Day fourteen --
Question -- from "Ave94" -- Even if you could get billions of people to speak the same language, it wouldn't last. Soon it would diverge into separate dialects, then mutually unintelligible languages. After all the work teaching everyone the language, you'd have to start all over again.
Answer -- This is a very important point. Yet, it was posted so early in this thread, and I was unable to answer it without laying out my position first. Now, I can answer it with an "Attractor theorem".
"Attractor theorem" -- if set R is an attractor (unifying set or a converging set) of some arbitrary set A, B, C,...; set R is also the attractor of all its (set R) descendant sets, set S, T, U, ...
In the case of our discussion, the Set R is the PreBabel set, and A, B, C ... are all natural languages. Set S, T, U,... are all dialects of the PreBabel, derived from a diverging force which is described in your post. Yet, the PB set will also be the converging point of all those dialects.
There are many examples in nature following this Attractor Theorem. For example, ocean is the attractor for all water. Hurricane is a diverging force for ocean, which sucks up billions tons of water away from ocean each time, and it moves those water to a very far away mountain. Yet, most of the water will eventually go back to the ocean. In short, if R is an attractor for a system L, there is no diverging force in L which can create a diverged set in L not converging to R.
Of course, if the PreBabel is not an attractor in L, then the Attractor Theorem does not apply. If the PreBabel is an attractor in L, then your question is answered.
Some of you are not happy about the way that I encoded some words, such as, laser. Of course, anyone can always do a better job. Yet, the key point is that whether the "Regression Encoding Procedure (REP)" works or not, regardless of whether a particular word was encoded good or poorly.
Before an actual PB set on hand, whether a system L (all natural languages) has an attractor (universal language) or not can only be discussed on a theoretical level, by proving many mathematic theorems, such as,
1. Isomorphic theorem -- all natural languages are isomorphic to one another.
2. Existence theorem -- there is, at least, one Set R is the root word set which is able to encode one natural language. Set R is called a PreBabel set or PB set in short.
3. Universal theorem -- if Set R can encode one natural language, then Set R can encode all natural languages.
4. Uniqueness theorem -- If R1 and R2 are both PB sets, then R1 and R2 are isomorphic.
5. Fuzzy set theorem -- If R is a PB set, then R is a fuzzy set.
Those above theorems can be proved mathematically even without knowing the content of the set R. When an actual set R is on hand, the situation "changed completely". All above theorems become "TESTABLE," not just provable. That is, we can actually try out one premise at a time.
We can actually try to encode all natural languages with the current PB set by using the Regression Encoding Procedure (REP). "How good or how bad an encoding is" is not an issue. "Can we do it?" is the issue.
There is a major difference between a theorem and a law. Theorem is provable by deduction. Law is testable, generally by induction. Now, the whole issue is testable, and there is a "Law 2" after the construction of the PB set.
Law 2: When every natural language is encoded with a universal set of root words, a true Universal Language emerges.
Why b-o-o-k is book? If it is arbitrary assigned, what is so terrible to assign (engineer, light) for laser? It is not wrong although it may not be the best. The whole concept about the PreBabel is to link all natural languages with a "single set" which provides mnemonic recognition beyond and above the alphabet level. This is the "Law 1".
Law 1: Encoding with a closed set of root words, any arbitrary vocabulary type language will be organized into a logically linked linear chain.
As a law, it is testable.
Question -- form "Ave94" -- I really don't see how this will make it any easier to learn a foreign language, except Chinese. If I want to learn, say, Swahili, I wouldn't learn PreBabel, then learn how it is mapped to Swahili. I don't want to have to learn another language just so I can learn Swahili. Even if languages are all isomorphic, which I don't believe is true, I still have to learn two languages instead of one.
Answer -- The using of dishwasher is, indeed, increasing the dishwashing steps, from a two step job (washing, put it away) to five steps (rinsing, loading, close door, turn power on, unloading). Yet, no one will disagree that dishwasher is an easier way of doing the dishwashing. Of course, with the PreBabel, we are learning two instead of one. The point is that which one uses less energy, the memory energy. This again is a testable issue. We should design one or many tests for this.
I did discuss the "Minimum complexity theorem". Before the minimum complexity is reached for a system, there is always a way to reduce its complexity by using a device which re-organizes the system, such as the dishwasher. Unless all current natural languages are all at their minimum complexity state, we can always introduce a device to reduce their complexity. In the extreme case, that they all are, indeed, at their minimum complexity state themselves, the total system (encompass all natural languages) is still at a random state, as none of the language (excluding the dialects or a family language) is linked in any fashion. That is, the complexity of this total system can be reduced although all its members are all at the minimum state themselves. Thus, the concept of the PreBabel could still be helpful even under this circumstance. Whether that the current PB set is the best candidate for this job or not is not truly an issue.
PreBabel is the true universal language, it is available at