Linguistic typology Greenberg has been described as a founder of modern
linguistic typology. His publications contributed to the development in the field through the 1960s and 1970s. Greenberg's work in
synchronic linguistics focused on identifying
linguistic universals. During the late 1950s, Greenberg began to examine languages covering a wide geographic and genetic distribution. He identified a number of proposed linguistic universals and cross-linguistic tendencies. In particular, Greenberg conceptualized the idea of
"implicational universal", which has the form, "if a language has structure X, then it must also have structure Y." For example, X might be "mid front rounded vowels" and Y "high front rounded vowels" (for terminology see
phonetics). Many scholars adopted this kind of research following Greenberg's example and it remains important in synchronic linguistics.
Mass comparison Greenberg rejected the opinion, prevalent among linguists since the mid-20th century, that
comparative reconstruction was the only method to discover relationships between languages. He argued that genetic classification is methodologically prior to comparative reconstruction, or the first stage of it: one cannot engage in the comparative reconstruction of languages until one knows which languages to compare (1957:44). He also criticized the prevalent opinion that comprehensive comparisons of two languages at a time (which commonly take years to perform) could establish language families of any size. He argued that the number of possible classifications increases exponentially with the number of languages compared (1957:44). For comparison, the proposed
Niger–Congo family is said to have some 1,500 languages. He thought language families of any size needed to be established by some scholastic means other than bilateral comparison. The theory of mass comparison is an attempt to demonstrate such means. Greenberg argued for the virtues of breadth over depth. He advocated restricting the amount of material to be compared (to basic vocabulary, morphology, and known paths of sound change) and increasing the number of languages to be compared to all the languages in a given area. He hypothesised that this would make it possible to compare numerous languages reliably. At the same time, the process would provide a check on accidental resemblances through the sheer number of languages under review. . Greenberg used the premise that mass "borrowing" of basic vocabulary is unknown. He argued that borrowing, when it occurs, is concentrated in cultural vocabulary and clusters "in certain semantic areas", making it easy to detect. With the goal of determining broad patterns of relationship, the idea was not to get every word right but to detect patterns. From the beginning with his theory of mass comparison, Greenberg addressed why chance resemblance and borrowing were not obstacles to its being useful. Despite that, critics consider those phenomena caused difficulties for his theory. Greenberg first termed his method "mass comparison" in an article of 1954 (reprinted in Greenberg 1955). As of 1987, he replaced the term "mass comparison" with "multilateral comparison", to emphasize its contrast with the bilateral comparisons recommended by linguistics textbooks. He believed that multilateral comparison was not in any way opposed to the comparative method, but is, on the contrary, its necessary first step (Greenberg, 1957:44). According to him, comparative reconstruction should have the status of an explanatory theory for facts already established by language classification (Greenberg, 1957:45). Most historical linguists (Campbell 2001:45) reject the use of mass comparison as a method for establishing genealogical relationships between languages. Among the most outspoken critics of mass comparison have been
Lyle Campbell,
Donald Ringe,
William Poser, and the late
R. Larry Trask.
Genetic classification of languages Languages of Africa Greenberg developed a classification system for the
languages of Africa, which he published from 1949 to 1954. He revised the book and published it again during 1963, followed by a nearly identical edition of 1966 (reprinted without change during 1970). A few more changes of the classification were made by Greenberg in an article during 1981. Greenberg grouped the hundreds of African languages into four families, which he dubbed
Afroasiatic,
Nilo-Saharan,
Niger–Congo, and
Khoisan. During the course of his work, Greenberg popularised the term "Afroasiatic", originally suggested by
Maurice Delafosse, to replace the earlier term "Hamito-Semitic", after showing that the
Hamitic group, accepted widely since the 19th century, is not a valid language family. Another feature of his work was to establish the classification of the
Bantu languages, which occupy much of Central and Southern Africa, as a part of the Atlantic–Congo family, rather than as an independent family as many Bantuists had maintained. Greenberg's work on African languages has been criticised by
Lyle Campbell and Donald Ringe, who do not believe that his classification is justified by his data and request a re-examination of his macro-phyla by "reliable methods" (Ringe 1993:104).
Harold Fleming and
Lionel Bender, who were sympathetic to Greenberg's classification, acknowledged that at least some of his macrofamilies (particularly the Nilo-Saharan and the Khoisan macrofamilies) are not accepted completely by most linguists and may need to be divided (Campbell 1997). Their objection was
methodological: if mass comparison is not a valid method, it cannot be expected to have brought order successfully out of the confusion of African languages. By contrast, some linguists have sought to combine Greenberg's four African families into larger units. In particular, Edgar Gregersen (1972) proposed joining Niger–Congo and Nilo-Saharan into a larger family, which he termed
Kongo-Saharan.
Roger Blench (1995) suggests Niger–Congo is a subfamily of Nilo-Saharan.
The languages of New Guinea, Tasmania, and the Andaman Islands During 1971 Greenberg proposed the
Indo-Pacific macrofamily, which groups together the
Papuan languages (a large number of language families of
New Guinea and nearby islands) with the native languages of the
Andaman Islands and
Tasmania but excludes the
Australian Aboriginal languages. Its principal feature was to reduce the manifold language families of New Guinea to a single genetic unit. This excludes the
Austronesian languages, which have been established as associated with a more recent migration of people. Greenberg's
subgrouping of these languages has not been accepted by the few specialists who have worked on the classification of these languages. However, the work of
Stephen Wurm (1982) and
Malcolm Ross (2005) has given a small amount of mainstream support for his hypothesis. Wurm stated that the lexical similarities between
Great Andamanese, West Papuan and Timor–Alor–Pantar families "are quite striking and amount to virtual formal identity [...] in a number of instances." He believes this to be due to a
linguistic substratum.
The languages of the Americas Most linguists concerned with the
native languages of the Americas classify them into 150 to 180 independent language families. Early on, Greenberg (1957:41, 1960) became convinced that many of the language groups considered unrelated could be classified into larger groupings. In his 1987 book
Language in the Americas, while agreeing that the
Eskimo–Aleut and
Na-Dené groupings as distinct, he proposed that all the other Native American languages belong to a single language macro-family, which he termed
Amerind.
Language in the Americas has generated lively debate, but has been criticized strongly; it is rejected by most specialists of indigenous languages of the Americas and also by most historical linguists. Specialists of the individual language families have found extensive inaccuracies and errors in Greenberg's data, such as including data from non-existent languages, erroneous transcriptions of the forms compared, misinterpretations of the meanings of words used for comparison, and entirely spurious forms. Historical linguists also reject the validity of the method of multilateral (or mass) comparison upon which the classification is based. They argue that he has not provided a convincing case that the similarities presented as evidence are due to inheritance from an earlier common ancestor rather than being explained by a combination of errors, accidental similarity, excessive semantic latitude in comparisons, borrowings, onomatopoeia, etc. However, Harvard geneticist David Reich notes that recent genetic studies have identified patterns that support Greenberg's Amerind classification: the "First American” category. "The cluster of populations that he predicted to be most closely related based on language were in fact verified by the genetic patterns in populations for which data are available.”
The languages of northern Eurasia Later in his life, Greenberg proposed that nearly all of the language families of northern
Eurasia belong to a single higher-order family, which he termed
Eurasiatic. The only exception was
Yeniseian, which he attached to a wider
Dené–Caucasian grouping, also including
Sino-Tibetan. During 2008
Edward Vajda proposed a relation between Yeniseian and the
Na-Dené languages of North America as a
Dené–Yeniseian family. The Eurasiatic grouping resembles the older
Nostratic proposals of
Holger Pedersen and
Vladislav Illich-Svitych by including
Indo-European,
Uralic,
Turkic,
Tungusic, and
Mongolic. It differs by including
Nivkh,
Japonic,
Korean, and
Ainu (which most Nostraticists had excluded from comparison because they are single languages rather than language families) and in excluding
Afroasiatic. At about this time, Russian Nostraticists, notably
Sergei Starostin, constructed a revised version of Nostratic. It was slightly larger than Greenberg's grouping but it also excluded Afroasiatic. ==Selected works by Joseph H. Greenberg==