Language is both a product and a process of evolution. It has mechanisms by which its continuing evolution is kept from running out of control. Those mechanisms ensure that every average user of a language uses words, phrases and pronunciation which, despite changes over time, conform to the expectations of every other average user. The natural laws of conformity give rise to the patterns in language that we call grammar.
We learn language as babies in a home environment. We are not taught language, nor do we learn it in any conventional sense. We just acquire it. In a way not yet understood, our brains extract the natural rules of language from overheard samples of speech. Our first use of language is modelled on our home environment. For most of human history it was mothers who contributed most to a child's way of speaking.
Outside the home, childrens' speech tends to adjust itself towards an average of the word selection, syntax and pronunciation used by the child's peers. As a child joins new groups, new means of expression are discovered, together with nuances of meaning and pronunciation. Eventually the child, as a member of the local speech community, has acquired a regional accent, regional grammar and regional speech biases. There seem to be laws of conformity at work so that all communicators in a group come to share a communal pool of methods for the construction of new sentences.
There are two opposing forces at work in the evolution of any language. There is a tendency for a language to drift in its pronunciation, vocabulary and grammar. Children can never copy their parents' language exactly - just as every human body is subtly different, so every human voice is subtly different. Each generation acquires a subtly changed language: every baby acquires language as a copy of a copy of a copy ad infinitum. In opposition to the natural drift of language is the set of communal rules of conformity. These are not social rules but natural rules - a product of nature. A child whose use of language differs excessively from the average will not be understood. Communication requires that at least two communicators share a common code. Evolution, through rules of conformity, maintains the commonality of the code in the local group.
Left to its own devices, a language will evolve so that all of its users conform to a shared grammar. There need be no conscious decision making by anybody. As travel becomes easier, groups will encounter each other and there will be a sharing and mixing of regional language customs. Eventually, over a large region bordered by natural barriers, an outside observer would note a single language with relatively minor regional variations.
One of the natural laws of conformity works very much like a gene expression mechanism in biology. In the mixing of two languages, the words taken from a second language L2 will be modified to conform with the norms of the first language L1. If sufficient L2 words are incorporated, there is a remarkable effect: L1 words will tend to drift towards the L2 model. This has occured with the English language: the incorporation of so many Norman French words into the English core caused a drift in pronunciation and spelling of both L1 and L2 towards a middle ground. Almost, one might say, modern English is neither L1 nor L2 but a hybrid L3 arising from the bi-directionality of the laws of conformity.
The linguist, or descriptive grammarian, being a scientist, merely notes the action of natural laws of conformity as a phenomenon worthy of study. The prescriptive grammarian, however, may well throw up his or her hands in horror on encountering this "atrocious grammar" and "appallingly bad" pronunciation. From a purely scientific perspective, language use is only 'bad' if it fails in its evolutionary purpose: communication. The prescriptive grammarian adopts this idea as pseudo-science: the purpose of language is communication: it follows that uniformity of speech will enhance communication. That is only true of the bland command-oriented speech of officialdom. In the realm of ordinary speech it is the very colour and flavour attacked by pedants that give a language much of its poetic and literary value.
People come to learn quite naturally that if they use a strongly regional variant of language, or mumble, they will not be understood. There is a law of conformity tending towards clarity of enunciation. The action of that law may well be enhanced if people are consciously aware of it, but there is a danger of placing pronunciation and enunciation together in the same political slot. There was a time when schools would try to force all children to adopt a 'correct' accent. Actors and BBC newsreaders would adopt this accent. Old newsreels and government documentaries from the 1930s to 1950s show this quite clearly.
Today we find these accents laughable. We don't have to be robots with cloned accents in order to communicate. The artificial rule of conformity has been abandoned. Not content with merely abandoning this enforced norm, the politically correct have, since the 1960s, actually sought to reverse it. Mumbling, laying on a regional accent with a trowel, wearing an idiolect grammar as a badge of pride: all of these make life difficult for the hearer, most especially the partially deaf hearer. But the right of the speaker to mumble in an exagerated dialect have now been allowed to ride roughshod over the right of the partially deaf listener to hear intelligible speech from the TV and radio.
Nature takes the middle ground with language. Nature does not require an extreme uniformity of pronunciation, so people with speech difficulties are not excluded from a role in the language community. Nature tends to move the extremities of pronunciation towards a middle ground, so people with hearing difficulties are not excluded from a role in the language community. When it comes to human rights in the language community there is no need for political interference - mother nature knows best.
I would like to credit Don Hucks for triggering my thoughts on language merging through his article about genome merging, whilst taking upon myself any blame for unscientific ranting.