Rishon Rishon

Socially optimal orthography in Hebrew

Amritas welcomes me back to the blogosphere with a question:

Would that description of a "socially optimal orthography" apply to Hebrew writing?

He is referring to this description, by Joel at Far Outliers:

I would suggest that a socially optimal orthography might get by with even fewer alphabetic distinctions. People could write fewer vowels and consonants than would be optimal in isolation, while relying instead on sentential, semantic, or social context to reduce ambiguity. But this approach would make linguists feel rather less useful.

Well this is a pretty interesting way to get back to posting! (I wanted to write something about the insights coming from my most recent culture clash in the US. But that can wait.) In order to answer the question I had to figure out what a “socially optimal orthography” is. It took me quite a long time, starting with Joel’s post, and following some of his links. Eventually I decided on the following: a socially optimal orthography is:

1. Easy to learn
2. Easy to read
3. Easy to write

The last two points can be grouped together as: easy to use. Now, any software engineer with some experience, who as written software for human beings (as opposed to software that interacts only with machines) should know that ease of learning and ease of use are two completely different things. But in the end, ease of learning usually beats ease of use. That’s why mouse-driven editors eventually replaced keyboard-driven editors even though nothing beats a good keyboard editor (like VI) for ease of use – once you’ve learned it. In general I would put it this way: if the feature is critical to a professional, go for ease of use; but if it is either not critical or not for a professional, go for ease of learning.

So, the first question is: are written language users professionals? We certainly use reading and writing a lot. On the other hand, it could be argued that learning to read and write is more important than ease of use – if we don’t do that, we can’t do anything. Joel makes his comment with respect to Marshallese, in which consonants come in three forms: palatalized (raising the front of the tongue), velarized (raising the back of the tongue), and rounded (rounding the lips), with consequent fronting, backing, or rounding of the neighboring vowels. The question he raises is: should the orthography be phonemic, containing three forms of every consonant and four vowels, or should the vowels be written phonetically (indicating absolute sound) not phonemically (indicating language-meaningful units of sound) with one form of every consonant?

This question is doubly interesting to me because it parallels an observation I’ve made myself with respect to Hebrew and Russian. Hebrew, like Marshallese, has phonemic velarization, and Russian, also like Marshallese, has phonemic palatization. But Hebrew (like Arabic) indicates velarization in the consonants, while Russian indicates palatization in the vowels. In both cases the solutions are appropriate, in my opinion. Hebrew and Arabic, of course don’t fully indicate vowels, so indicating velarization in the vowels is inapplicable. However, imagining that they did, it would obscure the root-and-pattern morphology of the language to do so, making it both harder to learn and harder to use. Russian, however, doesn’t have this type of morphology, but it has 21 consonants and 5 vowels (before indicating palatization) – indicating palatization in the vowels adds only 5 new letters, while doing so in the consonants would add 21. (Actually it would be somewhat less, not all consonants can be palatalized.)

Getting back to Amritas’s question: Is Hebrew orthography socially optimal? Fifty years ago there were all kinds of proposals for orthographic reform, including latinization, none of which were taken seriously by the public. The simple fact was that everybody already knew the traditional system, so any reform violated point #1. In any case, I would certainly say that point #1 is fully covered by the traditional system: fully pointed Hebrew is very easy to learn. I know many people who have tried to teach their children to read simultaneously in Hebrew and English, and Hebrew wins hands down. Though Modern Hebrew has merged the pronunciation of many consonants and vowels, it remains the case that a fully pointed Hebrew word can be pronounced only one way.

Regarding reading and writing, I have more difficulty answering. Ease of writing is clearly impacted by technology: the answer is different if you are writing by hand, using a typewriter, or using a computer. It so happens that Modern Hebrew writing became standardized during the typewriter age, that is: after the typewriter became popular, but before the computer was invented. Typewriter technology strongly favors discreet characters written in a single line, making it difficult, if not impossible, to type pointed Hebrew. (It was also inappropriate for written Chinese, Japanese, and Arabic. Probably Devanagari too.) Thus, the standard written form of the language uses no pointing whatsoever. I have wondered if Hebrew writing were standardized today, if at least some points would be written – computers have no trouble with pointing.

In any case, writing is no problem. Hebrew has 22 letters, 5 of them have final forms, and is no trouble to write or type. The big question is reading unpointed Hebrew. This is much less problematic than you might think, as I have discussed before:

The real reason that the system works is that only two things are needed to uniquely identify a Hebrew word: root and pattern. You can always identify a word’s root because roots consist of consonants alone. And it turns out that the degree to which vowels are indicated is enough to disambiguate almost all patterns – the little ambiguity that remains can easily be determined by context, much the way English speakers [not readers] disambiguate the words to, too, and two.

However, “optimal” is a pretty high standard, and I’m not so sure about that – but I do know it works.

UPDATE: Joel just posted again on socially optimal orthography – this time regarding Yapese.

Posted by David Boxenhorn at August 30, 2004 02:35 AM