This is one of the most important books I have read in the past ten years. Reading this work is like reorienting oneself to the field anew. It is necessary reading in creole studies and general linguistics, since it convincingly discards several still received notions in creole studies that never had any substantial empirical data to support them (e.g., decreolization and the “abrupt” or “broken” transmission of languages), and it correctly brings creole studies in line with general linguistic studies instead of insisting that so-called creole languages are exceptional cases of language emergence.

Most of this important book has already appeared as individual chapters or articles in other forums, and in that sense it is more like a collection of revised articles. The book is in three parts and fourteen chapters. After the Prologue (1–7) is part 1, “Population dynamics and language evolution”, which consists of “Language evolution: The population genetics way” (11–28), “Population movements and contacts in language evolution” (29–58), “How population-wide patterns emerge in language evolution: A comparison with highway traffic” (59–73), “What do creoles and pidgins tell us about the evolution of language” (74–92), and “Race, racialism, and the study of language evolution in America” (93–112). Part 2 is called “Competition, selection, and the development of creoles” and consists of “Competition and selection in language evolution” (115–32), “Transfer and the substrate hypothesis in creolistics” (133–59), “Grammaticization in the development of creoles” (160–79), and “‘Multilingualism’' ‘creolization,’ and ‘indigenization’” (180–201). Part 3, “Globalization and language vitality,” has chapters on “Language birth and death” (205–24), “Globalization and the myth of killer languages: What's really going on?” (225–52), “Myths of globalization: What African demolinguistics reveals” (253–72), and, last, “A case study: The ecology of Gullah's survival” (273–84). The book is well indexed by subject and author.

Unfortunately, the book's title is a misnomer (and Mufwene readily points out many other misnomers among other scholars and subdisciplines of linguistics, e.g. on 2, 180). The book has nothing to do with the evolution of language in incremental genetic changes that occurred among prehumans over enormous timespans (significantly more than just the several millennia that even some of the world's currently spoken languages have endured) and that conferred a species-specific reproductive advantage (16). Instead, Mufwene uses “language evolution” as a general synonym for emergence and change, which makes the book's subtitle redundant. It is common among the general public to use the terms “evolution” and “change” interchangeably; however, this lack of rigor robs the former word of any real distinctive meaning and makes one wonder why Mufwene chose to entitle his book Language evolution (other than perhaps because the word evolution is often interpreted as polemical), when he really means something like language contact and change, as specified in the book's subtitle. He also has borrowed the terms “competition” and “selection” from evolutionary studies, and, in a kind of back-formation, “evolution” piggybacks on top of those terms.

One of Mufwene's many lucid arguments that fly in the face of received wisdom is that English is not a killer language destroying all languages in its path (225–52). This belief, common in linguistics and anthropology, is dismissive of and disrespectful to the local, motivated choices speakers make, particularly in regard to language shift, while negotiating their daily lives on the ground. Mufwene also presents evidence that in many African settings it is largely local urban vernaculars and regional lingua francas that are threats to ethnic languages, not languages associated with European colonization (253–72).

Another strong case Mufwene makes is that so-called creole languages can teach researchers of non-creole language histories much about their objects of study. Nothing really unusual happened among the earliest generations of so-called creole speakers in terms of contact between speakers of different languages and cultures, and the record of artifacts regarding these contacts (since documented creoles are no more than 500 years old) is considerably broader and more detailed than the records we have for older languages. Similar population movements, contacts, and socioeconomic dynamics have occurred throughout human history. This is what Mufwene means by uniformitarianism – an approach that brings the same assumptions and perspectives to all human languages rather than marking one group of languages as unique or exceptional. To that end, Mufwene makes the case that so-called creole languages are just the latest stage of the dispersal of a subgroup of the Indo-European family of languages that began 10,000 to 12,000 years ago. In other words, the English-derived, French-derived, etc., varieties spoken in the Americas, Africa, and Asia are dialects or “daughters” of their lexifiers in the same way French and Spanish, for example, are daughters of Vulgar Latin spoken by the Romans in the administration of their empire. However, every language has what Mufwene calls its own specific language ecology – that is, the history of who came in contact with whom, the linguistic variants available, and the social meanings attached to those variants, as well as the specific ethnographic circumstances (54). In that sense, “Languages evolve in non-uniform ways” (14).

Mufwene's criticisms of historical linguistics are accurate and quite damaging to the field. Traditional genetic linguistics has largely ignored the influence of any language not thought to be in a direct line of descent from the “mother” language in any particular family tree. That is, substrate and adstrate influences, which would substantially reveal the diversification of “sister” languages, have largely been ignored. Of course, the “mother-daughter” analogy in diachronic linguistics was always myopic, or at best only partly revealing (as all analogies are; see below) since any child (the “daughter” in the analogy) has two genetic parents, not just a mother, and languages often have at least two “parents.” Yet because of what Mufwene describes as misguided notions of language purity (30; a notion he later connects to racial purity – more below), this basic fact was ignored in traditional historical linguistics.

This criticism applies to historical linguistics (chap. 3) while pushing creolists to view so-called creole languages in similar terms as dialects descended from their lexifier languages (taking into account the contributions of other languages as well), and to include creoles as Indo-European-descended languages. Other views that Mufwene deconstructs and refutes, still canonical in creole studies, have largely to do with the purported uniqueness of creoles. They include (i) the view that creoles emerge from pidgins, (ii) the view that creoles are the results of abrupt emergence resulting from breaks in “normal” language transmission, and (iii) the view that creoles offer a unique insight into language emergence in humans. Mufwene makes a sound case that creoles do not necessarily derive from pidgins, and that the emergence of these new European-language-derived dialects are the result of regular language contact and change.

Mufwene also tackles head-on several fanciful assumptions still perpetuated in creole as well as ethnic and multicultural studies about the death of languages in the Americas and the United States specifically (214, chap. 11). Despite the persistence of the notion that slaves were forbidden to speak their ancestral languages by Europeans and their descendents (and why would Africans obey this command?), this has had little or nothing to do with the disappearance of African languages in the Americas. Subsequent generations of descendants of original African slaves could often find few, if any, people with whom to speak their ancestral languages, since slave repositories such as plantations and small farms often brought Africans into contact with speakers of languages they never would have been in contact with in Africa. This factor, combined with the fact that African-descended slaves accurately identified European varieties spoken locally as the languages of power that would be necessary for negotiating their future survival, assured that whatever languages were originally spoken by Africans landing in the Americas would only survive a century or two at most as communal languages. Any remaining fragments of these languages today are largely restricted to ritual varieties. Mufwene reaches this sensible conclusion by comparing language attrition among European immigrants in the late 19th and early 20th centuries and noting that their languages are also declining as communal languages generation after generation.

Another point that is often overstated has to do with the influence of African languages in the emergence of vernaculars in the Americas. Clearly there is influence at the level of lexicon, but what about the extravagant claims made by the substratists in creole studies, particularly in the 1980s and 1990s? Mufwene has charted a reasonable path between superstratists and substratists. He admits that there is relatively little structural evidence of direct African language borrowing at the level of syntax, since most vernacular features of European languages spoken by descendents of Africans and Europeans in the Americas can be reliably traced to varieties spoken by colonists and settlers who were already present during the centuries in question. He concludes that the influence of African languages is largely restricted to convergence with European models. That is, when the African languages spoken in a particular location contained a feature with a similar structure in the local European language, then that convergence statistically increased the possibility that the feature would be selected or maintained in the subsequent grammar (133–59).

What Mufwene has brought to the discussion that has largely been ignored in linguistic studies is the crucial role of idiolects. Terms such as language and dialect are really abstractions that conveniently stand for bundles of idiolects that speakers synchronically display in their grammars. When these speakers' idiolects share a critical mass of features, linguists call that a dialect or language. Consequently, the locus of language change is in the interactions of these idiolects in terms of the contact among specific features (e.g., 59–69, 125–26, 131–32).

Inexplicably, Mufwene borrows terms like “Invisible Hand” (e.g., 2–4, 7, 59–73, 92, 116, 119, 134, 161, 183, 208, just to name a few occurrences) from other scholars to no real purpose other than to suggest that language changes occur in a patterned but often unpredictable way. The term seems no more useful than assuming that, say, God is responsible for the relentless change that all humans and their languages manifest. That is, living spoken human languages (as well as human beings in general) are always in the flux of becoming. The Invisible Hand is never really defined (and perhaps it would be impossible to do so), nor is it ever made clear why the concept is necessary. Even less successful is the fact that so much of the conceptual framework of this collection is based on analogies. For example, Mufwene suggests that languages are like viral species, languages are like highway traffic, and languages are like footpaths. I am not sure what we gain by this approach. One would think that the analogy of languages to biological entities would have some reference to the innate (i.e. biological) predisposition by all humans to discover the meaningful units of the languages they are exposed to as children, as Chomsky claims, but it really does not, and there are few references even to Universal Grammar (18). This biological analogy of languages as species is only an interesting comparison, and one that fails on nearly as many levels as it succeeds. Mufwene himself admits that his analogies have “serious limitations” (67).

Mufwene suggests that all change in human languages is externally motivated (31–32) since “All causes of change in any language are external to its structure” (31). He seems to be saying that if a change is introduced to a grammatical system, then it must come from outside the grammar. However, the seeds of change are built into every human's grammar, as is evidenced by four-part analogy (1:2 :: 3:x, x = 4), when the first three components are already part of the grammar in question, and the original but common outputs one can hear, especially among children acquiring their native language(s). Internal language change may result from variation inherent in the grammatical system, which makes it possible for even one speaker to innovate a change. Furthermore, one could make the case that all language change is ultimately internal (it becomes part of a speaker's grammar) even if the catalyst for the change is external, that is, inspired or triggered by contact with another speaker. The spread of change from speaker to speaker is always external, since it requires contact, but since the results of that contact must also be internalized (i.e., it is must be made part of a speaker's grammar), one could argue that the trigger is external but the final act of making that new feature part of the grammar is always internal, especially when we consider grammatical interaction among idiolects (the real locus of language change, as described above).

Last, I'd like to broaden Mufwene's notion that the idea of creoles as “exceptional” languages (i.e., different from all other natural human languages) is rooted in the racism of largely European linguistics in the 19th century, no matter how wrongheaded and objectionable the notions of linguistic and racial purism of that day were. It seems to me that the general reason why many scholars have been misled into thinking of so-called creole languages as “exceptional” languages today is that institutions of higher learning and granting agencies place a high value on what I'll call “the cult of uniqueness.” These agencies and institutions value the fragmentation of knowledge and viewing disciplines as disconnected from each other, paying largely lip service to buzzwords like “interdisciplinarity.” That is, one can make the case for funding only if one can demonstrate the uniqueness or exceptionalness of the area of investigation in question. The idea that only one discipline can tackle a given phenomenon or that a specific phenomenon is the intellectual “property” of a specific discipline has as much to do with the synchronic nature of creole studies today, in which these worthy objects of study are required to be viewed as “exceptional” instead of as other dialects among many. If creoles are dialects of their lexifiers (as they descriptively seem to be), then linguists may view them as more appropriately studied in departments of English, French, and so on, rather than in linguistics departments. However, the diachronic issues that Mufwene raises certainly are worth recognizing and may have converged with the issues I've noted.