Commonthroat: a language that only a dog could probably pronounce

[lurker]

qhjg

/huff, short rising weakening growl, short low weak growl/

Noun

Care, attention, reverence

The Bright Way distinguishes three levels of religious devotion, similar to the Christian distinction between adoratio/latria and veneratio/doulia. <qhjg>, literally care or attention, is the lowest of these three levels. It is given to sacred objects such as the star hearth, the bones of the deceased, and holy relics, as well as abstractions like the noosphere and even religious positions like that of hearthkeeper and research monk. Some Claravian theologians, especially research monks, include the universe itself and the laws governing it, known as the <rjGJDFg> Great Creation in this group.

Entities afforded this honor are not considered inherently holy, but only "holy by association." The star hearth is just a fusion reactor, but it is given reverence as an icon of an icon. Bones are just bones, but the bones of one's loved ones are special. This is especially true when considering religious offices. The office of hearthkeeper is accorded due respect, but a particular individual holding that office may be a despicable individual. Indeed, she is all the more culpable who abuses her office for worldly ends.

This concept is confusing to some humans, and even to many secular yinrih, who think, for example, that the Farspeakers' care for the body of the noosphere constitutes worship of the internet.

Along with other similar words like <kgg> praise, <rfDqg> interest, and <nLqg> honor, the word <qhjg> is used along with the verb <qcDr> give along with the proposition <rl> when used in a sentence.

rl h rFbrp b g rjsFl qcDr qhjg.
I reverence bones of my loved-ones.
Literally "I give attention to the bones of my loved ones"

The word is also used in a mundane sense. <qhjlNrg> medical care, <qhjsfcg> childcare, etc.

[TBPO]

qhjg

/huff, short rising weakening growl, short low weak growl/

Noun

Care, attention, reverence

The Bright Way distinguishes three levels of religious devotion, similar to the Christian distinction between adoratio/latria and veneratio/doulia. <qhjg>, literally care or attention, is the lowest of these three levels. It is given to sacred objects such as the star hearth, the bones of the deceased, and holy relics, as well as abstractions like the noosphere and even religious positions like that of hearthkeeper and research monk. Some Claravian theologians, especially research monks, include the universe itself and the laws governing it, known as the <rjGJDFg> Great Creation in this group.

Entities afforded this honor are not considered inherently holy, but only "holy by association." The star hearth is just a fusion reactor, but it is given reverence as an icon of an icon. Bones are just bones, but the bones of one's loved ones are special. This is especially true when considering religious offices. The office of hearthkeeper is accorded due respect, but a particular individual holding that office may be a despicable individual. Indeed, she is all the more culpable who abuses her office for worldly ends.

This concept is confusing to some humans, and even to many secular yinrih, who think, for example, that the Farspeakers' care for the body of the noosphere constitutes worship of the internet.

Along with other similar words like <kgg> praise, <rfDqg> interest, and <nLqg> honor, the word <qhjg> is used along with the verb <qcDr> give along with the proposition <rl> when used in a sentence.

rl h rFbrp b g rjsFl qcDr qhjg.
I reverence bones of my loved-ones.
Literally "I give attention to the bones of my loved ones"

The word is also used in a mundane sense. <qhjlNrg> medical care, <qhjsfcg> childcare, etc.

What are remaining two?

[lurker]

What are remaining two?

The vocab is not set in stone, but the highest level is worship proper, given only to the Uncreated Light. I may give it the word <kgg> glory, praise, renown mentioned above. Below that is the veneration given to saints and martyrs, which I may assign the word <nLqg> honor.

So it goes from <qhjg> reverence to <nLqg> honor to <kgg> praise.

[lurker]

bf rnP

Verb

To pass away of old age. Literally "to finish living".

bfrnPg: a deceased person who has died in such a way.



fb rnP

Verb

To die suddenly before one's time, as from accident or disease. Literally "to cease living"

fbrnPg: a casualty of war or natural disaster.

Edit: Also the victim of a crime or disease. The semantic space may be widening, or may have already widened, to include someone who has suffered misfortune of any kind, even if that doesn't involve the person dying.

[lurker]

qc

(co)verb, preposition

to face

qcg

/yip, short low strong whine, short low weak growl/

Noun

Face (anatomy), appearance, look, superficial qualities

qc qP

from qc + qP (block, hinder)

Verb

To oppose an enemy

Code: Select all

rlnqK sBrK g qcqn
rlnq-K  sBr-K      g   qc-qn
cut-DOG remove-DOG POS face-2
Imma cut yer face off!

An example of using the dogmatic mood as a threat and an example of verb serialization

Code: Select all

BC sjgqp BDsfsf.
BC  sjgq-p  BD-sfsf
all head-3D INT-MIR?
Gee, what are all those heads?

Inflecting a question in the mirative mood to show bewilderment

Code: Select all

h  qc-Mr   b  qkjGH-g
PL face-3P of evilness-3I
These are the faces of evil.

A simple declarative using the plural "these are the..." along with a prepositional phrase acting as an adjective

Code: Select all

G qgj sBF h BCMr.
G   qgj-0   sBF-0 h  BC-Mr
IMP fight-A win-A PL all-3P
You must conquer each.

Another example of verb serialization to express a more complex concept. fight win means [/i]to conquer[/i], and a (still tentative) use of a quantitative noun. <BCMr> = all of this (mass), while <h BCMr> = each of these (count).

NGL also another excuse to reference 2009 YouTube culture.

[lurker]

So it turns out a lot of software is case-insensitive, including Excel and anything that uses hyperlinks. I may have to redo the romanization for Commonthroat to get rid of the case distinction. Not sure how I'll do that. It needs to be easy to type and not look horrible.

[lurker]

OK, so my first romanization scheme used vowels, and I didn't like that because it gave the impression that this was a human-pronounceable language, and it used both capital <i> and lowercase <L>. My second romanization which I have been using for the majority of this thread simply removed the vowels and shifted the corresponding sounds to match.

Both of these systems use letter case to signify vowel length. That's bad because a lot of software is case-insensitive, such that Foo, FOO, foo, and fOo etc. are considered the same string. This includes the Windows filesystem, Excel, and anything that uses HTML hyperlinks, including applications that use Markdown.

To that end, I'm trying to come up with a third romanization scheme that doesn't rely on capital letters. Here are a few candidates.

Here's a sample text in the second romanization.

G rnL rDB rLPqp rNlr hgq rLPqp rNlrGHp rNlrGHp.
G rnL rDB rLPqp qCDr hgq rLPqp qCDrGHp qCDrGHp.
G rnL rDB rLPqp snp hgq rLPqp rNPqp b BC snpGHp.


Option 1: use macrons for long vowels, and replace huffs, chuffs, and yips with modifier commas to make it resemble the actual Commonthroat script a bit more.

ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʼn̄lʼ hgʻ ʼl̄p̄ʻp ʼn̄lʼḡh̄p ʼn̄lʼḡh̄p.
ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʻc̄d̄ʼ hgʻ ʼl̄p̄ʻp ʻc̄d̄ʼḡh̄p ʻc̄d̄ʼḡh̄p.
ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʽnp hgʻ ʼl̄p̄ʻp ʼn̄p̄ʻp b b̄c̄ ʽnpḡh̄p.

Option 2: Same as option 1, but keeping the consonants as they are in the current scheme.

ḡ rnl̄ rd̄b̄ rl̄p̄qp rn̄lr hgq rl̄p̄qp rn̄lrḡh̄p rn̄lrḡh̄p.
ḡ rnl̄ rd̄b̄ rl̄p̄qp qc̄d̄r hgq rl̄p̄qp qc̄d̄rḡh̄p qc̄d̄rḡh̄p.
ḡ rnl̄ rd̄b̄ rl̄p̄qp snp hgq rl̄p̄qp rn̄p̄qp b b̄c̄ snpḡh̄p.

Option 3: following long vowels with an <x>. It's easy to type but doesn't look good.

gx rnlx rdxbx rlxpxqp rnxlr hgq rlxpxqp rnxlrgxhxp rnxlrgxhxp.
gx rnlx rdxbx rlxpxqp qcxdxr hgq rlxpxqp qcxdxrgxhxp qcxdxrgxhxp.
gx rnlx rdxbx rlxpxqp snp hgq rlxpxqp rnxpxqp b bxcx snpgxhxp.

Option 4: Same as 3, but with <w> instead of <x>. Not sure how this is any better.

gw rnlw rdwbw rlwpwqp rnwlr hgq rlwpwqp rnwlrgwhwp rnwlrgwhwp.
gw rnlw rdwbw rlwpwqp qcwdwr hgq rlwpwqp qcwdwrgwhwp qcwdwrgwhwp.
gw rnlw rdwbw rlwpwqp snp hgq rlwpwqp rnwpwqp b bwcw snpgwhwp.

Option 5: similar to option 2, but uses a combining overline instead of a macron. It's supposed to make an unbroken line when two consecutive long vowels are used, but it doesn't render very well.

g̅ rnl̅ rd̅b̅ rl̅p̅qp rn̅lr hgq rl̅p̅qp rn̅lrg̅h̅p rn̅lrg̅h̅p.
g̅ rnl̅ rd̅b̅ rl̅p̅qp qc̅d̅r hgq rl̅p̅qp qc̅d̅rg̅h̅p qc̅d̅rg̅h̅p.
g̅ rnl̅ rd̅b̅ rl̅p̅qp snp hgq rl̅p̅qp rn̅p̅qp b b̅c̅ snpg̅h̅p.

Option 6: This is literally just the braille code I posted last year and that's in my sig. The idea is that this isn't a language humans can pronounce, so the very idea of a romanization makes no sense.

⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠜⠞⠊⠜⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠜⠞⠊⠜⠏⠯⠺⠀⠜⠞⠊⠜⠏⠯⠺.
⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠌⠭⠝⠜⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠌⠭⠝⠜⠏⠯⠺⠀⠌⠭⠝⠜⠏⠯⠺.
⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠬⠚⠺⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠜⠞⠾⠌⠺⠀⠉⠀⠍⠭⠀⠬⠚⠺⠏⠯⠺.

Option 7: This is a hack to fake case-sensitivity in case-insensitive software. Each capital letter is followed by a zero width space. That way the strings are different.

G​ rnL​ rD​B​ rL​P​qp rN​lr hgq rL​P​qp rN​lrG​H​p rN​lrG​H​p.
G​ rnL​ rD​B​ rL​P​qp qC​D​r hgq rL​P​qp qC​D​rG​H​p qC​D​rG​H​p.
G​ rnL​ rD​B​ rL​P​qp snp hgq rL​P​qp rN​P​qp b B​C​ snpG​H​p.

Option 8: similar to option 1, but with dots rather than macrons. The dots are supposed to be more aesthetically pleasing because letters with ascenders have dots below rather than above. (<F> being the exception because I couldn't find a precomposed character with an underdot).

ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʼṅlʼ hgʻ ʼḷṗʻp ʼṅlʼġḥp ʼṅlʼġḥp.
ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʻċḍʼ hgʻ ʼḷṗʻp ʻċḍʼġḥp ʻċḍʼġḥp.
ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʽnp hgʻ ʼḷṗʻp ʼṅṗʻp b ḅċ ʽnpġḥp.

Option 9: Not so much a romanization but a canadianaboriginalsyllabification. The shapes turn clockwise starting at 12 o'clock, going from low weak, to low strong, to high weak, to high strong. Closed triangles are whines, open triangles are growls, and U-shapes are grunts. A dot to the right indicates length, and the smaller glyphs used as finals in CAS are used for the consonants.

ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᒃᑘᑎᒃ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒃᑘᑎᒃᐽᑁᑕ ᒃᑘᑎᒃᐽᑁᑕ.
ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᒄᐓᐍᒃ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒄᐓᐍᒃᐽᑁᑕ ᒄᐓᐍᒃᐽᑁᑕ.
ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᐟᑌᑕ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒃᑘᑢᒄᑕ ᐃ ᐏᐓ ᐟᑌᑕᐽᑁᑕ.

[WeepingElf]

Option 8 looks best to me.

[lurker]

I think I'll go with option 2 for the third romanization, but continue using the current second romanization on the forum here. It's the least different from the current system and only uses a single diacritic that I can keep in the clipboard if I can't find/make a keyboard layout for it.

[thethief3]

You should use vowel letters for the commonthroat consonants

[HolyHandGrenade!]

What about doubling long vowels?

[lurker]

What about doubling long vowels?

Sometimes there's a hiatus between two identical short vowels.

<sCqgg> (novice healer). The first short low weak growl <g> is part of the root, and the second is the 3rd person indefinite suffix.

[HolyHandGrenade!]

What about doubling long vowels?

Sometimes there's a hiatus between two identical short vowels.

<sCqgg> (novice healer). The first short low weak growl <g> is part of the root, and the second is the 3rd person indefinite suffix.

Maybe use an apostrophe to indicate hiatus?

[lurker]

What about doubling long vowels?

Sometimes there's a hiatus between two identical short vowels.

<sCqgg> (novice healer). The first short low weak growl <g> is part of the root, and the second is the 3rd person indefinite suffix.

Maybe use an apostrophe to indicate hiatus?

Nah, the scheme with macrons is easy to convert to programmatically with a quick search and replace.

[lurker]

Back to the drawing board with the third Romanization and the transfer of my lexicon to Obsidian. For some odd reason Obsidian doesn't like combining diacritics. It messes with the search function for some reason. it doesn't think <-h̄ḡ> exists even though it clearly does. Right now the romanization with combining macrons is considered version 3.1, and the one with dots, which I devised earlier, is 3.0. V3.0 uses all precomposed characters, which should make searching easier, but will make typing harder, especially on Linux.

The upside is I have pronunciations for almost all of my dictionary entries, and (admittedly brief) etymologies for all my compound words.

The other problem is that Obsidian treats searches beginning with a hyphen as negations -XYZ = everything except XYZ. Because of the way Obsidian does linking the lemma has to be the filename, so rLPqg (light) has to be saved under rLPqg.md (adjusting for the case-insensitive romanization of course). Since the headword for suffixes contains a leading hyphen, it becomes tricky to search for them.

Obsidian is a much more flexible platform than Excel for this, and will hopefully allow me to add more rich and lore-dense vocab entries like I do here on the forum. I'll also be able to link between entries, for example, link all the words that use a particular suffix to the entry for that suffix.

Edit: Obsidian seems to convert charters with combining diacritics into their precomposed versions in some situations but not others. I think that's why the files are messed up. Specifically, Ḡḡ has a precomposed form. Obsidian automagically converts instances of <ḡ> into their precomposed forms, but only in filenames, which are treated as the title of the note, so a note titled "foo" will be saved as a markdown file "foo.md". The lemma is the filename in my case, so if the word contains a long low weak growl, it'll get converted to its precomposed form in the title/filename, but any links to that file will use the two-character g + combining macron version, which Obsidian won't recognize as the same string and hence not the same file.

[TBPO]

OK, so my first romanization scheme used vowels, and I didn't like that because it gave the impression that this was a human-pronounceable language, and it used both capital <i> and lowercase <L>. My second romanization which I have been using for the majority of this thread simply removed the vowels and shifted the corresponding sounds to match.

Both of these systems use letter case to signify vowel length. That's bad because a lot of software is case-insensitive, such that Foo, FOO, foo, and fOo etc. are considered the same string. This includes the Windows filesystem, Excel, and anything that uses HTML hyperlinks, including applications that use Markdown.

To that end, I'm trying to come up with a third romanization scheme that doesn't rely on capital letters. Here are a few candidates.

Here's a sample text in the second romanization.

G rnL rDB rLPqp rNlr hgq rLPqp rNlrGHp rNlrGHp.
G rnL rDB rLPqp qCDr hgq rLPqp qCDrGHp qCDrGHp.
G rnL rDB rLPqp snp hgq rLPqp rNPqp b BC snpGHp.


Option 1: use macrons for long vowels, and replace huffs, chuffs, and yips with modifier commas to make it resemble the actual Commonthroat script a bit more.

ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʼn̄lʼ hgʻ ʼl̄p̄ʻp ʼn̄lʼḡh̄p ʼn̄lʼḡh̄p.
ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʻc̄d̄ʼ hgʻ ʼl̄p̄ʻp ʻc̄d̄ʼḡh̄p ʻc̄d̄ʼḡh̄p.
ḡ ʼnl̄ ʼd̄b̄ ʼl̄p̄ʻp ʽnp hgʻ ʼl̄p̄ʻp ʼn̄p̄ʻp b b̄c̄ ʽnpḡh̄p.

Option 2: Same as option 1, but keeping the consonants as they are in the current scheme.

ḡ rnl̄ rd̄b̄ rl̄p̄qp rn̄lr hgq rl̄p̄qp rn̄lrḡh̄p rn̄lrḡh̄p.
ḡ rnl̄ rd̄b̄ rl̄p̄qp qc̄d̄r hgq rl̄p̄qp qc̄d̄rḡh̄p qc̄d̄rḡh̄p.
ḡ rnl̄ rd̄b̄ rl̄p̄qp snp hgq rl̄p̄qp rn̄p̄qp b b̄c̄ snpḡh̄p.

Option 3: following long vowels with an <x>. It's easy to type but doesn't look good.

gx rnlx rdxbx rlxpxqp rnxlr hgq rlxpxqp rnxlrgxhxp rnxlrgxhxp.
gx rnlx rdxbx rlxpxqp qcxdxr hgq rlxpxqp qcxdxrgxhxp qcxdxrgxhxp.
gx rnlx rdxbx rlxpxqp snp hgq rlxpxqp rnxpxqp b bxcx snpgxhxp.

Option 4: Same as 3, but with <w> instead of <x>. Not sure how this is any better.

gw rnlw rdwbw rlwpwqp rnwlr hgq rlwpwqp rnwlrgwhwp rnwlrgwhwp.
gw rnlw rdwbw rlwpwqp qcwdwr hgq rlwpwqp qcwdwrgwhwp qcwdwrgwhwp.
gw rnlw rdwbw rlwpwqp snp hgq rlwpwqp rnwpwqp b bwcw snpgwhwp.

Option 5: similar to option 2, but uses a combining overline instead of a macron. It's supposed to make an unbroken line when two consecutive long vowels are used, but it doesn't render very well.

g̅ rnl̅ rd̅b̅ rl̅p̅qp rn̅lr hgq rl̅p̅qp rn̅lrg̅h̅p rn̅lrg̅h̅p.
g̅ rnl̅ rd̅b̅ rl̅p̅qp qc̅d̅r hgq rl̅p̅qp qc̅d̅rg̅h̅p qc̅d̅rg̅h̅p.
g̅ rnl̅ rd̅b̅ rl̅p̅qp snp hgq rl̅p̅qp rn̅p̅qp b b̅c̅ snpg̅h̅p.

Option 6: This is literally just the braille code I posted last year and that's in my sig. The idea is that this isn't a language humans can pronounce, so the very idea of a romanization makes no sense.

⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠜⠞⠊⠜⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠜⠞⠊⠜⠏⠯⠺⠀⠜⠞⠊⠜⠏⠯⠺.
⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠌⠭⠝⠜⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠌⠭⠝⠜⠏⠯⠺⠀⠌⠭⠝⠜⠏⠯⠺.
⠏⠀⠜⠚⠎⠀⠜⠝⠍⠀⠜⠎⠾⠌⠺⠀⠬⠚⠺⠀⠫⠋⠌⠀⠜⠎⠾⠌⠺⠀⠜⠞⠾⠌⠺⠀⠉⠀⠍⠭⠀⠬⠚⠺⠏⠯⠺.

Option 7: This is a hack to fake case-sensitivity in case-insensitive software. Each capital letter is followed by a zero width space. That way the strings are different.

G​ rnL​ rD​B​ rL​P​qp rN​lr hgq rL​P​qp rN​lrG​H​p rN​lrG​H​p.
G​ rnL​ rD​B​ rL​P​qp qC​D​r hgq rL​P​qp qC​D​rG​H​p qC​D​rG​H​p.
G​ rnL​ rD​B​ rL​P​qp snp hgq rL​P​qp rN​P​qp b B​C​ snpG​H​p.

Option 8: similar to option 1, but with dots rather than macrons. The dots are supposed to be more aesthetically pleasing because letters with ascenders have dots below rather than above. (<F> being the exception because I couldn't find a precomposed character with an underdot).

ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʼṅlʼ hgʻ ʼḷṗʻp ʼṅlʼġḥp ʼṅlʼġḥp.
ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʻċḍʼ hgʻ ʼḷṗʻp ʻċḍʼġḥp ʻċḍʼġḥp.
ġ ʼnḷ ʼḍḅ ʼḷṗʻp ʽnp hgʻ ʼḷṗʻp ʼṅṗʻp b ḅċ ʽnpġḥp.

Option 9: Not so much a romanization but a canadianaboriginalsyllabification. The shapes turn clockwise starting at 12 o'clock, going from low weak, to low strong, to high weak, to high strong. Closed triangles are whines, open triangles are growls, and U-shapes are grunts. A dot to the right indicates length, and the smaller glyphs used as finals in CAS are used for the consonants.

ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᒃᑘᑎᒃ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒃᑘᑎᒃᐽᑁᑕ ᒃᑘᑎᒃᐽᑁᑕ.
ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᒄᐓᐍᒃ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒄᐓᐍᒃᐽᑁᑕ ᒄᐓᐍᒃᐽᑁᑕ.
ᐽ ᒃᑌᑚ ᒃᐍᐏ ᒃᑚᑢᒄᑕ ᐟᑌᑕ ᐳᐱᒄ ᒃᑚᑢᒄᑕ ᒃᑘᑢᒄᑕ ᐃ ᐏᐓ ᐟᑌᑕᐽᑁᑕ.

Option 1.

[lurker]

אבגדהוזחטיכלמנסעפצקרשת

I toyed with doing a Hebrewization of CT since it would capture the RTL direction of actual written Commonthroat, and I could use the vowel points as the breathing marks, but there aren't enough letters :( The paucity of ascenders and descenders might also make it hard for me to parse.

[TBPO]

אבגדהוזחטיכלמנסעפצקרשת

I toyed with doing a Hebrewization of CT since it would capture the RTL direction of actual written Commonthroat, and I could use the vowel points as the breathing marks, but there aren't enough letters :( The paucity of ascenders and descenders might also make it hard for me to parse.

Use Chinese alphapet!

[lurker]

אבגדהוזחטיכלמנסעפצקרשת

I toyed with doing a Hebrewization of CT since it would capture the RTL direction of actual written Commonthroat, and I could use the vowel points as the breathing marks, but there aren't enough letters :( The paucity of ascenders and descenders might also make it hard for me to parse.

Use Chinese alphapet!

The point in me trying Hebrew was to capture the right-to-left text direction and (reverse) abjad nature of the Commonthroat script. But yes, in the end since this isn't even a human-pronounceable language the very idea of a "romanization" is as arbitrary as using any other system, and whatever I come up with is merely a way for me to type it conveniently.

𝙸𝚗 𝚝𝚑𝚎 𝚎𝚗𝚍 𝙸 𝚝𝚑𝚒𝚗𝚔 𝙸 𝚊𝚖 𝚐𝚘𝚒𝚗𝚐 𝚝𝚘 𝚞𝚜𝚎 𝚝𝚑𝚒𝚜 𝚖𝚘𝚗𝚘𝚜𝚙𝚊𝚌𝚎 𝚞𝚗𝚒𝚌𝚘𝚍𝚎 𝚋𝚕𝚘𝚌𝚔. 𝙲𝚘𝚖𝚙𝚞𝚝𝚎𝚛𝚜 𝚛𝚎𝚌𝚘𝚐𝚗𝚒𝚣𝚎 𝚝𝚑𝚎 𝚞𝚙𝚙𝚎𝚛 𝚊𝚗𝚍 𝚕𝚘𝚠𝚎𝚛 𝚌𝚊𝚜𝚎 𝚌𝚑𝚊𝚛𝚊𝚌𝚝𝚎𝚛𝚜 𝚊𝚜 𝚍𝚒𝚜𝚝𝚒𝚗𝚌𝚝. 𝚃𝚑𝚒𝚜 𝚠𝚊𝚢 𝙸 𝚌𝚊𝚗 𝚔𝚎𝚎𝚙 𝚞𝚜𝚒𝚗𝚐 𝚝𝚑𝚎 𝚌𝚞𝚛𝚛𝚎𝚗𝚝 𝚛𝚘𝚖𝚊𝚗𝚒𝚣𝚊𝚝𝚒𝚘𝚗.

(In the end I think I am going to use this monospace unicode block. Computers recognize the upper and lower case characters as distinct. This way I can keep using the current romanization.)

[lurker]

I think I've finally finished porting the lexicon to Obsidian from Excel. Here is a picture of the graph view. The nodes clustered around the perimeter are monomorphemes with no internal etymology. Words composed of more than one morpheme are linked to their component morphemes. Because of the way I've decided to structure the lexicon, homophones and highly polysemous terms are grouped under a single note, so 𝙱 /long low weak whine/ is a single file 𝙱.md, with sub-entries for the verb "try" or "attempt" as well as the preposition "from"/"away from".