6

I'm curious about how writing systems like Burmese or Thai - the characters of which look to my untrained eye far more similar than Latin or Japanese characters - are distinguished by native readers, and non-native readers.

Have there been any studies done into whether the characters in certain writing systems are more or less visually distinct?

By this I mean, many of the glyphs are easy to differentiate from one another, as opposed to b, p, q and d, for example, or <シ> and <ツ> in Japanese.

I'd be interested to see whether anyone has come up with a way of comparing different writing systems according to their distinctiveness.

Adam Bittlingmayer
  • 7,664
  • 25
  • 40
Lou
  • 1,280
  • 11
  • 22
  • I have heard of (but not read) a study along these lines that seemed to show that the Devanagari abugida used for (e.g) Hindi is the best-designed (close to optimum) in terms of recognizability. But I can't give the reference, sorry. – jlawler Jun 12 '18 at 23:49
  • 1
    I'm sure there is research because distinctiveness is very important for those with dyslexia. – curiousdannii Jun 13 '18 at 00:48
  • @a-m-bittlingmayer I have seen the claim made for Armenian (and my time at the Matenadaran confirmed it to me), but you would be able to speak to it more authoritatively than me. – Nick Nicholas Jun 13 '18 at 23:59
  • 2
    Tangentially I believe Sampson's Writing Systems: A Linguistic Introduction has a section on this topic as well but I have the book at home. Maybe for now have a look at https://linguistics.stackexchange.com/a/27453/1668 – tripleee Jun 14 '18 at 09:06
  • 1
    @NickNicholas (@ did not work by the way.) I have heard that claim but I am not sure I believe it. It is possible that to someone who was never exposed to Latin (or Cyrillic or Greek which are close to it), it looks like soup too, but in my world there are very few such people. The other factor here is that the stylised versions of the same alphabet vary wildly. The Georgian alphabet(s) are an example of this, they are even considered separate alphabets, but really they were just versions used similarly to italics and casing. – Adam Bittlingmayer Jun 14 '18 at 12:38
  • 2
    @NickNicholas I do know that developers frequently increase the font size for Arabic and Armenian, but that is arguably just because the font designers made them too small. – Adam Bittlingmayer Jun 14 '18 at 12:38
  • 1
    So I am skeptical, anybody making a strong claim must account for those variables: size on the web, stylised versions and level of exposure. – Adam Bittlingmayer Jun 14 '18 at 12:40
  • 1
    I think effects would be greatly "dulled" by adaptability in any case; our brains so quickly specialize to identify distinguishing details from any set of data once importance is attached to them. (Also compare distinguishing individuals among humans vs. other species!) The effect would be likely be strongest at the edges, e.g. dyslexia as @curiousdannii mentioned or even viewing in small fonts or at a distance. – Luke Sawczak Jun 14 '18 at 14:10
  • @jlawler It sucks to be unable to verify the methodology or even existence of such study, because that seems just wrong to me. I mean, प म य . “Recognizability” seems quite subjective to me, so unless they used well-thought mathematical models in order to model and measure visual distinction, I'm not buying it. Devanagari, like other Brahmi-derived scripts, is rather complex, with many subtleties such as compounds and conjuncts that make things quite difficult for non-native readers. – Locoluis Jun 14 '18 at 21:25
  • 1
    A hypothesis: all writing systems with the course of time become equally efficient in terms of paper space per meaning. If some artifact (e.g. a stroke in a glyph) is not crucial for comprehension, it becomes abandoned and unused. OTOH, if two glyphs are visually similar, the developers of the fonts deliberately exaggerate the differences to make it distinguishable. Also: Doug Cooper, "How do Thais Tell Letters Apart?" (PDF). Also, Deciphering a handwritten script. – Be Brave Be Like Ukraine Jun 14 '18 at 22:23
  • @bytebuster Yes, you formulated my thoughts. I would also look at the evolution of characters that were actually the same, like j from i, or í from i, or â from a. They were added to make some subtle distinction, but they are often substituted intentionally or unintentionally, because the distinction is not always critical to understanding. Importantly, the substitution is non-random. – Adam Bittlingmayer Jun 15 '18 at 06:55
  • I think there is Eskayan script with some first consonants which have similar structures – WiccanKarnak Jun 20 '18 at 11:27

1 Answers1

1

The most confusing (lacking distinction among letters) writing system is the the oldest variety of the Arabic script, usually known as Rasm, 7th - 11th centuries AD.
The modern Arabic script has 28 letter of which only 18 letters have their distinct shapes, the rest are distinguished by the diacritic dots, 1 to 3 of them either below (1-2 of them) or above (1-3) the letters.
In the Rasm tradition, the letters were written without those dots, so, for example,
با [baː],
نا [naː],
يا [jaː],
تا [taː],
ثا [θaː],
were written the same way, that is, without the dots, like this:

enter image description here

Despite its ambiguity, the script was pretty legible, especially for the learned Muslim scholars.

Yellow Sky
  • 18,268
  • 39
  • 65
  • 2
    (Surely you mean [j], not [y]?) – Draconis Jun 21 '18 at 16:37
  • An ambiguous individual glyph can be disambiguated in context, but languages also differ in how much context they supply. I know Hebrew matres lectionis were introduced as a countermeasure when the historically (ostensibly) sufficient context no longer worked for new generations, but I can't even begin to articulate a proper model for how to assess, much less quantify this phenomenon. – tripleee Jun 25 '18 at 06:47