I have found several more classic linguistics books in my garage, written before the last one that I read (Chomsky's Syntactic Structures) so once again I am pushed backwards before going forwards.
Jespersen's Language is a popularly written book which, after a long first part (about 100 pages) giving a history of nineteenth-century linguistics, goes on to treat of a number of previously underemphasized and some very speculative topics. The second part is on the child's acquisition of language and its probable influences on the history of language; the third part is on the effects of foreigners or second-language speakers and their influences, and contains chapters on pidgin languages and on what are now called genderlects (the differences in usages between male and female speakers of a given language) — some of his attitudes and assumptions are almost comic from a contemporary perspective; the last part deals with the development of languages and the origin of language, and is totally speculative. Among other things, he argues that the development of language has been progressive rather than a deterioration from a more perfect origin as most linguists at the time assumed, going in general from very complex but also very irregular earlier languages (think of Sanskrit, classical Greek and Latin) to more efficient and flexible analytic languages. Not surprising given that Jespersen was Danish and wrote mainly in (and about) English, he sees English and Danish as the most perfect languages.
He then extrapolates this argument backwards to claim that language began in a state in which very long, musical or poetic words represented vague and largely emotional total meanings which later had to be analyzed to convey more rational thought.
All of this was fascinating to read, but I am not sure how much of it would still be accepted after a hundred years; I suspect that what has held up best, apart from his emphasis on language acquisition as fundamental to understanding language change, is his common-sense demolition of other writer's speculations, rather than his own equally speculative theories with which he tries to replace them.
The book is probably most important for those who are interested in the history of linguistics as a discipline; I am following it up with the same author's Philosophy of Grammar published a couple years later.