09.8.29.23:59: KHITAN KHUIZ:  QUESTIONS

Robinson's Hadan-ka script reminds me of the Khitan Large Script which appears to have been an attempt to write Khitan with logographic characters representing words rather than syllables or segments (consonants or vowels) without any regard for semantics*. Although Chinese writing has also been described as logographic, most Chinese characters and compounds contain phonetic elements whereas most of the c. 1000 Khitan Large Script characters are not "composed of multiple elements with large numbers of strokes": i.e., they probably lack phonetic elements (or at least complex phonetic elements).

Similarly, Hadan-ka does not seem to have any compounds with phonetic elements. Its compounds are semantic: e.g.,

TREE + MOUNTAIN + PLAINS = PLACE

for the Kasgen word us 'place'.

Robinson has compared the semantics of us with Japanese kuni 'land, country, state'. Some think kuni is a loan from Middle Chinese 郡 *gunh 'prefecture' (in China; the Japanese use a different character). 郡, now pronounced jùn in Mandarin, is a typical Chinese combination of a phonetic element with a semantic element:

君 jūn 'gentleman' (phonetic) +阝 yì 'town' (semantic) = 郡 jùn 'prefecture'

Why does Hadan-ka lack phonetic elements? Robinson explained that

Each single image [Hadan-ka character] contains meaning that can be deciphered without knowledge of character pronunciation. In simple, stand-alone form, Hadan-ka characters can be "read" by a speaker of any language educated in Hadan-ka.

This reminds me of Shavkunov's (1963) and Toyoda's (1964) suggestions that Khitan Large Script could represent words in the Jurchen as well as the Khitan language. Janhunen (1994: 7) wrote,

Due to its logographic nature, this [Khitan large] script was equally suitable for writing both Khitan and Jurchen, and any text written in it could, in principle, be read in either language, possibly even in Chinese.

I doubt that for reasons I won't go into here. For now I merely intend to note that Hadan-ka is meant to be a script that transcends spoken languages.

One might expect such a script to be pictographic so that it "can be deciphered without knowledge of character pronunciation." The elements of Hadan-ka are pictographic: e.g.,

SUN is three lines representing the radiant rays of the sun

PLAIN is a horizontal line

PERSON is a stick figure

TREE is a stylized picture of a tree (a trunk, three branches, and a semicircle outlining its leaves)

MOUNTAINS is an M-shape representing the contours of a mountain range

Such iconic elements are presumably meant to be easy to learn.

Let's return to Khitan Large Script, a real-life analogue of Hadan-ka. How easy would it be to learn? Look at the following KLS characters and try to guess what they mean:

1. 女

2. 牙

3. 午

4. 来

5. 人 with a dot 丶 beneath it

6. 矢 with a dot 丶 on the lower right

7. ナ atop 土 with a dot 丶 on the lower right

8. 将 (but with a 亅 that doesn't go above the horizontal)

9. 仁

10. 至

11. 火 atop 日

12. 日 atop 廾

13. 田 atop 人

14. 尸 atop ホ

15. ユ

16. ユ atop 山

I'll reveal the answers next time.

(8.30.00:40: Characters 3, 4, 5, 14-16 added to the quiz.)

*Chinese writing is almost entirely syllabic (with the sole exception of 兒 for Mandarin -r as well as the syllable er), but homophonous syllables with different meanings are generally written differently: e.g., 于 yu 'in' and 魚 yu 'fish' are distinct in writing even though they are pronounced identically.

09.8.28.23:59: PERSON DOG SHEEP I PARTICLE

Does that sequence make you think of

伏羲

Fu Xi?

伏 fu < Old Chinese *bək 'lie down'

亻 'person' +

犬 'dog'

羲 xi (used to write Fu Xi's name) < Old Chinese *hŋaj

義 yi < OC *ŋajs 'righteous' (phonetic)

in turn consisting of 羊 'sheep' + 我 OC *ŋajʔ 'I' (phonetic)

兮 xi < OC *gi (a sentence-final or metrical filler particle)

Although 羲 xi and 兮 xi are homophonous in Mandarin today, they were not homophones when 羲 was created. Therefore 兮 is not phonetic in 羲 , but I can't figure out how it could be semantic. It seems like an arbitrary addition to tell the reader that 羲 OC *hŋaj is not read like 義 OC *ŋajs.

He [Bouvet] proposed that Fu Hsi [= Fu Xi] was the same as the ancient Egyptian Hermes and also Zoroaster, the ancient Persian sage. Bouvet pointed out correctly [no! - A] that Fu Hsi's name means "Dog Man,"and that Hermes was represented by the ancient Egyptians as a man with the head of a dog.

伏羲 Fu Xi is not 'dog man'. I suspect that it is a phonetic transcription of Old Chinese *bək hŋaj which has nothing to do with dogs, sheep, me, or the particle *gi.

8.29.0:39: I cite this as an example of how non-ideographic Chinese writing is. The stylized pictures that form the characters 伏羲 are irrelevant to Fu Xi unless the phonetic values of their combinations are taken into consideration.

09.8.27.23:36: CAN IDEOGRAPHY EXIST?

The idea of a 'logical', pictorial script has been around for centuries, and I'm not talking about hieroglyphics or Chinese characters. I'm referring to Leibniz' 'alphabet of human thought':

The alphabet of human thought is a concept originally proposed by Gottfried Leibniz that provides a universal way to represent and analyze ideas and relationships, no matter how complicated, by breaking down their component pieces. All ideas are compounded from a very small number of simple ideas which can be represented by a unique "real" character.

René Descartes suggested that the lexicon of a universal language should consist of primitive elements. The systematic combination of these elements, according to syntactical rules, would generate "an infinity of different words." In the early 18th century, Leibniz outlined his characteristica universalis, an artificial language in which grammatical and logical structure would coincide, which would allow much reasoning to be reduced to calculation. Leibniz acknowledged the work of Ramon Llull, particularly the Ars generalis ultima (1305), as one of the inspirations for this idea. The basic elements of his characteristica would be pictographic characters representing unambiguously a limited number of elementary concepts. Leibniz called the inventory of these concepts "the alphabet of human thought." There are quite a few mentions of the characteristica in Leibniz's writings, but he never set out any details.

I'm not surprised he never went into details. I wonder if he ever realized that not everything can be drawn. What does something abstract like 'logic' look like?

Here are some sample characters of his alphabet of human thought: 'earth', 'water', 'air', 'fire'.

Blissymbols (1949-) are much simpler:

Blissymbols or Blissymbolics were conceived of as an ideographic writing system consisting of several hundred basic symbols, each representing a concept, which can be composed together to generate new symbols that represent new concepts. Blissymbols differ from all the world's major writing systems in that the characters do not correspond at all to the sounds of any spoken language.

Can you guess what this Blissymbolic sentence means?

I don't believe in 'ideographic' writing systems. Hieroglyphics and Chinese characters are actually phonetic, not ideographic. John DeFrancis debunked the 'Ideographic Myth' in his 1984 book The Chinese Language: Fact and Fantasy:

The concept of ideographic writing is a most seductive notion. There is great appeal in the concept of written symbols conveying their message directly to our minds, thus bypassing the restrictive intermediary of speech. And it seems so plausible. Surely ideas immediately pop into our minds when we see a road sign, a death's head label on a bottle of medicine, a number on a clock. Aren't Chinese characters a sophisticated system of symbols that similarly convey meaning without regard to sound? Aren't they an ideographic system of writing?

The answer to these questions is no. Chinese characters are a phonetic, not an ideographic, system of writing, as I have attempted to show in the preceding pages. Here I would go further: There never has been, and never can be, such a thing as an ideographic system of writing ...

Ideographic writing, however, requires mastery of the tens of thousands or hundreds of thousands of symbols that would be needed for ideographic representation of words or concepts without regard to sound. A bit of common sense should suggest that unless we supplement our brains with computer implants, ordinary mortals are incapable of such memory feats.

But when Chinese memorize about 3,000 characters to be literate, aren't they doing what DeFrancis says is impossible? No, because those characters represent sounds rather than ideas. DeFrancis doesn't think people can memorize complex writing systems without a link between sound and symbol. Phonetics is a necessary bridge between memory and writing.*

The theory of an ideographic script must remain in the realm of popular mythology until some True Believers demonstrate its reality by accomplishing the task, say, of putting Hamlet or at least Lincoln's Gettysburg Address into English written in symbols without regard to sound.

Has anyone tried to do such feats with Blissymbols? Coming up with 900 pictorial symbols is one thing; translating nontrivial texts into those symbols without reference to a spoken language is another. The sample Blissymbols sentence I linked to is in fact not purely ideographic, as it relies on English grammar. The order of symbols matches the order of words in its English translation:

Blissymbols: I WANT GO CINEMA.

Translation: I want to go to the cinema.

If a Japanese speaker had invented Blissymbols, the order of symbols might be based on Japanese:

I CINEMA GO WANT

Or 'I' might not even be in the sentence, since Japanese speakers don't use pronouns as often as Anglophones do.

A true 'ideography' would not be dependent on any particular spoken language and its symbols would be in an order understandable to all. But I have never seen such a thing, and I doubt it could ever exist.

*8.28:00:06: But what about the literate deaf? Most are still dependent on phonetics because they use writing systems based on phonetics.

Writing systems based on sign languages like Stokoe notation and SignWriting would still not be ideographic because sign language is not entirely iconic:

Although it often seems as though the signs are meaningful of themselves, in fact they can be as arbitrary as words in spoken language. For example, a speaking child may often make the mistake of using the word 'you' to refer to themselves, since others use that word to refer to him or her. Children who acquire the sign YOU (pointing at one's interlocutor) make similar mistakes – they will point at others to mean themselves, indicating that even something as seemingly explicit as pointing is an arbitrary sign in ASL [American Sign Language], like words in a spoken language ....

The theory that signs are self-explanatory can be conclusively disproved by the fact that non-signers cannot understand fluent, continuous sign language. The majority of signs are opaque.

Does the Stokoe notation for 'snake' in ASL look like a snake to you? All languages have some degree of arbitrariness.

09.8.26.23:59: IS HADAN THE HUMAN ORIGIN?

I mixed up Robinson's glosses for ha and dan in Kasgen so I interpreted Hadan as

ha 'person' + dan 'origin' = 'original person'

but in fact it is

ha 'origin' + dan 'person' (makes me think of Vietnamese 民 dân 'people', borrowed from Middle Chinese *min with *m-!*)

and could be interpreted as possessor-possessed: 'origin's person' = 'orignal person'

So parts of the last two posts are wrong.

I considered making Kasgen a consistent modifier-modified language like Japanese with adjectives before nouns (e.g., Hadan 'original person') but that wouldn't explain why the second syllable of Hadan is stressed. Why not stress the first syllable to emphasize that he's the original person rather than just any old person? (Cf. the stress on White in White House.)

I briefly hypothesized that Kasgen has vowel length and that syllables with long vowels are stressed. Hence Hadan has final stress because it's really Hadaan. However, that spelling contradicts the long-established spellings Hadan and Hadanus. Moreover, if Hadanus were actually Hadaanus, why does it have initial stress instead of medial stress?

Here's my attempt to reconcile everything so far:

Hadan is a noun-adjective compound. dan can be an adjective 'having to do with people' as well as a noun 'person'. Kasgen adjectives and nouns can be homophonous: cf. English human which can be either an adjective or a noun. In speech, the adjective part of a noun-adjective compound is stressed, whereas the possessed in a possessor-possessed compound is unstressed:

Hadan [həˈdan] lit. 'origin person' (noun + adj) = 'human origin'

not just any origin, but the living human source of all that the Kasgen value

hadan [ˈhadən] lit. 'origin person' (possessor + possessed) = 'human of origin' = 'original person' (translated as adj + noun in English, but really noun + noun in Kasgen)

referring to humans from Dead Earth?

These compounds are distinguished from phrases which have stress on both words:

ha dan [ˈhaˈdan] 'origin person' (noun + adj) = 'human origin'

not Hadan, but in contexts like 'is this of human origin?'

dan ha [ˈhaˈdan] 'person origin' (noun + adj) = 'original human'

synonym of hadan referring to humans from Dead Earth?

this phrase or a compound Danha [dənˈha] is probably preferred to hadan for Old Humans because they avoid near-homophony with Hadan.

8.27.0:27: The Kasgen may regard Hadan and Danha as fitting mirror-image names for all they consider to be good and evil: the greatest man and the worst species - ours!

I expect noncompound possessive constructions to contain an unstressed genitive marker equivalent to Japanese no which I indicate as X between two stressed nouns:

ha X dan 'origin 's person' = 'original person' (whose position was later taken by another?)

dan X ha 'person 's origin' = 'origin of a person; background'

X is unnecessary in compounds because the stressed-unstressed pattern indicates the roles of the two nouns.

Hadanus means 'place of Hadan'. The first syllable of possessor-possessed compounds is stressed, and this rule can even apply to possessors with more than one syllable:

Nadus 'Nad's place' (monosyllabic possessor Nad)

Hadanus 'Hadan's place' (disyllabic possessor Hadan)

and perhaps someday

Ladaneus 'Ladane's place' (trisyllabic possessor Ladane**)

Since Kasgen has initial stress, I interpret it as a possessor-possessed compound of an unknown root plus gen 'mountain': 'mountain of ...?' I suggest that kas may mean 'power' and that the name reflects how the bodybuilding Kasgen view themselves as walking muscular mountains of power.

8.27.0:27: Kasgen refers only to the Superiors, whereas dan is a general term for any human.

*8.27.0:59: Middle Chinese *m- before *i may have become *mj- in the southern dialect known to the Vietnamese in the 10th century. Labial-*j-clusters may have briefly shifted to labial-fricative clusters before hardening to stops by the 17th century:

10th century Vietnamese *pj- > *pɕ- > *ps- > *s- > Middle Vietnamese t- > Modern Vietnamese t-

10th century Vietnamese *phj- > *phɕ- > *phs- > *sh- > Middle Vietnamese th- > Modern Vietnamese th-

10th century Vietnamese *mj- > *mʑ- > *mz- > *z- > Middle Vietnamese d- > Modern Vietnamese d- [z]

I don't know of any language with a distinction between Cɕ- and Chɕ-. Perhaps the distinction was really between *ps- (> later t-) and *pɕ- (> later th-) which would parallel these more certain changes without labials:

*s- > t-

*ɕ- > th-

**8.27.1:09: I don't know how Robinson pronounces Ladane. I'm guessing the name has initial stress

Ladane [ˈladəni]

and is a possessive compound with stress on the initial syllable of the possessor:

(la-dan)-e 'Ladan's e'

The meaning of the noun e is unknown.

This is the same pattern as Hadanus:

(ha-dan)-us 'Hadan's place' = 'Hadan's planet'

Robinson has pronounced a short form of this name as Ladan [ləˈdan] with final stress, implying that it is a noun-adjective compound 'human ...'. The meaning of the noun la is unknown.

All etymologies for native Kasgen words that Robinson has shown me so far contain monosyllabic roots. The only unanalyzable polysyllabic words are of foreign origin: e.g.,

Duwaranir 'Dwar' < Dwar Dwar-a-nir 'of the Dwar' (one morpheme in Kasgen, three in Dwar; if the suffixes -a and -nir exist in Kasgen, they don't mean 'of' and 'the')

hiyakal 'hyena-like creature' < Dwar hyakl 'id.'

Polysyllabic words contain one stressed syllable which may have one of five vowels [a i u e o]. All other syllables are unstressed and in normal speech can have only one of two vowels [ɪ ə]. This reduction of vowels is not reflected in the romanization of Kasgen which is based on etymology rather than phonetics: e.g.,

[ˈgentən] 'capital of the Kasgen' is romanized as Gentun with -u- rather than as Gentən with -ə- or some more typeable symbol for [ə] because it is a compound of gen 'mountain' + tun 'city'. There is no Kasgen root tən, since roots can only have one of the five stressed vowels.

The reduction of vowels is similar to the Russian phenomena known as

ikan'je: shift of unstressed e to i-like [ɪ]: e.g.,

jemu [jɪˈmu] 'to him'

cf. Kasgen Kasgen [ˈkasgɪn]

akan'je: shift of unstressed o to a-like [ɐ] before a stressed syllable and [ə] elsewhere (in Kasgen, unstressed u as well as a and o merge into [ə])

xorosho [xərɐˈʃo] 'well'

cf. Kasgen Hadanus [ˈhadənəs]

There are five types of stress patterns in polysyllabic Kasgen words. I already mentioned the first four in my last post.

In categories 1 and 2, the specifier (adjective or possessor) is stressed and the specified noun is unstressed.

1. Stressed adjective (following the noun it modifies)

gendan [gɪnˈdan] 'mountain' + 'original' = 'the original mountain'

usgen [ʔəsˈgen] 'place' + 'mountain' = 'the mountainous lands'

(I am assuming all syllables written with initial vowels actually have an unwritten glottal stop [ʔ] unless preceded by another consonant: e.g., Nadus is [ˈnadəs] and not [ˈnadʔəs] because d precedes u.)

Note that the presence or absence of stress and vowel reduction as well as location indicate whether a root (e.g., gen 'noun') is a noun or an adjective.

2. Stressed possessor (preceding the possessed)

genha [ˈgenhə] 'mountain' + 'person' = 'people of the mountain'

haus [ˈhaʔəs] 'person' + 'place' = 'rank' (i.e., a person's place)

3. Initial stress in unanalyzable loanwords

naryuh [ˈnaryəh] 'Urdreh biological weapon' (borrowed from Urdreh naryuh)

4. Post-initial stress in unanalyzable loanwords which originally had consonant clusters absent from Kasgen

durun [dəˈrun] 'a kind of metal' (borrowed from Dwar drun; the word cannot be Urdreh since it lacks the expected Urdreh number/gender/case ending)

I just came up with a fifth pattern: initial stress in native reduplicated forms:

gengen [ˈgengɪn] 'mountain' + 'mountain' = 'mountains'

haha [ˈhahə] 'person' + 'person' = 'people'

cf. Japanese hitobito 'people', a reduplication of hito 'person'

Unlike Urdreh which has a system of number affixes

-y-: singular

-w-: dual

-l-: countable plural

-r-: collective plural

Kasgen simply doubles a noun to indicate that there is more than one of it: e.g.,

Urdreh Remreh [rəmrəh] < Rem + -r- collective plural + -e- masculine + -h absolutive

'enslaved Urdreh' (plural)

is equivalent to

Kasgen Remrem [ˈremrɪm]

Kasgen plurals are rarer than Urdreh or English plurals: cf. the low frequency of Japanese plurals.

Kasgen compound stress is like English compound stress. The Kasgen noun-adjective stress pattern is like a mirror image of the English blackboard or White House (written as if it were a phrase, but otherwise compound-like) with an adjective-noun pattern. In both languages, this pattern contrasts with noncompounds containing two stressed components:

Hadan [həˈdan] 'the original person; Hadan'

ha dan [ˈhaˈdan] 'some original person (not necessarily Hadan)'

blackboard 'a chalkboard (not necessarily black)'

black board 'some board that is black'

White House 'residence of the president of the United States'

white house 'some white house'

Notice how the compounds are not necessarily the sum of their parts; they are bahuvriihi compounds.

English orthography uses spaces and capitalization (or their absence) to indicate a distinction between compounds and phrases.

The Kasgen script may use a different technique (e.g., enclosing elements of a compound in a box?) to make the same distinction.

09.8.24.23:59: UNSTRESSED VOWEL REDUCTION IN KASGEN?

W. Robinson Mason III released the first half of his Hadanus audio book The Burrowing of Fort Nadus today. Listening to Robinson's pronunciations of Kasgen names, I realized that Kasgen has a stress accent with reduction of unstressed vowels:

Examples (ˈ precedes a stressed syllable in bold):

Unstressed e: Kasgen [ˈkasgɪn]

Unstressed a: Hadan [həˈdan], Darad [dəˈrad], Ladane [ləˈdan] (sic; should this be re-romanized as Ladan, or is -e a feminine suffix that Darad drops?)

Unstressed u: Nadus [ˈnadəs], Gentun [ˈgentən], oskul [ˈaskəl] (sic; should this be re-romanized as askul?)

I didn't hear any examples of unstressed i or o, but I assume they would be similar to unstressed e and u.

Monosyllabic words can only have stressed vowels:

Rem 'enslaved Urdreh' is [rem], not [rɪm]

fen 'a kind of broth' is [fen], not [fɪn]

us 'place' is [us], not [əs]

The stressed syllable may indicate which part of a compound is more important:

Hadan < ha 'person' + dan 'origin' = 'original person' (as opposed to some other kind of person)

Nadus < Nad (a personal name) + us 'place' = 'place of Nad' (as opposed to someone else's place)

Note the resyllabification of Nad-us as Na-dus.

Gentun < gen 'mountain' + tun 'city' = 'city of mountains' (as opposed to some other city)

These compounds indicate that Kasgen has monosyllabic roots and Tangut-like word order:

noun-adjective

possessor-possessed

Hence Hadan [ˈhadən] with a stressed second syllable would be a possessive compound: 'person's origin'. 'Origin of Hadan' might then be Hadan hadan [həˈdan ˈhadən], perhaps with a possessive marker equivalent to English -'s (or Tangut jie or Japanese no) between the two words.

8.25.2:33: Urdreh ([ˈurdʒrə] in Anglicized pronunciation) would presumably be [ˈurdrəh] in Kasgen with a nonsilent final -h. Kasgen has no stressed vowel like Urdreh e [ə] (not [e]!), so Urdreh [urdrəh] was borrowed as [ˈurdrəh] which would be pronounced as [ˈurdrah] in overenunciated Kasgen speech with clearly pronounced vowels in all syllables. Kasgen a is the stressed vowel closest to Urdreh e [ə]. (Cf. the Japanese use of a to approximate English [ə].)

Urdreh e [ə] would be borrowed as a [a] in monosyllabic content words which are always stressed in Kasgen: e.g.,

'Urdreh male figurehead leader': Urdreh Jeh > Kasgen Jah

Kasgen e is [e], so a hypothetical Kasgen Jeh would not sound like Urdreh Jeh with e [ə].

While Kasgen approximate Urdreh e [ə] in stressed syllables as a, the Urdreh conversely approximate Kasgen e [e] as either i or e [ə]: e.g.,

'Drone; enslaved Urdreh': Kasgen Rem > Urdreh Rimyeh or Remyeh [rəmyəh]

-yeh is a singular masculine absolutive ending; a female Rem would be a Rimyah or Remyah in Urdreh

'Urdreh biological weapon': Urdreh naryuh > Kasgen naryah [ˈnaryəh]

Perhaps all loanwords into Kasgen have stress on the first syllable: e.g.,

'an Urdreh male name': Urdreh Yussuyeh > Kasgen Yussayah [ˈyussəyəh]

Exceptions may involve unstressed vowels inserted to break up foreign initial consonant clusters:

'forest people': Dwar Dwaranir > Kasgen Duwaranir [dəˈwarənɪr] (careful [duˈwaranɪr])

-a-nir are actually Dwar suffixes meaning 'of the' which Kasgen misinterpreted as integral parts of the name

'hyena-like creature': Dwar hyakl > Kasgen hiyakal [hɪˈyakəl]

Highly assimilated Rem may speak Urdreh with Kasgen vowel reduction which causes endings with e, a, u to sound alike. Three of four derivatives of the root na- 'parent' can become homophonous in Kasgenized Urdreh:

8.25.2:45: I was surprised by Robinson's pronunciation of Hadanus as [ˈhadənəs] with initial stress instead of [həˈdanəs] which is how I've been pronouncing it for almost 20 years.

Why would Hadanus and Hadan have different stress patterns if one is derived from the other? Perhaps Hadanus is not 'Hadan's place' but is 'place of the origin of people':

us 'place'

dan-us 'origin place' = 'place of origin'

ha-dan-us 'person-origin-place' = 'place of the origin of people'

Although the New Humans actually originate from the spaceship Hope, the name Hadanus may reflect a desire to regard Hadanus as the New Humans' point of origin.

09.8.23.23:59: DETERMINING THE DUCTUS (PART 2)

It's impossible to verbalize every variable in the ducti of writing systems. Yet I think an attempt at an explicit description of the ductus of an fictional writing system is worthwhile. Knowing what are and aren't acceptable elements (graphemes) in the system enables the creator to

- easily generate new characters that are consistent in form with all other characters

- create a coherent-looking system that does not look like a grab bag of random shapes

The ductus of cuneiform can be easily formulated. When the script became highly abstract, characters consisted of combinations of only five different wedges:

horizontal

vertical

downward diagonal

upward diagonal

Winkelhaken

Characters typically consisted of five to ten wedges. The first four types of wedges could vary in length for compositional purposes.

It is harder to concisely describe the ductus of Chinese characters. 永 is often cited as an example of the eight different stroke types in Chinese characters, but there are actually more. These stroke types occur not only in Chinese characters and their descendants (e.g., katakana and chữ nôm) but also in the 'siniform' (Chinese-shaped) scripts of the Khitan, Jurchen, and Tangut.

This is not to say that it is impossible for a person without training in Chinese or Tangut to distinguish between the two scripts. Chinese has many characters with small and large numbers of strokes whereas Tangut characters all tend to have around ten strokes. Compare Chinese

一 'one' (one stroke), 二 'two' (two strokes), 三 'three' (three strokes)

with their six, nine, and thirteen-stroke Tangut equivalents:

Conversely, Tangut has no characters like

龘

'dragons in flight' (three 16-stroke dragons = 48 strokes)

or this 64-stroke character for 'loquacious' made up of four 16-stroke dragons.

Tangut has no box-shaped elements unlike Chinese (e.g., 口 'mouth', 品 'product'). The Tangut character element (not an independent character) derived from Chinese 口 'mouth'

resembles the unrelated Chinese character 反 'turn over' because its bottom right strokes have been rearranged to open the box. Some Chinese characters consist of other characters enclosed in a large box: e.g., 國 'country' < 囗 'enclosure' + phonetic 或. Tangut has elements that cover three but never four sides of other elements: e.g., radical 114 'silk'

>

The hangul alphabet seems to share a ductus with Chinese characters with one major exception: the use of circles. The hangul letters ㅁ m and ㅅ s look like the Chinese characters 口 'mouth' and 人 'person', but no Chinese character contains a circle like hangul ㅇ ng orㅎ h. (Early hangul had other letters with circles: e.g., ㆆ ʔ.) An East Asian text with prominent circles other than zero is likely to be in hangul. (Japanese uses smaller circles ゜ in characters for p-initial syllables: e.g., hiragana ぱ pa, katakana パ pa.)

The ductus of the Kasgen script should be unlike any ductus ever seen on Dead Earth.