Answers to the Haiku Numbers Questions
William J. Higginson
(This is a companion piece to the essay "Haiku by the Numbers, Seriously" available online at: http://www.2hweb.net/haikai/haiku/haiku-by-the-numbers.html.)
| Question (Calculations to 7 significant figures, rounded to 1000ths.) |
Ratio of Japanese On | Ratio of English Stresses | Percentage Difference 100 x (E–J) ÷ J |
|---|---|---|---|
| 1. How do the ratios of a shorter line or phrase to a longer line or phrase, that is, 5/7ths (Japanese) and 2/3rds (English), compare? | 0.714 |
0.667 |
– 6.667 % |
| 2. How close are the ratios of the shorter line or phrase to the overall length to each other, that is, 5/17ths (Japanese) and 2/7ths (English)? | 0.294 |
0.286 |
– 2.857 % |
| 3. How close are the ratios of the longer line or phrase to the overall length to each other, 7/17ths (Japanese) and 3/7ths (English)? | 0.412 |
0.429 |
+ 4.082 % |
| 4. How close are the ratios of combinations of a shorter line or phrase plus a longer line or phrase to the overall length to each other, 12/17ths (Japanese) and 5/7ths (English)? (This is called the "assymetry ratio" below.) | 0.706 |
0.714 |
+ 1.190 % |
Question 1, above, examines the difference between the ratio of the shorter line or phrase to the longer line or phrase within a 5-7-5 "sound" (called on) Japanese haiku (5:7 or 5/7ths) to the same ratio within a 2-3-2 "beat" English-language haiku (2:3 or 2/3rds). The result is that the English ratio is 6.667 % less than the Japanese ratio. Put another way, the ratio of shorter to longer elements in English haiku in this 2-3-2 form is about 93 % of the similar ratio for Japanese haiku in their 5-7-5 form.
Question 2, above, looks at the difference between the ratios of the shorter line or phrase to the overall length, that is, 5/17ths for the Japanese, and 2/7ths for the English. The result is that the English ratio is 2.857 % less than the Japanese ratio. From the alternative perspective, the ratio of the shorter line or phrase to the overall length in English haiku in this 2-3-2 form is about 97 % of the similar ratio for Japanese haiku in 5-7-5 form.
Question 3, above, looks at the difference between the ratios of the longer line or phrase to the overall length, that is, 7/17ths for the Japanese, and 3/7th for the English. The result is that the English ratio is 4.082 % more than the corresponding Japanese ratio. In other words, the the ratio of longer element to oveall length in the English 2-3-2 form is about 104 % of the similar ratio for Japanese haiku in 5-7-5 form.
Question 4, above, examines the difference between the ratios of a combination of one shorter and one longer line to the overall length, in each language, these ratios being 12/17ths in Japanese and 5/7ths in English. The result is that the ratio for English is 1.190 % larger than the similar ratio in Japanese. In other words, the ratio of a combination of one shorter and one longer line to the overall length in an English 2-3-2 haiku is about 101 % of the corresponding ratio in a 5-7-5 haiku in Japanese.
What these data show is that the internal ratios within each of the two forms match up quite well. So the symmetry and assymetry of the two forms are nearly equivalent. It seems to me particularly noteworthy that the "assymetry ratios" of 12:17 and 5:7 are particularly close, basically 0.71 or 71 %. This means that the two forms, when broken by a "cut" (kire) after the first or second line or phrase, have almost identical assymetries in terms of the perceptions of rhythm among speakers of the two very different languages. Since the overwhelming majority of haiku exhibiting a cut have this break at the end of the first line or phrase or at the end of the second, this parallel assymetry suggests that Blyth's recommendation of a 2-3-2-beat rhythm for haiku in English is particularly apt.
I'm trying to show here that each form, the Japanese 5-7-5 on count and the 2-3-2 count of stressed syllables in English, offers an opportunity for similar aesthetic relationships between the internal rhythms of haiku in their respective languages. The fact that the most basic unit of Japanese prosody, other than the single on ("sound") has been twelve of these, and that one of the most popular line-lengths in English prosody has been the five-beat or pentameter line gives us another point of similarity.
The fact is that no purely syllabic count in English allows any similar basis for developing a "traditional haiku form" in our language, because of the much greater variability in the lengths of English syllables when compared with the more or less uniformly short, clipped Japanese on.
Note that all this proves nothing about the relative lengths of the two forms, overall. (Methods for examining this question and my and another translator's conclusions about it are presented in my book, The Haiku Handbook, on pages 101-102.)
Conclusions
The sum of all these kinds of experiments usually results in finding that, on average, Japanese and English are about similar as to the amount of semantic content in haiku of similar time duration, and that typically 10-12 syllables of English “reads” in the same length of time it takes to recite a traditional 5-7-5 haiku in Japanese. Renowned haiku translator and interpreter R. H. Blyth figured this out more than forty years ago. (Review his comments on this in the main article, here: http://www.2hweb.net/haikai/haiku/haiku-by-the-numbers.html#bottomlinereality.)
I leave it to readers to supply their own answers for the two "why" questions at the end of the paragraph in "Haiku by the Numbers" that brought them to this page.