Two papers

First, Japanese has syllables: A reply to Labrune (2012), by Shigeto Kawahara:

In a provocative article, Labrune (2012b) argues that there is little phonetic or psycholinguistic evidence for syllables in Tokyo Japanese (henceforth Japanese), and that phonological phenomena which have been hitherto analyzed in terms of syllables can be reanalyzed by deploying a distinction between a "regular/full mora" and a "deficient/special mora". She concludes that Tokyo Japanese does not have syllables, and as a further theoretical consequence of this view, she argues that not all prosodic levels are universal, extending on the suggestions by Hyman (1985, 2008). Although this proposal is very thought-provoking and its theoretical consequence is an important one, it does miss some of the previous experimental findings about the existence of syllables in the prosodic organization of Japanese. Therefore, this reply article summarizes evidence that Japanese does show evidence for syllables both phonetically and psycholinguistically.

I like Kawahara's directness here, from the title of the paper to its conclusion ("In conclusion, Japanese has syllables"); for what it's worth, I also found Labrune's arguments unconvincing at the time (I haven't read the 2012 paper Kawahara specifically refers to, but she discussed the issue in The Phonology of Japanese (2012), which I did read, and it doesn't sound like she used substantially different arguments there.)

Kawahara also recently published a couple of papers on "maids," i.e. employees in maid cafes, which may be of somewhat less universal interest although I found them good reading: "The phonetics of Japanese maid voice I: A preliminary study" and "The sound symbolic nature of Japanese maid names", both available here (along with the rest of his papers, it seems).

Two experiments show that obstruents are associated with tsun-type maids, whereas sonorants are associated with moe-type maids. [...] Morever, the identified sound symbolic relationships exemplifiy cases of emergent sound symbolic relationships, not based on conventionalized rules, as the notion of "tsun" and "moe" are new notions.

Second, Helen J. S. Lee's Writing Colonial Relations of Everyday Life in Senryu (2008).

Life in the colonies for Japanese settlers was — as life everywhere is — largely shaped by class status. Peter Duus provides an insightful distinction of class membership by identifying two types of settlers: immigrants and colonists. The working-class Japanese settlers depicted in the poems belong to Duus's definition of "immigrants," which he opposes to "colonists" whose presence in the colonial territory was subsidized by the state and who had membership in the dominant stratum of the host society. The elevated and privileged status of "colonists" engendered markedly different colonial experiences from those of their countrymen who went to colonial territories as "immigrants"; the latter were "oppressed, assimilated, and rejected by the host society," although they still ended up contributing to the exploitation of the colonized in one form or another.

[...] Senryu poetry illustrates the diverse colonial experiences of Japanese immigrants who took up a wide range of occupations, such as day laborers, peddlers, shop clerks, and factory workers. Rather than assert a monolithic portrayal of these working-class Japanese settlers, this article explores the colonial dynamics rendered in each poem as a way to offer insight into the multifaceted lived daily experiences of Japanese immigrants in colonial Korea.

The details and close reading get denser towards the end of the paper. It's worth the buildup.

The book Lee uses as her main source, Chōsen Senryū 朝鮮川柳, is actually available online (albeit in hated .djvu format). This allowed me to confirm that kahetamanma on page 617 of Lee's article is indeed the weird transcription error it appears to be. The original is 咬へたまんま, which although a bit obscure orthographically (the "standard" Japanese reading for 咬 is kamu) is almost certainly to be pronounced kuwaeta manma — cf 鉛筆なんか咬へて pronounced enpitsu nanka kuwaete here.

Popularity factor: 10


> Morever, the identified sound symbolic relationships exemplify cases of emergent sound symbolic relationships, not based on conventionalized rules, as the notion of "tsun" and "moe" are new notions.

I kind of don't get what he's getting at with this statement. "Tsun" and "moe" may be new notions, but the subjective feelings that they're based on – harshness and cuteness – are as old as humanity itself; and sound-symbolism with, respectively, obstruents and sonorants (in this case) is ~exactly~ the synæsthetic association you'd expect, given all that's know about sound-symbolism both in Japanese and universally. The author is aware of this point and discuss it at length at the Conclusion; so why claim that the maid gitaigo are emergent and not based on rules? They're just as emergent as any other sound-symbolic datum.

> If this hypothesis is on the right track, then our results predict that front or acute
vowels like [i, e] are associated with tsun maids and back or grave vowels like [a, o, u] are associated with moe maids. This prediction is to be tested in a future study.

I wonder if Tokyo Japanese will see a less pronounced effect on /u/, since it's often devoiced.

Great find btw :)


My interpretation is that when they say "conventionalized rules", they don't mean abstract sound-symbolic rules, like "spikey <=> obstruent", but rather more concrete (and, perhaps, more arbitrary) rules, like "girl <=> sonorant". So the association between spikiness and obstruents did already exist, true, but "tsun" was not associated with anything, and the fact that it got hooked up with obstruents suggests that obstruents generally (in Japanese) are linked with spikiness as an abstract concept, rather than just a list of things that happen to be spikey which Japanese speakers learn as children.

Vilhelm S:

I guess as just a random person on the internet I don't get to have an opinion, but for the maid voice paper I wish he had also published the recordings themselves. I have no idea what a "maid voice" sounds like, so it would be interesting to hear the contrast between the natural and the maid style.


It would be fun to hear women switching rapidly between regular voices and in-character voices as they pronounce strings of vowels, but it might have been harder to get their consent (for just 500 yen!). Searching YouTube for "maid cafe" and the like reveals a lot of what I assume they are talking about.

Of course the real question is whether/how "maid voice" differs from the regular "public female voice" expected of women in certain professional situations (most visibly, customer service and phone duties).


In general it would be cool if researchers got into the habit of always publishing the data files, not just the analyses.

Re: voices – also what about the voice of maids at crossdressing maid cafés?

Tim Martin:

Interesting paper on syllables! Thanks for sharing.

Have you ever blogged about the question of whether "ga" is really a "subject marker"? I found a paper several years ago that argues against this, and it makes a lot of sense to me (though I'm not a linguist). I'd be interested to hear what you think: http://www.linguistics.ucsb.edu/faculty/sathomps/cv/pdfs/Ono-etal2000Ga.pdf


I'm not wading into the "ga" thing until I get at least a year or two of funding to investigate it...

I like that paper, though. I'm a big fan of the actually-record-and-analyze-what-people-say approach to linguistics, so the paper has that going for it. I've also felt for a long time now that particles like "ga" and "o" aren't "dropped" in spoken Japanese so much as "added" when they are present, so I agree with the authors there. I dunno about rejecting ALL connection to subjecthood, though -- they argue that it appears for pragmatic rather than case reasons, but since those pragmatic reasons apparently never result in indirect objects, places, etc. taking "ga", I wonder how meaningful that distinction is.

I also would have liked to see more exploration of non-matrix clauses, because "ga" has historically had interesting relationships with those, but maybe there weren't that many in their sample.


Browsing the Google Scholar list of papers citing Ono et al's, I found this one by Rosenthal (PDF) which seems to agree with your intuitions:

> In order to address the issues raised by Ono et al., whose analysis is based actual data from conversation […] this paper looks strictly at conversation. Evidence is adduced from three corpora of native adult-adult Japanese conversation.¹ This paper takes a “conservative” approach in that it sees "ga" as a case particle and that it therefore needs to be understood in relation to other case particles. However, it takes the radical view that "ga" does not signify subject per se, but is the case particle used with subject nouns because other case particles would specify ‘non-subject’.² This paper is also novel in that it takes an integrated view of "ga" as being conditioned by two sets of factors: the participant-predicate relationship and the pragmatic environment.

In other words, "The choice of 'ga' over other case particles in conversation is a reflection of the participant-predicate relationship [who's doing what to whom], while the actual appearance/non-appearance of "ga" is a reflection of the pragmatic environment".


Thanks, that's a nice find! And yeah, that paper is arguing something a bit closer to what I intuit. My intuition has to be suspect though because I do so much reading, and of premodern varieties too... #kobunproblems

Tim Martin:

The weirdest use of "ga" I ever heard was on an episode of SMAP X SMAP (though I still thought the meaning of the sentence was completely obvious).

The members of SMAP were participating in some sort of batsu game, and Goro had to do an embarrassing public performance of a song. He's backstage getting ready, acting very nervous the whole time, when suddenly the music he'll be performing to starts playing out near the stage. At which point he says to himself, うわ、この音楽が緊張する!, with extra emphasis on the word "ongaku".

My understanding of the meaning was that he was saying "Even this music makes me nervous!" It made complete sense at the time, but if you repeat the sentence to people without any context, it seems totally weird.

Anyway, I'll have to check out this paper by Rosenthal!

Comment season is closed.