Unlike the formidable "How to Kill a Dragon" by Calvert Watkins, which has stuff a total layman can understand, this book is pretty impenetrable (as the other reviewer noted). Not only are the linguistic references very difficult to follow, the author's train of thought is also on a very-high level. ("As we have already seen with regard to the dative in Tokkarian-V," would be a pastiche.)

This is not meant to criticize the book, or say that it is bad. I just want to point out that there are just possibly 100 people on the globe who could read this book. :-)

This is no more a book for the casual reader than is Newton's _Principia_; but, like the _Principia_, it leaves its subject transformed forever. Greenberg argues that the Indo-European language family should be seen as part of a superfamily that also includes the Uralic, Altaic, Yukaghir, Gilyak, and Chukotian families; Korean, Japanese, and Ainu (seen as distantly related members of a single family); and the Eskimo-Aleut languages, another family. Plus Etruscan. This volume concentrates on "grammar"--mostly pronouns, suffixes and prefixes with grammatical functions, and other formatives; a second volume on vocabulary is planned.

Greenberg's methodology, focusing on the assessment of degrees of probable relationship rather than the quasi-mathematical demonstration of relationship via laws of sound change, is controversial. Yet he makes a strong case supporting the claim that the patterns he demonstrates are stronger than any of their individual data points. Even a small subset of the evidence he presents (for example, the material on first- and second-person pronouns and verb endings) is hard to account for except by genetic relationship of the languages involved.

A virtue of the book is the testability of the relationships he alleges: it opens the way for further study which can strengthen or weaken his case.

It is hard to imagine that a common ancestor for Finnish, Sanskrit, Japanese, and the Eskimo languages--and most of the languages in between--could be more recent than the last ice age. I find it wonderful that elements of English that we use every day, in almost every sentence--the "m" of "am" and "me," the "g" of "ego" (buried just under the surface of "I"), the "th" of "the" (transformed from an earlier "t"), and the"sc" of "crescent" and "fluorescent"--could be shared across the whole northern cap of the planet, passed down to us from linguistic ancestors who witnessed perhaps ten thousand years of history.

Perhaps the most provocative element of the title is the word "closest." Greenberg argues here for only one linguistic superfamily, equal in status to a number of others--one galaxy, as it were, in the starry heavens. What, then, is the closest other galaxy to ours? The American Indian languages, from Canada to Patagonia.

The approximately 1,000 native languages of the Americas are classified into approximately 200 different language families by mainstream scholars, a very high level of diversity in comparison with, e.g., Europe, which contains only three language families (Indo-European, Finno-Ugric, and Basque). In this book, Greenberg claims to demonstrate that there are only three language families in the Americas: Eskimo-Aleut, Na-Dene, and Amerind. Eskimo-Aleut is uncontroversial, as is Na-Dene but for Greenberg's inclusion of Haida, for which the evidence is generally considered unconvincing. The argument that most of the languages fall into an "Amerind" family is the main point of the book.

Greenberg's claim has been rejected by the great majority of specialists for two reasons. First, the method used, superficial lexical comparison, is known to be unreliable. The similarities presented may well be due to chance; even if they are not, the method cannot exclude borrowing as the source of similarties. Second, Greenberg's data have been shown, in a number of published studies (including one by this reviewer), to be riddled with errors. An additional problem is that, although Greenberg offers a subclassification of Amerind, he presents no evidence whatever in support of it.

In sum, this book does not provide either reliable information on the classification of the languages of the Americas or an example of valid historical linguistic methodology.

This book poses a mighty challenge to the orthodox view that there are up to 200 or at least several dozen independent families of indigenous languages in the Americas, by asserting that there are no more than three: Eskimo-Aleut which is related to the Eurasian macrofamily extending across Europe and Northern Asia; Na-Dene which is concentrated in the south- and northwest of North America, and Amerind, comprising about 90% of American languages. In seeking to reconstruct the evolution of language groups and the relationships among their component languages, linguists have become accustomed to comparing a few languages across many words, but Professor Greenberg's approach is the opposite -he looks at a large number of languages across a smaller number of words. This book examines a vast amount of lexical material, mostly Amerind. Chapter 3 treats each of the 11 proposed subgroups of Amerind in a separate section with an enumeration of its languages and their classification, plus a brief history of previous taxonomic hypotheses and a set of characteristic etymologies. The following two chapters present the evidence for Amerind as a single macrofamily, with Chapter 4 providing about 1900 lexical etymologies common to two or more subgroups and Chapter 5 pointing out more than 100 grammatical features found across the subgroups. Other chapters discuss the unity and bounds of Amerind, classification methodology and the problem of Na-Dene. The final chapter deals with the historical implications for the settling of the Americas by three waves of peoples. The book contains 3 maps, 4 appendices, a bibliography and 3 indices: Amerind Etymologies, Language Names and General. Further interesting studies on language in the Americas can be found in Merritt Ruhlen's "On The Origin Of Languages" and in "Sprung From Some Common Source," edited by Sydney M. Lamb, both available here on amazon.com

The classification of African languages into 4 major families (Khoisan, Nilo-Saharan, Niger-Kordofanian and Afro-Asiatic) was the work of Greenberg who published this research in the early 60s. Of course it was initially treated with scepticism by the "splitter" linguists, Greenberg's classification is now universally accepted. There are some very sound arguments in favour of his hypothesis of 3 macro-families in the Americas (Eskimo-Aleut, Na-Dene & Amerind). Christy Turner's dental studies demonstrate only 3 distinct shapes of teeth in the native people of the Americas, corresponding with Greenberg's classificaton. Genetic studies of native Americans also indicate the same three groups (see the work of Luigi Luca Cavalli-Sforza). Most archaeologists believe that modern humans first entered the Americas only about 12 000 years ago. If this is so, then the "splitter" linguists have to explain how so many (up to 200) language families arose in such a short time.

Language In The Americas is a very valuable book and I have no doubt that Greenberg's Amerind hypothesis will gain the widespread acceptance now enjoyed by his work on the language families of Africa.