Guy Deutscher in The Unfolding of Language devotes several chapters to answering this question. First he lists and dismisses three "obvious" reasons for languages to change.
The first is to keep up with new technology and ideas. While this accounts for many new words, especially during the last few decades, it utterly fails to account for the vast majority of seemingly random changes which have occurred over many millennia. Language change has been recorded, over the space of one or two generations, among stone age tribes whose technology hasn't changed in over 30,000 years.
The second "obvious" theory is that languages change from contact with neighboring tribes by borrowing or imitating vocabulary and grammar. As I showed in the previous post, the French speaking Normans and later Latin speaking clerics and clergy had a significant influence on the English language. But change happens even in isolated peoples. The island of Papua New Guinea has 850 indigenous languages (about 14% of the world's 6,000) precisely because the tribes live in isolated mountain valleys.
The third theory is that people seek novelty and like to change things just for the sake of changing them. But most people actually fear change and try hard to prevent it from happening. This explains why, even though pronunciation of English has changed much in the last 400 years, spelling has changed little, resulting in the problems discussed in my last post.
So if these theories don't explain language change, then what does? It's only in the last few decades that linguists have developed satisfactory answers. Before providing three motives for change, Deutscher addresses the question of who does the changing. The answer of course is that we all do, albeit unintentionally. I refer you to the quote on my blog header. One cow wanders off the path a little - perhaps to crop a fresh bunch of grass, or to follow a butterfly - and the one behind follows. As an example, I've noticed people lately using the word "perfect" at the end of a transaction as a new way of saying "thank you". I overheard a really absurd situation in the bank last week when the teller apologized to a customer for not having enough American dollars to complete the requested cash withdrawal for her planned holiday to the States, and asked her to come back in a few days. The customer answered "perfect" as she turned away empty-handed. My point here is that as silly as this expression seemed to me the first few times I heard it, I now catch myself using it sometimes. I admit it - I am a human cow. In summary, we all are the modifiers of our language, unintentionally and often unknowingly. Now - what motivates us to make changes?
Deutscher lists three motives for language change: economy, expressiveness and analogy. Economy refers to the human propensity for energy conservation - especially our own. In other words, we are just plain lazy. We pronounce words and phrases with the least effort possible to still be understood. "I don't know" comes out as "dunno"; "what's up?" something like "tsup?". These examples, from my youngest son's telephone conversation earlier this evening, involve omitting one or more syllables from a phrase. The words "don't" for "do not" and "what's" for "what is" are examples of the very same thing which have become formalized as proper English.
A countering motive to economy is expressiveness where we add more words for emphasis. Instead of replying with a simple "yes" we might say "by all means" or "I'd be delighted". I explained in the August 25 post on Grammaticalization how this led to the development of the French word "pas" to mean "not". As words become overused (an "awesome" salad), the intensity of their meaning is reduced (similar to the value of money with inflation) and new words must be found to replace them.
The third motive, analogy, is a force that returns order to language. Irregularities in grammar over time become more regular. Children learning to talk do this unintentionally, and are usually quickly corrected. Classic examples are "I goed" or "two foots". Sometimes these "mistakes" catch on. Deutscher gives the following examples: the plural of "book" used to be "beek" but changed to "books" in the 13th century; "eyes" replaced "eyen" as the plural of "eye" in the 14th century; and "kine" was replaced by "cows" for the plural of "cow" in the 16th.
These changes do not occur everywhere at once. They start at a particular place and time - perhaps with a 4 year old! Others copy and it starts to spread. Both versions would co-exist, for several generations, with one gradually growing in popularity until it becomes the standard. Sometimes we can find in literature examples of two versions co-existing. Chaucer in "The Merchant's Tale" uses the word "maked" referring to God's creation of Adam (chosen to rhyme with "naked") and four lines later uses "made" for God's creation of Eve (chosen to fit the line's meter). Note - this is an example of economy, not analogy. A word which uses the regular past tense form ("maked") is in the process of being changed into an irregular one ("made") just because it's easier to pronounce. A similar current situation is "dove", the past tense of "dive", which, at least in North America is replacing the regular "dived", perhaps following "drive" and "drove". Both forms are in current use.
Different languages change in different ways, which has resulted in the 6,000 extant languages today plus the tens of thousands of extinct and intermediate languages of which we can only surmise.