What do the Chinese call 'flat lice'?

扁蝨 Mandarin bianshi 'ticks', also known as 壁蝨 bishi 'wall lice'.

Where do 'flat lice' live in Japan?

In valleys. Dani 'tick' sounds like tani 'valley' (which becomes -dani in some compounds).

Who are 'old lice'?

Teachers. Mandarin 老師 laoshi 'teacher' (lit. 'old master') could be misspelled as 老蝨 laoshi 'old tick'. Note, however, that the 師 -shi of laoshi 'teacher' has a neutral tone (though by itself it has the same tone as shi 'louse'). (Googled tonight to find others have independently coined this pun.)

What are 'sons of lice'?

Lions. 獅子 shizi 'lion' is homophonous with 蝨子 shizi 'lice'. Although the -zi of 'lice' is a noun suffix not meant to be taken as 'son', one could misinterpret it that way.

Next: Why don't I drink beer? (This is not a non sequitur.)

*5.17.2:12: 'Wind' is 風. Deleting the left-hand stroke 丿 results in 虱, the short form of 蝨 'louse'.

風 'wind' consists of 凡虫 'ordinary bugs'. 凡 is phonetic** and 虫 'bug' is presumably semantic, though I don't understand its function. Chinese characters can be as puzzling as Tangut characters.

**5.17.2:31: In Old Chinese, 風 'wind' was something like *prəm (or *plom?, if it's the source of Siamese ลม lom 'wind') and 凡 was *bram. (The *br-cluster was also in 梵 Late Old Chinese *bramh, which was devised to transcribe Brahma.) The final consonants shifted later to avoid a sequence of consecutive labials in the same syllable(*p/b-m):

風 OC *prəm > Middle Chinese *puŋ > Mandarin feng

凡 OC *bram > Middle Chinese *bɨam > Mandarin fan

梵 Late OC *bramh > Middle Chinese *bɨamh > Mandarin fan

East Asians call Sanskrit 梵語 'Brahma language'.

5.17.2:36: Chinese raised on simplified characters won't see much of a resemblance between 'wind' and 'louse' since 風 'wind' has become 风 which only shares 乁 with 虱 'louse'. FOUR FLAT LICE

has been on my mind tonight. Here are four questions about them:

What do the Chinese call 'flat lice'?

Where do 'flat lice' live in Japan?

Who are 'old lice'?


In my survey of Western Slavic reflexes of earlier *rʲ, I noted that this palatal r still existed as ŕ in Lower Sorbian. The acute accent atop consonants is associated with palatals in Sorbian and Polish orthography.

Slovak also has an ŕ in its orthography, but this turns out to be a long syllabic [r̩r̩]. In Slovak as well as Czech orthography, acute accents on vowels signify length: e.g., á = [aa]. Hence Slovak ĺ is a long syllabic lateral [l̩l̩]. Slovak long liquids behave like regular long vowels in declension: e.g., the stem-final lengthening rule applies to both syllabic liquids and vowels in the genitive plurals of some neuter stems:

kladivo 'hammer' > kladív 'of hammers'

srdce 'heart' > sŕdc 'of hearts' (cognate to heart)

jablko 'apple' > jabĺk 'of apples' (cognate to apple)

Sanskrit has syllabic ṛ, ṛṛ, and ḷ. The subscript dots represent syllabicity. Sanskrit-based alphabets also include a character for long ḷḷ (e.g., devanagari ॡ), but no such sound exists in any real Sanskrit word*.

One might think that Slovak is extremely conservative because its syllabic liquids sometimes correspond to those of Sanskrit: e.g.,

'heart': Slovak srdce : Skt hdaya (cognate to heart)

'dead': Slovak mŕtvy : Skt mta (with a short syllabic ṛ; cognate to murder)

'wolf': Slovak vlk : Skt vka (original l can become Skt r; both cognate to wolf)

Like Slovak, Sanskrit has stem-final lengthening in some genitive plurals: e.g.,

Skt pit 'father' : pitṛṛṇaam 'of fathers' (cognate to father; note that the lenghthening rule applies to masculine and feminine as well as neuter nouns: e.g.,

maat 'mother' [f.] : maatṛṛṇaam 'of mothers')

cf. Skt pada 'step': padaanaam 'of steps' (cognate to foot)

(See above for Slovak examples.)

However, Slovak syllabic liquids originate from Proto-Slavic vowel-liquid sequences and their length is secondary (Short 1993: 537). Those vowel-liquid sequences in turn partly originated from Proto-Indo-European syllabic liquids (Shenker 1993: 64). Slovak has gone full circle.

Moreover, the lengthening of liquids in Sanskrit genitive plurals was by analogy with the lengthening of a in a-stems like 'step' above. In Beekes' (1995: 177) Proto-Indo-European reconstruction, the genitive plural of 'father' ended in *-r-om without a short or long syllabic nasal.

Schenker (1993: 64) reconstructed both long and short syllabic nasals and liquids in Proto-Indo-European:

*m̩ *m̩m *̩n *n̩n̩ *ṛ *ṛṛ *ḷ *ḷḷ

but Beekes only reconstructed short syllabic consonants:

*m̩ *̩n *ṛ *ḷ

How many other languages distinguish long and short syllabic consonants? I don't know of any other than Slovak and Sanskrit. I can't find a way to search for syllabic consonants in UPSID.

Googling "long syllabic consonants" led to only 18 hits. Most involve Slovak or PIE. One is a description of ... Spanish as spoken in the US! G- BEFORE Z(H)-: SEGMENT OR TONE?

Nishida (1964) and Tai (2008) both reconstructed ʁz-type clusters in Tangut without corresponding noncluster z-type intitials. I don't know of any other language with ʁz- but without z-. Does g- in Tibetan transcriptions like g-z- and g-zh- represent this ʁ-? g- is close to ɣ- which is close to ʁ-. (Perhaps g- was [ɣ] in preinitial position in the Tibetan dialect[s] of the transcriber.) Although that is phonetically plausible, Arakawa (1999) and Tai (2008) proposed that preinitial g- is a transcription for the Tangut level tone. Does the g- in g-z- and g-zh- correspond to level tones? If it did, I would ideally expect

Tibetan transcription g-z- g-zh-
Nishida (1964) ʁz-, ʁ- ňž-
Tai (2008) ʁz- ʁź-
This site z- ʒ-
Tangut level tone syllables all all
Tangut rising tone syllables none none

but the actual numbers in Tai (2008) are

Tibetan transcription g-z- g-zh-
Tangut level tone syllables 21 6
Tangut rising tone syllables 27 (!) 0

One could propose that g-z- represents ʁz- regardless of tone, but I still doubt that a language could have ʁz- without z-.

(5.14.0:15: There are no transcriptions of ʒ-syllables with the rising tone, so I do not know whether such syllables would have been transcribed with g-zh- or zh-. Hence the figures in the g-zh- column cannot be used to defend either a segmental or tonal interpretation of transcriptional g-.) TANGUT LIQUIDS: HERE COMES THE SUN

Why would the Tangut group z-sounds with the liquids l- lh- r-?

In Early Middle Chinese, there was only a single liquid *l- (symbolized by 來 *ləj 'come'). There was also a palatal nasal *ɲ- (symbolized by 日 *ɲit 'sun').

In northwestern Late Middle Chinese, nasal initials often partly denasalized: e.g.,

*ɲ- > *ɲdʑ- > *ɲʑ-

EMC voiced sibilant fricatives devoiced:

*z- > *s-

*ʑ- > *ɕ-

(Hence Tangut borrowings from Chinese after this shift have s- and ʃ- corresponding to EMC *z- and *dʑ- > *ʑ-.)

The loss of *ʑ- left a gap to be filled by the 'sun' initial which lost its nasality:

*ɲʑ- > *ʑ-

In modern NW Chinese dialects, the 'sun' initial has become retroflex:

*ʑ- > *ʒ- > *ʐ-

In the Late Middle Chinese-based tradition of consonant classification, the 'come' initial was a 半舌音 'half-tongue sound' and the 'sun' initial was a 半齒音 'half-front-tooth sound'. 半 'half' may imply some degree of phonetic similarity between the two. Pulleyblank reconstructed them as *l- and *r- in LMC.

Perhaps the 'sun' initial was *r- in the dialect of the unknown person who devised this classification system, but there is no evidence for it being *r- in the Tangut period NW dialect known to the Tangut. The Tangut transcribed the local version of the 'sun' initial with a Tangut initial transcribed in Tibetan as (g-)zh- (not r-). The Tibetan transcription implies a Tangut initial like ʑ-, ʒ-, or ʐ-. (The initial g- in the Tibetan transcription may indicate a level tone rather than the ʁ- reconstructed by Nishida and Tai. I'll look into this later.)

The Tangut may have learned that l- and z- sounds go together because a Chinese consonant classification system grouping *l- and *r- together as 'half' sounds was interpreted as grouping *l- and *ʒ- together in the local Chinese dialect. In other words, the Tangut were unaware of the original logic behind the grouping but followed it anyway instead of grouping z- and ʒ- with their voiceless counterparts in chapters VI and VII of Homophones. Voiced sibilant fricatives were grouped with their voiceless counterparts in the dialect underlying the Chinese consonant classification tradition, but the Tangut did not realize that because in the Chinese dialect they knew,

*ts- *tsh- *dz- *s- *z-

*tɕ- *tɕh- *dʑ- *ɕ- *ʑ-

had shifted to

*ts- *tsh- *tsh- *s- *s-

*tʃ- *tʃ- *ʃ- *ʃ- *ʃ-

and the only remaining voiced sibilant fricative in that dialect was the 'sun' initial. TANGUT LIQUIDS: THE DVOŘÁK HYPOTHESIS (PART 4)

Given how Czech developed the exotic sound ř, I would expect the exotic Tangut liquids to have similar origins:

*rʲ > ř

*lʲ > ɮ

I would also expect Tangut words with these initials to have liquid-initial cognates in other languages. However, Guillaume Jacques (2006) proposed gDong-brgyad rGyalrong cognates with nonrhotic, nonlateral fricatives and affricates:

'juniper tree': rG ɕɤɣ : my current Tangut reconstruction řiw (formerly ʒiw)

'leopard': rG kɯ-rtsɤɣ : my current Tangut reconstruction ɮeʳw (formerly zeʳw)

'long': rG kɯ-zri: my current Tangut reconstruction ɮ (formerly ziʳ)

This led me to propose that Tangut ʒ and z (now ř and ɮ) were partly lenitions of earlier and *ts in intervocalic position.

I see three scenarios:

1. Guillaume's proposals and my lenition hypothesis are wrong. The true cognates, if any, of those Tangut words should have initial liquids. Can anyone find such forms in other languages? Offhand I can think of

Old Chinese 柏 *prak 'cypress' (a juniper is a kind of cypress)

Old Chinese 豹 *prlaws or *rplaws 'leopard'

but it's not known if their initial clusters consist of stop prefixes plus liquid root initials.

2. Guillaume's proposals are correct, but my lenition hypothesis needs to be revised from

*CVʃV (or *CVʂV?) > ʒV

*CVtsV > zV


*CVʃV (or *CVʂV?) > řV

*CVtsV > ɮV

However, I cannot think of any language in which *ʃ/ʂ became an r-like sound or *ts became an l-like sound.

3. Guillaume's proposals, my lenition hypothesis, and my earlier reconstructions ʒ and z all are correct, even though z-sounds aren't liquids like the other initials in Chapter IX of Homophones.

None of the three are satisfying, but I will revert to my earlier stance for now.

Next: The 'solar solution' might explain why the Tangut grouped z-sounds with liquids instead of sibilants. TANGUT LIQUIDS: THE DVOŘÁK HYPOTHESIS (PART 3)

How did Czech develop the exotic sound ř? One can guess by looking at these correspondences:

Czech r : Polish r : Russian р [r]

Czech ř [r͡ʒ]: Polish rz [ʒ] : Russian рь [rʲ]

Apparently original *rʲ came to be pronounced with more friction in pre-Czech and pre-Polish:

*rʲ > *r͡ʒ

Short (1993: 463) states that this shift occurred in Czech in the thirteenth century.

Czech more or less maintained *r͡ʒ, but Polish simplified it:

*r͡ʒ > rz [ʒ]

Modern Polish [ʒ] has two spellings, rz < *rʲ and ż (the only dotted letter in Polish) from nonrhotic sources. I presume the Polish spelling rz dates from a period before *r͡ʒ lost its rhotic quality.

Stone (1993: 764) describes Cassubian rz as "[ɼ], a post-alveolar fricative trill" which may sound like Czech ř.

The frication of *rʲ did not occur in all western Slavic languages. Sorbian (not to be confused with Serbian) has ʀʲ ~ (spelled r ~ rj; also ŕ in Lower Sorbian) without frication*. The extinct Polabian language also had without frication (Polański 1993: 799). Slovak is very similar to Czech but it has r(i) without frication corresponding to Czech ř: e.g.,

'speech': Slovak r : Czech ř

river': Slovak rieka : Czech řeka

Short (1993: 541) mentions "depalatalization" of earlier *rʲ but does not refer to defrication, so I presume there was never any ř-like intermediate stage.

Next: Does the origin of Czech ř have any parallels in Tangut?

*However, *rʲ in the cluster *trʲ shifted to Upper Sorbian [tsʲ] and Lower Sorbian [tʆ]: e.g.,

*trʲi 'three' > US i, LS i
cf. my English tree [tʃrii]

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2009 Amritavision