Unexceptional: Part 5 -- Are Japanese Syllables Too Short?

Unexceptional: Part 5 –

Are Japanese Syllables Too Short?

One of the bases for the idea that the Japanese language is uniquely unique is the brevity of the Japanese syllable.  The implication is that the Japanese syllable is so short that in comparison to English the difference results in a qualitative distinction.  This is one of the reasons why some ELH Haijin will refer to Japanese as using ‘morae’ (plural of ‘mora’) as opposed to English which uses ‘syllables’.

For example, in a 1993 paper by Anne Cuttler and Jacques Mehler, Mora or Syllable? Speech Segmentation in Japanese, the authors write in the Introduction:

“When Japanese poetic forms such as the haiku are rendered in other languages, an approximation to the prescribed form is usually achieved by specifying the number of syllables per line . . . But morae do not necessarily correspond to syllables.  Consider the first line of the haiku: shinshin-to.  Although it has the prescribed 5 morae (shi, n, shi, n, to), it has only 3 syllables (shin, shin, to).  The mora is a subsyllabic unit . . .”

To my mind, and not a few linguists, this is simply confused thinking.  As I have noted in previous posts in the Unexceptional series, the concept ‘syllable’ is not defined by the specific acoustic features of any particular language; that is to say the fact that Japanese count acoustic features, such as a concluding ‘n’, that English does not count does not mean that the Japanese are not counting syllables.  What it does mean is that Japanese count sounds that English speakers do not count.  But that is true of many languages, as I have noted in previous posts.

To my way of thinking the idea of ‘mora’ is a distinction without a difference.  If morae are subsyllabic, then what about the English syllable ‘it’; a very brief syllable.  Does the brevity of ‘it’ make it subsyllabic?  If it does not make it subsyllabic, then why would a Japanese ‘n’ be subsyllabic?  And what about other languages that also have shorter syllables than English; languages such as Spanish and Italian?  Does Spanish and Italian use morae or syllables?  Seriously; is this distinction of any help at all?  The idea of ‘mora’ is one of those academic conjurations that is little more than jargon posturing as insight.  It should be noted, though, that there is little agreement among linguists about this topic; if you look up papers on syllabic timing and usage you will find a whole range of differing opinions.  There does not seem to be any consensus.

There is a lot of literature in linguistics on the topic of syllable duration in various languages.  Much of it is highly technical.  But I don’t want to get side-tracked into a technical discussion when I believe that, for the most part, these technical considerations do not clarify; rather, in my opinion, they create a conceptual fog.  However, I think a few non-technical remarks are worth considering.

First, Japanese is not the only language which flows by at a more rapid rate than English.  Spanish and Italian, for example, are also more rapid than English.  When I say ‘more rapid’ I mean that the average syllable duration of Italian and Spanish is briefer than the average syllable duration of English.  In fact there are a great many languages that are more rapid than English.  Japanese, once again, is not unique in this regard; Japanese is unexceptional in its pacing.

There are also languages which are slower than English; tonal languages tend to be slower than English because speakers need time to enunciate the tone.  For this reason Chinese flows by, generally speaking, at a slower rate than English.

In other words, English is roughly in the middle when it comes to how fast the sound units, or syllables, of various languages flow by.  It is neither the slowest nor the fastest.  And speaking of being in the middle range; while Japanese is more rapid than English, from the studies I have read there are languages that are more rapid than Japanese.  Again, from the studies I have read, Thai consistently rates as the most rapid.  In other words, Japanese seems to occupy a speed that is only slightly faster than Spanish and Italian, but is surpassed by other languages such as Thai.

This puts both English and Japanese in the mid-range for speed of syllabic flow.  In other words, English and Japanese share the middle ground; neither of these languages is at the extreme.  Japanese is middle-fast, while English is middle-slow; but neither English nor Japanese is exceptional in its pace.

Before going further, I want to make a few remarks about the studies I have read.  First, many of the studies use a very small sampling; more than a few use a single speaker.  Statistically this leads me to be skeptical of the value of these studies.  I suspect that larger samplings would yield different results.

Further, some of the information about speed that is used by ELH Haijin is completely anecdotal, including some of the most often cited supporting claims.  Anecdotal reporting has its place and should not be dismissed out of hand.  But anecdotal reporting needs to be supported by further investigation if one is going to use these reports to make broad claims about language speed.

In addition, I have found that few studies take into account dialect variation.  Some studies I have read take British English as their standard of the pace of English syllable flow.  British English may differ from other types of English usage in terms of syllable pace.  Anecdotally, I strongly suspect that this is true.  When I was working construction, many years ago, I worked with many men from Texas as well as those from the Louisiana Bayou.  The long, slow-paced drawl of the Bayou dialect, I suspect, is significantly slower than standard British English.  On the other hand, I once had a forewoman on an assembly line who was Scott by birth.  Her English flew by at a rapid clip that took me some time to key into.  Again, I suspect the pace of Scottish English differs from that of standard British English.

The effect of dialect is not a minor consideration.  Millions of Indians speak English either as their primary language, or as a significant second language.  The steady flow of Indian English seems to me to differ, and to be more rapid, than that of standard British English.  Given what appears to me to be the wide variation in the pace of syllabic flow among different English dialects, I am reluctant to accept the generalizations about the pace of English verses the pace of Japanese when they are based on very small samplings of a single type of English.

Dialect further complicates the situation when one takes into account that there are regional dialects in Japan.  The question is, do these regional dialects effect the pace of the flow of the Japanese syllables?  Again, studies that do not take this into account make me inclined to regard their findings as of limited value; not valueless, but limited in terms of what kind of conclusion we can draw from their studies.

And finally, I think it is worth pointing out that context will effect pace.  If a subject knows they are being studied for the purposes of determining the pace of their language, this will inevitably effect their performance.  It will not be an example of a ‘natural’ interaction.  In addition, people change the pace of their output depending on the situation.  Speaking to children people tend to speak more slowly.  Speaking to long-term friends, people might speak very rapidly in comparison to their normal interaction.  When speaking to an audience this will also effect pace.

This last remark about audience is significant because some of the anecdotal stories about how rapidly people speak are derived from poetry readings.  Poetry readings are a highly specialized situation; they are not a normative use of a language and it is unlikely, I feel, that significant generalizations can be drawn from anecdotal reports of how people speak when reading their poetry.

None of this is to dispute the general conclusion that Japanese syllables are, on average, briefer than English syllables.  All the studies I have read support this conclusion.  The reason I bring up the above caveats is that even though it is true that Japanese is more rapid than English, one should not exaggerate the differences in the pacing of the two languages.  English and Japanese both occupy, as mentioned above, a middle ground when it comes to pacing.  They are not really so far apart when one looks at the full spectrum of language pacing.  I think this point needs to be emphasized because when reading some of the claims by ELH haijin one gets the impression that Japanese syllables are super-fast, or extraordinarily brief.  But that simply isn’t the case; rather Japanese occupies the middle ground with some languages just as fast and some faster.  And English also occupies this middle ground with some languages, like French, at roughly the same rate, some, like German or Chinese, slower, and some, like Italian, Spanish, and Japanese, faster.

All of this is very interesting, no doubt, for linguists.  But I would like to suggest for the readers’ consideration that all of this discussion about syllable duration is, quite simply, colossally irrelevant.  Consider the transmission of poetic forms from one culture to another, from one linguistic context to another.  When Latin poets in Rome began to write in hexameters they were imitating Greek; no one worried about the relative duration of Latin and Greek syllables.  When the Sonnet moved from Italy to England and France and Spain, no one felt that syllable duration was a factor of concern; it is never mentioned.  And perhaps most telling, when Japanese poets wrote in Chinese forms, using Chinese characters, what are called Kanshi, the Japanese were unconcerned with the fact that Chinese syllables are longer than Japanese syllables (and Chinese syllables are longer than English syllables as well).  Only in the case of the transmission of Japanese forms such as Haiku and Tanka has this issue been raised.

I find this revealing; and it only makes sense if one views the Japanese language as uniquely unique, as so completely different from other languages, from any other language, that one is compelled to treat it differently.  In other words, the foundation for the idea that the Japanese don’t count syllables, that they count something else, is, once again, nihonjinron, the highly problematic, and highly suspect, idea that the Japanese people and culture are estranged and distinct from the rest of humanity.

But there is another reason why the relative pacing of syllable duration is irrelevant: my view is simply that this idea that English (or other languages) should match the duration of Japanese poetic forms is a misunderstanding of how we should comprehend duration.  I would suggest for the readers’ consideration, especially for readers that have bought into the idea of the significance of Japanese syllable duration, that what we should be looking at is the relative duration within each linguistic context, not the absolute duration of the Japanese as a measure for non-Japanese languages. 

If Haiku in Japanese are written in 17 Japanese syllables and Haiku in English are written in 17 English syllables, the relative duration within each linguistic context is the same.  It is not that English and Japanese Haiku have the same absolute duration; in this they differ.  Rather it is that within each linguistic context their relative duration within their respective contexts is the same.  And, I would suggest, it is this relative duration that matters.  Notice, though, that if one approaches the issue this way, then all the discussions about syllable duration, syllable versus mora (or onji, or jion, or sound unit, or whatever term is currently in fashion) simply fall away.  Things become much simpler, much less cloudy, much more direct.

This is why I regard all these studies on duration to be a kind of huge mistake: it is completely unnecessary and does nothing to clarify the relationship between Japanese and non-Japanese using Japanese poetic forms.  In fact, there is a great advantage to the relative duration view: it is that for each language the same count will apply.  It will apply equally to German, Russian, Spanish, Hindi, Zuni, and Bantu.  The Haiku in all these languages would simply be constructed using the count of 5-7-5 syllables however each language counts, whatever sounds they count as a syllable.  Very simple.

In other words, and in conclusion – it doesn’t matter if Japanese syllables are shorter, longer, or the same as English syllables.  It is an irrelevant consideration.  It is time to simply put this line of reasoning aside.