The question of how creole languages come into being has a long tradition. For a brief moment, with Bickerton’s bioprogram hypothesis, this question became the focus of attention not just of linguists but also of psychologists and biologists. The question of the subsequent development of creoles, by contrast, has attracted far less attention, as it does not lend itself to grand generalizations. Siegel’s book serves as a reminder that neither the question of origins nor that of development is a logical one, but that they both require painstaking attention to data. Siegel’s long experience as a fieldworker has greatly benefited this book. Even those who question his interpretation of the data will profit from reading it.

The title of Siegel’s book somewhat misleadingly suggest that it deals with the entire spectrum of pidgin and creole languages, whereas in actual fact it is heavily biased toward the English-based pidgins and creoles of the Pacific, particularly Melanesian and Hawai’ian Pidgin/Creole and languages resulting from European colonial expansion in the Pacific region (Pidgin Fijian and Hiri Motu). The book has little to say about the numerous precolonial Pacific pidgins and creoles, such as Pidgin Yimas and the Hiri trade languages of New Guinea. But even with its focus on a relatively small number of languages that arose in similar social conditions, generalizations are not easy to make, and Siegel concludes that “the development of pidgins and creoles shows language to be a multifaceted, ever-changing system of communication that is highly adaptive to the requirements of its use and the environment it is used in” (p. 279).

This conclusion would seem to suggest that a characterization of pidgin and creole development needs to integrate environmental factors with structural ones. Somewhat surprisingly, this is rarely done in the text, though such integration is crucial to finding explanations for the discrepancy between input and intake in creole genesis and development. Siegel mentions that “sociolinguistic factors may include the number of speakers of the various languages, their prestige and power, and the amount of contact and frequency of use” (149), but these named factors are only the tip of the iceberg. Relevant parameters are indefinitely many, and the relationship between their order of magnitude and the magnitude of their effect is far from straightforward. Again, there are many unmentioned linguistic factors other than transparency, regularity and lack of markedness – a concept whose usefulness is limited to situations “where the structures of the substrate languages do not coincide substantially” (157). The matter is complicated by the fact that pidgins and creoles are more changeable and dynamic than most “natural” human languages. This poses problems with labels such as Tok Pisin or Tayo, which suggest some kind of identity over time.

Siegel’s approach, despite his concluding remarks, is concerned with the linguistic/structural factors that determine the shape of pidgins and creoles, and it thus differs little from many previous studies concerned with linguistic factors that bring these languages into being and determine the direction of their development. Siegel’s differs from most other studies in concentrating on the constraints that prevent transfer rather than seeking an unfolding developmental program. Identifying constraints is not an easy task, and the success of this approach depends on reliable developmental data and claims that can be empirically falsified. Siegel concentrates on pidgins and creoles that are reasonably well described and that are familiar to him. Even so, available descriptions are often normalized to an extent that renders them problematic for the job Siegel wants to undertake. In particular, emergence and development require longitudinal data, which are not readily available. However, such longitudinal studies have been made for Melanesian Pidgin English by Sankoff Reference Sankoff1980. I suggest that development is not a one-dimensional or unidirectional process. On the contrary, parallel competing developments can coexist for considerable periods of time, and seemingly successful and productive grammar can come to an abrupt halt. Typically, previously grammatical regularities – such as reduplication to mark intransitives in Tok Pisin (waswas ‘ to bathe’, singsing ‘to dance’, puspus ‘to copulate’, tingting ‘to think’) – become lexical irregularities (Mühlhäusler Reference Mühlhäusler, Carle, Heinschke, Pink, Rost and Stadtlander1982). Another finding is that variability does not necessarily decrease over time, and that different strategies for relativization can coexist in the same communication network (Aitchison Reference Aitchison, Gerritsen and Stein1992).

Siegel’s careful attention to descriptive detail is evident, but there are a number of smaller issues on which I disagree. First, it is problematic to refer to ol as a 3rd person plural pronoun (89). It certainly did not function in this role when the plantation workers in German Samoa and New Guinea sang Olo boi I limlimbur tumas ‘the black workers take too much time off’ and used em or em ol (with people only) as 3rd person plural pronoun. Second, it is true that Melanesian Pidgin English (MPE) has a smaller set of subject-referencing pronouns (or predicate markers or anaphoric pronouns; I am not convinced that Siegel’s descriptor covers all users and uses, or indeed all of his examples) and thus could be regarded a simplified version of Melanesian substratum languages (90). However, whereas in Melanesian languages 3rd person marking is more marked than 1st and 2nd person marking, in MPE the 3rd person is the least marked context, suggesting that anaphoric he is the primary model. Third, Siegel’s statement, on p. 153, that b(i)long is always used in MPE to mark possession is contradicted by examples such as kapa pinga ‘fingernail’ and manki masta ‘indigenous male servant of a European employer’, which were used as grammatical phrases before becoming lexical compounds. Nor is it the case that the distinction alienable/inalienable is not made in MP. Alienable givim bel ‘impregnate’ has to be distinguished from inalienable givim bel bilong mi ‘to surrender my heart (to Jesus)’.

Throughout the book under review, Siegel takes a strong anti-biological (and at times anti-universalist) position, with which I sympathize. However, his arguments would have been stronger if he had spent less effort (66–78; 91–104; 133–134; 277–279) flogging a dead horse, Bickerton’s language bioprogram hypothesis, while ignoring more sophisticated and potentially more interesting biological models of language. When it comes to explanations, Siegel emphasizes the importance of non-European substratum languages with social and environmental factors facilitating their transfer rather than being seen as directly influencing linguistic structures. This raises a couple of issues. First, Silverstein’s (Reference Silverstein and Hymes1971) notion of pidgins as “convergent generative systems” suggests that apparently identical surface structures can be generated from diverse deep structures, reflecting pidgin users’ different first languages. A Tok Pisin sentence such as em I haisim ap plak ‘he hoisted (up) the flag’ is analyzed as either verb chaining V haisim ‘to hoist’ and V ap ‘to be above’ or as single verb plus particle. Second, Siegel’s approach raises the question of similarity across languages. Similarity is appealed to throughout this book, but I am rather less confident than Siegel that this notion can be sufficiently constrained to do actual explanatory work.

One has to avoid the circular argument of selecting one’s substratum languages on the basis of perceived similarity with a pidgin or creole, but this is precisely what appears to have prompted Siegel (193 ff.), when he argues that “in nearly all cases there are structural or semantic parallels between the variants that ended up in each dialect of MPE and corresponding features of substratum languages of that geographic area.” It is misleading, however, to pick just six closely related Austronesian languages of the New Britain/New Ireland area (among them Tolai) as the substratum of Tok Pisin. Although they were important during its formative years, Tok Pisin was also used from the early days by speakers of Papuan languages, such as the Baining (neighbors and traditional enemies of the Tolais). Brenninkmeyer Reference Brenninkmeyer1924, who resided among the Baining for 15 years, produced a 35-page typescript on the grammar of Tok Pisin on the basis of this experience. It is also of interest to note that Mosel Reference Mosel1980 compares Tolai and adjacent Austronesian languages with Tok Pisin and arrives at the conclusion that there are relatively few similarities. I am not sure what kind of evidence could disconfirm the claim that (constructions of) two languages are similar.

Chapter 8 deals with the prediction of substratum influence in Tayo and Kriol, and at first sight the accuracy of the predictions seems impressive. However, there are potential issues, including possible historical links between Reunion Creole and Tayo (Speedy Reference Speedy2007) and the presence of speakers of non-New Caledonian languages in St. Louis. The argument about reflexive marking in Kriol raises the problem of what it means for a verb to be reflexive (semantically rather than just formally). While all western Germanic languages have formal ways of signaling reflexives, they differ considerably in the use of such reflexives. Compare Dutch zich herinneren with English to remember or German sich erholen, sich regenerieren ‘to regenerate’. Is the semantics of Kriol reflexives similar to English, or similar to local Aboriginal languages?

Generally speaking, I found this book a worthwhile read not because it answered my questions about the emergence of pidgins and creoles, but because it serves a useful reminder that, in spite of much creolistic research over the past three decades, there is still scope for a lot more investigation.