The foreign origins of the lion dance and words for "lion" in Sinitic

« previous post | next post »

Here at Language Log, we have shown how the most common word for "lion" in Sinitic, shī 獅, has Iranian and / or Tocharian connections (see "Selected readings").  The etymological and phonological details will be sketched out below.  For a magisterial survey, see Wolfgang Behr, "Hinc [sic] sunt leones — two ancient Eurasian, migratory terms in Chinese revisited", International Journal of Central Asian Studies, 9 (2004), 1-53.  This learned essay has appeared in multiple guises and many places (I knew it originally and best while it was still in draft, perhaps back in the 90s), so I don't know which one the author considers to be the most authoritative version.  Perhaps he will enlighten us in the comments to this post.

shī 獅

Originally written as . Earliest written attestations in the Eastern Han era include the Book of Han [before 111], the surviving portion of the Dongguan Hanji [c. 150], and Lokakṣema's translation of the Aṣṭasāhasrikā Prajñāpāramitā [179]. The term is possibly of Iranian origin. In the Book of Han, the lion was described as indigenous to the historical 烏弋山離 (OC *qaː lɯɡ sreːn rel, “Alexandria”), possibly Alexandria Prophthasia (Yu, 1998), which was part of the Parthian Empire at the time. The Dongguan mentions a lion as a gift from the Shule Kingdom in the year 133, where a Saka language was spoken then. Earlier definite dates associated with the lion mentioned in the Book of the Later Han include the year 87, when a gift lion from the Yuezhi was recorded.

Compare Proto-Iranian *cárguš (lion). Possibly related to 狻猊 (OC *sloːn ŋeː).

Meanwhile, Adams (2013) postulates possible Tocharian connections; cf. Tocharian B ṣecake and etymologies.

(Zhengzhang): /*sri/


The previous notes have already introduced the term suānní 狻猊, which is another way to refer to the lion or a lion-like creature.  It is even older than shī 獅.

suānní 狻猊

The (legendary) beast was mentioned in the ancient Chinese glossary Erya as , a "devourer of tigers and leopards". It was also attested in the Mu Tianzi Zhuan (nominally from the Warring States period; however the textual history of the book was problematic). If correctly identified as a legendary account of the lion as later texts do, it may be a (partial) borrowing, possibly from a historical Iranian or Indo-Iranian language.

The use of 狻猊 (OC *swar ŋe) possibly predates that of (OC *sri, “lion”). Compare Proto-Indo-Iranian *sinȷ́ʰás (lion), Proto-Iranian *cárguš.

(Zhengzhang): /*sloːn  ŋeː/


The famous lion dance was imported to China from Central Asia.

There has been an old tradition in China of dancers wearing masks to resemble animals or mythical beasts since antiquity, and performances described in ancient texts such as Shujing where wild beasts and phoenix danced may have been masked dances. In Qin Dynasty sources, dancers performing exorcism rituals were described as wearing bearskin mask, and it was also mentioned in Han Dynasty texts that "mime people" (象人) performed as fish, dragons, and phoenixes. However, lion is not native to China (a species found in Northeast China Panthera youngi had long become extinct), and the Lion Dance therefore has been suggested to have originated outside of China from countries such as India or Persia, and introduced via Central Asia] According to ethnomusicologist Laurence Picken, the Chinese word for lion itself, shi (獅, written as 師 in the early periods), may have been derived from the Persian word šer. The earliest use of the word shizi meaning lion first appeared in Han Dynasty texts and had strong association with Central Asia (an even earlier but obsolete term for lion was suanni (狻麑 or 狻猊), and lions were presented to the Han court by emissaries from Central Asia and the Parthian Empire. Detailed descriptions of Lion Dance appeared during the Tang Dynasty and it was already recognized by writers and poets then as a foreign dance, however, Lion dance may have been recorded in China as early as the third century AD where "lion acts" were referred to by a Three Kingdoms scholar Meng Kang (孟康) in a commentary on Hanshu. In the early periods it had association with Buddhism: it was recorded in a Northern Wei text, Description of Buddhist Temples in Luoyang (洛陽伽藍記), that a parade for a statue of Buddha of a temple was led by a lion to drive away evil spirits. An alternative suggestion is therefore that the dance may have developed from a local tradition that appropriated the Buddhist symbolism of lion.

There were different versions of the dance in the Tang Dynasty. In the Tang court, the lion dance was called the Great Peace Music (太平樂, Taiping yue) or the Lion Dance of the Five Directions (五方師子舞) where five large lions of different colours and expressing different moods were each led and manipulated on rope by two persons, and accompanied by 140 singers. In another account, the 5 lions were described as each over 3 metres tall and each had 12 "lion lads", who may tease the lions with red whisks. Another version of the lion dance was described by the Tang poet Bai Juyi in his poem "Western Liang Arts" (西凉伎), where the dance was performed by two hu (胡, meaning here non-Han people from Central Asia) dancers who wore a lion costume made of a wooden head, a silk tail and furry body, with eyes gilded with gold and teeth plated with silver as well as ears that moved, a form that resembles today's Lion Dance. By the eighth century, this lion dance had reached Japan. During the Song Dynasty the lion dance was commonly performed in festivals and it was known as the Northern Lion during the Southern Song.


The suānní 狻猊 came to my purview today as part of an exhibition of art treasures that is being held at the National Palace Museum in Taipei (10/26/21-1.26.22) in the form of a 15th c. painting depicting this animal.  It looks remarkably like a lion to me.

Nearly every government building in China is guarded by a pair of lions flanking the front door.  The same is true of major banks and other public buildings and even private houses in many other countries around the world.  This is a custom that may be traced back to Hattusa, the capital of the Hittite empire (c. 1650 BC-c. 1178 BC), where an imposing pair of lions guards the gateway leading to the palace.  Bear in mind that Hittite is the oldest attestable Indo-European language, older yet than Tocharian and Sanskrit.

"Hic sunt leones!"


Selected readings

[h.t. Geoff Wade]


  1. Victor Mair said,

    January 14, 2022 @ 1:24 pm

    From Miriam Robbins Dexter:

    "In Qin Dynasty sources, dancers performing exorcism rituals were described as wearing bearskin mask"

    Cf. girls playing the Brauron bear in Attica (Aristophanes, "Lysistrata" 645; Whence the Goddesses 115)

  2. John Swindle said,

    January 14, 2022 @ 8:04 pm

    With a bit of costume the Chinese lion dance is eminently accessible for children.

    "Right, the cow goes 'Moo!' And how does the lion go?"

    "I wanna be the head! I wanna be the head!"

  3. Martin Schwartz said,

    January 14, 2022 @ 10:29 pm

    I cannot stress enough how, for the history of the 'lion' word,
    one must read the excellent exhaustive study by Václav Blažek.
    "Hic erant leones", JIES 2005, which importantly expands the scope
    of the problem. My only disagreement with VB is a minor one;
    I don't think that the Anc. Egyptian name of the lion-headed
    goddess is relevant; rather it amounts to 'strong, mighty'.
    VB's presentation of the Iranian data is admirable. The reconstruction *carguš given in the present LL is problematic.
    The affricate *c implies that the word belongs to the oldest stratum of Iranian; in fact it is useless, since it does not indicate the alternation of s- (Khotanese and Khwarezmian) vs. š- in Persic
    and Sogdian. Also, it is an unadvisable habit to cite forms
    in the nominative (actually unattested for 'lion' in Iranian,
    but that's niot my point) rather than the stem; note that
    Khot. sarau makes likely a strong stem in -āu-, from acc. sg.
    or nom. pl. I once speculated that the Persic š- was due to a collocation seen in the alliterative Classical Persian cliché
    šēr-i šarža 'fierce lion', connecting the adj. with the apparent Ir.
    LW Armenian šarž- 'to move rapidly' etc., perhaps *sarž-
    < PIE √*sk'elg referring to the motions of beasts, but no I'm not sure.
    Martin Schwartz

  4. martin schwartz said,

    January 14, 2022 @ 10:45 pm

    Re Picken's derivation, it should be kept in mind that Cl. Pers.
    and Late Middle Persian šēr derive from Middle Persian
    šaGr (G = gamma), with -Gr- representing a methathesis
    of *-rg-.
    Martin Schwartz

  5. Chris Button said,

    January 15, 2022 @ 8:14 am

    The use of 狻猊 (OC *swar ŋe)

    (Zhengzhang): /*sloːn ŋeː/

    An issue with 狻 is the evidence for a lateral onset. Hence Zhengzhang's reconstruction of *sloːn as opposed to the *swar option above.

    My hunch is that we might be talking about a shift of *ɬ- to s-, which is supported within Sinitic (albeit with an apparent preference for type-B syllables) and typologically (e.g. Canaanite and Phoenician just came up here:

  6. Brian Spooner said,

    January 15, 2022 @ 12:05 pm

    Interesting. We need more research on the prehistory of Central Asia (from today's Iran to today's China) and on the history of some particular species, like lions.

  7. Victor Mair said,

    January 16, 2022 @ 7:18 am

    From Hiroshi Kumamoto:

    For the word "lion" in Chinese and Iranian, John Brough's article in Henning Memorial Volume (referred to in passing by Blažek) is a good place to start from.

  8. Hiroshi Kumamoto said,

    January 16, 2022 @ 8:23 pm

    On the final -n in Chinese for the foreign final -r, see further Baxter- Sagart, Old Chinese. A New Reconstruction, Oxford 2014, 263.

  9. Chris Button said,

    January 16, 2022 @ 11:00 pm

    @ Hiroshi Kumamoto

    I've never really bought the idea of a simple merger of -r to -n. A merger of – r with -l and a sometime alternation of -l with -n seems more likely to me. But that then requires the reconstruction of an -l coda, which is something I follow Pulleyblank in advocating, but it can throw off the symmetry of other reconstructions.

  10. Chris Button said,

    January 17, 2022 @ 9:17 am

    From the Mu Tianzi zhuan: 狻猊野馬走五百里 "The suanmi and wild horse run 500 li".

    Any chance we're dealing with a kind of fleet-footed cat (e.g. panther, leopard, cheetah) rather than a lion? The word family around the phonetic 允 (with the *l- onset) itself suggests a semantic field similar to that of PIE *ḱers- "run".

    There are couple of other notably speedy animals in the xiesheng series too: 駿 "fine horse" and 㕙 "hare".

  11. Chris Button said,

    January 17, 2022 @ 10:45 am


  12. Chris Button said,

    January 17, 2022 @ 1:28 pm

    Might also be worth noting the 狁 component of the supposedly highly mobile Xianyun 獫狁 nomads.

RSS feed for comments on this post