In mid-2016, I got to thinking about about the evolution of written languages (I think I was reading The Selfish Gene at the time). Human writing developed from pictographs and glyphs to today’s mostly phonetic systems (where each letter or combination represents a sound). The trend among websites to use symbols in place of words — in order to more easily reach multilingual audiences, certain icons are becoming universally recognized to mean “menu”, “settings”, “share”, “like”, etc. — seems like a backward trend to me.
I began to think about alphabets, and how we simply have to memorize the letters’ relationships with sounds and words. The shape of the letter “A”, for example, doesn’t inherently evoke the sounds that A makes in our language. We just have to be taught that “A” makes any number of sounds, depending on usage — pan, pane, and pauper illustrate three different sounds “A” can make. The same is true for all the letters in all languages (with the possible exception of “O”).
If the shapes themselves have no bearing on the sounds they represent, then the shapes are immaterial. They are also rather random — look at them and compare them to shapes in other alphabets that make the same or similar sounds. Cyrillic, Arabic, Hebrew, Korean (hangul), Latin (ours), and others are different enough to illustrate my point. Any person or group who develops any phonetic alphabet can use any kind or number of shapes if building an alphabet from scratch.
It made me wonder:
What if scientists and/or linguists attempted to design an alphabet from scratch today? What form would it take? How would they begin?
If I were called upon to do it, I would set a couple of goals for my alphabet: (1) it should be as simple as possible to teach and learn, and (2) it should be as difficult as possible to mistake one character for another.
These goals are almost self-evident, since the entire point of an alphabet is communication. But think of all the times we’ve mistaken a one (“1”) for a lowercase L (“l”) or an uppercase i (“I”), which look identical in many fonts and similar in others. Also the letter O and the number 0. The lowercase n and lowercase h are only slightly different. Z and a 2 can be confused when handwritten, as can 5 and S.
Would it make sense to begin with basic geometric shapes and branch off from there? Examples would be dots, circles, lines, triangles, squares…
I easily came up with 22 symbols. The Rotakas language only has 12 letters and modern Hawaiian has only 13. But this alphabet of shapes and symbols doesn’t represent all the possible sounds that a language can use. English alone has 44 phonemes (distinguishable units of sound).
Some languages — Mandarin, for example — use distinguishing marks to indicate pitch. The word “ma” in Mandarin can have five distinct meanings, depending on pitch: “mom”, “hemp”, “horse”, “scold”, and an interrogative particle. Even in English, we use stress to change meanings of words, though it’s not reflected in our alphabet (inVITE is a verb; INvite is a noun). “Envelope” is pronounced entirely differently depending on whether it’s a noun or a verb. Many languages use their alphabet markings to reflect these sounds. Others include many clicks, pops, glottal stops, and other sounds possible with human speaking apparatus.
Should our new alphabet have markings to indicate all these? It should, if it’s going to be a universal one. Using basic geometric shapes, we begin to run out of possibilities. A rectangle can easily be confused with a square, if handwritten. You might write a small circle and a reader will think it’s a dot. A two-line symbol could be mistaken for repetition of the one-line symbol.
My next thought was to take a single geometric shape, such as a circle or square, and then decorate it differently for separate sounds. For example, start with a circle. There are many possibilities: filled circle, empty circle, circle within a circle, circle with a dot, bisected circle (horizontal, diagonal, vertical), circle with an X or + in it, half-filled circle (vertical or horizontal), half a circle, circle with a line on the top or bottom or side, intersecting circles, etc.
Though I couldn’t find code for all the examples I wanted, I again came up with 23 or more possibilities, just in a few seconds. It would be similar with a square (and, I think, squares are easier to draw well than circles). Perhaps combining both circles and squares would give us the 44+ sounds we want to represent.
Then I remembered Braille. It’s just two columns of three rows of raised dots (six dots), any of which can be used in combination with any other, for 64 possibilities in each cell. Later, another row was added, giving us eight dots, or 256 possibilities. It’s enough to represent every possible human phoneme, as well as punctuation marks and other symbols and shortcuts.
I looked at the digital alarm clock next to my bed and saw that each character has seven bars (four vertical and three horizontal) that can be lit or dark:
I wondered how many possibilities were available with just those seven bars.
Thinking about this brought me to my idea of a 3×3 grid (like a tic-tac-toe board but with borders). There are nine squares, any of which can be empty or filled in combination with any or all of the others. How many possibilities does this give us?
I know 29 is 512 (each of the nine squares has two possibilities: filled or empty). Or you can do the math the difficult way, as I first attempted:
All filled or all empty is two choices.
All filled except one gives us nine more options.
All empty except one gives us nine more choices (we’re at 20 now, if you’re keeping score).
All filled except two adds 36 options (56 total).
All empty except two adds 36 options (92 total).
I quit trying right there, realizing it was going to get difficult and time consuming. The point I quickly realized is that a 3×3 grid is easier to draw than the 2×8 expanded Braille grid, and gives you one more square. It could be drawn either as a grid of squares (my original idea) or a grid of dots — like Braille — and still mean the same thing. See the following two examples, which both utilize spots 1, 5, 6, and 8.
What was the point of all this thinking, you might ask?
Nothing at all. I was just lying in bed, hoping to fall asleep. After this had run through my mind for 10 minutes or so, I scribbled a note in a notebook about it, dating it 2016.07.13, and then fell asleep. But I record the idea here, in case anyone has thought along similar lines.