The author of this work attempts to present the wide spectrum of linguistic variation in Arabic as an "integrated whole" in order to give the reader a "snapshot" of Arabic as a language undergoing changes due to this linguistic variation. In those parts of the book which are detailed descriptions of these interactions he succeeds in his mission, and does impart to the intended audience (advanced students of Arabic and general linguists) a very good sense of what Arabic language variation is all about. However, at several points the work is flawed, either because of unclear presentation, careless organization, or simple mistakes.
The book is divided into an introduction and eight chapters, ranging from an overview of the history of Arabic, through a description of variation at specific linguistic levels (phonological, morphological, syntactic, semantic, and lexical), and concluding with a useful description of the mechanics of code switching and its use and significance in modern literature. Overall, the most successful parts of the book are those chapters dealing primarily with variation, while the weakest are those presenting analyses of different aspects of the verbal system (sections 3.3-3.5 on the phonology of the verb and sections 6.2 on verbal aspect, mood and tense). There is one general deficiency, namely the absence of maps. Considering the extent to which this work deals with a variety of Arabic dialects, maps are needed to indicate some of the major dialect areas and isoglosses in the Arab world.
The first chapter provides a good overview of current opinions on the development and spread of Arabic. Although the author describes the situation as well as anyone has, when he comes to draw his conclusions from the description (pp. 28-30), they do not fit his own facts. For example, his emphasis on Arab immigration and settlement as the most important factor in arabicization is a great improvement on previous vague notions about why certain Islamic countries became Arabic speaking while others did not. I have only two criticisms about his presentation here: the first concerns his comments on how the linguistic changes came about, while the second deals with his summary comments on the role of Islam in arabicization. In dealing with the early linguistic results of the conquests (p. 19), Holes gives a good account and rebuttal of Versteegh's pidginization and creolization hypothesis, with which I agree. However, the most solid evidence against it is textual, which, being textual, as even Holes notes, weakens it somewhat, and nothing is offered in its place. Holes portrays Arabic as developing within a continuum of multilingual variation ranging from the pidgin/creole extreme (of very limited duration and impact, if any), through the lingua franca stage (short-term, urban), through bilingualism (long-term, rural), and finally, with increasing islamicization, to the monolingual stage; but there is nowhere offered an explanation of how these different linguistic situations brought about changes in Arabic, especially as it relates to interaction with the contact languages. Some comment would have been useful.
To address the second point, on pp. 28-29, Holes summarizes the four common strands in the spread of Arabic as follows: (1) pre-Islamic contact with Arab tribes (except in North Africa); (2) Islam; (3) urbanization, i.e., the establishment, by invaders, of towns which became regional centers of power and in which Arabic became the main language; and (4) migration and assimilation. He maintains that points (3) and (4) were the most decisive, whereas (1) only predisposed certain areas to arabicization, while (2), Islam, "initially was the least important influence in the spread of Arabic, though it came to acquire a central educative (and hence linguistic) role later on" (p. 29). It is this last comment that I take issue with.
To say that "Islam" was the least important factor is misleading, and does not reflect the situation he describes. First, his quote from Wansbrough is instructive not just because the language chosen reveals a certain strange negative bias toward Islam (the conquests diffused Arabic in a "natural, uninhibited" way, while Islam diffused it in an "artificial, restricted" fashion) but also because of its textual bias; Wansbrough is referring to Arabic as a written, not as a spoken language, since his examples of both "natural" and "artificial" diffusion of Arabic refer to written materials, while spoken Arabic is the main emphasis of Holes' work. Considering the extremely high rate of illiteracy throughout this period - which Holes mentions elsewhere - Wansbrough's comments would appear to be less than relevant. Second, using the term "Islam" here is misleading since throughout his description he uses the term "islamicization." If by "Islam" in point (2) he means "islamicization," then it is clear why it was the least important factor initially, since like arabicization, there was so little of it. But to emphasize that "Islam" or "islamicization" was not important in the beginning, and not to call attention to its importance later on is misleading, since his description of the process repeatedly makes the point that they were interdependent. For instance, in clarifying point (3), Holes notes that "the rural peasant communities were more slowly islamicised, [and] were much slower to give up their original languages" (p. 29), while in point (4), "assimilation" means quite simply conversion to Islam.
Diminishing the importance of Islam or islamicization in the process of arabicization is perhaps due in part to the fact that elsewhere conversion to Islam took place without concomitant arabicization, most notably among the Iranians, Kurds, and Berbers. However, this approach overlooks an important difference between these situations and those in arabicized lands - after the Islamic conquests the former societies were mono-religious ones, while the latter were multi-religious ones. In the multi-religious societies the indigenous languages were identified with the dhimmi religions, either Christianity or Judaism. If specific languages were associated with specific faiths, conversion to Islam could only mean "conversion" to Arabic, since holding on to the dhimmi tongue would imply an identification with the dhimmi religion. In other words, arabicization and islamicization were inextricably linked. One could derive this point from Holes' presentation, but it is not made explicit, and is even more obscured in his summary. Thus, despite the fact that Holes' account of arabicization is superior to that of others due to his emphasis on the importance of Arab immigration and settlement, his failure to make the link between arabicization and islamicization leads him to ignore the importance of Arabic as a marker of individual and group identity for Muslims in the multi-religious societies of Iraq, Syria, and Egypt. He thus misses out on an important insight into the reasons for this arabicization.
Chapter two provides an overview of the phonology of MSA (Modern Standard Arabic) and the dialects, and is best when it details the variation to be found among the dialects on this point, e.g., in sections 2.2, and 2.3. Here, Holes presents a review of those features which are "salient markers of geography or identity," such as variation in the pronunciation of interdentals, velars and uvulars, the alveolar fricative /j/, and the various reflexes of /q/. It is at this point especially that one feels the lack of maps, since the features discussed form the bases of the most important isoglosses among the Arabic dialects. Section 2.1, which deals with the phonology of MSA, is less than successful primarily because of the way stress is handled. For example, it is unclear which dialect stress pattern is being used to pronounce the MSA words given as examples on p. 51: in (a) the MSA words are pronounced with a Cairene stress pattern (madrasa / baa-Hathu), while the words in (b) appear with a Levantine stress pattern (muHtawayaatuhu / munaasaba). Holes then notes that Cairenes stress (b) differently, but one of the examples he gives is wrong, viz., muHtawayaatuhu, which should be stressed muHtawayaatuhu.(1) Also, he does not note that Levantine stress is different for (a). Such information has been a part of the linguistic literature on stress at least since Mitchell 1960.(2)
The treatment of stress in the dialects (2.2.2, p. 64) is equally disappointing. The author presents very succinct formulations of pan-dialect stress rules and principles, but they do not stand up to reality, as is apparent from the exceptions which really are the rule, at least for the major dialect areas. He makes stress completely dependent on syllable structure, failing to see that syllable structure and stress are interdependent. This points to a further problem in this section and in the following chapter on morphology. The author talks about these phenomena as if they were processes but fails to make the process explicit by identifying a base and an output. The "three principles which govern word-stress in Arabic dialects virtually without exception" are: (a) stress is predictable and automatic and determined by syllable structure; (b) stress assignment is not lexically fixed but based on the "phonological word"; (c) stress is non-distinctive. He then sets forth the pan-dialect stress rules as follows: (1) if the last syllable of a polysyllabic word is long (Cvv, CvvC, CvCC), stress it; (2) if not long, stress the penultimate. But these elegant rules and principles, while fine for the Djidjilli and Bahraini dialects, do not work for the most widely spoken dialects - Cairene and Levantine. He acknowledges this in part. For example, in both Cairene and Levantine katabu is stressed katabu not katabu, while in Levantine madrasa is stressed madrasa, not madrasa. In addition to these rule violations, the principles also admit of important exceptions: stress may be distinctive, as he himself notes (sakat vs. sakat[t]), violating principle (c), and it may not be dependent only on syllable structure, as he notes for some Maghrebi dialects ([neHla] "bee" vs. [neHla] "I become sweet"), violating both principles (a) and (c). One can easily think of other exceptions to these rules and principles as well, which would make them moot.
Part of the problem with these rules and their formulation has to do with the fact that Holes does not clarify what the stress rules apply to (i.e., what is the base and what is the output) and at what level they apply, nor does he note the concomitant fact that syllable structure is not independent of and prior to stress placement, but interacts with it between the different linguistic levels. This is in part because stress can be assigned at any level, from stem to word to syntactic phrase to sentence, and the syllable structure of a word may change depending on a change in the placement of stress at word-level or phrase-level. Therefore, it may not be distinctive at the lexical level, as we have it in English, but it may be at the morphological or syntactic level. Unless one clarifies where the rule takes place, and what it applies to, and what its output is, then it is useless.
With the exception of the final section (3.6) on dialectal verb morphology, chapter three is the weakest chapter in the book. It begins with a summary of Greenberg's generalization on consonantal roots in Semitic, which would have been clearer if the term "homorganic" were explained, and if a few examples were given. Morphological patterns are then divided into two types: morphosemantic and morphosyntactic, the former including the derived verb patterns in Arabic, while the latter include verbal inflection and nominal patterns indicating agent, patient, instrument and plural (p. 82). However this characterization of patterns contradicts the one at the beginning of chapter four (p. 119), where these nominal patterns are termed "morphosemantic," not "morphosyntactic."
Apart from this, I have two general objections in this chapter: (1) the principle of "root and pattern" morphology is overgeneralized, such that other important features of the morphology are overlooked; and (2) the vowel and consonantal patterns are not integrated in the presentation, which again leads Holes to overlook significant features and generalizations. As regards (1), Holes presents Form I as being "derived," but "unaugmented," which is misleading, since most Form I verbs are basic, non-derived words. Also, in describing the other derived (= augmented) forms, his presentation makes clear that most (five out of nine) do not involve interdigitation or manipulation of the consonantal pattern (i.e., root and pattern), but rather prefixation; yet this non-root-and-pattern aspect of the morphology is not noted by Holes. As regards point (2) above, by focusing on consonantal patterns here Holes fails to point out the most important feature of Form IV, namely the sukuun (or absence of vowel) on [C.sub.1]. A similar lapse occurs later on with regard to verbal morphosyntactic patterns (p. 87) - Holes does not note the changed C-pattern in the imperfect inflection of Form I verbs, because in his emphasis there on the vocalic pattern, he ignores the consonantal pattern.
The nomenclature used for the verb forms is ambiguous and misleading, as is the presentation of the facts surrounding their form and function. Holes eschews the traditional terms "perfect" or "preterit" and "imperfect," and relies instead on abbreviation of "suffix stem" (s-stem) to indicate the perfect and "prefix stem" (p-stem) to indicate the imperfect. There is likely to be some confusion among readers not accustomed to such terminology who might associate "p-stem" with "perfect" or "past." Such confusion is compounded by the unclear presentation of other aspects of verbal inflection, in which the s-stem and p-stem are put in varying positions implying relations of precedence and antecedence which conflict with one another. For example, on p. 87, in the first figure, the p-stem is on top of the s-stem implying a derivational precedence for the p-stem, while in the other figures, the s-stem is before the p-stem, implying a derivational precedence for the s-stem. But on p. 88 the p-stem is put before the s-stem, implying a different precedence, while the active and passive voices are presented simultaneously, with the implication being that they are somehow equally basic in some way. In general, it is unclear what he is describing here; from the textual description it appears to be a process, but from the representation it appears to be a static relationship of some sort, where notions of precedence and antecedence do not matter.
The attempt to summarize the vowel changes in final weak verbs (sect. 3.5, p. 92) is also open to criticism. On the face of it, the rules given appear to be a very succinct statement of a complex process, but it is so succinct that it is unclear even to the author, since he forgets that he is dealing with final weak verbs (where [C.sub.3] = /w/ or /y/) in this section and refers in his comments on these rules to examples of nouns involving hollow roots (where [C.sub.2] = /w/ or /y/). Such a presentation would confuse even those in the intended audience who know Arabic, especially since Holes makes no reference to the traditional terminology (e.g., assimilated, hollow, etc.). In addition, the statement of the rule in the first figure on p. 92 contains as a possible sequence one with a final short /i/ (row 2), which is not found in Arabic verbal inflection - hence his generalization concerning the "sonority hierarchy" of the different vowels is called into question. As in the section on stress, linguists who have worked on these questions and have come up with valid generalizations are not mentioned, and the result is a confused analysis of dubious validity.
The fourth chapter is for the most part a good and useful description of noun morphology in MSA and the dialects, as is chapter five which deals with pronouns and deictics. Chapter six deals with the syntax and semantics of noun phrases and verb phrases, and in dealing with the former, Holes is very successful, providing a good comparative overview of MSA and the dialects in this regard. However, the treatment of verb phrases is full of contradictions and unnecessary complications which revolve around attempts to avoid recognizing the temporal nature of the perfect-imperfect verb contrast. In a sense, this may come down to a professional difference of opinion regarding whether Arabic verb forms encode a time reference or not, but I believe that Holes' approach contains unrecognized contradictions and unanswered questions serious enough to call it into question. The presentation is not coherent; the points are presented in several different places in the book, with slightly different wordings and emphases, and often without any examples of the core uses or meanings of the verb forms in Arabic. Only with the greatest difficulty would the reader be able to determine how these meanings work together and how valid the analysis is, since the exposition of them is so scattered and lacking in real Arabic examples. This causes problems even for the author since there are many gaps in the presentation, with certain terms being left unexplained, while the contradictions they give rise to are ignored.
For example, the "s-stem" (hereinafter the perfect) and the "p-stem" (hereinafter the imperfect) are first presented on p. 86, where it is noted that the imperfect is marked for mood and has temporal and aspectual uses. However on p. 176 we learn that the perfect-imperfect distinction "is not one of tense but of verbal aspect only." On pp. 89-90 we find out that the perfect indicates completed action while the various inflections of the imperfect are given, one of which, the imperfect jussive (p-stem, base set), is said to indicate the negative of completed actions. On p. 176 we learn that the perfect indicates both complete and factual actions, while the imperfect indicates "non-completed/on-going/notional" ones. However no mention is made at this point of the imperfect jussive, whose use to indicate the "negative of completed actions" given on p. 90 contradicts the definition of the imperfect given here as indicating "incomplete" actions. Further on in the section on "mood and modality" (p. 182) we learn that the imperfect indicative serves to indicate "factuality," which contradicts the previous formulation of the perfect-imperfect contrast as one between factual (= perfect) vs. notional (= imperfect) on p. 176. Neither of these terms ("factual" and "notional") are adequately defined, and no further mention is made of the rather ambiguous term "notional." Furthermore, the term "factual" is applied to the perfect, despite its uses in conditional and optative contexts (exemplified on pp. 177-78), while the use of the imperfect jussive in these same contexts (p. 183) is sufficient proof that it indicates "non-factuality."
The scarcity of examples in Arabic throughout most of these sections is simply astonishing, and the net effect is to conceal the temporal nature of these oppositions. For instance, there are no examples of the core meanings of the perfect (i.e., as the main verb in a non-contextualized sentence). Rather, on p. 176 the author gives examples of certain non-core uses of the perfect which are problematic to a tense analysis, viz., its use in conditionals and optatives, and with present perfect connotation. This is despite the fact that the author is forced to note several times throughout the presentation that perfects of "dynamic" verbs (i.e., those denoting an action, probably most of the verbs in the language) are "necessarily" or "intrinsically" or "inherently" time marked as past. The author explains this rather disingenuously as due to their being "completed at a time anterior to the time of utterance" (p. 177), not to their simply being anterior to the time of utterance. Furthermore, there are no examples of the core, unmarked uses of the imperfect indicative. Rather, on p. 176 the different meanings of Arabic imperfect verbs are indicated with reference to sentences in English which supposedly exemplify these meanings in Arabic. The net effect of this is again to conceal the temporal nature of these forms: not being able to see the Arabic imperfect used in Arabic sentences as a main verb, the reader cannot see that the "past" interpretation of imperfects is due to their being embedded under a past tense of the verb 'to be' (kaana) or in a past circumstantial context.
On p. 183 the author attempts to set up a system of moods for the imperfect forms in the dialects which closely parallels that in MSA. Mood in the dialects is marked by the verbal prefixes which are elsewhere described as "aspectual"; that is, in Cairene the bi-imperfect is equivalent to the indicative imperfect ("marks factual statements"), the ha-imperfect indicates a brand-new mood known as "proximate intent," while the bare imperfect fills in for the MSA subjunctive. But these moods are not anything like the MSA ones, since all of the dialectal "moods" can occur as the main verb in an unmarked declarative sentence, while of the MSA ones, only the indicative may do so. Furthermore, the moods in MSA are primarily syntactic in nature: the indicative is used in verbs which are not directly subordinate to other verbs or particles, while the subjunctive is used in verbs which are directly subordinate to other verbs or particles, with the jussive being primarily an archaic form used in a restricted set of well-defined contexts.
The final section on the verb phrase (sect. 6.2.3, pp. 188-94) contains some very interesting material on the paragraph structure of Arabic and the interaction of the various verb forms. Given the preceding analysis, however, in which any indication of verb tense is maladroitly concealed, its insights are limited. It would be useful here to compare how English verbs and Arabic verbs are analyzed in those contexts which contradict their temporal meanings; if one concentrated on these peripheral uses, or on the stylistics of English tense arrangement in paragraphs, one might well be convinced that English verbs are not tenses. For example, in the above it was noted that the Arabic perfect was not a past tense because it could be used in conditional and optative contexts, as well as with a present perfect connotation (pp. 177-78). However, the use of past tense forms occurs in similar contexts cross-linguistically (as a "past irrealis"), and in English is exemplified in the subjunctive form of English which is closely related to the past tense form, and which for almost all verbs is identical to it. Thus we have examples such as: 'If I sat here quietly, would you leave?' Such a use of the past tense in English does not preclude it from indicating pastness elsewhere, nor should the use of the perfect in Arabic in conditional contexts preclude its marking pastness elsewhere. Also, the notion that the perfect in Arabic is used in these conditional contexts because the "action/state in the main clause is envisaged as dependent on the prior occurrence of another action/state" is simply not believable, since one can easily think of counter-examples to this in which the action in the protasis (if-clause) may occur after the action in the apodosis (then-clause):(3)
idha xarajtu issaa ata l- aashirata, fa- aakulu aadatan issa ata t-taasi ata: If I leave at 10 a.m. then I generally eat at 9 a.m.
With regard to the use of the Arabic perfect with a present tense connotation (sections [b] and [c] on pp. 177-78),(4) it is important to note, as Holes does in several places, that the more literal translation of these verbs of emotion and cognition is as a present perfect, not as a simple present in English, thus karih-t ("I hate") is literally "I have come to hate." One of the primary functions of a present perfect is to indicate the present relevance of a past action. In English the pastness or anteriority is provided by the (non-inflected) past participle, while in Arabic it is provided by the perfect verb. This usage therefore does not contradict the temporal nature of the perfect, since the present perfect connotation is dependent on the past time reference contained in the perfect verb. Also, it must be noted that the verbs for which this reading is relevant form a much smaller class of verbs than do the "dynamic" verbs whose perfect is "necessarily" understood as past.
English narrative discourse is replete with examples of tense forms being used in contexts which contradict their normal, unmarked temporal meaning. An example of this is the use of the historical present in English, especially in narrative contexts, and most especially in colloquial narratives:
You know John? Well yesterday I see him on the train and I go up to him and say hello and he just ignores me. Finally he turns to me and nods and says...
Another example of this is the English pattern of sequence of tenses, in which subordinate clause verbs are made to agree with the past tense of certain main clause verbs, often giving rise to time adverb collocations which normally would be ill-formed:
I thought that you finished tomorrrow. Clinton said that he was now the president. Now we were worried.
My point here is not to deny that English and Arabic verbal systems are different (they are), nor to deny that the way they structure paragraphs or "discourse chunks" is different (it is different), but rather to point out the different ways in which English and Arabic verbs are treated. In analyzing English verbs as tenses, the researcher first emphasizes the unmarked, main clause uses based in large part on collocational restrictions with time adverbials. Then attention is directed toward the marked or peripheral usages, in subordinate and modal contexts, which may (or may not) contradict the unmarked usage. In Arabic, on the other hand, the researcher looks at the marked or peripheral usages first, and on that basis determines that Arabic does not have tenses, ignoring the unmarked, main clause usages and how they are used with time adverbials. If one examines the collocational restrictions with time adverbials in unmarked, main clause contexts, one would find that the perfect in Arabic cannot be used with future-time adverbials, while the imperfect (indicative) in MSA cannot be used with past-time adverbials, while all verb forms can be used with a present-time adverbial denoting 'now' (al- aana, dilwa ti, etc.). This, together with native speaker intuitions regarding these forms, as well as with the native grammatical tradition which terms the 'perfect', al-maaDi, or 'the past', indicates that these forms in Arabic do grammaticalize time reference, i.e., they are tenses: the perfect indicates past, while the imperfect indicates non-past. Notions such as 'complete' or 'incomplete' which have been applied to the Arabic verb are not really aspectual in nature, and are mere cover terms for the temporal notions of 'anteriority' and 'simultaneity'. The real aspectual distinctions in the base forms themselves (perfect and imperfect) are minimal (they are 'event' forms, expressing the underlying lexical aspect of the verb as transparently as possible), with additional aspectual distinctions in the imperfect in the dialects being added by the use of verbal prefixes such as bi- in Cairene Arabic. All of these aspectual meanings, however, are dependent on the lexical aspect of the underlying verb (its Aktionsart) and on the adverbial context; thus, a bi-imperfect is not inherently progressive or habitual in meaning - that depends on the lexical aspect of the verb and on the precise context of use. Nor is the perfect inherently 'punctual' or 'non-iterative' since for most lexical classes of verbs it can be used with durative or iterative time adverbials with no ill-formedness. None of these questions is taken up by Holes, who continues the Semiticist tradition of analyzing verb forms in Arabic as being non-temporal, despite the existence of data which contradict it.
Chapter seven is a fine and at times very interesting and detailed discussion of sentence and paragraph structure. Holes' use of English translations rather than the Arabic original is sometimes effective, but often it is overdone, especially when there is no opportunity at all to view the Arabic in the original. Also, one point which may lead to some confusion is the use of the abbreviation COMP to mean 'complement' rather than 'complementizer', as it ordinarily does. Chapter eight, on lexical and stylistic developments in MSA and the dialects, is also very well done and interesting, as is chapter nine, which deals with language levels and code-mixing in Arabic. This area is one which holds a great deal of interest for the intended audience, and it is clearly and well presented. ! have only one criticism to make on this chapter, which has to do with Holes' discussion of the term diglossia and his failure to recognize in his analysis of code-mixing the relevance of an essential difference between MSA and the dialects given by this term, viz., that the latter are natively spoken and learned, while the former is not.
First, Holes misrepresents (p. 278) the notion of diglossia as presented in Ferguson 1958; Ferguson did not claim that only the high variety (H) is used in political speeches and only the low variety (L) is used with friends and colleagues, but that each is typically required in certain contexts, and he did note the mixing of varieties. Also, despite the fact that there are "epistemological problems" in identifying levels (due to fluctuating native speaker impressions of them), Holes has no problem in identifying them in the section following on Nasser's speeches and the songs of Fayruz. It is also important to note that the epistemological problem he refers to (in Parkinson 1991 and 1993) was to a certain extent based upon native speakers' expectations regarding the type of language which they would normally find in a certain context. Thus a few words of colloquial in a written MSA context might be sufficient for a native to identify it as colloquial, while the use of a few words or phrases of MSA provenance in a spoken context would be sufficient for a native listener to evaluate it as "elevated." If anything, this shows that Ferguson's original formulation of the varieties as being associated with certain contexts of use still has a great deal of validity.
More important, Holes' negative reaction against the notion of diglossia leads him, I believe, to mistakenly equate the abstractness of 'H' and 'L', and thus he fails to understand the reasons for the asymmetrical nature of code-switching which his examples point to. This begins early on; on p. 39 he claims that the "concept of Arabic as 'diglossic' language (if ever accurate) is now a misleading oversimplification," since there is no such thing as pure MSA or pure dialect. Therefore, he claims that both ends of the linguistic spectrum are equally abstract. This, however, is a drastic oversimplification, as indicated by his observation a few pages later (p. 46) that there are no native speakers of MSA. This fact is ignored, to the detriment of the analysis of code-mixing, both in chapter two (section 2.3.2, on phonological variation in the dialect) and in chapter nine. In both places, but especially in chapter nine, Holes details the asymmetricality of code-mixing between MSA and the dialects. Simply stated, there are fewer restrictions on using MSA forms in dialect contexts than vice versa. All of his examples, which are quite interesting, point to this essential difference between the dialects and MSA which the diglossic approach emphasizes: dialects are real, natively spoken and learned languages, therefore native speakers have a greater freedom and creativity in borrowing items into it; while MSA is a related but learned second language of sorts, whose users do not possess the same freedom or creativity of borrowing that they do in their native dialect. The data that Holes brings out in this final chapter on code-switching clearly indicate this, but he does not notice it, having rejected the notion of diglossia and the insights it gives with regard to the different status of the dialects vs. MSA.
In sum, while Holes presents in general a great deal of useful and valid information on variation in Arabic, he is too often defeated in his analyses of certain aspects of this variation by his own intransigent stance regarding certain points which run counter to his own presentation of the facts. Such was the case with his comments on the role of Islam in arabicization, on stress, on the generality of the root and pattern system, on the nature of the perfect/imperfect contrast, and finally on diglossia and code-mixing. The book suffers thereby, and makes recommending it difficult.
1 The stressing of muHtawayaatuhu as 'muHtawayaatuhu' is not to be found in Tomiche 1964. The rule on which Tomiche bases his stressing refers to three syllable words of the type madrasa, not six syllable words like muHtawayaatuhu.
2 T. F. Mitchell, "Prominence and Syllabification in Arabic," BSOAS (1960) 23:369-89.
3 On p. 238 there may be a contradiction. Initially Holes states that the protasis "may occur before or after the apodosis." But later, on the same, he notes that "the s-stem has the function of simply projecting Zayd's death forwards or backwards in time as a fact from which consequences have flown/might flow/or would have flown."
4 The use of the perfect as a performative is quite limited, and in any case they are a subset of the present perfect usage.
|Printer friendly Cite/link Email Feedback|
|Author:||Eisele, John C.|
|Publication:||The Journal of the American Oriental Society|
|Article Type:||Book Review|
|Date:||Apr 1, 1997|
|Previous Article:||On the authenticity of the "preface" to the collection of poetry written at the Orchid Pavilion.|
|Next Article:||I Nomi di luogo dei testi di Ebla (ARET I-IV, VII-X e' altri documenti editi e inediti).|