Larry Trask was a widely respected linguist, an expert on Basque and on the way languages change over time. WHY DO LANGUAGES CHANGE is an introduction to language and history meant for a wide audience with no prior training in linguistics. It was left unfinished when Trask died in 2004, but it was revised by Robert McColl Millar and published six years later.
HOW DO LANGUAGES change is divided into seven chapters based on common questions: 1) How do languages change? 2) Why are languages always changing? 3) Where do words come from? 4) Skunk-Leek - my kind of town: what's in a name? 5) Where does English come from? 6) Why is American English different from British English? 7) Why is English spelling so eccentric? 8) Which is the oldest language?
Trask answers these questions with a great deal of trivia. English examples are drawn almost entirely from US English and British English (Australian English gets brief mention in the chapter on placenames), so residents of those two countries are the intended audience of the book. There are also examples from many other languages around the world, and Trask even discusses the rise of pidgin and creole languages. Trask was known for his sense of humour, and I think his tone here is often friendly and not drily academic.
The most prominent book with this kind of approach is Bill Bryson's The Mother Tongue, which appallingly has factual errors and misunderstandings on every page. There is a real need for a introduction accessible to the general public, but written by someone who actually knows what he is talking about. To a degree, HOW LANGUAGES CHANGE will prove informative.
Unfortunately, I don't think Trask's book is ideal. He doesn't explain the technique that helps linguists reconstruct ancestral languages, the so-called "comparative method". There is an index entry for "comparative method", but on the page referred to, that term is not found. Consequently, I fear that Trask's listing of reconstructed Proto-Indo-European terms is likely to seem voodoo science to most laymen. Furthermore, Trask's answer to the last question ("No language is 'older' than any other language, and there is no 'oldest' language") comes after such a long and convoluted explanation that a lot of readers may get lost.