The Forbidden Experiment

In the ancient world, people speculated that humans had one original language, lost from most people due to being parented in their parent's native tongue. They figured that a baby raised without language used would speak this original language, which could then be compared to existing ones.
So, many people tried it. They would raise sets of babies whose caretakers were forbidden from talking to them. the results were startling.
There was no primal language. Many of the children just plain outright died. While the new-age psychology book that first taught me about the Forbidden experiment claimed that they had felt unloved and willed themselves to die, this seems an unlikely explanation for the phenomenon. More likely, the infant did not manage to communicate its needs well enough and wound up infected or malnourished. The survivors grew up incapable of understanding language, and with it, civilization. The caretakers had handfuls of feral children who had the intellectual capacity of a puppy at best. An embaressingly human-shaped puppy who tended not to grasp ideas like not pooping on the floor, and wearing clothes in front of other people.
This is why it is now called the forbidden experiment -- it has an awful human toll, and proved that the base hypothesis was blatantly wrong. Other developments since that have concerned people who could not learn language for other reasons -- the deaf, and neglected feral children recovered from the wild.
Studies of deaf children confirms the original discovery: there is no natural human language, and we need exposure to it at a young age to understand it at all. Also, children with no language exposed to each other, tend to invent some form of language. This gives me my hypothesis on language.
My hypothesis is that language was invented some ten million times, independently, across the globe, wherever humans gathered. Languages have since been refined by exposure to neighboring languages, by grammatical simplification over time, and mispronunciations and misspellings becoming correct by force of habit. Languages have been abandoned, amalgamated, and mutated since then, to fit the needs of the people who speak them.
A scholar of ancient languages has confirmed to me that older languages are in fact clunkier in nature. Their grammar involves obtuse and excessively complicated rules. They are unreasonably lengthy, and often awkward in construction. So the trend in language is one of improvement.