I have been considering the curiosity of tonal languages lately. My mother's native tongue is Sango, which has three tones. Mandarine has four. Attic Greek had, I think, three.
Tonal languages are those in which the tone of the voice has as much to do with the meaning of a word or phrase as does the pronunciation of the vowels and consonants. Thus, in the Mandarin language, the word "ba," if pronounced in a steady normally pitched voice, means the number eight. However, if it is pronounced with a falling and then a rising in the tone it becomes the verb "to hold."
English is generally considered a non-tonal language. For the most part I think it is. The same written sounds do not mean different things depending on the tone. Consider, however, the many tone differentiations with which we can say, "Yeah, I think you're right." There could be a great many different meanings to this simple phrase simply based on the tone (or for English we could consider it an emotional element) that is superimposed over the phrase. Tones mean as much in English as do the words themselves.
But that does not make it a tonal language. The differences in meaning in that phrase are not yet entirely new words. They may take on opposite meanings depending on the tone, but yet the tone only tells us with what attitude the speaker is saying the words. It may be straightforward, it may be sarcastic, it may be hesitant, etc. But it is the same words that are understood. We always understand the words themselves as "Yeah, I think you're right," no matter what slant is put on them by the tones and attitudes of the speaker.
Still, there seems to be the hint of tonality in English speach.
At least in my region of the English speaking world, we do not finish all of our words with the letters that are written down.
Take for example the phrase "I went to the store." When speaking this in common usage I, and my neighbors, not only ignore the "t" in went, but we don't even quite pronounce the "n" that came before it. Yet we don't totally ignore the "n." Instead we say something like "I we to the store" in which there is actually a sound (though not an "n" sound) after the "we." We do not totally drop the "n." Instead we vowalize it, causing the voice to rise just a bit after the short "e" sound. So our "went" comes out as a "we" (not like the word "we" with a long "e" sound) with a rise in tone that comes late after the initial pronunciation of the short "e" sound, and then an abrupt glottal stop.
Thus, at least where I live, the word "went" is really used and pronounced more as a tonal word than as a phonetic word. If I were to say "I we to the store" with no tonal indication of the "n" that is missing, people would not only think I was daft, they may not even understand me. At least it would take them a moment to reconstruct what I had said. But with the rising of the tone on the very tip end of the word, no one even realizes that I left off two sounds from the end of the word. That slight raising of the voice hints at what sounds it has omitted, just as an apostophe hints at the omitted letters.
While English is not yet really a tonal language (at least not here where I live), it has the beginnings of becoming such a language. The mystery of how tonal languages get started is not really a mystery. It is all part of the natural inclination to make our languages as efficient as we can. Why should I move my tongue more than I need in order to get my ideas across?