Legs, bridges and running: the parallel between Japanese and Korean

  1. 7 years ago


    Jun 2014 Administrator
    Edited 7 years ago by Zachary

    While looking up the origins of some Japanese words, I came upon a most curious parallel between Japanese and Korean. In both Japanese and Korean, the words for "leg", "bridge" and "run" all seem to share the same shape and could possibly have the same roots, although the link between the Japanese set and the Korean one isn't clear.


    • あし【足】 ashi "leg; foot"
    • はし【橋】 hashi "bridge"
    • はしる【走る】 hashi-ru "to run; to travel"


    • 다리 dali/dari
    • 다리 dali/dari
    • 달리다 dalli-da/darri-da "to run"

    Although it would have been tempting to reject the link between ashi "leg; foot" and hashi "bridge" in Japanese due to the second having an initial consonant, the addition of hashiru "to run", whose connection to "leg; foot" seems plausible, and the parallel between the three forms in Korean makes the relation possible. Unfortunately, the initial and medial consonants, which would lend to the Old Japanese form *pasi, are quite different from the initial d- and medial -l-/-r- found in the Korean forms.

    Or are they?

    Deciphering the initial consonant

    Interestingly, in Japanese, the Kanji 足 has another native "kun" reading, ta-, that's used in the verbs たりる【足りる】 tariru "to be enough" and たる【足る】 taru "to be enough". Unfortunately, the meaning of these words seems to have nothing in common with legs, bridges or running, so it's unlikely that they're related or that they're cognate with the similar-looking Korean forms. However, given that Korean scholars would have historically taught the Japanese the Chinese writing system, it's not unreasonable to think that tariru and taru were spelled with 足 as a result of the root sounding the same as the Korean word for "leg".

    That said, Korean uses a different word for "foot": 발 bal. Could bal be related to the Old Japanese form *pasi? It's possible, given that the initial consonant shares the same place of articulation and that the vowel is the same. But this leads to the next question: could Korean -l- be linked to Japanese -s(i)?

    Deciphering the medial consonant

    Given other words that have been proposed as cognates between Japanese and Korean, it's not unreasonable to believe that there exists a correspondance between Korean -l(i) and Japanese -s(i). Two of such words include Old Japanese 星 *posi (modern hoshi) "star" and Korean 별 byeol "star", as well as Old Japanese 年 *toshi "year" and Korean dol(s)/tol(s) "first birthday; cycle", nominalized as /tols-i/ and pronounced [to.li].

    I therefore consider the relation between Korean -l(i) and Japanese -s(i) as possible. In fact, and interestingly enough, 足 in Japanese is also used in two other words: たす【足す】 tasu "to add; to top up" and たし【足し】 tashi "top up". Again, the meanings are irrelevant, but the Kanji usage and variance between tar- and tas- in Japanese itself lends further credit to the idea that /r(i)/ and /s(i)/ could share an origin.

    If we accept this proposal, then Korean 발 bal "foot" could very well be linked to the Old Japanese root *pasi

    What about dali/dari?

    Unfortunately, with the available data, we cannot posit a relation with the Old Japanese root *pasi- due to the phonological difference in the initial consonant, even if they both have two syllables and semantically it entirely matches up. The Korean form would lead us to expect the Japanese reflex *tari/*tashi, which we do not see. So if there really once was a connection, then this raises a few questions. How did *d- evolve into *p- by Old Japanese? Or conversely, how did Korean gain the initial *d-?

    If we were to assume the first question was correct, then the quickest way to the modern Japanese form would have been through aspiration: /t/ → /tʰ/ → /h/. This would effectively mean that the reconstructed Old Japanese form *p- would be incorrect. However, a later shift from /h/ to /ɸ/ would not have been impossible. If we were to assume the second question was correct, then the change remains a mystery to me.

    Having said that, in an earlier thread , I remarked on the curious similarity between the Japanese and Korean words for "teeth" (OJ *pa vs. MK ni) and "leaves" (OJ *pa vs. MK nip). It may be worth noting that here too, the difference is between a bilabial consonant (/p/) and an alveolar consonant (/n/).


    Independently, within both Japanese and Korean, the words for "leg", "bridge" and "run" share the same form and could be cognates. However, when the words in each language are compared to those of the other, the phonology fails to add up. The Old Japanese word *pasi "leg; foot" could very well be related to modern Korean 발 bal "foot", as the phonology is similar and the changes needed are supported by similar cognate proposals. Nonetheless, we cannot deny the parallelism that exists between Japanese and Korean, but the mystery remains as to how *pasi could be etymologically related to Korean dali/dari. Further studies on the two languages would need to compare the possible relationship between Old Japanese *p- and modern or middle Korean d- and n-.

  2. Miyukwa

    Jul 2014 Daejeon, Republic of Korea

    If Middle Korean tali "leg" and toli "bridge" shares a same root, then it must be the root of tol- (< *dVg-Vr-) "to hang", toli- "to bring someone along", and tah- (< *dVg-) "to reach". Japanese has turu, turara, tuk- 4, tukaf- 4, and tur- 2e (glosses omitted). Japanese words such as tuna, tudura, tug- 4, tunag- 4 (once again, sorry) seem to belong to another root (in my reconstruction, Proto-Korean-Japanese *diɲ- LH), namely that of Korean nis- "to connect". I am not sure about tukur- 4 "to build" and yuf- 4 "to plait, to weave". Phonetically and semantically similar yuk- 4 "to build" is most likely to be a cognate of Korean ni- "to go".

    Modern Korean talli- "to run" is from Middle Korean tolgi- "to make run" < tot- "to run" + -gi- "(causative suffix)".

  3. Zachary

    Jul 2014 Administrator

    Do you know what shift led to Middle Korean /o/ becoming Modern Korean /a/ in the words for "bridge" and "run"? Was it actually /ɔ/ > /a/?

  4. Miyukwa

    Jul 2014 Daejeon, Republic of Korea
    Edited 7 years ago by Miyukwa

    Middle Korean vowel ㆍ (o in Yale Romanization) is now considered to be /ʌ/ by many scholars. In Yale, it is written such because the relationship between ㆍ /ʌ/ o and ㅡ /ɨ/ u is just like that between ㅗ /o/ wo and ㅜ /u/ wu.

    Common Modern Korean reflexes of /ʌ/ are ㅏ /a/, ㅓ /ə/ (as in 벌 RR beol = Yale pel < MK pol "plain"), and ㅡ /ɨ/ (mostly in non-initial syllables). In some cases involving labial consonants, it have also became ㅗ /o/ or ㅜ /u/ as in 혼자 RR honja = Yale hwonca < MK hoboza "by oneself".

or Sign Up to reply!