<1>

In the mountainous landscape of northeastern Morocco, a fairly large geographically contiguous Berber-speaking area with a few million inhabitants is to be found, called the Rif. Its people are generally described as speaking a single Zenati Northern Berber language, Tarifit, apart from in its southwestern corner, where another non-Zenati Northern Berber language, Senhaji, may be distinguished. While such a description is well-founded as far as it goes, it often represents a great oversimplification, as Mena Lafkioui's work reveals. Historical linguists looking at innovations within Berber, typologists examining diversity, syntacticians looking at microvariation, anthropologists seeking to trace migrations - all will benefit greatly from this work, which makes it both possible and imperative to resist the temptation to say “Tarifit does so-and-so” in favour of a more nuanced account of variation within the Rif. Even many elements commonly thought of as characteristic of the Rif, such as the sound change l > r, are revealed to be in truth characteristics of a particular central area within the Rif, with older forms still conserved along the outskirts, or to be just a particular case within a large array of forms.

<2>

The 356 maps are accompanied by a text outlining the basics of Rif phonology and grammar, and commenting on the distribution of, and sometimes explanation for, both lexical/phonological and grammatical variation. Each map examines the expression of a particular variable across 141 villages selected to include all tribes of the area, making them readily comparable to existing ethnographic maps such as Coon (1931); at each village, an attempt was made to ask people of both sexes and all classes, selected for minimum external contact. Large urban areas and their newly emerging koinés are deliberately not included in the maps, although they are occasionally commented on in the text. No exact figures are given for how many people were questioned, and little detail of how the data was obtained is provided; this vagueness about method mars the otherwise excellent work.

<3>

The phonological variables examined are in many cases ones familiar from variation within other Berber areas, such as Kabylie (e.g. spirant vs. non-spirant /b/ or [ww] vs. [gg]), although some, notably those relating to /r/ and /l/, are Rif-specific. The grammatical variables examined cover the morphology quite thoroughly, revealing a number of interesting facts, but the paucity of information on syntactic variation, while very understandable given the greater difficulty of collecting it and the time constraints, is regrettable. The lexical variables examined are well-chosen, often permitting comparison with Basset's work for Algeria; future work will hopefully expand the set. The maps are visually pleasing and easily readable for most variables; however, where variation is particularly great (e.g. 2nd person pronouns, p.143), it becomes quite difficult to understand the distribution, and a compositional representation of such complex variables would have made such maps more comprehensible.

<4>

Before the expansion of Arabic, the Rif was part of a much larger Berber-speaking continuum, and it is worth noting how some variables examined link parts of the Rif to Berber areas beyond it. Thus the tendency to turn yt into kt/št/ etc. (84) begins rather sharply within the Rif at the river Nkur and to its east, but continues into western Algerian varieties, such as Djebel Bissa (Genevois & Reesink 1973), despite the hundreds of kilometres of Arabic-speaking areas now intervening between them, and many Senhaja Berber words link it with Middle Atlas Tamazight as against the rest of the Rif, though the two are now some distance apart. Much variation, however, turns out to be less in keeping with the broader distribution, or what, in the absence of more detailed dialect maps, seems to be the broader distribution; for example, the sound shift t- > h-, common in north Algerian Zenati varieties like Chenoua and Chaouia, is found in the Rif (57), but only in its southwestern corner rather than to the east. While the broader picture of variation across the whole of Berber is well beyond the scope of this work, it would have been nice if some of the closest small islands of Berber speech, such as Ghomara or (if it is still spoken) Beni Snous, could have been included.

<5>

The speech of the Rif, like other northern Berber varieties, has undergone strong Arabic influence, and some of the most interesting variation to be observed here involves contact effects; Arabic dialectologists should take note. A particularly striking case is to be observed in negation, a part of the grammar where significant areal convergence is observable in the Maghreb (Chaker & Caubet 1999). In the Rif and in Zenati varieties in general, for the most part, the verb is negated by preposing a reflex of wer and optionally postposing a secondary negative element, usually ša or bu . Different texts have reported several different forms for the Rif, and decomposing the map on p.235 into some of its several component variables makes it clear why. The second negative element is ša throughout a long central belt; this is a retention, to be derived (like its equivalents in a number of other northern Berber varieties) from proto-Northern Berber kra '(some)thing' through the common Berber and Tarifit sound change k > š. However, in the more strongly Arabic-influenced varieties both to its east and to its west, the superficially very similar second negative elements š/ši(y)/šay are used; as both the vowel and the distribution indicate, these must derive not from Berber but from Maghrebi Arabic š/ši/šay 'some(thing)', as used in the Maghrebi Arabic negative construction ma... š/ši, and ultimately from Arabic šay' 'thing'. The coincidental great similarity, phonetic and semantic, between these two elements must undoubtedly have facilitated the borrowing of this element (compare perhaps doublets like Berber ddu vs. Arabic ɛdu 'walk', p.274, and Berber ali vs. Arabic ɛelli 'go up', p. 231); yet it is surprising to see, effectively, half of a negative construction being borrowed while the first half is retained. In the Senhaja region, as the maps show and the text notes, the first negative element has in fact become optional and the second obligatory, approaching stage 3 of Jespersen's Cycle. Here, as in several other cases, the map cries out to be accompanied by a similarly detailed one showing the same variable for the surrounding Arabic dialects: do the Arabophone Senhaja likewise drop ma? How well does the distribution of the form ši(y) (found in the western Rif but not the eastern Rif) correspond to its distribution in the surrounding Arabic dialects? Likewise, p.279 reveals a straightforward partition of the Rif on the variable 'need' between a western zone using ḥdažž and an eastern one using ḥwažž, both Arabic loanwords; does the line dividing these continue unbroken into Arabic-speaking regions? Conversely, many Arabic dialects surrounding the Rif display substantial Berber influence, which it would have been illuminating to view plotted on the same map (as in Behnstedt 2002). In a contact situation as intense as that of northern Berber, stopping at language boundaries often feels arbitrary, although practical constraints may require it.

<6>

This work is, to the best of my knowledge, the most extensive atlas of dialect variation in Berber ever published, covering a far wider range of variables than Basset (1929, 1936, 1939), and will be an essential reference for anyone interested in linguistic variation within Berber or across North Africa as a whole. However, it whets the reader's appetite for even more - more syntax, more vocabulary, more locations, (one) more language. It is to be hoped that future work will expand its coverage even further, and more generally expand the project of mapping variation across North Africa.

References

Basset, André 1929

Etudes de geographie linguistique en Kabylie. Paris: Leroux

Basset, André 1936, 1939

Atlas linguistique des parlers berbères (Algérie du nord) . Algiers: Université d'Alger

Behnstedt, Peter 2002

'La frontera entre el bereber y el árabe en el Rif'. Estudios de dialectología norteafricana y andalusí 2002:7-18

Chaker, Salem and Dominique Caubet (ed.) 1996

La négation en berbère et en arabe maghrébin . Paris: L'Harmattan

Colin, Georges S. 1929

'Le parler berbère des Ġmāra', in Hespéris 9:43-58

Coon, Carleton S. 1931

Tribes of the Rif. Cambridge MA: Peabody Museum

Destaing, Edmond 1907

Etude sur le dialecte berbère des Beni-Snous. Paris: Ernest Leroux

Genevois, H. and P. Reesink 1973

'Djebel Bissa : prospections à travers un parler encore inexploré du Nord-Chélif.' Le Fichier périodique 117,1

Jespersen, Otto 1917

Negation in English and other languages. Copenhagen: Host