This book begins with some history of the development of the method of historical linguistics, then it shows how some families were proven with successful use of the methods.
Then it also briefly discusses how some alleged families, like Nilo-Saharan, Khoisan, and Australian, were unsuccessfully proven. It has a lot of quotes of linguists who doubt those alleged families. Though I found two quotes from Mikkola's work were misleading. Mikkola did say that Nilo-Saharan cannot be proven with normal methods, but he thought most of Nilo-Saharan is proven by his statistical methods. I wish the authors noted that, but instead they left the impression that Mikkola did not believe in Nilo-Saharan.
Then the book has in-depth criticism of several alleged families, like Eurasiatic, Nostratic, Altaic, Amerind and Indo-Pacific. In some cases it disproved families by noting lack of use of regular sound correspondences. Though with Nostratic the book notes the use of sound correspondences and shows why it's not adequate. This discussion can be also useful to show why others' use of sound correspondences for other alleged families is not adequate.
Clearly the book has its main emphasis on disproving the methods used by some to establish families that cannot be proven, but in one chapter it also discusses the opposite, Dixon's idea that families can converge so much that they now look related when they are not. Clearly they show how Dixon is wrong.
I agree with much in the book, though when it deals with so many languages, clearly errors can creep in, like a claim that Hungarian has numeral classifiers. Also I sometimes disagree with what families they think are proven, Sino-Tibetan is in my opinion not proven. The book shows why it is hard to prove Sino-Tibetan, but yet it claims it is proven. Maybe here it relies too much on the predominant concensus. The book also claims North Caucasian is not disputed, which is far from true, some correctly dispute this alleged family.
I have some quibble with the overall methods, they claim we need grammar, basic vocabulary and regular sound correspondences to prove a family. But there are some isolating languages where only the latter two can be used, like with the Tai-Kadai family. Also there is the interesting case of Abkhazo-Adygean, where there is plenty of basic vocabulary similarity, but we cannot demonstrate sound correspondences, due mainly to too many consonant phonemes. And finally there are cases where there is not much vocabulary in common, sound correspondences cannot be demonstrated, but yet there is so much grammar in common, including verb paradigms, that the family is almost proven. This is true of both Afroasiatic and Chukotko-Kamchadal. The book claims that Afroasiatic is not proven, and it is true that even paradigms can be borrowed (see Alsea), but that is very rare, so I think most of Afroasiatic is close to proven, or at least very likely.
Also I disagree with their dismissal of lexicostatistics as a preliminary method of demonstrating or disproving family relationship.
But in general the book is very good in dismissing such popular pseudofamilies like Altaic, Nilo-Saharan, Khoisan, Australian and Amerind.
So it is worth it to buy it.
無料のKindleアプリをダウンロードして、スマートフォン、タブレット、またはコンピューターで今すぐKindle本を読むことができます。Kindleデバイスは必要ありません。
ウェブ版Kindleなら、お使いのブラウザですぐにお読みいただけます。
携帯電話のカメラを使用する - 以下のコードをスキャンし、Kindleアプリをダウンロードしてください。
著者について
著者をフォローして、新作のアップデートや改善されたおすすめを入手してください。
他の国からのトップレビュー
Thomas Martin
A good review of methods of historical linguistics.
2009年2月26日にアメリカ合衆国でレビュー済みAmazonで購入
