THE BLACKWELL HISTORY OF THE LATIN LANGUAGE by James Clackson and Geoffrey Horrocks finally brings us a reference based on all the great research into historical linguistics over the last few of decades, superseding to some degree Palmer's The Latin Language. Note that while the authors gloss each Latin word, some previous experience with Latin is necessary to make much use of the book, and you'd be better off with more knowledge of basic linguistics than the average Classics student has.
The book consists of seven chapters, "Latin and Indo-European", "The Languages of Italy", "The Background to Standardization", "'Old' Latin and its Varieties in the Period c. 400--150 BC", "The Road to Standardization", "Elite Latin in the Late Republic and Early Empire", "Sub-Elite Latin in the Empire", and finally "Latin in Late Antiquity and Beyond".
The first two chapters, which track Latin's evolution from Proto-Indo-European and describe some of its neighbouring Indo-European languages, requires some basic knowledge of Indo-European linguistics. Happily, at the same time that this book was published, Clackson's new textbook Indo-European Linguistics: An Introduction appeared. The material is quite mainstream and elegantly presented. My only complaint is that the authors use the Praenestine Fibula as evidence of Early Latin without mention the controversy over its authenticity--it's been twenty years now since a good case was made that it's a forgery.
The chapters on the development of Latin in the late Republic and in the Empire make much use of modern sociolinguistics and of J.N. Adams' recent work on bilingualism in the classical world. The authors include a great number of non-literary attestations of Classical Latin, such as letters and epigrapha, which I never got to encounter when I read Classics as an undergraduate.
The last chapter gives the standard account of the phonological developments from Classical Latin into late Latin/Proto-Romance. However, I was very unhappy with the limited scope of the chapter. Early in the chapter the authors write, "Throughout the whole period [the first millennium] speakers within the Romance area were able to communicate with each other." This does not hold for Romanian, which lost contact with the other Romance languages early on. And indeed, for the rest of the chapter Eastern Romance is completely neglected, with Romanian getting only one mention (and one possibly erroneous), and Dalmatian nothing at all. It's a real pity that a book which generally moves beyond the limitations of earlier histories of Latin nonetheless concentrates only on Western Romance.
In spite of my new complaints, I nonetheless think that this work of Clackson and Horrocks is worth reading for Indo-Europeanists and anyone with some basic training in linguistics who wants a full view of the historical development of Latin. Readers interested in the Indo-European portion may want to supplement it with Baldi's Foundations of Latin