A Braille Orthography for Na’vi

Started by Zángtsuva, December 19, 2021, 08:45:09 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.


As someone who is interested both in the Na'vi language and in Braille systems in general, I've given some thought to how one might go about writing Na'vi in Braille. Note that I approach this as a writing systems nerd who has done research on the subject and not as someone who has had significant contact with blind people or anyone else who frequently works with Braille, so I welcome criticism or other feedback from anyone who has deeper familiarity with the things I'm discussing. I'm especially looking for feedback on how to present a table illustrating the correspondence between Braille and Latin characters in a way that isn't awkward for a screen reader.

With visual writing systems, even if they are alphabetical and people start out reading one letter at a time, in the hyperliterate societies of the modern world it is normal for readers to recognize the shapes of entire words at a time and thereby read much more efficiently than the appearance of an alphabetical system would suggest. With tactile writing systems, this kind of efficiency is limited by how fast a person can read information with the fingers. From what I've read on the subject, it would seem that in practice the human brain reads most efficiently when using two fingers at a time, taking the information in units of six or at most eight bits (Braille dots). Apparently anything beyond this strains the human's ability to reliably distinguish between tactile dot patterns.

Therefore it is particularly desirable for a Braille writing system to present information with as little redundancy as possible. A system that makes full use of the amount of information that can be packed into a Braille cell, in exchange for some additional study and practice during the learning phase, prevents reading from remaining an unnecessarily time-consuming and tedious task for one's entire life. Thus a system such as German Braille, which has numerous special characters for common letter combinations like GE and SCH, is much more user-friendly than for example French Braille, which for cultural reasons insists on using a one-to-one mapping to recreate for the blind reader the authentic experience of standard French orthography. (Note that the inventor of this kind of tactile writing system, Louis Braille, who was himself a blind French speaker, actually used a more phonetic approach before the modern standard took hold.)

For the Na'vi language I tried designing a syllabic system. I don't remember the exact number of possible syllables allowed by Na'vi phonology (including the distinction between stressed and unstressed syllables) but I found it to be on the order of 216. So I designed a system where each syllable was represented by 16 Braille dots, which in practice meant pairs of 8-dot Braille cells. Aside from the problem that one would have to keep track of which Braille cells were to be read as the first half of the syllable and which as the second half (not to mention the use of 8-dot Braille cells which is already a higher level of complexity than most Braille-reading fingers are trained to work with), it turned out that this system is not really more efficient than an alphabet because, even though Na'vi syllables can be as complex as skxawng, most of the time they're much simpler than that, just V or CV (which is no more than two characters even in an alphabetical system). So in effect the syllabic system accepts considerable disadvantages merely for the sake of marking where consonants could be but rarely actually are. In other words, the great disparity between maximum syllable complexity and average syllable complexity that contributes to Na'vi's unusual (alien) sound also means that, at least in the context of Braille, it is best written using an alphabet.

Now, ideally an alphabet would not use completely arbitrary symbols but rather follow some kind of internal logic that reflects the phonetic similarities or phonological relations of the sounds to be transcribed. So for example in Na'vi ideally the shapes of the letters representing sounds that can result from lenition would be related to the those of their unlenited counterparts in a systematic way. If I were imagining that a Na'vi Braille alphabet had been invented by the Na'vi themselves then this is certainly how I would go about it. However, considering that in the Avatar universe it would make more sense for Braille to have been taught to the Na'vi by humans, and, more importantly, for the sake of any blind people in the real world who are interested in studying Na'vi, I think in practice it makes more sense to use a Braille mapping that is in line with the international standard that forms the basis for French Braille, English Braille, German Braille, and most other Braille systems in use around the world today.

Thus the basic system should look something like this:

⠀    ⠁A   ⠃   ⠉TS   ⠙   ⠑E   ⠋F   ⠛NG   ⠓H   ⠊I   ⠚   ⠈ˊ   ⠘
⠄'   ⠅K   ⠇L   ⠍M   ⠝N   ⠕O   ⠏P   ⠟    ⠗R   ⠎S   ⠞T   ⠌Ì   ⠜Ä
⠤   ⠥U   ⠧V   ⠭KX   ⠽Y   ⠵Z   ⠯PX   ⠿LL   ⠷RR   ⠮    ⠾TX   ⠬    ⠼
⠠   ⠡AW   ⠣    ⠩AY   ⠹EY   ⠱EW   ⠫    ⠻    ⠳    ⠪    ⠺W   ⠨    ⠸
   ⠂    ⠆   ⠒   ⠲   ⠢    ⠖    ⠶    ⠦    ⠔    ⠴    ⠐   ⠰

Note that Ì and Ä, in a lovely coincidence, are already next to each other in the established system of French Braille as combinations of letter and diacritic that are not used in French proper but are potentially useful for foreign words. The accent character ⠈, which is used for example in English Braille as a generic representation for diacritics in loan words like café and piñata, is useful in Na'vi for marking stress. I chose ⠄to represent the tìftang because it not only represents the apostrophe in for example English Braille but is even used for the phonemic glottal stop in for example Arabic Braille. The choices for ts and ng are based on Frommer's own old transcription for these sounds as c and g. To generate the ejectives and pseudovowels I used the pattern of adding the bottom right dot to the corresponding "basic" letter. I chose the characters ⠡and ⠩for aw and ay because they are used for the same sounds in German Braille, and since ⠁and ⠑are a and e it is then elegantly logical to use ⠱and ⠹for ew and ey.

Now, this system is entirely adequate for transcribing the Na'vi language, and for the beginner I would not recommend anything more complex than this, but for anyone who has reached a higher level of proficiency, it would be a shame not to follow the example of German Braille and English Braille in making use of the remaining dot patterns to enable more efficient reading. Deciding the details of this is more arbitrary than laying down the basis, but one can get fairly far by looking for similarities between characters and thinking about what combinations of sounds are frequent in the language. I found it most intuitive to focus on affixes and especially infixes.

⠀    ⠁A   ⠃PE   ⠉TS   ⠙TSA   ⠑E   ⠋F   ⠛NG   ⠓H   ⠊I   ⠚ÌY   ⠈ˊ   ⠘
⠄'   ⠅K   ⠇L   ⠍M   ⠝N   ⠕O   ⠏P   ⠟FRA   ⠗R   ⠎S   ⠞T   ⠌Ì   ⠜Ä
⠤   ⠥U   ⠧V   ⠭KX   ⠽Y   ⠵Z   ⠯PX   ⠿LL   ⠷RR   ⠮SY   ⠾TX   ⠬US   ⠼IV
⠠   ⠡AW   ⠣SÄ   ⠩AY   ⠹EY   ⠱EW   ⠫FNE   ⠻ER   ⠳TÌ   ⠪OL   ⠺W   ⠨ÌM   ⠸LE
   ⠂AM   ⠆KE   ⠒ATS   ⠲NÌ   ⠢AWN   ⠖FÌ   ⠶ÄNG   ⠦TU   ⠔EI   ⠴YU   ⠐   ⠰

Some of these are inspired by existing Braille systems. For example using ⠻ for er is borrowed directly from German and English Braille, and using ⠬ for us is a sort of grammar joke as the same character is used for ING in English Braille. (The fact that it's also similar to both U and S just goes to show how rich such connections often are in Braille—it's not just boring dots ;) )
While one might be inclined to reserve these "affix characters" for transcribing actual affixes, the example of English Braille and German Braille suggests that it is no disadvantage to use such characters wherever the corresponding letter sequences occur regardless of etymology. Similarly, while the purist might insist that for example ⠩ should be used exclusively for the diphthong ay and not for a sequence of a and y across syllables, until the day when Frommer declares a minimal pair between these things, I see no problem with letting the Braille system conflate the two.
If you're wondering why I didn't use all of the possible Braille dot combinations in this more efficient mapping, that is because the remaining ones are traditionally reserved for punctuation, which however I do not find interesting enough to lay out in detail.


That looks really interesting! Great work :)