Day sixty-four -- About heavily inflecting or agglutinating languages
[quote="Khagan"] Question 1: Do you have any real test cases for PreBabel dealing with heavily inflecting or agglutinating languages?
Question 2: Do you understand that the words "duzen/Siezen" "tegez?/mag??" (using the familiar form of you/using the formal form of you, in German and Hungarian respectively) are semantically unique and independent words--not ciphers or variations of "you" (whether the "you" of English or Chinese)?
Question 3: Do you understand that in an agglutinating language, even a huge word like "megszents?telen?hetetlens?esked?" can be a proper, unconjugated and undeclined word of its own right, with a rightly separate place in the dictionary/semantic mindspace from its ultimate root word ("szent" for the preceding)? [/quote]
[quote="Trailsend"] So then, when ciphering languages besides Chinese, why do you take all the various semantic usages of a word and pick only a "core case" to use for its encoding? (Even on its surface, that's a problem, since in many situations selection of a "core case" will be necessarily arbitrary.) Would it not be more consistent to do as you have done with Chinese, and assign different encodings for the various semantic meanings?[/quote]
[quote="sangi39"] Then there's always the question of language like Arabic and Hebrew which indicate grammatical information by means of internal vowel changes, ...[/quote]
If allow me, I would like to summarize the above into three issues.
1. A word can be like a thousand hand Budhisattva with each hand having its own magic. That is, a word has a huge semantic space.
2. Different words are sitting on the same toilet doing the same job at the exact same time.
3. A word can be absolutely unique, and no member in the Vx set can encode it.
In the Vx set, the syx can be a word, a phrase or a sentence. So, whether the "megszents?telen?hetetlens?esked?" is a word, a phrase or a sentence, it is simply a syx.
The PB procedure (a. regressive encoding, b. progressive ciphering, c. mixing pot settling, etc.) is a mapping or a series of mappings. Furthermore, this PB mapping is a "masking" function, that is, it will not change the essence and the features of the items which are masked although their "appearances" could be altered. I truly do not see a problem from the case 3.
Furthermore, in math, a few members of a domain of a mapping could cause the mapping to go berserk, such as,
F = 1 / (1-x), x is a real number.
When, x = 1, F goes haywire. Yet, there are two ways to deal with this problem.
a. remove x = 1 from the domain
b. give F (x = 1) = b, a special definition.
In the above case, x = 1 is called a singularity. In calculus, the singularity is cut out from the Riemann intergradation. That is, even a singularity can be deal with.
As long as the "megs...sked?" is a syx, a member of the Vx, we can always deal with it.
i. cut it out from the PB mapping
ii. give a special definition in the PB mapping
Furthermore, I do not think that "megs...sked?" is a singularity in that language. Thus, regardless of whether it is divisible or not and however unique a syx it is, it is pointing to (...), and this (...) can always be described by other members of the Vx. Then, we can reduce that description to two syx (syx1, syx2), and this becomes its cipher which will act exactly as it was (whatever the magic it was playing).
Note: this new cipher must be learned. Anyone who did not learn it will never know the new item. Of course, an old, old, old issue arises again. Will PreBabel becomes easier after all these? I think that if this becomes a wide spread problem in the language x, it might truly cause some problems. But, I don't think that this kind of unique situation takes up 10% of any given language. It will be a success if a language x can be 90% PBlized.
When the worse comes to the worst, we can simply un-PBlized it (cut it out). Over 1,100 Chinese characters in Japanese are not Japanese-lized. If one or two languages are truly outside of the "scope" of PreBabel, I will settle for excluding them (cut them out) in PreBabel. Even if we could only PreBabelize two languages (Chinese and English), I will still be happy about it at this early stage on the development of the PreBabel. We should not smash the toilet because that one piece of ... is too big to be flashed down.
The issue 1 and 2 are seemingly different but are truly inter-connected, one with thousand hands and thousand hands in the same pot.
For PreBabel (language x), it is an 100% internal operation in the language x without any concern about any other languages. Every word in language x will be ciphered. That is, the thousand hands in the same pot will have 1000 ciphers. Yet, one word with thousand hands will be ciphered with only one cipher which can indeed be arbitrary (whether core or not) selected. However, if this is not adequate or not convenient, then we can split it into two or three. When the worse comes to the worst, we can, of course, cipher all those 1000 of them. There is no reason too stingy about it. Of course, I do not believe that we will ever come to that point.
The "splitting" operation to accommodate some different semantic needs is done in many languages. In English,
1. (I, me)
2. (he, him)
3. (she, her)
4. (they, them)
The two words in ( ) above are identical words with a splitting operation to encompass different semantic spaces. The following pairs are words which did not split.
5. (you, you)
6. (it, it)
So, if PB splits a few more words during the PreBabelization process, it will not be a major crime. [quote="answer Khagan from Tienzen"] If your concern is about "the same word in different languages," then it is the same question as Trailsend's -- the PB (proper) will fail to drop out of the mixing pot. I will discuss this issue later. [/quote]
The same word in different languages in the PB (proper) has the same situation as case 2 (different words are sitting on the same toilet doing the same job at the exact same time) in the language x. As I answered the second part in my previous post, the first part is also answered although I did not point it out explicitly at the time.
Signature -- PreBabel is the true universal language, it is available at