Martin’s classic article Speech Levels in Japan and Korea (first presented in 1958) doesn’t seem to be on the Internet. Being a fan, I dug around for Dell Hymes’ Language in Culture and Society, an anthology that includes it, and …Dude! What a book! Big! Heavy! Dusty, musky, hefty in the kind of cloth cover that makes it feel like a movie prop! And all the celebrities are in there: Hymes himself, and Boas, Lévi-Strauss, Pike, Malinowski, Firth, Mauss, Evans-Pritchard, Haas, Sapir, Whorf… Academia has this way of getting you to know the names (and general ideas) of the previous paradigm, without actually ever reading them. I’m sorry, generative linguists, but you’ll never have the sheer amount of cool of anthro people; pondering over hermetic syntactical operations while redrawing trees like a comp-sci major simply doesn’t compare in sexiness to living with exotic peoples to chart their fascinating cultures, then switching societies like jackets and coming home to tell the story. Finding this book was like going to a bar and meeting Indiana Jones (or Oriental Jones…)
Intellectual fetishes aside, yeah, Martin on Keigo. First of all, due to the Ancient Traditions of Linguistics, anyone who ever mentions Martin’s “structuralist” analysis is required by custom to draw a little cross to explain his two axes (as in “of evil”, not as in “firefighter’s”). Here’s my version, annotated with a few common terms for the various levels, and an example each:
The idea is that all Japanese utterances must specify two orthogonal (independent) variables:
In the “axis of address” (horizontal in the chart), you establish your distance towards the person you’re speaking with (plus your audience). One has basically two choices in this axis (similar in many ways to the T-V distinction of many European languages):
- “Plain”, “direct”, or “normal” forms (Japanese 普通体 Futsûtai, 常体 Jôtai). Familiar (or rude) speech, using dictionary forms like da or kuru.
- “Polite” or “distal” forms (丁寧語 Teineigo, 敬体 Keitai). Showing respect (or distance) by using the special morphemes desu and -masu. In writing, exceptionally, this level creates a conversational tone and can be felt as more intimate/casual than the plain.
In the “axis of reference” (vertical in chart), you establish your position towards what you’re talking about (—notice that it’s perfectly possible to talk about the person you’re talking with). In the article under discussion, Martin says the point of reference is the subject of the sentence, which is the common case; but, later (in the Reference Grammar), he notices that respectful speech can also be triggered by the object, or by more subtle connections to the referent. In this axis, you can choose to be neutral, or not; and, if non-neutral, you have to make use of two kinds of forms:
- “Humble” speech (謙譲語 Kenjôgo) to refer yourself and things related to yourself, including your in-group;
- And “honorific” or “exalted” forms (尊敬語 Sonkeigo) to anything related to the referent.
So one always lowers one’s level and raises theirs, which means that these morphological forms also double as something like persons of speech (i.e. humble forms always refer to me or my group, honorific forms to they and their group).
Therefore, this model predicts 2×3=6 possibilities (see the wiki for more examples of each form). Martin doesn’t discuss what some call “Minus Keigo” (マイナス敬語), rude or offensive forms (such as Temee! Monku aru kai?) that would extend the axes the other way (and are absolutely necessary when reading manga!). There used to be a higher level of addressee-axis politeness, marked by gozaimasu instead of arimasu (and de gozaimasu for desu), which Martin calls “deferential”; but, already when he wrote this, it was fading out from usage, and the gozaimasu forms came instead to be used as humble speech. (Addressee-axis gozaimasu still appears in some formal contexts, such as the tea ceremony.)
I’m interested in “really existing Keigo”; Keigo as she is spoke, as opposed to how the Japanese think it should be spoken. This curiosity, of course, steems from my basic linguistics training, where we learn to be scientific and therefore descriptive, not prescriptive. But, in this case, it’s not as clear-cut as usual; in many important regards, Keigo is prescription—it’s a deliberate, artificial norm, cherry-picked from prestige usage, assembled by commitee, and kept alive by manuals and style guides. It wouldn’t even count as “language” for those who have a strict nativist definition: it certainly isn’t present in the language acquired by five-year–olds, and requires conscious training to learn. It’s like a small artificial mini-grammar—a mini-conlang, if I might be so bold—designed to be grafted into the “natural” language. I’ve been calling these artificial components deliberately inserted into natural languages their “cyborg parts” (wink wink), and I believe they deserve more attention.
But this doesn’t mean there’s no value in taking a descriptive approach to Keigo; only that we cannot afford to ignore the prescriptive rules. Keigo is a sociolinguistic phenomenon; we have to describe real-world usage and formal expectations, to contrast them.
Unfortunately I haven’t yet had access to any large-scale corpus-based studies, but there’s a number of easily-available historical and sociological studies like Wetzel’s. Questionnaires are comparatively easier to find, and as long as we keep in mind that they reflect the natives’ opinions and not fact, they’re quite informative. I wanted to write this post not to show Martin’s axes (which are easy to find elsewhere) but his summary of the 1957 report Keigo to Keigo Ishiki (“Keigo and Keigo consciousness”), made by the National Language Research Institute (Kokuritsu Kokugo Kenkyûjo) and including “attitudes in non-standard dialect areas”. Below are some points highlighted by Martin, quoted in full (I keep his pitch-accent marks; and “honorific” below refers to Keigo in general, not strictly exalted/Sonkeigo):
- Honorific forms incorporating negatives (analogous to our “wouldn’t you like to”) are generally felt to be more polite than those without negatives.
- An individual does not show marked idiosyncratic preference for a particular one of several competing honorific forms in his usage.
- The longer the utterance, the more polite it is felt to be (sóo nan desu “(it is a fact that) it is so” somehow just “sounds” more deferential than sóo desu “it is so” (cf. Gengo Seikatsu, n. 82, July 1958)). [I’ve heard somewhere that this is proposed to be a sociolinguistic universal, but couldn’t find it again.]
- Utterances with local dialect in them are considered less polite.
- Utterances with at least a few Chinese loanwords in them are considered more polite than those without.
- Politeness of usage seems to be in inverse proportion to feeling that one has the upper hand in a situation.
- Actual usage is often at variance with ideal usage. [How exactly??]
- Strangers are accorded more polite usage than acquaintances.
- Distinctions of usage toward the addressee are more finely drawn as one moves geographically from East to West.
- One’s sex [gender] is the most important social factor determining one’s honorific usage; one’s age is the least important.
- There is a tendency for men to discriminate different situations calling for honorific usage, and for women to use honorifics all the time.
- Knowledge of honorific forms is primarily controlled by education background. One’s sex has little to do with it.
- Attitudes toward honorifics differ greatly with age level.
- In general, people like the more polite honorific forms.
- Honorific speech incorporating dialect forms is displeasing, both from oneself or from others.
- Those who favor the more polite honorific forms are more polite in their own usage.
- Situations of address where more politeness is expected are: women to men, the young to the old, lower classes to upper classes.
- Of these three factors, class difference is generally felt to be the overriding one.
- One’s sex is no factor in one’s ability to discriminate honorific speech behavior on the part of others.
- There is considerable psychological resistance to the rule suggested by the Ministry of Education’s guide Kore kara no Keigo that one should not use honorifics in speaking to outsiders of one’s own work superiors (bosses).
- There is considerable opposition in the abstract to the overuse of the deferential prefix o-, but in actual conversation situations the resistance weakens.
- In speaking of one’s own relatives, there is a strong consensus that honorifics are inappropriate; yet actual usage contradicts this.
- Young people are more easy-going in their usage of honorifics. (Cf. Gengo Seikatsu: “It is difficult to smile when you say gozaimásu”).
- People who strive to be polite tend to use honorific forms.
- The “rigid” personality type (i.e. the one slow to grasp a change of situation in a psychological test) is poor at using honorifics.
Notice 20 & 22 are still taught in the abstract as proper Keigo, and 20 in particular has been cited as proof of “the Japanese group mentality”; it’s interesting to learn that there was resistance to it.
Another important point of Keigo as a mini-grammar is that, contrary to what I just claimed, some of its real-world usage has nothing to do with sociological issues (which can be very confusing for learners trying to make sense of it!). In particular, the so-called “overuse of the o- honorific prefix” probably has little to do with “a Buddhist respect for everyday things”, as some have claimed; it would perhaps be possible to imagine a special respect for tea (o-Cha), but then why not for powdered tea (Matcha), fine-grade tea (Gyokuro-cha), or, say, for Shoyu or Miso? And why would farts always be referred “respectfully” as o-Nara, and (low-prestige) effeminate men as o-Kama? It’s more reasonable to think that these words are simply spoken like this, with a conventional neutral prefix o-. In other cases (like the infamous o-Biiru “beer”) the o- isn’t mandatory, but rather than expressing respect or politeness, it reflects at best a desire to speak elegantly (what some call 美化語 Bikago, beautified language). Miller says that, rather, this has to do with euphonics, but I ran out of time right now; any of these days I’ll try to search for this argument and post a summary.
(Pardonnez-moi, Korean and Okinawan/Uchinaa enthusiasts, for not writing on Martin’s comparative description of the speech levels in these languages; I was afraid of running too close to copyright infringement already.)