How Long Is a Tanka?

First published in the Tanka Society of America Newsletter 4:4, December 2003, pages 2–3. Excerpted from my “President’s Message” of that issue.

In the previous issue of this newsletter, Bill West wrote a brief letter to the editor noting his surprise that the first-place winner in the annual Tanka Society of America tanka contest was given to a poem with more than 31 syllables (by Sanford Goldstein):


from my hospital window

I see across a bare field

in the morning rain

a yellow silk umbrella

on its solitary way


This poem weighs in at 33 syllables, though personally I wouldn’t have guessed it had that many, as it feels natural and unpadded. But Bill raises an interesting question. How many syllables is too many for a tanka? One could also ask, how many syllables is too few? It may not be merely length that answers whether the poem is a tanka, of course, because a preponderance of other characteristics come into play, meaning that length (let alone a precise length) is not the only characteristic that defines the tanka genre in English. But I would like to offer this response: Though a traditional Japanese tanka has 31 sound units in a pattern of 5-7-5-7-7, and that tanka in English are typically shorter than 31 syllables in no set pattern, there is no rule for tanka in English that it should always be fewer than 31 syllables. The Tanka Splendor Award chooses to ask for poems of 31 or fewer syllables, but that rule applies only to that contest (which is perfectly fine) and I would suggest that it should not be taken as a prescription for all tanka. Though both tanka and haiku in English typically have a number of syllables that is usually fewer than 31 and 17, respectively, there are plenty of haiku that are longer than 17 syllables. Why should tanka be any different, relative to 31 syllables? In other words, there’s nothing magic about 31 as a maximum number of syllables for tanka in English, and what greater authority on the matter could there be than tanka pioneer Sanford Goldstein who won our 2003 contest with a 33-syllable poem?

        For what it’s worth, I’ve checked the syllable count of our previous first-place contest winners, and here they are: Edward J. Rielly (2000): 29; David Rice (2001): 29; Carol Purington (2002): 29. I think it’s mere coincidence that each one is 29 syllables. In English, where we can generally say the same things as Japanese but in fewer syllables, I would suggest that we say in our tanka what needs to be said in a clear, rhythmic, and lyrical manner, letting the number and pattern of the syllable count fall where it may.

        Out of curiosity, I also reviewed the syllable counts of each of the poems in the previous newsletter’s “Tanka Café” column, and found that the number of syllables ranged from a low of 15 (Doris Kasson) to a high of 34 (Thelma Mariano). Three poems had fewer than 20 syllables (Pamela A. Babusci, Doris Kasson, Michael Dylan Welch), and two poems had more than 31 syllables (Thelma Mariano, Sanford Goldstein). Single lines ranged from 2 to 11 syllables in length (tanka certainly seems malleable!). Here’s the complete distribution of syllable counts from the last “Tanka Café” column (total syllable count followed by the number of poems with that count): 15:1; 16:1; 19:1; 21:1; 22:1; 23:2; 24:2; 25:1; 26:7; 28:7; 29:4; 33:1; 34:1. The total of 773 syllables divided by 30 poems yields an average syllable count of 25.77. While the 30 poems in the previous issue’s “Tanka Café” column may not be a representative sampling of tanka in English, I suspect they come fairly close. A tanka with more than 31 syllables in English is perhaps rare, but the same reasons that allow us to write with a lesser number also allow us to write, on occasion, with a greater number. It’s always the poetry that matters most. After all, as the great architect Louis Sullivan once said, “form follows function.” The same is true, I believe, for tanka in English.