This book "asserts that the origin and spread of languages must be examined primarily through the time-tested techniques of linguistic analysis, rather than those of evolutionary biology" and "defends traditional practices in historical linguistics while remaining open to new techniques, including computational methods" and "will appeal to readers interested in world history and world geography."
SUMMARY This book is the first full-length grammatical description of the Domari dialect spoken in Jerusalem. Domari is an Indo-Aryan language spoken throughout the Middle East by isolated service providing populations whose self-appellation is Dom. The volume will be of great interest to students of Indo-Aryan languages and more generally anyone interested in the linguistic diversity of the Middle East. All the speakers of Jerusalem Domari being bilingual Domari/Arabic, this description offers also an insightful case study in contact linguistics and as such should interest scholars involved in the study of contact phenomena. The book consists of 14 chapters covering most levels of linguistic analysis. The four last chapters include discussions of the Arabic component, a selection of texts and a lexicon.
In Chapter 1, “Introduction”, Matras discusses background information such as ethnographic data about the Dom and the sociolinguistics of the language. The dramatic state of endangerment of Jerusalem Domari is striking: it is estimated that only 10 to 20 fluent speakers are still alive. Matras further discusses previous work on the language, largely restricted to word lists dating back to the 19th century and a grammatical sketch of the very same variety published at the beginning of the 20th century (Macalister 1914). Fragmentary data about other varieties are also dealt with and the author finally attempts to characterise the differences between southern and northern dialects of Domari. Also discussed is the relation between Domari and Romani. Romani is the Indo-Aryan language spoken by the Roma, and Domari has often been referred to as a variety of Romani. This view clearly appears to be inaccurate and there is “no evidence that Domari and Romani ever constituted a single language, at any period in their development; but there is on the other hand plenty of evidence that they underwent shared developments as a result of sharing the same geo-linguistic environments during successive periods” (27). Fieldwork was carried out in Jerusalem between 1996 and 2000. The linguistic data on which the description is based consists of elicited material, narratives and conversations. Although no particular theoretical model is adhered to, two broad assumptions underlie this work. The first is that pragmatics is considered a “method of analysis” and not merely a “single component of language” (33). The second assumption pertains to Matras’ characterisation of what a language is. Domari being in many ways fusional with its contact language Arabic, Matras defines it “as the repertoire components that speakers activate when they define their discourse as ‘Domari’” (34-35).
In Chapter 2, “Phonology”, a structural account of Jerusalem Domari phonology is provided. A system of seven short vowels is posited: /a, e, i, o, u, ɔ, ʌ/. Although minimal pairs are hard to find, length is distinctive and five long vowels are identified: /ā, ē, ī, ō, ū/. The consonantal system largely reflects that of the contact language, Arabic. This materialises in the presence of a set of two pharyngeals (/ħ/ and /ʕ/) and a set of pharyngealised consonants. Interestingly, pharyngealisation also spread to pre-Arabic items. Nevertheless, Domari maintained distinctions absent from Arabic such as /b/ - /p/ and a three way distinction between /k/, /g/ and uvular /q/. Suprasegmental phonology is dealt with only in a paragraph on stress. Matras concludes with an insightful paragraph about historical phonology, situating Domari in the Indo-Aryan group.
In Chapter 3, “Parts of speech and grammatical inflection”, the author departs from a categorisation purely based on semantic-pragmatic criteria because it blurs the relation between function and morphosyntactic properties and adopts a characterisation based on inflection. Word classes are thus identified according to their inflectional properties. Matras isolates eight paradigmatic categories: gender, number, Layer I case marking, Layer II case marking, person inflection, TMA markers, an indefiniteness marker and non-verbal predication markers. Combining these paradigmatic categories with inflectional properties, Matras identifies the following parts of speech: noun, demonstrative, interrogative, indefinite, personal pronoun, local relation expression, adjective, numeral, gerund, participle, 3SG past tense verb, finite verb, adverb, and particle. None of these categories are inflectionally identical (particles remain uninflected). In the remainder of the chapter, Matras reviews these paradigms considering the following macro-categories: inflection, TMA, gender-number-person and non-verbal predication.
In Chapter 4, “Nouns and nominal inflection”, Matras starts with nominal derivation and identifies a limited number of suffixes such as -īš to create verbal nouns from verbal roots, and -wāy to derive abstract nouns from adjectives.. On the whole, the only truly productive derivational suffix in Jerusalem Domari is the feminine ending -ī and derivation is a marginal procedure in expanding the lexicon. Broadening is sporadically attested: qarwi “bitter (F)” > “coffee”, nohra “red” > “tomato” and “Englishman”. There are two genders: masculine and feminine. Masculine nouns most often end in -a and feminine nouns in -ī. Nouns can also end in a consonant, in which case gender is not predictable. Gender governs agreement patterns of demonstratives, adjectives and past-tense verbs in the 3SG. Matras also makes clear that previous statements that Domari kept the neuter are inaccurate. Plural is marked on nouns by the ending -e and extends to demonstratives, adjectives, and numerals. Plural marking is also instantiated in oblique cases through the ending -an. This ending is, according to Matras, also used to mark other categories such as the free 3SG pronoun pandži pl. pandžan bound pronouns -o/-i-m-an (1PL), -o/-i-r-an (2PL), -o/-i-s-an (3PL) . Jerusalem Domari also imports the plural form of items borrowed from Arabic and frequently marks them with the ending -e: dakākīn-e (< Arabic dukkān “shop” plural dakākīn; Arabic relies mostly on non-concatenative morphology to form the plural). Gender marking is neutralised with inherited numerals (dī “two” and taran “three”), while numerals imported from Arabic below ten impose plural agreement (as in Arabic). The language exhibits an indefiniteness marker -ak, similarly found in other Indo-Aryan languages and in Kurdish, a former contact language of Jerusalem Domari. Definitness is overtly expressed only in differential object marking whereby the noun in object position is marked for oblique case if it is definite. The remainder of the chapter deals with case marking and possessive bound pronouns. Case marking in Domari is very similar to what is found in other Indo-Aryan languages, most notably Romani. Matras identifies three layers. Layer I markers developed directly from Old Indo-Aryan and mark a nominative/oblique opposition. Layer II markers are a set of markers that arose out of the integration of Middle Indo-Aryan postpositions, while Layer III consists of borrowed Arabic postpositions. Layer I oblique marking is sensitive to gender, roughly masculine -as and feminine -a. Gender is neutralised in the plural with a common marker -an. Layer I marks definite objects and mediates between the noun and Layer II markers. Matras identifies five Layer II cases and six markers (two of them alternate). Domari marks possession on nouns by bound pronouns that also attach to verbs and a closed set of “local relations expressions” (164). When bound pronouns are attached, oblique marking is apparent in the singular, but not in the plural: bar-om “my mother (NOM)”, bar-im- “my brother (OBL)”, bar-oman “our brother (NOM/OBL)”.
In Chapter 5, “Noun modifiers”, Matras starts with genitive constructions. The main constituent order in present-day Jerusalem Domari is head-modifier, while older sources refer to modifier-head, now marginal. This change is explained by convergence towards Arabic in which the order head-modifier prevails. The rest of the chapter deals with demonstratives, numerals, adjectives and quantifiers. The lengthiest part treats numerals, since Jerusalem Domari has borrowed all numerals above five wholesale from Arabic. The outcome is an entire Arabic grammatical sub-system integrated into Jerusalem Domari, making a synchronic account complex. Matras opportunely summarises the agreement patterns saying that “The first, consisting of inherited ‘2-3’, neutralises morphological plurality on the noun. The second, consisting of Arabic-derived ‘4-10’, requires morphological plurality on the noun. The third, also from Arabic, covers numerals above ‘10’, and again neutralises plurality marking on the noun.” (200). Matras identifies a class of adjectives that inflect for gender and number. Here also the older order modifier-head is disappearing in favour of the Arabic order head-modifier. A striking point is the reliance of speakers of Jerusalem Domari on Arabic items for comparative forms: tilla “big” but Arabic akbar “bigger”.
Chapter 6, “Pronominal categories” starts with personal pronouns. These can be free or bound. Intricacies arise when describing case marking on personal free pronouns in Jerusalem Domari, as these are the outcome divergent historical processes involving free and bound forms, Layer II markers and a set of “local relations expressions” whose etymology is rather opaque. Pronominal demonstratives are treated in this section (adnominal use is described in Chapter 5). Two series, “proximate” and “remote”, that both inflect for gender, number and case are identified. Opposition of distance is neutralised only in the nominative case for F.SG. and PL. Matras further provides a rather thorough pragmatically oriented description of the use of demonstratives. Interesting is the existence in Jerusalem Domari of third person “enclitic subject pronouns” (225) that attach to the interrogative kate “where” and the presentative haṭe: kate-ta “where is he?” and haṭe-ta “there he is”. Also noteworthy is the use of what Matras calls “Arabic referential devices” (226). These are Arabic bound pronouns whose selection is compulsory with the borrowed Arabic particle iyyā- (a pronominal object carrier) and the Arabic pseudo-verb bidd- “want”. The interrogatives are mostly inherited, although qadēš (< Arabic qaddēš) “how much” and waqtēš “when” (< rural Palestinian Arabic waktēš) are commonly found. For indefinites, Jerusalem Domari relies mostly on Arabic material, except ekak “someone”, kiyak “something” and šinak “a little”. The same goes for the expression of reflexive and reciprocal constructions.
Chapter 7, “Verb inflection, modals and auxiliaries” is a detailed analysis of verbal morphology. The Domari verb consists of a lexical root to which derivation, aspect/modality, subject, object, and tense markers are suffixed in a linear order. Jerusalem Domari derives verbs from non-verbs by incorporating the markers -ka(r)- and -(h)o-/-(h)r- from the verbs ka(r)- and (h)o-/(h)r- meaning respectively “do” and “become”. This process is used to integrate Arabic verbs. The use of the former or the latter is governed roughly by transitivity. Along with Romani, Domari is the only New Indo-Aryan language that has kept a present tense conjugation based on the suffixation of subject markers descending from Middle Indo-Aryan, while the past tense conjugation was remodelled from participial forms. Jerusalem Domari has three markers used in existential constructions: the root ho- to express a change of state, ašti to express existence and possession, and a set of predication markers. The rest of the chapter deals with the expression of tense, aspect and modality. Domari has no infinitive and, like other languages of the region, relies on subjunctive marking in embedded non-factual predication. In the area of modals and auxiliaries, it is striking to see that Jerusalem Domari draws almost exclusively on Arabic material with the sole exception of the inherited root saka- “to be able”.
In Chapter 8, “Local and temporal relations”, the author describes how Jerusalem Domari expresses spatial, temporal relations, thematic roles because the devices used in the expression of these semantic categories largely overlap. These are Layer II markers, a set of “inherited spatial expressions” that exhibit a variety of syntactic behaviours, “person-inflected case markers” and up to 24 Arabic prepositions (294). The expression of the subject and object roles are first covered then peripheral source and association, and spatial and temporal relations.
In Chapter 9, “Clause structure”, Matras first examines “Nominal clauses” whose canonical form consists of two nominal components. The first “takes on the subject-topic role” (313) and is usually sentence initial whereas the second is marked with a predication marker. In a pragmatically oriented way, information structure is described as a tripartite division of an utterance: pre-verbal field, finite verb and post-verbal field. Information structure in Jerusalem Domari largely reflects that of the contact language. Matras further describes interrogative and imperative clauses. Complex clauses are shown to rely exclusively on Arabic conjunctions and connectors. There is thus wholesale replication from Arabic for relativisation, complementation, causality, and conditionals. The last part deals with negative clauses. Jerusalem Domari exhibits split negation in the present tense, with the prefixation of the inherited marker n- and the accented suffix -eʔ. The marker n- often drops -eʔ is left alone to mark negation. Such a pattern also appears to be a case of convergence towards the Palestinian Arabic pattern. Negation in the subjunctive/imperative and the past tense relies also on n- and is interchangeable with Arabic ma-. Jerusalem Domari also borrowed other Arabic negation morphemes such mišš, mostly used in non-verbal predication, and the second element of the Arabic split-negation morpheme -š.
In Chapter 10, “Adverbs and particles”, are defined as a “fuzzy and ill-defined category” that includes elements whose “common feature is their lack of inflectional morphology and their tendency to have modifying scope over an entire propositional content” (352). Here also it is striking to see that Jerusalem Domari relies heavily on Arabic, with the exception of a limited number of inherited items such as ghay “well”, bol “much”, ihni “so”, and other local and temporal adverbs. An interesting borrowing is the focus particle gēna “also”, derived from Turkish. Other particles such as interjections, quotation particle, modal particles, fillers and tags are all replicated from Arabic. Matras traces the “quotation particle” (359) qal back to the Arabic verb qāl “he said” and its grammaticalisation path from an inflected verb to an uninflected particle is specific to Domari and unattested in Palestinian Arabic.
Chapter 11, “The Arabic component”, reviews all the grammatical structures that, though replicated from Arabic, “constitute a stable and integral part of the structural inventory of Domari” (368). Matras being also very active in the field of contact linguistics (Matras 2009), it comes as no surprise that this chapter draws primarily on his research and is well integrated into the grammatical description. Categories that are prone to replication are treated first. These are lexicon, phonology and discourse markers. Amongst the borrowed lexical items, one finds also terms that overlap with the inherited lexicon. These are borrowed in their original phonological shape, making the Arabic inventory available to speakers. As far as discourse markers are concerned, Matras notes that there is now ample evidence that this category is amongst the most prone to replication. His main argument is that bilingual speakers in bilingual communities tend to reduce the burden of keeping to the two systems apart by generalising “just one set of interaction-regulating structures across the repertoire” (371). As far as morphological and syntactic borrowing is concerned, Matras reminds us that derivational morphology is more easily borrowed than inflectional morphology. It thus appears strange that Domari doesn’t seem to have borrowed any ‘derivational segment’ from Arabic. The first explanation for this is that speakers of Domari largely integrate Arabic roots into their speech, rendering the borrowing of inflectional morphology redundant. The second reason is that Arabic makes only limited use of concatenative morphology and isolating derivational morphemes is impossible. While the identification of Arabic-derived morphemes is impossible, the integration of free function words is widely attested (relativiser, complementiser, auxiliaries, and conjunctions). This also extends to the syntax and semantics of phrases, clauses, sentences, simple and complex. Matras calls this “convergence in form-function mapping” or “pattern replication” (374), as opposed to “matter replication” (Matras 2009). Domari is characterised as a language with “heavy borrowing” (377). Such a notion is, although somewhat impressionistic, justified in the case of Jerusalem Domari because the extent of borrowing goes beyond structures that are easily transferred in contact situations. We find thus the total replication of “Arabic local and temporal expressions” (377), Arabic core prepositions such as maʿ “with”, min “from”, and paradigm inflections of a group of “aspectual and modal auxiliaries”. From a “system-oriented” point of view (379), there are two different types of Arabic-derived material. The first is elements for which Domari has no internal alternative. The second type is categories for which Domari possesses inherited options. The question of when one is dealing with borrowing versus code-switching in the context of Jerusalem Arabic is quite complex and Matras suggests that it should be seen as a continuum. Typically, borrowings will be the first type, morphologically integrated into Domari, while code-switching is defined as “optional insertions of words as well as entire phrases and utterances, often intentionally for special conversational and stylistic effects” (381-382). In conclusion, although it may be tempting to consider Jerusalem Domari a mixed language, such a characterisation is improper because the Arabic and Indic components are not complementary and their use is stylistically contrastive. This massive intertwining of both components leads Matras to conclude that “languageness” in multilingual contexts should not be considered in terms of separate systems but rather focus on “speakers’ modes of negotiating their entire personal and collective repertoires of linguistic and communicative structures” (390).
Chapter 12, “Samples of talk”, contains five texts with translations and morphological glossing. Happily, the corresponding audio files are accessible on the publisher’s website.
Chapter 13, “Notes on the Domari lexicon”, discusses the nature of the lexicon in the recordings. Strikingly, only one third of the items are of pre-Arabic origin, comprising inherited (Indo-Aryan), Iranian and Turkic items. This is balanced however by the fact that basic vocabulary “shows much higher dependency on inherited lexemes” (428).
Chapter 14, “Domari vocabulary”, provides a glossary of the material retrieved from the corpus. Matras provides a translation, part of speech and language of origin (Arabic, Turkish, Kurdish and Persian).
EVALUATION This work offers a thorough description of the variety of Domari spoken in Jerusalem. The time-depth between the fieldwork (late 1990s) and this publication clearly indicates that the author provides us a mature analysis, patiently refined over a decade. Jerusalem Domari is a highly endangered Indic language in intense contact, whose last speakers are all bilingual in Palestinian Arabic and it exhibits layers of mainly Kurdish elements and more marginally Persian and Turkish. The author clearly demonstrates an in-depth expertise in the fields required to deal with this configuration: contact linguistics, intimate knowledge of Palestinian Arabic and previous contact languages of Jerusalem Domari, and knowledge of languages with parallel profiles such as Romani.
The coverage of grammatical structures is relatively comprehensive and the general function-to-form organisation will help typologists quickly find their way around. The only area that receives limited attention is suprasegmental phonology and a more detailed analysis of stress assignment and intonational patterns would have been welcome. For instance, Matras states that it is not possible to recognise borrowings on phonological grounds because the phonology of Arabic was integrated wholesale into Domari (381). However, the language seems to have kept a separate lexical stress pattern. Accordingly, items integrated into Domari should be stressed differently than in Arabic (as acknowledged on page 62). A discussion of stress assignment in verbs borrowed from Arabic could also provide further evidence about the degree of structural integration between the two components of complex verbs.
As far as the current contact language is concerned, the author clearly demonstrates a deep knowledge of Palestinian Arabic, which allows him to provide us with a thorough analysis of the interactions between Domari and Arabic. The extent of intertwining between the two languages is so great that one may be tempted to say that writing about the grammar of Jerusalem Domari is also about writing about the grammar of Jerusalem Arabic. Arabic dialectology is a very active field and many reliable descriptions covering most of the varieties of Arabic are available. References to these works may have given in some cases further explanations about the semantics of Arabic-derived material. This can be exemplified with the particle atāri (360-361), commonly found in the eastern dialects of Arabic, about which Matras says that “its original lexical meaning is one of approximation -- ‘just like’ --”. The meaning of this particle is indeed opaque but recent studies shed new light on it and describe it as an “evidential presentative” (Henkin 2010:141-142). Apart from these two minor remarks on phonology and references to works in Arabic dialectology, the book is written in a clear and rich style which makes it easy to read. It will benefit everyone interested in language contact, Indo-Aryan languages, including students of Romani, and other ‘Gypsy’ languages. Also valuable to the field of grammaticography is the author’s efforts to put pragmatics at the centre of linguistic description.
REFERENCES Matras, Yaron. 2009. Language Contact. Cambridge: Cambridge University Press.