4 vols. xlviii, 1504 pages. Oxford: Oxford University Press.
The Atlas of Pidgin and Creole Language Structures (Michaelis et al. 2013, henceforth apics) is an extremely welcome contribution to the field of contact linguistics, in that it provides the most comprehensive descriptive work on contact languages to date (incidentally, its publication coincides with that of another, more theory-oriented anthology on contact languages, Bakker and Matras 2013). It thus fills a gap in language contact research by making available a wealth of empirical documentation on high-contact varieties. The discipline has namely been in sore need of such a reference tool in order to, among other things, investigate the differences and commonalities between a broad variety of contact languages, which has in fact been the impetus for creole studies since their very inception – with the notable innovation that many pidgins and creoles not based on Indo-European languages are also included in the APiCS. Moreover, it provides a typological lens through which contact languages can be incorporated into the wider pool of variation displayed by the languages of the world.
The apics consists of four volumes for a grand total of 1504 pages. The first volume, the Atlas proper, printed on glossy high-quality paper, describes some 130 (mostly structural, and some lexical) features in 76 contact languages worldwide, based on various lexifiers. Naturally, pidgins and creoles which have already been studied in depth, for instance those of the Atlantic and Indian Ocean (e.g. Jamaican or Mauritian Creole), have been included. In addition, many lesser-known, or even extinct varieties (such as Pidgin Hindustani spoken on Fiji, or Batavia Creole, formerly spoken in Indonesia) are covered. Each chapter includes a colour map showing the distribution of apics languages, as well as the frequency of occurrence of a construction in cases where an alternative strategy is possible. The features are reasonably spread so as to cover all major areas of grammar: 12 features relate to word order (features 1–12), 24 to nominal categories (13–36), six to nominal syntax (37–42), 14 to verbal categories (43–56), 16 to argument marking (57–72), 19 to clausal syntax (73–91), eight to complex sentences (92–99), seven to negation, questions and focusing (100–106), 11 to lexicon (107–117) and 13 to phonology (118–130). Each subsection is richly illustrated with a myriad of examples. The features were inspired in part by the World Atlas of Language Structures (Haspelmath et al. 2005; Dryer and Haspelmath 2013, henceforth WALS) and the Comparative Creole Syntax volume (Holm and Patrick 2007). Also, features traditionally discussed in the literature (such as serial verb constructions and systems of tma marking) are included, as well as a number of other typological traits. These were selected in order to provide a synchronic snapshot of each individual language.
As a side note, the editors have to be commanded for using the Gall-Peter’s projection (as opposed to the more popular alternative, the Mercator projection) for their maps in the printed volumes (although the choice was not carried over in the online edition due to compatibility issues with the wals). This choice makes sense in light of the fact that the greatest linguistic diversity is found around the Equator and, as pointed out by the editors in the introduction (Michaelis et al. 2013: xlv), a majority of contact varieties worldwide are also concentrated in circum-equatorial regions, in line with traditional languages, and biological species as well for that matter.
The other three volumes (the Surveys) focus each on a particular group of contact varieties according to their lexifier. The first describes 29 Germanic-based languages (Volume I, English-based and Dutch-based Languages), the second focuses on 29 varieties based on Romance languages (Volume II, Portuguese-based, Spanish-based, and French-based Languages), while the third surveys 18 contact languages not derived from an Indo-European source (Volume III, Contact Languages Based on Languages From Africa, Australia, and the Americas). Each volume is structured in a consistent way, providing information on each individual language. A short introduction opens the chapter, with a black-and-white map situating the language geographically. Some information on the socio-historical background that led to its emergence follows, as well as a summary of its current socio-linguistic situation. A short grammatical sketch is provided for every language, covering phonology and syntax, to which a short glossed text is added. For a majority of languages, a sound sample is also available via the online version. An index at the end of each volume facilitates the task of finding particular languages involved in the relevant contact situations, and very conveniently, all known substrate languages are listed individually as well.
Although the title suggests that the apics only includes varieties traditionally viewed as pidgins or creoles, other types of contact varieties, such as mixed languages, are also taken into account. Since most proposals for classifying high-contact varieties have not been accepted unanimously by the creolist community, no language in the apics was assigned to a particular category. This choice was deliberately made by the editors of the apics, who, aware of the controversies surrounding the status of some of those languages, have left it for the users to decide to which category a particular language belongs. As the editors mention in the introduction, their main goal was not “to engage in theoretical or ideological debates” (Michaelis et al. 2013: xxxv). One can easily understand their decision not to classify the included contact varieties under a specific label, so as not to misrepresent somebody’s position or unjustly misclassify a language. After all, the status of a majority of apics languages is rather uncontroversial, e.g. nobody would ever claim that Haitian or Saramaccan are not creoles, and everyone would agree that Michif represents a completely different type of outcome of a high-contact situation, but for several borderline cases such as Kikongo-Kituba or African American Vernacular English, a consensus has not been reached.
The editors of the apics have strived to make their survey maximally compatible with that offered by the WALS (Haspelmath et al. 2005; Dryer and Haspelmath 2013). Hence, more than a third of apics features (48 features unevenly distributed across the grammar) have a counterpart in the wals. However, some substantial differences between the databases have to be mentioned. Firstly, for each language of the apics, an expert was asked to fill out the values (from 2 to 9 fixed values) for each feature. In contrast, each single feature in the WALS was examined by a specialist who extracted the information from reference grammars and other sources. Here again is another point of divergence. The data in the apics were predominantly gathered in naturalistic settings. Moreover, as mentioned above, for a number of features in the apics (43 in total), it is possible to indicate whether an alternative construction exists, whereas this option is not provided by the wals. For instance, for wals feature 87A on the order of noun and adjective, French receives the value ‘Noun-Adjective’, although it is a well-known fact that a subset of adjectives in French and other Romance languages are preposed to the noun. This kind of information can be explicitly encoded in the apics. Furthermore, in many instances, discrepancies between the feature values can be observed in the features common to both databases. In some cases, these do not seem to correspond, when they actually do (e.g. for apics feature 28 on definite articles, the equivalent in wals does not offer the same number of possibilities, but the feature values have been adjusted accordingly in the data), but in other cases, they do not (e.g. for apics feature 21 and its wals equivalent feature 46A on independent pronouns, both Tagalog and Zamboangueño lack indefinite pronouns and use an existential construction instead, but the two languages receive differerent values in the data). Besides, the apics option of indicating whether alternative constructions exist is detrimental to the compatibility of the two databases. Indeed, whenever a wals feature has a value ‘Mixed’, no such value is available for the apics equivalent, because the information is explicitly encoded as a shared value, and does not appear in the main feature value (as for instance for apics feature 1 on word order, six values are available for denoting the six basic word orders, but a number of apics languages have at least two preferences with regard to basic word order, and a single language (Michif) allows all six possibilities; these languages would have received a value ‘Mixed’ in the wals). The level of description in the two databases is also significantly skewed. While a staggering 97.34% of the data points are filled in the apics (or 9607 data points, out of a total of 9880 possible data points), the wals ‘only’ provides 14.84% of the possible information (for 192 features surveyed in 2676 languages, or roughly 75.000 data points out of the half a million possible). For these reasons, direct comparisons between the two databases is not completely straightforward and can at times be quite cumbersome. But at least the possibility is there now.
Finally, the online edition (available at <http://www.apics-online.info>) provides additional functionalities not included in the print edition. Most are inherent to the digital medium, such as sorting languages alphabetically or according to geographical region or lexifier. Another convenient feature is that direct comparisons of apics and wals maps are possible in a user-friendly manner, as combined maps can be displayed side by side in the same browser window. Another major difference is that the online edition contains additional information on various lects and on earlier stages of the language for many languages, thus adding a diachronic dimension to the work. Moreover, supplementary phonological and sociolinguistic features and ipa charts of the relevant languages are also included for a variety of languages. On the other hand, some features will only be available in the printed edition of the apics, for instance, all the chapter texts on individual languages in the Surveys, as well as the detailed feature descriptions and corresponding examples contained in the Atlas are not available online.
This tremendous piece of scholarship could not have been made possible without the efforts of a team of 88 committed contributors (several of whom are native speakers of the variety they describe) and to the long-term dedication of the editorial team who coordinated them. Since the idea of compiling this atlas of contact varieties arose some seven years ago, the apics was intended to be a companion to the now widely-used and appraised typological database, the World Atlas of Language Structures (Haspelmath et al. 2005; Dryer and Haspelmath 2013). It is to be hoped that the quality of the apics will enable it to become as widely acknowledged as the wals now is. In any case, it will undoubtedly become an inevitable reference work for present and future generations of creolists and language contact specialists.