What do you base your conlang romanizations off of?

A forum for all topics related to constructed languages
Posts: 780
Joined: 25 Dec 2014 20:17
Location: Virginia

What do you base your conlang romanizations off of?

Post by Birdlang » 06 Jun 2015 20:06

And do you use any diacritics? I use diacritics or digraphs and I base some romanizations off of a Birdlang Phonetic Alphabet (BAP). Also. What are the special characters you use the most?
Ꭓꭓ Ʝʝ Ɬɬ Ɦɦ Ɡɡ Ɥɥ Ɫɫ Ɽɽ Ɑɑ Ɱɱ Ɐɐ Ɒɒ Ɓɓ Ɔɔ Ɖɖ Ɗɗ Əə Ɛɛ Ɠɠ Ɣɣ Ɯɯ Ɲɲ Ɵɵ Ʀʀ Ʃʃ Ʈʈ Ʊʊ Ʋʋ Ʒʒ Ꞵꞵ Ʉʉ Ʌʌ Ŋŋ Ɂɂ Ɪɪ Ææ Øø Ð𠌜 Ɜɜ Ǝɘ

User avatar
Posts: 5292
Joined: 22 Aug 2010 18:46

Re: What do you base your conlang romanizations off of?

Post by Xing » 06 Jun 2015 20:15

Birdlang wrote:And do you use any diacritics? I use diacritics or digraphs and I base some romanizations off of a Birdlang Phonetic Alphabet (BAP).
Waku doesn't use diacritics, simply because there isn't any need for 'em, due to its relatively small number of phonemes. But even in a language with a limited number of phonemes, diacritics can be a handy way of indicating tones or stress. I'm currently half-working on a project in which the circumflex indicates high tone.

In Nizhmel, an underdot is used to indicate a certain tone/pitch pattern in stressed syllables. <ẹ>, <ụ>, <ạ> etc. In an earlier version, mid-close vowels were indicated with a circumflex: <ê> <ô>. Together with the tone marker, it could generate glyphs like <ệ> and <ộ>.

User avatar
Posts: 5773
Joined: 20 Oct 2013 02:57
Location: Toma-ʾEzra lit Vṛḵaža

Re: What do you base your conlang romanizations off of?

Post by Ahzoh » 06 Jun 2015 20:24

I base my romanizations off of Semitic, since I find it most aesthetically pleasing...
Image Ӯсцьӣ (Onschen) [ CWS ]
Image Šat Wərxažu (Vrkhazhian) [ WIKI | CWS ]

User avatar
Lao Kou
Posts: 5665
Joined: 25 Nov 2012 10:39
Location: 蘇州/苏州

Re: What do you base your conlang romanizations off of?

Post by Lao Kou » 07 Jun 2015 05:33

Birdlang wrote:And do you use any diacritics? I use diacritics or digraphs. What are the special characters you use the most?
I've stated elsewhere, Géarthnuns romanization is rather Anglo in its approach, perhaps a victim of its time. Early Géarthnuns was handwritten and never committed to printed hardcopy because that would have had to be done on a US electric typewriter (Remember those? Typing French term papers on a US keyboard was a bit of a drag, because you had to go back in afterwards and add all the diacritics by hand (and the prof always caught the one you missed :mrred: )). What diacritics there were were informed by French and German.

Then along came personal computers, and I internalized the ALT codes for French and German (and to a lesser extent, Spanish) pretty early on. So when the time came to commit Géarthnuns to computer files, <ü>, <í>, <ö>, <é>, and <öi> crossed over seamlessly. The original handwritten romanization for [ɕ] (ugh, :roll:) changed over to <ç> either before or as a result of computerization, I'm not sure, but it was an ALT code I knew. I do remember actively searching for an appropriate diacritic "z" for [ʑ], a relative latecomer to the phonology, but -- maybe it was early Unicode days, I'm not that tech savvy -- it seemed to get one(s) I thought might be okay, you had to change fonts to Latin2 or Latin3, wink twice, and kiss your sister. Screw that! And so it became <zç>.

It's a romanization I can live with, and it gives Géarthnuns, like it or no, its romanized "look". There are a lot more diacritic bells and whistles now than there were then, but a romanization reshuffle seems much more trouble than it's worth. The only thing that really sticks out as a sore thumb for me (and Des picked up on it once) is <ou> for [aʊ]. But whaddaya gonna do?

User avatar
Posts: 4455
Joined: 20 Nov 2014 02:27

Re: What do you base your conlang romanizations off of?

Post by qwed117 » 07 Jun 2015 05:44

My (horrible awful) vowel romanization was based off of English
e /i/, a /a/, oo /u/, and i /ai/

I don't know where my consonant romanization came from.
My minicity is Zyphrazia and Novland
What is made of man will crumble away.

User avatar
Posts: 3512
Joined: 22 Nov 2014 04:48

Re: What do you base your conlang romanizations off of?

Post by elemtilas » 07 Jun 2015 06:11

Birdlang wrote:And do you use any diacritics? I use diacritics or digraphs and I base some romanizations off of a Birdlang Phonetic Alphabet (BAP). Also. What are the special characters you use the most?
Romanisations for the languages of The World are fairly straightforward continental / average European. Why completely reinvent the whole concept of locomotion when you can simply use a ready-made wheel?

I assume that the conlangs themselves will contain within them "odd sounds" and "nuances" that don't exist in English. I am not a phonology-phreak so don't really give a fiddlers fart whether the actual sound is [ʊ] or [ɪ̈] -- they sound close enough to me, and I'm just not going to lose any sleep over that. Now, when it comes to etymology or odd grammatical bits and pieces, I take much more interest -- I'm pretty pleased with Mentolation word chain-derivation, for example.

I do use diacritics: most heavily used are chapeau, acute and grave. I like to avoid diaeresis/umlaut (mostly cos I don't remember the alt-codes); and have only rarely used tildes, macrons and tipped chapeaux. Chapeaux mark varying degrees of long vowels: e = short, ê = long, êê = overlong. Acute and grave will typically denote primary and secondary stress or else rising / high and falling / low tone. Tilde will usually denote nasalisation. I use digraphs for diphthongs, of course, and have flirted with them for umlaut. Doubled consonants sometimes mean doubly articulated / long consonants: ttalle = [t.tal.le]; sometimes a different form of a sound: ssarq = [ɕar̥q].

I generally restrict special characters to thorn. Natively, the languages from The World have their own writing systems -- from cuneiform ideo-syllabaries to runes to tactile vine-script, so some kind of relatively easy-to-use-and-remember romanisation is important

User avatar
Posts: 850
Joined: 10 Oct 2013 00:56
Location: California

Re: What do you base your conlang romanizations off of?

Post by Sḿtuval » 07 Jun 2015 07:01

I hate digraphs so much. In my opinion they look messy, although too many diacritics can evoke that feeling for me as well.

I don't count doubling letters as digraphs if they're used to indicate length or geminates.

I use carons a lot in my langs. Usually it's š for /ʃ/, ž for /ʒ/, ť for /θ/, and ď for /ð/.

I sometimes use the acute for stress (à l'espagnol) but I also (rarely) use it to indicate vowel height (i.e. <e> for /ɛ/ but <é> for /e/) as well.

Whenever I use macrons they usually indicate vowel length.

Graves I often use for "reduced" vowels (like with <è> for the "schwa" sound and <ì> for /ɪ/).

One of my langs use the acute for irregular stress and grave for "reduced" vowels. When a "reduced" vowel would be marked with the acute I use the circumflex. So an irregularly stressed <è> becomes <ê>.

I also use the circumflex for vowel height (much more than the acute) in other languages and in one I use it instead of the grave for "reduced" sounds (i.e. <û> for /ʊ/).

In Kauzasian I use the dot diacritic (it's my first lang to use that diacritic AFAIK and is the only one that does so far, other than its daughterlangs) for retroflex sounds.

I rarely use apostrophes. A couple langs use it for the glottal stop and another uses it to indicate geminate consonants.

There was one (which already had a shitty orthography without the apostrophe) that kind of used it strangely. It wasn't anything like "zo'on'lity'u", where it's completely meaningless; instead it was used on articles and most prepositions, which were attached to the word they modified. It was like some sort of shortened clitic, which never had anything shortened in the first place. I guess it's a little like English's possessive suffix. I have no idea what's being shortened to return the apostrophe in <'s> (and there might be a good reason for it).

I use special characters as well but rarely. Kauzasian uses the Latin gamma, and a couple langs used thorn and edh.
I have absolutely no idea what I'm doing.

User avatar
Posts: 110
Joined: 08 Feb 2015 01:51

Re: What do you base your conlang romanizations off of?

Post by Darvince » 07 Jun 2015 08:20

So, I wrote the whole post out, but then it got deleted and didn't get saved, so I am now very demotivated to make a long post.

English's use of the apostrophe is usually indicating the beginning of the clitic since all of English's clitics are enclitics, with the exception of <n't>.

I usually use an orthographic system based on what sounds I can associate that symbol or digraph as, so even if it's inconvenient I may use <q χ> instead of the simpler <q qh> since I more easily can associate <χ> with /χ/ than with <qh> or <rh> (those are all uvular fricatives btw). Recently I have found myself bending more towards making it easier to type than easier to figure out instantly what sort of sounds the language has, though, but since I've made myself quite an excellent keyboard for typing letters with diacritics I don't find myself often needing to bend to that.

I do have a lot of random phoneme inventories hanging around that I will probably never use or develop at all, but when I romanize them I've also seen that I am leaning more towards developing something of a personal phonetic alphabet (much like what Birdlang said in their post), always using the same diacritics/digraphs for the same phonemes or phoneme sets.

Mainly I use diacritics however, since I like that they prevent ambiguity from arising, although this can be avoided if you're careful with what romanization you use. Like making sure you don't use both <xhw> and <hw> and <xh> in the same romanization.
CBB forum is of Polandball
My minicity

:usa: Fluent
:esp: C1
:con: Himalian, A1

User avatar
Posts: 3319
Joined: 12 Aug 2010 01:53
Location: North Yorkshire, UK

Re: What do you base your conlang romanizations off of?

Post by sangi39 » 07 Jun 2015 12:14

My romanisations are largely inspired by the likes of languages like Polish and Czech, but also on some romanisations for languages like Arabic, Berber and some of the North American languages. I also, sometimes, use Mandarin romanisation as a base, as well as taking into account in-world historical considerations.

I try, though, to keep my romanisations fairly consistent across languages, which is why you might see, for example, the use of say <ż> or /ð/ without either <z> or the overdot appearing elsewhere in the romanisation, and <> for /q/ in a language that doesn't have <q> appearing in the rest of the romanisation for any other sound (I tend to reserve <q> for /ʔ/).

I use digraphs, though, at least kind of. I'll write ejective and implosive consonants as C', with an apostrophe (one of the reasons I keep <q> for /ʔ/) and I'll write prenasalised plosives as NC, separating clusters of N+C with an interpunct <·>.

I'll tend to use more and more digraphs the simpler the syllable structure is, see Lesi Kirra vs. Proto-Skawlas as an example, where Lesi Kirra uses no diacritics at all while Proto-Skawlas makes much heavier use of diacritics, especially in relation to the vowels, containing no digraphs as all. Proto-Sirdic, on the other hand, uses just one diacritic.
Last edited by sangi39 on 07 Jun 2015 12:35, edited 1 time in total.
You can tell the same lie a thousand times,
But it never gets any more true,
So close your eyes once more and once more believe
That they all still believe in you.
Just one time.

Posts: 641
Joined: 14 Aug 2010 13:28

Re: What do you base your conlang romanizations off of?

Post by Nortaneous » 07 Jun 2015 12:34

Whatever is easiest to type, unless there are large consonant clusters (Kannow) or there's no good way to fit it into 26 characters (Gehui).

User avatar
Posts: 7783
Joined: 13 May 2012 02:57

Re: What do you base your conlang romanizations off of?

Post by Lambuzhao » 07 Jun 2015 13:20

Yeah, bud. I feel yer pain. Who buoy!

As for Rozwi, originally it was straight up Englishification for the "prickly" letters (i.e. pre-IPA digraphs such as sh, ch, zh, th, kh, uu, oo/ou, ee).
But then I looked back at Old Persian/Akkadian romanizations, and also Czech/Slovak.
Then it became carons and hats, plus a bar-d (ščžđx ûôê). In handwriting, though, I preferred to use macrons for long vowels, and Greek /θ χ/ for 'th' and 'kh'. With the advent of ASCII and computers, the dream became reality (as did /ʔ ʕ/ -w00T!) [:)]

As far as vowels went, dealing with typewriters made me early on say to myself: "Self, are gonna go back and write in all those umlauts and accents and tildes - FUDGE NO!" Vowels have been pretty straightford and without whatchamacallits in Rozwi. The only weird thing was that early on,
/ai/ and /ei/ were separate sounds, but both became [ej]. I still preserve the spelling difference in words, though it is largely just a fossilization. Another is /oo/ and /ou/. Originally two distinct sounds as well, both became [o:] in pronunciation. In this case, though, nowadays I use o-macron /ō/.

Sadraas, being a Germanic zonelang, has an Englishy/Germanish/Dutchish alphabet. I have said almost from the beginning that it suffers from excessive döëtïnchëm (Adams, Meaning of Liff), no doubt influenced by "Teutonesque" spellings such as Würm (fave song by YES), Blue Öyster Cult (Don't fear the reaper!), Hüsker Dü (a crazy old band from MN), Mötley Crüe, and who could forget Deathtöngue, Ren Höek or Spümcø?

Yauchuan, Rozwi's Far Westron sisterlang, follows more or less Japanese Romaji conventions, but also uses /ü ǔ/, and /x/ as in Pinyin.

Çedara, a Latin-lang, originally used the unadorned Roman alphabet. Over time, certain letters popped up, though.
/ç/ appears more and more for [ʧ]. /ñ/ is replacing the older digraph /nn/ for [ɲ].

User avatar
Posts: 1796
Joined: 11 Feb 2015 11:23

Re: What do you base your conlang romanizations off of?

Post by gestaltist » 07 Jun 2015 15:55

For me it’s usually „whatever is easier for me to type“. Alternatively I sometimes try to make the romanization somewhat „intuitive“ for a random reader (not that I have any).

Both of these rules make me favor digraphs over diacritics, although proto-Ardanian has the engma which is not convenient to write at all.

User avatar
Posts: 2044
Joined: 12 Aug 2010 16:42
Location: 가매

Re: What do you base your conlang romanizations off of?

Post by masako » 08 Jun 2015 02:12

What do you base your conlang romanizations off of?
I tried to keep it simple with Kala. I wanted people to be able to type it quickly and without much effort.
And do you use any diacritics?
No. See above.
What are the special characters you use the most?
See above.

for reference: http://i.imgur.com/r2iW1F8.png

Posts: 262
Joined: 30 Oct 2014 18:24

Re: What do you base your conlang romanizations off of?

Post by xroox » 08 Jun 2015 19:16

I use no digraphs or diacritics for Tnaap's phonemes, since there are only 12. Acute is used for unexpected stress.
For Vior I use the under dot from Sanskrit for retroflexes and acute over s, z, n for alveolapatals. Eng stands for the velar nasal.
I don't like diacritics a lot, they're usually difficult to type.
Other projects included t and d with strike for retroflexes. I use a lot ñ for a dorsal nasal, since my keyboard is in Spanish.

User avatar
Posts: 32
Joined: 31 Dec 2014 20:50
Location: Harrogate, North Yorkshire, UK

Re: What do you base your conlang romanizations off of?

Post by Ketumak » 08 Jun 2015 19:21

A variety of factors. Like several people above I look for easy typing, especially now as I do more work on a tablet with a restricted character set. I could add more of course, but it's a tablet and extra software reduces performance.

I think about IPA but also about common European usages. For example, I tend to favour< y> for /j/ as it's intuitive for speakers of widely-spoken languages like English, French, Spanish and Portuguese.

I also try to avoid digraphs where possible, but prefer them to diacritics. Again, I look for established usages and only use my own as a last resort. My conlang Õtari has /j tʃ dʒ ʃ/ which I represent as < y c j x>. This means I only need one digraph: < ny> for /ɲ/

I don't think there's a satisfactory Romanisation for /ɂ/. Apostrophe is well-used in natlangs, but I've used < q> before now, as it frees apostrophe for abbreviation.

I like fonts, so have been known to let the look influence the phonology in places.
Good: :fra: :esp: :por: | OK :ita:


User avatar
Posts: 2680
Joined: 27 Aug 2010 08:17
Location: nowhere [naʊhɪɚ]

Re: What do you base your conlang romanizations off of?

Post by Omzinesý » 08 Jun 2015 21:48

I try that the letter is used similarly in some natlang.
<z> can be used to mark /z/ (Polish) or /ts/ (German)
In Vtayn, I had to combine the two and use <z> for /dz/ (In Italian /dz/ and /ts/ are allophones both marked with <z>). In Vtayn, /dz/ contrasts with /ts/ written with <c>.

Sometimes, I try to make the othography have a favor of a natlang.
In a Saami-based conlang, I have already forgotten, I just couldn't use <y> for /j/ although the language had /ʒ/ as well and <y> was free.

I normally try to avoid diacritics, but I often have some. /ŋ/ is especially hard to mark without diacritics or special marks.

If I use digraphs, I try to make them somehow mirror the history of the conlang, not English or other natlangs.
In Suyra /ʃ/ is marked by <sc> but that is a track of a sound change /st͡ʃ /-> /ʃ/, which is still somehow visible in the morphology of the language.

Posts: 256
Joined: 01 Sep 2010 15:31
Location: UK

Re: What do you base your conlang romanizations off of?

Post by Curlyjimsam » 09 Jun 2015 11:28

I don't consciously base my system off anything in particular, and things vary quite a bit. I'll most usually use carons for postalveolars (e.g. š ž č), but sometimes I'll use digraphs with h instead (sh zh ch), and sometimes other things entirely. Long vowels I'll generally use either an acute or a macron - I also really like the circumflex in theory, though for some reason don't use it all that often in practice. If the native script represents a long vowel by doubling, I'll do the same in the romanisation. Front rounded vowels are pretty much always written with an umlaut (ö ü). /k/ is pretty much always k, never c.

These are probably the main sounds which repeatedly cause problems for me; lots of languages have sounds that it's also not so straightforward to make a decision for, but I tend to resolve these on a rather more ad hoc basis.
Twitter: @jsbaker750
Website: seven-fifty.net

Posts: 2387
Joined: 24 Jun 2012 07:33
Location: Tokyo

Re: What do you base your conlang romanizations off of?

Post by clawgrip » 10 Jun 2015 15:03

My Romanizations are generally just based on English. I have no conlangs that are natively written in the Roman alphabet, and since English is my native language, and all the places my languages are presented, such as here, the main language is English, it seems appropriate enough, since it makes it easier for people to take a guess at the pronunciation even if they don't know the language's phonology. Himmaswa's Romanization as a bunch of digraphs and trigraphs, and only one diacritic (ñ) because my earliest dictionary was in the DOS text editor, and typing alt-codes all the time was annoying. Even now I have a habit of avoiding diacritics if there's an obvious digraph, though for less Englishy-sounds I will probably use diacritics.

Posts: 220
Joined: 09 Oct 2012 03:37

Re: What do you base your conlang romanizations off of?

Post by wintiver » 10 Jun 2015 18:57

I find the Americanist Phonetic Notation pretty aesthetically pleasing. Many of my languages have overlap with the sounds of the Pacific Northwest, so transcribing with <ƛ> instead of <tl> is usually preferred if I'm looking for succinctness in such sounds.

Generally though, I tend towards maximalist phonologies which is troublesome, and forces my hand to use any method I can for a (fairly) consistent transcription.

Digraphs are fine in my opinion if they are warranted. But I have a strong aversion to trigraphs.

Posts: 578
Joined: 28 Nov 2013 14:47

Re: What do you base your conlang romanizations off of?

Post by Squall » 10 Jun 2015 20:49

Romanization actually is not always good, because the reader would have to learn the spelling rules to know the pronunciation. It is not different from learning easy alphabets, such as Greek and Cyrillic. If there are diacritics, digraphs or the language has more than five vowels, the accuracy of the guessed pronunciation will be bad. C has a different pronunciation in each language. H may be silent.

It is fun to create romanization rules based on the phonology. It is better when the conlang is based on a natlang and we use a irregular or complex set of spelling rules (as long as it is phonetic).

I use romanization because I can type it easily. I usually am reluctant to use diacritics because of the trauma caused by old ASCII computers. So many people ignore diacritics.

I use <x j> for /ʃ ʒ/ instead of <ŝ ẑ> or <sh zh>. In another language, I used <c> for /x/ and <ç> for /ɣ/.
Affricates are represented with digraphs, so <dj> is /dʒ/.
/ŋ/ may be represented with <ñ> or <q>. <q> is a useless letter.

Vowels are a problem because I usually have more than five vowels.
Symmetric ways to distinguish /æ/ and /ɒ/ forbid the use of <a> alone to represent one of them.

I usually have to distinguish /ɛ ɔ/ and /e o/. If I use a diacritic to mark the stress, I will have problems to distinguish vowels. If I used vowel digraphs, it would be bad to put the stress diacritic on the digraph. Moreover, it is difficult to put multiple diacritics on the same letter, but there is a cheat that I can use: e /e/; é stressed /e/; è /ɛ/; è stressed /ɛ/.
If I used the combination <eq> to distinguish them, it would be aesthetically ugly even though <q> is useless.

I may represent /j w/ with <i u> or <y w> and use <ï ü> to distinguish then when needed.

I use apostrophes when I want to break digraphs into individual letters.

I put <n> after a vowel to nasalize it. Most of my languages lack pauses and nasal glides. So single n is part of the next syllable, double n nasalizes the previous vowel; and apostrophe cancels the interaction between n and the vowel.
anna /'ãna/, manmáin /mã'mãj/, man'in'a /mã'ĩ.a/, man'ián /mã'jã/
qwed117 wrote:My (horrible awful) vowel romanization was based off of English
e /i/, a /a/, oo /u/, and i /ai/
That is very inaccurate because English is not phonetic. [o.O]
clawgrip wrote:and since English is my native language
I really thought that you were Japanese.
Last edited by Squall on 10 Jun 2015 21:51, edited 1 time in total.
English is not my native language. Sorry for any mistakes or lack of knowledge when I discuss this language.
:bra: :mrgreen: | :uk: [:D] | :esp: [:)] | :epo: [:|] | :lat: [:S] | :jpn: [:'(]

Post Reply