Anyway, the back story for this, is that for my morphology class I had to write a final paper about anything morphology-related, save for a few exceptions like English and Spanish. I chose to work with Japanese, since I was also in a Japanese Linguistics course at the time, so I had a nice textbook to work with (Tsujimura's An Introduction to Japanese Linguistics), and I chose compounding since it was the longest part of the morphology chapter which would help me meet my page requirements. I put so much effort into this paper, I can't believe he only gave me a 90%. I know that's not bad, but that was below the class average, meaning that he gave lots of people way higher scores, and even my friend who said she B.S.ed her paper got an 89. I worked so hard on thiiiiis T_T
Anyway, mind you, I wrote this paper with the expectation that the reader knows next to nothing about Japanese and only the basics (if any) linguistics. So anyone (who can read English XD) should be able to read it! This should be interesting to you whether or not you know Japanese!!!! Even if you are a native Japanese speaker, I implore you to read it- most people don't actually know or think about the intricacies of their own language, so maybe this will be enlightening! :D
I'm sorry, but I have to switch back to the smaller font just for this one. :P
EDIT: When I posted, I noticed not all of the formatting stayed the same as my paper, so things don't line up properly. Please bear with it, as formatting with this blog is...reeeally difficult? Sorry!
This paper is my own intellectual property save for where sources given. Do not reduplicate.
Japanese
Compounding
Mary Gebbie
1. Introduction:
Japanese is a fascinating language
rich in structure and vocabulary. Within this expansive vocabulary
exist a variety of compound words, which is what I will cover in this
short paper. First I will give a quick background of the Japanese
language, as well as historical information regarding the influence
of Chinese in regards to writing and pronunciation as this is
pertinent when analyzing compounds. The subsequent section describes
the morphological compounds, as well as the phonological rules and
patterns that accompany compounding. Finally, I will cover certain
syntactic and semantic constraints for the compounding rules that
will be discussed.
2. Background:
Many people are familiar with Japanese
enough to know that, like Chinese, the writing system includes many
characters, called kanji, picture-like symbols which represent
the meaning of the word as opposed to representing a phonological
utterance. This is true, but there is more to the story than that.
Japanese, Ryukyuan, and Ainu, all of
which are spoken on the islands of Japan, are somewhat a mystery
regarding their origins. Japan in an archipelago in the Pacific Ocean
near the North Eastern coast of the Asian continent. The original
inhabitants of the island are considered to be the Yayoi- a
people believed to have been of Mongoloid descent, leading some to
consider Japanese a part of the Macro-Altaic language family (Inoue,
242). Although almost exclusively spoken in Japan (with only a few
migrant communities in other countries, such as America and Brazil),
with over 125 million speakers, it is the ninth most spoken language
in the world (“Most Widely Spoken Languages”).
Prior to the influence from Chinese,
the people of Japan had no writing system. It wasn't until “552
A.D., as it is officially dated, Buddhism was introduced to the
Yamato court...thus initiating 300 years of active cultural exchange
between China and Japan” (Inoue, 244). One of the most influential
additions of this exchange for Japan was the introduction of the
Chinese characters. Not only did the Japanese take the meanings of
the characters and apply them to words they already had, but they
also incorporated the Chinese pronunciation (although they have been
altered to suit the Japanese phonetic constraints). It is for this
reason that almost every single kanji has at least two ways of
being pronounced- although there are often times more (Inoue, 244).
As the Chinese grammar structure is
very simplistic, it is not unreasonable to use a series of a few
characters to represent a sentence. However, Japanese is much more
complicated in regard to the use of particles and verb and adjective
conjugation. Over time, the Japanese scholars began simplifying the
kanji that represented particles, then eventually simplifying
one kanji to represent each sound in the phonetic inventory as
a simple symbol, ultimately creating the Japanese alphabet, called
hiragana (Inoue, 247). Since then, Japanese has been written
using a combination of kanji and hiragana, kanji
being used to represent nouns and the roots of verbs and adjectives,
hiragana being used for certain adverbs, particles, and
conjugation. There are two other writing systems in Japanese as well,
katakana- another kanji based writing system used
primarily for loan words and mimetics- and romaji- the roman
alphabet, which is commonly used on public signs- but these systems
are only important in the fact that they are not used in Japanese
compounds with other writing systems.
3. Phonological and
Morphological Features:
Since Japanese is an agglutinating
language, seeing a word that is a mix of kanji and hiragana
is a sign the word will be pronounced with the kanji's
kun-reading, the
reading of traditional Japanese pronunciation. However, when the word
is part of a compound, the on-reading
is most often used, which is the pronunciation based off the
Chinese reading of the kanji. Figure 1 compares the two
readings (although some examples many have more than the two readings
shown). Kanji is in bold. Data is from my own knowledge of
Japanese.
1. Kanji Kun and On Readings
Kanji | Kun-reading | Kun-reading example | On-reading | On-reading example |
変 – change, strange, difficult | ka | 変える
– kaeru to change (v.) |
hen | 変化 – henka change (n.) |
山 - mountain | yama | 息吹山
– Ibuki yama Mt. Ibuki (name) |
san | 火山
– kazan volcano (n.) |
楽 – comfort, ease, music | tano | 楽しい
– tanoshii fun (adj.) |
raku | 楽園 –
rakuen paradise (n.) |
急 - hurry, sudden, emergency | iso | 急いで
– isoide hurriedly (adv.) |
kyuu | 急に
– kyuu ni suddenly (adv.) |
Note that although the on-reading
example for 急
does include the hiragana に(ni),
this is realized separately from the kanji word, as it is a
particle ending which indicates manner (ie. turning an adjective or
nominal adjective into an adverb).
Rendaku
Rendaku, or sequential voicing, is a phonological phenomena
that occurs when compounding in Japanese. In general native Japanese
words tend to start with voiceless consonants, but when compounded,
the second word of the compound changes from voiceless to voiced.
This rule applies almost exclusively to native Japanese words, but
there are prominent exceptions, such as san becoming kazan
above in Figure 1 (Otsu 1980). A common explanation for Sino-Japanese
and other loan words becoming exceptions to Rendaku is because they
“occur frequently enough to be considered native words”
(Tsujimura, 53). The following examples illustrate the use rendaku
in Japanese words compared to non-Japanese native words, which
include some data from Otsu (1980). Others are constructed examples.
2. Rendaku Application
1a. ato + kaki → atogaki
after writing → afterword
(of a book); postscript
b. ato + kin → atokin
money → balance;
left over money
2a. te + kami → tegami
hand paper → letter
b. te + sou → tesou
aspect → palm
reading
3a. yasu + heya → yasubeya
cheap room → cheap room
b. yasu + hoteru → yasuhoteru
hotel → cheap
hotel
In each a example, the second word of the compound is a native
Japanese word, which undergoes rendaku, whereas in each b
example, the second word is a Sino-Japanese word or other loan word.
As for why heya of 3a becomes beya, historically,
present day /h/ was originally the voiceless bilabial stop /p/, so in
cases of sequential voicing, it changes to /b/ (Tsujimura, 51).
Besides not being applicable to non-Japanese native words, rendaku
exhibits other constraints. One such constraint is Lyman's Law, which
states that if the second part of a compound contains any voiced
obstruents within it, ie. stops, fricatives, or affricates, rendaku
is blocked (Tsujimura, 53). This law is illustrated in Figure 3 (Otsu
1980: 210).
3. Lyman's Law
1a. oo + kata → oogata
big size → big size
b. oo + kaze → ookaze
wind → big wind
2a. juzu + tama → juzudama
rosary beads → prayer beads
b. juzu + tsunagi → juzutsunagi
tie, fasten → roping
together
There exist other constrains for the application of rendaku,
however they are syntactic and semantic in nature, and will be
addressed later.
N-V Compounds
Looking at the morphology of compounds, it is clear that there exist
many types regarding different parts of speech and what types of
words they are made up of. Noun-Verb compounds are one such type. In
these compounds, the noun always proceeds the verb, however the part
of speech is determined by the Right-Hand Head Rule; when the verb
maintains the categorical status of a verb, the whole compound is a
verb. If it is a deverbal noun, the compound is treated as a noun
(Tsujimura, 165).
Verbs in Japanese are easily recognized in their dictionary form as
they always end in the letter 'u'. The following is a list of N-V
compounds that are verbs, and as such, can be conjugated:
4. Verb Noun-Verb Compounds (Tsujimura, 165)
Iro + color (n.) |
Aseru →
to fade (v.) |
Iroaseru to discolor (v.) |
Abura + grease, oil (n.) |
Shimiru →
to permiate, to soak in (v.) |
Aburajimiru to become greasy, to be oil-stained (v.) |
Tema + time, labor (n.) |
Toru →
to take (v.) |
Temadoru to take time, to be delayed (v.) |
As Japanese verbs use the kun-reading, rendaku takes
affect in the latter two compounds.
When verbs in Japanese are deverbalized into nouns, they will end in
either [ i ] or [ e ], depending on how it is conjugated. The
following figure contains N-V compounds that are nouns:
5. Noun Noun-Verb Compounds (Tsujimura, 165)
Kan can (n.) |
Kiru → kiri →
to cut (v.) → (n.) |
Kankiri can opener (n.) |
Sake alcohol (n.) |
Nomu → nomi →
to drink (v.) → (n.) |
Sakenomi heavy drinker (n.) |
Futsuka second day (n.) |
You → yoi →
to become intoxicated (v.) → (n.) |
Futsukayoi hangover (n.) |
Hi sun (n.) |
Yaku → yake →
to grill, to burn (v.) → (n.) |
Hiyake sunburn (n.) |
These type of N-V compounds are much more common. Also, for
syntactic/semantic reasons, the second word in the compound does not
undergo rendaku, as seen in the kankiri example.
V-V Compounds
(pg. 169 in Tsujimura)
Verb-verb compounds in Japanese are extremely extensive in not only
number, but also syntactic and semantic constraints. However, they
are relatively easy to create.
Tsujimura
explains V-V compounds are rather common in Japanese, and can serve
several functions. Syntactic compounds express aspectual meaning,
such as such as the start, stop, or continuation of an action.
Lexical compounds showing semantic function can be an action that
would normally be two verbs in English, such as kake-agaru,
literally meaning “run-go up”, to describe running uphill.
Another function of lexical compounds is what in English would be a
resultative expression, such as “shake someone awake”, yuri-okosu
(shake-wake someone), or
“cut [a tree] down”, kiri-taosu
(cut-fell something) (169). Because the list of V-V compounds and
their semantic and syntactic constraints is so vast, I will only
cover the morphological aspects of their formation.
For
all verb-verb compounds, the stem form of the first verb is used-
which looks like the polite form of the verb minus an inflected
ending. This is attached to a complete form (as opposed to stem form)
of the second verb, which can be conjugated and inflected
accordingly. I am using the word stem and not root, since in the
majority of cases, it is not simply the root of the verb that is
used. The stem form of verb roots that end in a consonant include an
epenthetic vowel. For example, the verb kaku,
to write, is the root kak,
and the present tense inflection u.
The stem form, however, is kaki,
which comes from kakimasu;
kak,
the root, mas being
the polite ending, u
again inflecting present
tense. The i between
the root and the polite ending does not carry meaning; it is
presumably added for ease of articulation, as Japanese is a mora
based language and has certain constraints regarding consonant
clusters.
Syntactic
compounds are very productive, and their meaning is predictable based
on their components (Martin 1987: 438). Their productivity comes from
the fact that the second word of the compound must be a verb from a
finite group of words that expresses aspectual meaning, while the
meaning of the main verb remains unaltered. The following are
examples of syntactic v-v compounds.
6. Syntactic V-V Compounds (Tsujimura 170)
Main Verb | Aspectual Verb | Compound |
Kaku – to write | Hajimeru – to begin | Kaki-hajimeru – to begin writing |
Kaku – to write | Owaru – to finish | Kaki-owaru – to finish writing |
Kaku – to write | Tsuzukeru – to continue | Kaki-tsuzukeru – to continue writing |
Kaku – to write | Naosu – to correct | Kaki-naosu – to re-write |
Kaku – to write | Sugiru – to exceed | Kaki-sugiru – to write too much |
Seemingly,
in lexical verb-verb compounds, instead of the first verb being the
main verb, the first verb acts as a description of manner, such as in
the previously given example, Kiri-taosu
(“to cut down”, lit.
cut-fell something). If we consider the second verb the main verb,
this theory makes sense since it is a resultative expression; If you
explain that you felled a tree by cutting it, the fact that you
felled the tree is overall more important, the fact you did so by
cutting it is only extra information. The following are more examples
of lexical V-V compounds. In regard to the first example, note that
/ts/ and /ch/ are allophones for [t] when before [u] and [i]
respectively.
7. Lexical V-V Compounds (ex. 1-3 from Tsujimura 170-171)
Verb 1 | Verb 2 | Compound |
Utsu – to shoot | Korosu – to kill | Uchi-korosu – to shoot to death |
Taberu- to eat | Aruku – to walk | Tabe-aruku – to go and try food at various venues |
Ukeru – to receive | Toru – to take | Uke-toru – to accept |
Miru – to see | Suteru – to throw away | Mi-suteru – to abandon, to desert |
Hiku – to pull | Tomeru – to stop | Hiki-tomeru – to hold back, to restrain |
The manner-action order seems to work for all of the above examples,
albeit obscurely for some, as the meaning of the compound is not
always clear based on its components. For instance, to accept
something means not just that you are taking something, but that you
are doing so willingly, that you are receiving it, and to abandon
something implies there was volition behind the act of not returning
to it, which is shown in Japanese by the fact that you throw
something away even though you can see it. This sort of pattern works
for all lexical verb-verb compounds, regardless if their English
representation is a two-verb action or a resultative expression.
Dvandva
Compounds
Another set of compounds in Japanese are known as “dvandva
compounds”, a term meaning the two words are related semantically.
Although these copulative compounds are always nouns, they may be
formed using regular lexical nouns or deverbalized nouns. By nature,
copulative compounds do not have a head, so dvandva compounds will
not be subject to rendaku (and we will see why this is
important in the syntactic section of this paper). This is not
surprising for another reason, as it is the nature of compounds to
use the Chinese on-reading, but even in the cases of the
kun-reading being used (as in example a, e, and g-j),
sequential voicing does not occur. The following list (of my own
examples) provides some examples of dvandva compounds:
8. Dvandva Compounds
a. oya + ko → oyako
parent child parent and child
b. ten + chi → tenchi
heaven earth heaven and earth
c. dan + jo → danjo
male female boys and girls
d. nichi + ji → nichiji
day time date and time
e. te + ashi → teashi
hand leg arms and legs
f. anshin + anzen → anshinanzen
secure safe safe and secure
g. kachi + make → kachimake
win loss winning and losing
h. kashi + kari → kashikari
loan debt borrowing and
lending
i. yomi + kaki → yomikaki
reading writing reading and writing
j. iki + ki → ikiki
going coming coming and going
Examples a-f of figure 6 above are made up of lexical nouns. These
type are the more common of the two, and the list of possible
combinations is most probably endless, as it seems to be possible to
invent new compounds when the two nouns you are referring to are
related in the discourse, as well as when new words are added to the
Japanese lexicon via foreign language word borrowing. An example of
the latter would be something like beekon-eggu, meaning bacon
and eggs, or some words including the “and” of English, like
pointo-ando-shuuto - yes, all one word-, meaning point and
shoot. Examples g-j, however, are made up of deverbal nouns. Although
compounds most often use the Chinese based on-reading for the
kanji, since verbs use the native Japanese reading, their
deverbalized counterparts do as well.
4.
Syntacic and Semantic Issues:
As previously mentioned in section three, the phonological
phenomenon rendaku exhibits some syntactic and semantic
constraints as well as phonological ones. Just as in English,
ambiguity may exist with Japanese words, for example, Tsujimura
writes in her book about the ambiguity of a noun phrase like
“lacquered chopsticks box” (p.53). There are two interpretations
of this phrase:
a) a chopsticks box that is lacquered
b) a box for lacquered chopsticks
Essentially, this problem can be solved by looking at the use of
rendaku. The Japanese pieces to make up the word are nuri +
hashi + hako, meaning “lacquered chopsticks box” in that order.
Rendaku will apply differently for each semantic
interpretation of the compound.
9. Rendaku and Semantics
Lacquered chopsticks-box Lacquered-chopsticks box
a) a chopsticks box that is lacquered b)
a box for lacquered chopsticks
In sentence a, chopsticks-box is considered one unit, or
constituent, so looking at the combination hashi + hako, we can see
the hako, as the second unit in the compound, becomes voiced
(recall that the “voiced” version of /h/ is /b/ for historical
reasons). Now the next constituent is nuri + hashibako. This is where
Lyman's Law comes into play, as now the second unit of the compound
contains a voiced obstruent, so rendaku cannot apply.
In sentence b, we want lacquered-box to act as a constituent, which
in Japanese, is nuri + hashi. In this case as well, rendaku is
applied to the second unit of the compound, making hashi become
bashi. Now, looking at nuribashi + hako, there is nothing to
stop rendaku from applying to the second unit of the compound,
so hako becomes voiced as well (Tsujimura, 55).
Another constraint had to be considered to account for any compounds
that follow this previously stated pattern in ambiguous noun phrases,
yet cannot be confined by Lyman's Law. One example would be nuri +
kasa + ire, lacquered + umbrella + case, which can become nurigasaire
or nurikasaire due its ambiguity, even
though Lyman's Law should not stop rendaku from
happening in the latter interpretation as kasaire contains no
voiced obstruents (Tsujimura, 56). This final constraint is called
The Right Branch Condition, and proposes that “rendaku
applies only when a potential rendaku segment is in a right
branch constituent” (Otsu 1980: 219). This not only explains the
nurikasaire exception, but it explains how one could get both
nurihashibako and nuribashibako from the example above.
In tree form, that would look like this:
- Right-Branch Condition (Tsujimura, 57)
5.
Conclusion:
Although Japanese may be difficult for learners, as we can see
through the different methods of compounding, there is a
systematicity behind the formation of such words. Because of this,
Japanese compounds are quite productive, which is quite convenient
for learners such as myself, for if I am not sure of the correct way
to say a word, I can explain what I mean with a made up compound, and
odds are, I will be understood.
Bibliography
Inoue,
Kyoko. "Japanese: A Story of Language and People."
Languages
and Their Speakers.
Ed. Timothy Shopen. Cambridge, MA: Winthrop, 1979. 241-300. Print.
Martin,
Samuel E. "A Reference Grammar of Japanese." Journal
of Linguistics (1987)
"Most
Widely Spoken Languages." Ignatius.edu.
N.p., 26 May 2012. Web. 14 Apr. 2013.
<http://www2.ignatius.edu/FACULTY/TURNER/languages.htm>.
Otsu Y. “Some Aspects of Rendaku in Japanese and related
problems.” (1980)
Tsujimura,
Natsuko. An
Introduction to Japanese Linguistics.
2nd ed. Cambridge, Mass,: Blackwell, 2007. Print.