic inflections such as ‘Lo!’, etc.
• Most Japanese poems can be identified after three phones. However, many of the adaptations have a small number of poems with the same timing. Being aware of this, translators/collators can be careful to select more poems with starting words that share three phones. e.g., ‘wanting’/’ wander’ (wɑntɪŋ/wɑndər).
• Non-English translations of the poems lacked the presence of highly competitive starting phones. To mitigate this, adaptations should be careful to use a smaller selection of starting phones, while ensuring that the phones that follow them continue to be diverse.
Negligence to consider phonetic consistency might affect the game in several negative ways, including lower diversity in gameplay, delayed reaction speed and general friction in adapting to the international standard of the original one hundred poems in Japanese.
We hope that future adaptations of the game will consider the research presented here and our recommendations for improving adaptations of the game, and that the international interest in karuta continues to flourish and reach more people around the world.
Limitations of this research
Below, we acknowledge potential shortcomings of our research and items that may be taken into consideration for future research or development.
Vulnerability to IPA Translation Mistakes
Whilst efforts have been made manually to verify portions of the data produced by the Epitran or IPA-Dict software, the metrics gathered are vulnerable to mistranslation and errors in the software.
Subtleties of the spoken word
In a game of karuta, the pitch, tone, accent, and overall delivery of the poem by the reader can make a tangible difference that is impossible to analyse with text alone. As this study only considered the textual representations of these poems, rather than audio recordings, these extra factors were not accounted for.
As mentioned previously, Matsuda’s (2020) research details that some elite players are able to distinguish poems by the subtlety in their readings even before the unique phones are read. Using IPA, we were empowered to perform this research, but are limited also in what minute phonetic differences it encodes.
Research Coverage
The Hyakunin Isshu has been translated many times by many different authors with different interpretations and priorities. The set we present here is by no means exhaustive and there may exist adaptations with better gameplay approximations, however, our objective was to broadly sample the space and suggest recommendations for future adaptations, than to find the most appropriate existing deck.
In addition, the native poem selections in our study are all published by a singular company (Ogoola Karuta). Therefore, their representation in the study relies heavily on their methodologies and selection criteria which may differ if produced by a different source.
We acknowledge that this paper only has examined a handful of Indo-European languages. It would be interesting for future research to examine these characteristics in other language families.
Glossary
• Adaption—conversion of media for a local audience through techniques like translation and localisation
• Aite-Jin—The opponent’s territory
• Chihayafuru—A popular manga series that centres around the game of Karuta
• Field of Play—The overall area where karuta cards can be actively played and placed.
• Hakama—Traditional clothing often worn over a kimono.
• Hyakunin Isshu—An anthology of poems known as the One Hundred Poets.
• IPA—International Phonetic Alphabet
• Ji-Jin—The player’s territory
• Kimari-ji—Uniquely identifying sounds in each of the poems.
• Kimono—A traditional Japanese garment with long sleeves
• Kireji—A ‘cutting word’ in Haiku which signals a pause in the flow of a sentence
• Kyōgi Karuta—Competitive Karuta
• Localisation—Adapting content to fit a target audience, often using methods stronger than translation (e.g. converting measurements to local versions, replacing idioms and proverbs, etc.)
• Manga—Japanese comics
• Mora—A phonology unit similar to a syllable
• Makurakotoba—A type of Japanese “pillow word” employed in poetry to create subtle imagery and emotional depth.
• Phone—A tangible distinct phonetic unit in a language. (e.g. the ‘p’ in ‘pat’ and ‘spat’ are different phones (due to the difference in aspiration))
• Phoneme—An abstract unit of sound that can be pronounced slightly different in different contexts. (e.g. the ‘p’ in ‘pat’ and ‘spat’ are the phoneme /p/)
• Tanka—Short Japanese poems
• Tatami—A traditional Japanese flooring mat
• Teju-Jin—See: Aite-Jin
• Territory—The specific area in front of a player where they are in control of card arrangement.
• Torifuda—Cards containing the last two lines of each poem.
• Obi—A wide belt used to secure a kimono.
• Waka—A type of classical Japanese poetry.
• Yomifuda—Cards containing all lines of each poem
Notes
1. Readers might notice that the manga title is “Chihayafuru,” while the first line of poem 17 is
“Chihayaburu.” The author, Ariwara no Narihira lived 825-880 AD (Funatsu, n.d.). A series of script
reforms happened in later years, standardising characters and pronunciations. Dakuten, or the “゛” symbol
which indicates the change of “fu” sound to “bu,” was only enforced in the Meiji Era (1868-1912). The
manga has taken the first line directly from the poems, whereas other sources have adjusted it to modern
script. Readers of the poems may notice other instances that may seem strange to those only familiar with
contemporary Japanese.
2. Ironically, the translation of the Italian into English loses the extremely close sounds of the Italian, which is part of its message—how translation can betray the original’s qualities
References
Abdallah, J., 2018. Subtitled vs. Dubbed Anime: Viewer Perception of Japanese Culture, Dayton, Ohio: University of Dayton.
All Japan Karuta Association, 2020. Ogura Hyakunin Isshu festival 2020 in Tokyo (2020) The History of Karuta’ s card.. [Online] Available at: https://www.karuta.or.jp/2020/karuta2020/en/retro/index.html [Accessed 2 February 2024].
Allen, M. & Sakamoto, R., 2014. Globalizing Japanese Popular Culture: The coolness of Japan?. s.l.:Routledge/Taylor & Francis Group.
Anon., n.d. Karuta France. [Online] Available at: http:/francekaruta.wixsite.com/my-site [Accessed 15 February 2024].
Association of Japanese Classical Haiku, n.d. Brief Notes on "Kire-ji". [Online] Available at: https://web.archive.org/web/20070731130736/https://www.haiku.jp/haiku/nyumon_English_03.htm [Accessed 20 08 2024].
Basho, M., 1686. 古池や|蛙飛びこむ|水の音. s.l.:s.n.
Benjamin, W., 1923. "The Task of the Translator," In: H. Zohn, ed. Illuminations. s.l.:s.n.
Blyth, R. H., 1942. Zen In English Literature and the Oriental Classics. Tokyo: Hokuseido.
Bull, D., 1996. Karuta: Sports or Culture?. [Online] Available at: https://www.asahi-net.or.jp/~xs3d-bull/essays/karuta/karuta.htmlhttps://www.asahi-net.or.jp/~xs3d-bull/essays/karuta/karuta.html [Accessed 22 10 2024].
Columbia University, n.d. Japanese Syllabaries | Asia for Educators. [Online] Available at: http://afe.easia.columbia.edu/special/japan_600ce_syllabaries.htm [Accessed 03 February 2024].
Cutler, A. & Otake, T., 1994. "Mora or Phoneme? Further Evidence for Language-Specific Listening," Journal of Memory and Language, 33(6), pp. 824-844.
Doherty, L., 2023. Open-dict-data/IPA-dict: Monolingual wordlists with pronunciation information in IPA. [Online] Available at: https://github.com/open-dict-data/ipa-dict [Accessed 24 February 2024].
Eupedia, n.d. What languages in Europe have the most or least phonemes ?. [Online] Available at: https://www.eupedia.com/linguistics/number_of_phonemes_in_european_languages.shtml
Funatsu, J., 2014. An Annotated Translation of the Poems of Ariwara no Narihira from the Shinkekinshu. s.l.:University of Hawaii at Manoa.
Hasegawa-Feurst, A., 2016a. The Ogoola Karuta Poetry Game. Sweden: Olayaselection.
———2016b. Ogoola Karuta Deutsch. Sweden: Olayaselection.
———2016c. Ogoola Karuta Svenska. Sweden: Olayaselection.
Hiroshi, Y., Kazuya, M., Tsutomu, T. & Mashiko, T., 2017. "A Method for Determining the Moment of Touching a Card using a Wrist-worn Sensor in Competitive Karuta," Journal of Information Processing, Volume 26, pp. 38-47.
Honda, H., 1957. One Hundred Poems from One Hundred Poets.. s.l.:The Hokuseido Press.
Horton, H. M., 2018. "Making it Old: Premodern Japanese Poetry in English Translation," Asia Pacific Translation and Intercultural Studies, 4 5, 5(2), pp. 110-204.
Huaigo, 2016. A Study of the Relationship between Chihayafuru and the Increase in the Competitive Population. [Online] Available at: http://kakyo-karuta.com/data/chihaya1.html [Accessed 15 February 2024].
International Phonetic Association, 2021. Handbook of the International Phonetic Association: A guide to the use of the international phonetic alphabet. s.l.:Cambridge University Pres.
Jularbal, K. R., 2023. An Interdisciplinary Analysis of Matsuo Basho's 'The Old Pond' Haiku: Exploring Stylistic Nuances and Linguistic Sensibilities in English Translation. International Journal of Research Publication and Reviews, 4(2), pp. 1052-1055.
Lost in Translation. 2003. [Film] Directed by Sofia Coppola . United States: Focus Features.
MacCauley, C., 1917. Hyakunin-Isshu (Single Songs of a Hundred Poets) and Nori no Hatsue-Ne (The Dominant Note of the Law). Yokohama: Kelly and Walsh.
MacMillan, P., 2019. ‘WHACK A WAKA’. The world’s first English card game based on the ‘One Hundred Poets, One Poem Each’. s.l.:Kawada.
Matsuda, T., 2020. "Statistical Analysis of kimariji in competitive karuta," Japanese Journal of Applied Statistics, 49(1), pp. 1-11.
Mortensen, D. R., Dalia, S. & Littell, P., 2018. Epitran: Precision G2P for Many Languages. International Conference on Language Resources and Evaluation.
Mostow, J. S., 2023. Pictures of the Heart: The Hyakunin Isshu in Word and Image. s.l.:University of Hawaii Press.
Müller , T., 2021. Ogura Hyakunin Isshu Hundert Menschen, Jeder ein Gedicht Müller Till.. Norderstedt: BoD - Books on Demand..
Naude, J. A., 2004. "An overview of recent developments in translation studies with special reference to the implications for Bible translation," Acta Theologica, 13 10.22(1).
Otsu, the Karuta Mecca Promotion Office, 2020. Bringing Ogura Hyakunin-isshu Kyogi Karuta from Otsu, the Holy Land of Karuta, to the World Kyogi Karuta Trial Lessons for International Students. [Online] Available at: https://pressreleasejapan.net/2020/02/19/bringing-ogura-hyakunin-isshu-kyogi-karuta-from-otsu-the-holy-land-of-karuta-to-the-world-kyogi-karuta-trial-lessons-for-international-students-otsu-city-shiga-japan/ [Accessed 23 08 2024].
Sato, H., 1995. One Hundred Frogs. s.l.:Weatherhill.
Sieffert, R., 2012. De cent poètes un poème. s.l.:Éditions Verdier.
Sokolov, V., 2001. “Сто Стихотворений Ста Поэтов (Сборник Классических Японских танка, Составленный В 1235 Году фудзивара-но тэйка). [Online] Available at: http://www.lib.ru/JAPAN/japan100.txt [Accessed 10 February 2024].
Suetsugu, Y., 2007-2022. Chihayafuru. Japan: Be Love.
University of Virginia, n.d. The English translation of Ogura Hyakunin Isshu. [Online] Available at: http://jti.lib.virginia.edu/japanese/hyakunin/macauley.html [Accessed 3 February 2024].
Yamada, H., Murao, K., Terada, T. & Tsukamoto, M., 2018. "A Method for Determining the Moment of Touching a Card Using Wrist-worn Sensor in Competitive Karuta," Journal of Information Processing, 26(0), pp. 38-47.
Article copyright Ian Hunter