- Zygmunt FrajzyngierZygmunt FrajzyngierDepartment of Linguistics, University of Colorado Boulder
Afroasiatic languages are the fourth largest linguistic phylum, spoken by some 350 million people in North, West, Central, and East Africa, in the Middle East, and in scattered communities in Europe, the United States, and the Caucasus. Some Afroasiatic languages, such as Arabic, Hausa, Amharic, Somali, and Oromo, are spoken by millions of people, while others are endangered with extinction. As of the early 21st century, the phylum is composed of six families: Egyptian (extinct), Semitic, Cushitic, Omotic, Berber, and Chadic. There are some typological features shared by all families, particularly in the domain of phonology. Languages are also typologically quite distinct with respect to syntax and functions encoded in the grammatical systems.
Some Afroasiatic languages, such as Egyptian, Akkadian, Phoenician, Hebrew, Arabic, and Ge’ez, have a longtime written tradition, but for many languages no writing system has yet been proposed or adopted. The Old Semitic writing system gave rise to the modern alphabets used in thousands of unrelated contemporary languages. Two Semitic languages, Hebrew (with some Aramaic) and Arabic, were used to write the Old Testament and the Koran, the holy books of Judaism and Islam.
1. The Scope of the Article
The present article provides very basic information on individual Afroasiatic families and the Afroasiatic phylum as a whole. The choice of issues selected for presentation was dictated by what are considered to be important elements in understanding the language structures and functions found in languages of the phylum. These choices are to a certain degree arbitrary. References cited in this article include grammars of individual languages, where the reader will find expanded discussions of various issues.1
2. Subgrouping and History and Evidence of Genetic Relationships
2.1 When the Phylum Was Recognized as Such
For a thorough history of how the notion of an Afroasiatic phylum coalesced from the observation of different relationships among families that compose the phylum, see Cohen (1947), from which the following information is taken. It appears that the first author who postulated that Semitic-Galla (a Cushitic language, in contemporary classification), Berber, and Egyptian are related was C. Lottner (1860–1861, pp. 20–27, per Cohen, 1947, p. 9). The idea of the genetic relationship of these languages increasingly gained acceptance throughout the 19th century. Lepsius (1981) took Lottner’s composition of the family and added to it Hausa, a Chadic language. The family (this was the term for the unit used at that time) was given the name Hamito-Semitic by Müller (1876). The inclusion or exclusion of what have been referred to since the mid-20th century as Chadic languages has become the main source of differences regarding membership within the family. Marcel Cohen (1947, see also 1953) postulated that Hamito-Semitic consists of four families, namely Semitic, Egyptian, Cushitic, and Berber, and tentatively included Hausa. Although the name Hamito-Semitic may suggest a dichotomy between Semitic languages on the one hand and Egyptian, Berber, Cushitic, and Chadic on the other, Cohen claimed that there are no linguistic reasons to separate Semitic from other branches, and stated that he saw no internal subgrouping of the phylum. Nevertheless, such subgroupings have been proposed from time to time (see Diakonoff, 1988). Greenberg (1950 and other writings, e.g., Greenberg, 1963) included the Chadic family within the group and started calling the family Afro-Asiatic, a term coined by Delafosse in 1914 (cited in Newman, 1980). This term is widely used in the early 21st century. The term Hamito-Semitic is still used in some current literature. Other terms that have been proposed but did not gain acceptance include Semito-Hamitic, used by Diakonoff (1965); Afrasian (Russian afrazijskiye), used by Diakonoff (1988; in 1980 he used the term “Afro-Asiatic”); Lisramic, used by Hodge (1975); and Erythraean, used by Tucker (1967). Given that the Afroasiatic languages consist of a number of families, the group is referred to as a phylum.
There have been a variety of proposals regarding the internal division within the Afroasiatic phylum. One such proposal, which has gained some acceptance, is the division of the Cushitic branch into the Cushitic and Omotic families (see Amha, 2012).
2.2 Number of Languages
By some estimates, there are 375 Afroasiatic languages. At present one cannot provide a more precise statement regarding the number of Afroasiatic languages, primarily because many Chadic languages have not yet been described, and also because, for a number of languages, there is no necessary information about mutual intelligibility among related languages. Mutual intelligibility within the Berber family also requires systematic study.
2.3 The Evidence for Genetic Relationships
Claims concerning genetic relationships within the Afroasiatic phylum are based on two types of evidence: morphological correspondences and sound correspondences. Morphological correspondences, invoked by Marcel Cohen (1947), consist of the observation that in four families, namely Semitic, Cushitic, Berber, and Egyptian, there exist two means of adding subject markers to the verb, referred to in Afroasiatic linguistics as “prefix conjugation” and “suffix conjugation,” a distinction invoked by David Cohen (2014) as being an important typological characteristic of Afroasiatic languages. Moreover, in some languages across the phylum, the prefix conjugation is associated with the coding of the imperfective aspect, while the suffix conjugation is associated with the coding of perfective or completive aspect. While Chadic languages do not share this characteristic, the reconstruction of linear order in Chadic may indicate that the suffix conjugation that is present in some Chadic languages is older than the conjugation in which subject pronouns precede the verb. According to the analysis in Frajzyngier (1983), the formation of the imperfective aspect by means of the locative construction ‘X being located at Y,’ where Y is the name of the action, involved the presence of the subject pronouns before the verb.
Another morphological characteristic used by Cohen (1947) as evidence of genetic relationship was the identity of possessive suffixes and object suffixes. This characteristic is shared by all families of the phylum.
Also of interest was the demonstration by Greenberg (1955) of the existence of the infix a, which marks nominal and verbal plurality in several Afroasiatic families. The interest of this marker lies not only in the phonological and functional similarity but also in the formal similarity, as the marker a is inserted inside the consonantal root of the verb.
Cohen (1947) provided a large number of sound correspondences among the Semitic, Egyptian, Berber, and Cushitic languages, as well as a smaller number of correspondences involving Hausa, the best-known Chadic language at that time. Ever since, other linguists have provided even more morphological and phonological pieces of evidence for the genetic relatedness of Afroasiatic languages (Greenberg, 1966; Belova et al., 1994–1997, based on the work of Diakonoff’s team; Orel & Stolbova, 1995; and Ehret, 1995).
As of 2017, the unity of the phylum is noncontroversial, although there have been proposals to revise its internal structure, for example to combine the Cushitic and Omotic families, as well as claims that various families constitute an internal subgrouping within the phylum, e.g. Egyptian, Berber, and Semitic (Loprieno & Müller, 2012).
3. Geographical Distribution
3.1 The Current Areas
Languages of the Afroasiatic phylum are spoken in the Middle East and in North, West, East, and Central Africa. The total number of speakers is estimated to be in excess of 300 million. Some languages have a very large number of speakers, for example Arabic, spoken in North Africa and the Middle East; Amharic, the official language of Ethiopia; and Hausa, spoken in Northern Nigeria. There many languages that are endangered and have only a few thousand speakers left.
3.2 Hypotheses About the Geographical Origin of the Phylum
The current distribution of Afroasiatic languages posits the important question of their geographic origin. Diakonoff (1975, 1988, 1998) places Proto-Afrasian in the southeastern Sahara, between the Tibesti mountains in northern Chad and Darfur. Militarev and Shnirelman (1984) place the origin of Afroasiatic in the Middle East. Diamond and Bellwood (2003) place the origin of Afroasiatic in the Middle East, based on archeological data and the comparison of vocabulary pertaining to plants. This hypothesis is rejected in Ehret, Keita, and Newman (2004). The current consensus regarding the spread of Afroasiatic is that the Berber and Chadic families arrived in their present location from areas further east.
The facts in the following account of Egyptian are largely drawn from Loprieno and Müller (2012). Michael Avina has provided important corrections.
4.1 First Records and Stages
The earliest written records of Egyptian date from 3000 bce. The most recent records are represented in Coptic, the last stage of Egyptian, which is a liturgical language of the Coptic Church in Egypt. The five thousand years of written records, from 3000 bce to the present, provide a unique source of data that may be used to study language changes in morphology, syntax, and even phonology (Hodge, 1970). The establishment of various stages of Egyptian is based on both linguistic changes and changes in the writing system. The following stages are postulated (Loprieno & Müller, 2012): Old Egyptian (3000–2000 bce); Middle Egyptian (2000–1300 bce); Traditional Egyptian, which is linguistically not very distinct from Middle Egyptian but which has an expanded set of characters; Later Egyptian (1300 bce–1300 ce), divided into late Egyptian (1300–700 bce) and Demotic (7th century bce–5th century ce); and Coptic (starting in the 4th century ce.)
The hieroglyphic writing system consisted of over 700 pictograms, which served to convey the information by two means. One was a phonetic, and possibly phonological, means, whereby each phonogram represented a consonant or a sequence of consonants: “For example, from the representation of water *maw is derived the phonological value of this sign as /m-w/” (Loprieno & Müller, 2012, p. 106). Over time, the pictograms came to be used to convey the phonetic value of the first consonant in the word. In the earliest system, vowels were not represented. In the later systems, glides representing [w] and [y] were sometimes used to represent vowels. In later stages of Egyptian, during the Greco-Roman period, the number of pictograms increased to several thousand.
The other way in which the pictograms convey meaning is through semantic value, whereby a sign is associated with some broad area of meaning. For example, a sitting man () expresses the lexical realm of “man, mankind”; a sitting man touching his mouth () expresses the domain of “eating, speaking, thinking, sensing”’ (Loprieno & Müller, 2012, p. 107). The equivalent of a formal unit ‘word’ in Egyptian was represented by a character or characters providing the phonetic characteristics of the word and a character providing the broad semantic area to which the word belonged. A word could also have characters adding positive or pejorative connotation to a word.
Different stages of Egyptian had different phonological systems. Earlier Egyptian had three series of stops, voiced, voiceless, and ejectives, and six places of articulation, bilabial, alveolar, palatal, velar, uvular, and glottal. The language also had a series of continuants that included the following places of articulation: labiodental, alveolar, post-alveolar, palatal, velar, pharyngeal, and glottal; all were voiceless except for the pharyngeal, which had two phonemes, a voiceless and a voiced. The series of sonorants included labial, alveolar, uvular, three approximants, and one lateral. It is postulated that Earlier Egyptian had three short vowels, i, a, u, and the corresponding three long vowels.
In Sahidic Coptic, one of the contemporary descendants of Egyptian, a palatalized affricate was added to the velar series, in addition to the voiced, voiceless, and ejectives. Voiced labiodental (labial dental in some terminologies) and alveolar continuants were added to the series of continuants, and the alveolar affricate merged with the voiceless continuant. The vowel system of Sahidic Coptic was significantly enriched through the addition of front and back, short and long mid vowels. In unstressed position, these vowels were reduced to schwa and a.
The formal means in morphology included the consonantal pattern that carried the broad semantic meaning in both nouns and verbs, referred to as the “root” in Afroasiatic linguistic terminology. Roots could consist of one, two, three, or four consonantal segments, with two- and three-consonantal roots representing the majority. The morphological processes included reduplication of the root, prefixation, and suffixation. These formal means were characteristic of both verbal and nominal morphology and formed the derivational source for a variety of unrelated functions.
Nouns could be inflectionally marked for three numbers, singular, dual, and plural, and two genders. Nouns in early Egyptian could also have possessive suffixes, and possessive prefixes are found in Coptic. Early Egyptian had morphological derivation of adjectives. In later Egyptian, morphological derivation of adjectives was replaced by syntactic construction. Some scholars claim that Old Egyptian might have had case inflection (Loprieno & Müller, 2012, p. 121), but there is no evidence of it, as the potential carriers of case inflection, namely vowels, were not written down. The syntactic structure of Old Egyptian indicates a structural motivation for case marking on nouns: the nominal subject comes last, after prepositional phrases and object pronouns (Michael Avina, personal communication). All other nominal participants are marked by prepositions. The same preposition marked the addressee of a verb of saying and the agentive participant of a passive predication.
The verbal piece in early Egyptian consisted of the verb root; certain aspectual markers; the passive marker, if the clause represented the event from the point of view of the patient; and the subject suffix.
4.4 Linear Orders Across Stages
In older Egyptian, linear order was a means of coding head-dependent relations. Thus, the verb was in clause-initial position, and the nominal predicate was also in clause-initial position. The later form of Egyptian had the linear order SVO, applying to both nominal and pronominal arguments. This change is attributed to the grammaticalization of the verb corresponding to ‘do’ (Loprieno & Müller, 2012, p. 127).
4.5 Functions Coded in the Grammatical System of Egyptian
We do not have a list of functions coded in the grammatical systems of various stages of Egyptian. The following list of functional domains is an interpolation from the formal means contained in the descriptions of Egyptian.
Modality. The assertive modality seems to be the unmarked value in the domain of epistemic modality. The imperative mood was the unmarked form of the verb in the singular, but there was a suffix coding the imperative for plural addressees.
Aspect and tense. Loprieno and Müller (2012, p. 126) list a preterite, most likely a tense rather than an aspect, marked by the suffix -n; a perfective marked by the suffix -t, which also marked the prospective aspect with a few irregular verbs; and a prospective aspect, marked by the suffix -w with the majority of verbs. The prospective aspect designated an event that had yet to happen. It appears that these markers could not co-occur and that there was no distinction between the tense and aspectual system.
Relations between the predicate and noun phrases. Equational predication was marked by the linear order, whereby the first noun in the sequence of nouns was the predicate and the second noun was the subject. The verbal predication distinguished between active/dynamic predicates, which were unmarked, and stative predicates, which were marked by the stative/passive suffix added to the verb. The subjects of active predication were controlling. The subjects of stative predication were in the state described by the predicate. Thus, the subjects of stative verbs could be undergoers of what in the Indo-European tradition of linguistics are considered transitive verbs, and single arguments of what are considered intransitive verbs, such as the verb ‘go.’
With respect to the relationships between noun phrases, the number of potential semantic relations corresponded to the number of prepositions.
There were no overt markers of the relationships between clauses within a sentence, or clauses following each other in a sequence, in paratactic constructions. However, the relationships of sentences to other sentences within discourse were marked at the beginning of the clause.
4.6 For Further Reading
The tradition of the study of Egyptian dates from the decipherment of hieroglyphic writing and the first grammar and dictionary, written by Champollion nearly two hundred years ago. A brief typological account of the Egyptian family can be found in Loprieno and Müller (2012). For Ancient Egyptian, see Loprieno (1995); for Middle Egyptian, see Allen (2014); for late Egyptian, see Junge (2001); and for Coptic, see Layton (2011) and Lambdin (1983). De Gruyter Mouton has announced a new series on Egyptian and Coptic under the editorship of Eitan Grossman.
5.1 Geographical Location and Classification
The following presentation of Semitic is based mainly on Gragg and Hoberman (2012), with occasional use of other sources and with the author’s own interpretations.
Semitic languages are particularly important for western cultural tradition, as the foundations of Judaism and Islam were formulated in the Semitic languages Hebrew and Arabic. The Proto-Semitic writing system became the foundation of the alphabetic and syllabic writing systems used in a variety of unrelated languages of Europe, Middle East, Africa, Americas, and many parts of Asia. The original Proto-Semitic writing system was only consonantal, and the early writings did not indicate word or sentence boundaries. Later writing systems developed the marking of vocalic sounds and introduced the marking of word and sentence boundaries.
The existence of the Semitic family was recognized by scholars as early as the 17th century, and has not been challenged since. The Semitic family, along with Egyptian, is by far the most studied group among Afroasiatic languages, as measured by the numerous grammars and dictionaries devoted to individual languages and the hundreds of descriptive and comparative studies on all aspects of language structure, language change, and language use.
Semitic languages are spoken in North Africa (various dialects of Arabic), the Middle East (Arabic, Aramaic, Hebrew), Ethiopia (a large number of languages, including Amharic, Tigrinya, and Gurage), West and Central Africa (Arabic in Nigeria and Cameroon), and in other areas to which speakers of Semitic languages have immigrated (Aramaic in the United States and Sweden).
The internal classification of the Semitic family has been undergoing relatively minor changes over the last 50 years (see Gragg & Hoberman, 2012; Hetzron, 1974). The basic division is between East and West Semitic. The East Semitic branch is represented by Akkadian and Eblait, while all other Semitic languages belong to the West Semitic branch. The basis for this classification is morphological, in that East Semitic languages are assumed to preserve the coding of both the past tense (or stative) and the non-past tense by means of prefixes, while the West Semitic languages code the equivalents of past tense by means of suffixes, which is considered to be an innovation. The internal division within West Semitic is a subject of continuing debate.
The system of obstruents in Semitic languages is characterized by three series: voiceless, voiced, and a third series that is sometimes called ejective and sometimes called emphatic. The series of sonorants includes liquids, glides, and nasals. While these manners of articulation are shared by all Semitic languages, the number of places of articulation within a given language varies considerably. The consonantal systems of Old South Arabian and Modern Standard Arabic are considered to be the richest, potentially closest to the Proto-Semitic system. Old South Arabian had the following places of articulation: labial, interdental, dental, palatal, lateral (continuants and the liquid l), velar, post-velar, pharyngeal, and glottal (both a stop and a continuant). Akkadian most probably had labial, dental, palatal velar, and glottal places of articulation, the last represented by the glottal stop. Classical Arabic has labial, interdental, dental, palatal, lateral, velar, post-velar, pharyngeal and glottal places of articulation, the last one containing both the stop and a continuant. There is considerable fluctuation in the places of articulation even within one language, as is the case in Israeli Hebrew.
The vowel system of Proto-Semitic is considered to be close to that of Classical Arabic, consisting of three short and three long vowels, i, u, a, and two diphthongs, ay and aw. This system has undergone expansion in many languages, mainly through the addition of mid vowels and schwa.
5.3 Morphology of Semitic Languages
As in Egyptian, all verbs and many (but, importantly, not all) nouns have a consonantal root system that is associated with some general meaning. For example, k-t-b is generally associated with writing. Various vocalic and consonantal markers can be added to such a root to derive a variety of lexical categories such as nouns, verbs, and adjectives. Within each of these categories, another set of affixes may be added to derive narrower lexical classes, such as the name of the agent or a place. Inflectional markers may also indicate a variety of functions coded in the language, such as tense, aspect, and mood for verbs, and the relationship of the given lexical item to other constituents of the utterance, such as case markers added to nouns. It is important to note, however, that there are nouns in Semitic languages that are not derived through the consonantal root and vowel insertion method, for example yad ‘hand’ in Arabic. The consonantal root is a product of linguistic analysis; uneducated speakers have no explicit notion of root, even though they can recognize to what semantic category a word belongs and can analyze foreign words and extract consonantal roots from them. The main morphological means deployed in Semitic languages are grammatical vowel and consonant insertions within the consonantal root; suffixation; prefixation; gemination of consonants; and reduplication of a sequence of segments, sometimes involving syllables. Several of these means can be used at the same time on a single lexical item.
5.3.1 Morphology of the Noun
Nouns in Semitic are divided into two classes, labeled “masculine” and “feminine.” Membership in a class is not overtly marked on the noun itself but is reflected on other lexical categories in the clause through an agreement system. Noun classes are also reflected in the marking of nominal plurality through suffixation, as there are different suffixes for masculine nouns versus feminine nouns. There are three numbers, singular, dual, and plural, with the dual not being productive in a number of languages.
In Arabic (West Semitic) and Akkadian (East Semitic), there exist inflectional case marking, with a striking similarity of case markers across branches, whereby the suffix u marks nominative, a marks accusative, and i marks genitive. In some languages, for example Akkadian, the three cases are attested for singular nouns. In dual and plural, the system is reduced to two cases, nominative u and genitive/accusative i. Akkadian also has traces of two other case markings, which appear to have locative functions.
In Arabic, nouns with internal plural (the older form) have a three-case system, while nouns with suffixed plural have a two-case system. Given that Arabic and Akkadian belong to different branches of Semitic, linguists have been reconstructing these three case markings as a retention from Proto-Semitic (but see Owens, 1998 for a different approach to the status of case in Arabic). Classical Ethiopian, Ge’ez, has two cases, the unmarked nominative and the marked accusative. Languages that do not have inflectional marking of case, for example Hebrew, are considered to have lost this marking. Some contemporary Semitic languages, for example Amharic, have developed a new case-marking system, with the nominative unmarked and the accusative marked. The marking of the accusative is associated with the definiteness of nouns.
In addition to case system as described above, nouns in Semitic languages can also be marked for the category called “state” in the Semitic grammatical tradition. This marking is restricted to the relationships between nouns, between a noun a numeral, and a few other contexts. The “construct state” marks the head noun, which precedes the modifier. The modifier is unmarked. The construct state did not co-occur with other case markers, which is the evidence that it belonged to the same formal domain. The construct state is not a productive formal means.
There were two other morphological markers added to nouns, one consisting of the suffix –m (“mimation” in traditional Semitic terminology) and the other consisting of the suffix –n (“nunation” in traditional Semitic terminology). Both markers contrasted with the absence of a marker and most likely had functions within the domain of reference.
5.3.2 Morphology of the Verb
All Semitic languages have some mechanism for deriving nominal and adjectival stems from roots. The various means of derivation can be interpreted as a part of the process of deriving various lexical categories from one broad semantic notion.
Semitic verbs code the person, gender, and number of the subject through suffixes in one tense/aspect and through prefixation and suffixes in another tense/aspect, which is thought to have originally coded a state.
Tense, aspect, and mood are marked by a variety of means, which include vocalic patterns interspersed with the consonants of the root, a formal means common to all Semitic languages; vocalic and consonantal prefixes, suffixes, and infixes; and gemination of the second consonant of the root. Many Semitic languages combine the coding of tense, aspect, and mood on the verb with the use of auxiliaries. Some other languages, for example North-Eastern Neo-Aramaic, code the categories tense, aspect, and mood through auxiliary verbs.
All Semitic verbs share several morphemes, which in general terms can be seen as encoding semantic relationships between the noun phrases and the verb. These categories have different labels in different languages. One of the terms used for these categories in Arabic grammars is “theme,” followed by a numeral. Despite the existence of numerous grammars, functions of these markers have yet to be described within the totality of the system. Every verb in the Semitic languages must occur in either the unmarked form (“basic theme”) or one of the marked forms. These markers define the semantic role of the subject and sometimes of other arguments in the clause. They also code the manner of the event. Since these markers occur in all the languages, they most probably constituted a part of the Proto-Semitic grammatical system. Here is a brief list of the markers and a traditional, but very approximate, description of their functions. The functions listed, even in the best-described languages such as Arabic, require a thorough and rigorous analysis. Note that in traditional Semitic terminology reference is made to the whole stem even if just one morpheme is involved. Parentheses include traditional terms in Semitic scholarship.
Basic stem. This is a morphologically unmarked stem, whose function can be described only in comparison to other, marked functions.
Second consonant geminate (d-stem). This stem appears to derive verbs from nouns and creates an opposition between transitive predicates and intransitive predicate (e.g., ‘learn’ [basic] → ‘teach’).
First vowel lengthened stem (l-stem). There are a large variety of meanings associated with this stem in Arabic, including intensive, conative (requiring an effort), and transitive. The function of this stem in languages where it is in contrast with other stems requires a thorough explanation.
Reduplication of the second syllable (r-stem). This occurs in all branches of Semitic. In Amharic it has the frequentative meaning.
Causative (s-stem). The stem is marked by the prefix s, š, h or glottal stop . This stem transitivizes an intransitive verb.
Passive (n-stem). This stem is marked by the prefix n. In addition to indicating that the subject of the clause is the undergoer, the marker n has a variety of other functions.
Passive-reflexive (t-infix). This stem is marked by the infix t after the first consonant of the root. It is used to code reflexives and reciprocals.
Some derivational morphemes can be combined within one verbal theme. In addition, in some languages, for example in Arabic, each verbal theme may have a passive marker u added.
5.3.3 Linear Orders
Linear order is a coding means in equational clauses in the majority of Semitic languages where the first position in linear order codes the subject and the second position codes the predicate. This fact is quite important, as it correlates with the presence of the construct state to code the function of one noun modifying another noun.
Linear order is also a coding means in verbal clauses in Israeli Hebrew and in many contemporary Arabic dialects. In Biblical Hebrew, Classical Arabic, and Ge’ez the verb was in clause-initial position. In Akkadian, the verb was in clause-final position. The contemporary South Ethiopian Semitic languages also have the verb in clause-final position, most likely under the influence of the Cushitic and Omotic languages by which they are surrounded. Therefore, in the verbal clauses in these languages there was only one position available as a coding means, and order was not a coding means for relations between the verb and noun phrases. Most likely, these relations were coded by the thematic forms of the verb as described earlier, and possibly by case marking on nouns.
Most Semitic languages have prepositions. In some languages, for example in Amharic, some spatial configurations, for example ‘in,’ are coded by postpositions derived from nouns. Most prepositions in Semitic languages code spatial configuration and locative function; they are used whenever a noun is a locative complement, regardless of the nature of the predicate and the nature of the complement. This is different from the use of locative prepositions in some Chadic languages, where locative prepositions are not used if the complement is inherently locative. In Biblical Hebrew, the preposition et coded focus (see Joüon, 1947; Joüon & Muraoka, 2006). In Israeli Hebrew, the preposition has acquired the function of the direct object marker.
5.4 Functions Coded in the Grammatical System of Semitic Languages
The assertive modality is the unmarked modality in all Semitic languages. All languages code the imperative mood through the inflectional form of the verb.
5.4.2 Tense, Aspect, and Mood
All Semitic languages minimally code a three-way distinction, which Gragg and Hoberman (2012) refer to as Past-Present-Jussive, using tense terminology, or Perfective-Imperfective-Subjunctive, using aspect-based terminology. Individual languages have grammaticalized a much richer system of tense, mood, and aspectual distinctions. The terminological uncertainty displayed in Gragg and Hoberman (2012) reflects a deeper problem, that of the analysis of tense, mood, and aspectual distinctions in individual languages and in comparative studies. The issue is whether the forms referred to as “perfective” in Semitic languages code bounded events, whether forms referred to as “imperfective” code unbounded events, and whether there is indeed a distinction between tense and aspect. For an early study of aspect in Semitic, see Kuryłowicz (1973), and for a recent study of tense and aspect in Semitic, see Bubenik (2017).
5.4.3 Relationships Between Predicates and Noun Phrases
The existence of the category “subject” in Semitic is noncontroversial. The subject is coded through a dedicated set of prefixes or suffixes affixed to the verb. Similarly noncontroversial is the existence of the category “object.” Its existence is supported by the set of object pronouns and the accusative case-marking in Classical Arabic and Akkadian, Amharic, and Israeli Hebrew. There is a correlation in both Amharic and Hebrew between the overt coding of the object and the definiteness of the nouns serving as an object, in that the object marker is deployed if the noun is morphologically or semantically determined.
All Semitic languages have a passive marker that can be added to the verb, and the function of this marker is to indicate that the subject of the clause is undergoing the event and not controlling it.
Most Semitic languages have a causative marker, whose most general function, spread across a large number of languages, is to indicate that the subject of the clause is controlling the event but not participating in the process described by the verb. Other functions of the causative, such as transitivizing, may well follow from this general function.
Most Semitic languages have prepositions, and a few have postpositions which code spatial configuration and locative function at the same time. Some of these prepositions also code non-locative functions, such as instrumental and associative.
The key to the semantic functions of participants coded in Semitic languages lies in the verbal themes. As demonstrated in Gragg and Hoberman (2012), some of these markers have clearly delineated functions, such as passive or causative, but others, such as gemination of the second consonant of the verb, have their original function considerably diluted.
Relationships Among Nouns. Modification of one noun by another, sometimes referred to as “possessive construction,” is often coded by the construct state, in which the head precedes the modifier and the head is in the construct state, coded by phonological reduction.
Reference System. Proto-Semitic did not have a definite marker or indefinite marker, but there is a consensus that mimation and nunation did play a role in the reference system. It is entirely possible that the reference system of Proto-Semitic encoded categories other than definiteness and indefiniteness. Contemporary languages have grammaticalized a definite prefix or suffix but have not grammaticalized an indefinite article.
5.4.4 Relationships Among Clauses
Semitic languages have a clausal coordinating conjunction and a conjunction coding an unexpected event, corresponding to ‘but.’ In addition, Semitic languages distinguish between “parallel event” (Gragg & Hoberman, 2012, p. 232) and sequential events.
5.5 Further Readings
The tradition of linguistic studies of Semitic languages is almost 1,000 years old, and as a result the literature on individual Semitic languages and on comparative Semitic studies is very rich. For recent studies dealing with the totality of Semitic family, see Goldenberg (2013), Gragg and Hoberman (2012), Izre’el (2002), and Lipiński (2000). For an older introduction, see Moscati (1964).
6.1 Geographical Location and Classification
Chadic languages, members of the largest (between 150 and 160 languages) family within the Afroasiatic phylum, are spoken in Northern Nigeria, Northern Cameroon, Niger, and the Republic of Chad. The largest of these languages is Hausa, spoken by at least 20 million people in Northern Nigeria and also used by Hausa immigrant communities scattered over West and Central Africa. Less than ten Chadic languages are spoken by more than 100,000 people. Many Chadic languages are endangered and are spoken by less than 1,000. At present, only about 50 Chadic languages have been described. For most of these languages, there exists only one grammar. Hausa, however, has more than a dozen grammars, written in English, French, German, Russian, and Polish.
The genetic unity of languages classified as Chadic was first postulated by Greenberg (1966a). More recent classifications divide Chadic languages into four groups, West, Central, East, and Masa (Newman, 1977a, 1990), or three groups, West, Central, and East (Jungraithmayr, 1978), with the Masa branch incorporated within the Central group. While the subclassification of Chadic languages continues to be debated, the genetic unity of the family is generally accepted and is supported by numerous sound correspondences (Newman, 1977a; Jungraithmayr & Ibriszimov, 1994). Given that only one-third of the Chadic languages have been described, and that the majority of descriptions rely on analyses proposed by a single author, we lack the alternative analyses that are the norm for better-described languages. The characteristics listed should therefore be treated as tentative.
In addition to being the largest family within the Afroasiatic phylum, Chadic languages are also typologically the most diversified family. Some languages have gender and others do not; most languages have a number distinction within the nominal system, but a few languages do not; some languages have two tense and aspectual systems, and other languages have only one system; some languages deploy the linear order Subject-Verb, and others have Verb-Subject. Verbal pieces in some languages may have up to nine segmental morphemes, like polysynthetic languages, while others have verbal pieces with at most three morphemes. There are also significant differences with respect to functions coded in those languages. It is quite possible that the reasons for significant typological differences among Chadic languages lie in the multiple language contacts that Chadic languages had on their way from some place in East Africa to their present locations in Central and West Africa.
Newman (1977a, p. 9) reconstructs the Proto-Chadic consonantal system as consisting of three series of stops, voiced, voiceless, and glottalized, similar to the three series already seen in Egyptian and Semitic. Newman also reconstructs a number of voiceless continuants, including a lateral continuant, and just one voiced alveolar continuant, z. The language also had nasal and glide sonorants, and one liquid, r. Newman postulates the following places of articulation: labial, alveolar, palatal, palatalized velar, and labialized velar. Contemporary Chadic languages display a wide variation with respect to the consonantal inventory. Some have two lateral continuants, voiceless and voiced . Some have only one labial stop. Some languages have velar and palatal nasal in addition to labial m and alveolar n.
Newman (1977a) postulates that Proto-Chadic may have had a four-vowel system consisting of i, u, a, and ə, or possibly just two vowels, a central vowel and a. Some synchronic analyses postulate two-vowel systems, at the cost of postulating additional palatal and labial prosodies that would generate five or six vowels in phonetic realizations. In many languages, mid vowels are products of vowel raising or lowering. The central high or mid vowel schwa is often epenthetic, inserted to satisfy syllable structure condition and to prevent the emergence of disallowed consonant clusters. It must be noted, however, that every vowel can serve as an epenthetic vowel.
Some Chadic languages have vowel harmony systems operating across consonants whereby vowels are fronted, rounded, or raised if the word has a high front or high round vowel. In some languages, for example in Gidar (Frajzyngier, 2008), the vowel harmony system operates regardless of whether the triggering vowel is in the stem or in the affix. Vowel harmony may have in its scope constituents larger than a word, for example a preposition and a noun following it.
From various descriptions of Chadic languages, one can conclude that vowel addition or retention has been grammaticalized to code phrasal boundaries and vowel reduction has been grammaticalized to code phrase-internal position. A somewhat similar phenomenon has been observed in Omotic languages. Vowel reduction with the first constituent in the linear order is reminiscent of the construct state of nouns in Semitic languages. In Chadic, however, vowel reduction is not limited to nouns (Frajzyngier, 2016).
All Chadic languages have phonemic tones. The most frequent is a two-tone system, but three-tone systems are also found. The role of tone is much greater in the grammatical system than in the lexicon. Functions in a variety of domains are coded solely by tonal distinctions. A small selection of these distinctions includes the distinction between the assertive mood and the mood of obligation, the distinction between completive and imperfective aspect, the distinction between singular and plural number in nouns, the distinction between a locative preposition and the preposition coding both the focus and object function, and the distinction between a locative and non-locative complement. In all Chadic languages, one can find lexical items that differ only in their tonal structure, but such minimal pairs are not frequent.
6.3 The Formal Coding Means in Chadic
In addition to tonal changes, the following formal means are deployed in the coding of a variety of functions:
Vowel deletion, as described earlier.
Vowel retention and/or vowel addition.
Lexical categories: nouns, verbs, adjectives, adverbs, prepositions, particles. In many languages, there exist morphological means to derive nouns from verbs, to derive adverbs from other lexical categories, and to derive nouns from adjectives (Wandala, Frajzyngier, 2012b).
Reduplication of the first CVC or the second CV sequence in verbal and nominal stems.
Repetition of a lexical item to derive adverbs.
Repetition of deictics and anaphors having the function of intensifiers.
Repetition of a verb, with object if any, to derive some aspectual distinctions (Gidar, Frajzyngier, 2008).
Insertion of the vowel after the first consonant of the root (with verbal themes).
Infixation of the CV sequence in the verbal root.
Suffixation in nouns and verbs, including a large number of verbal extensions.
Prefixation (considerably less frequent than suffixation).
Serial verb constructions.
Prepositions and spatial specifiers.
Particles (a large group of morphemes coding a variety of functions and occupying varied positions within the clause).
Linear orders coding a variety of functions.
6.4 Nominal Morphology
In most Chadic languages, nominal morphology differs from verbal morphology in the constraints on the structure of the verb. While verbs cannot have a vowel in word-initial position, nouns (and free-standing grammatical markers) can. While verbs can have only the consonantal root to carry the broad meaning of the lexical item, nouns may have both consonants and unpredictable internal, and sometimes final, vowels, which, along with tone, constitute the underlying form of the noun.
Prefixation is a relatively rare process with respect to nouns. One prefix, mV-, coding names of places and names of agents, has been inherited from Proto-Afroasiatic. Other contemporary prefixes are products of relatively recent grammaticalization from compounding or from determiners fusing with the noun. This is the case with prefixes marking the gender of [+human] nouns in Mupun (Frajzyngier, 1993).
Suffixation is a frequent process among nouns in Chadic that codes a variety of functions, including the coding of number and the diminutive/feminine gender.
Reduplication of a syllable codes plural number in some language, for example Hausa (Newman, 2000).
6.5 Verbal Morphology
In many languages, verbs, unlike nouns and free-standing grammatical markers, cannot begin with a vowel. For many verbs, one can postulate that the underlying (unpredictable) structure consists of consonants only. The mid or high-central schwa in verb-initial or verb-medial position is almost always prothetic or epenthetic. While there are no specific semantic functions associated with the most frequent verb-internal vowel, a, there are some functions associated with the verb-internal vowel u and the verb-internal vowel i. In some languages, u is associated with the point of view of the affected subject and i is associated with the separation of an object from its place (Hdi, Frajzyngier & Shay, 2002). The importance of this fact is that it points to separate functions of the consonantal root of the verb, which carries a broad semantic function unique to each verb, and the vowels, which add the same component of meaning for different verbs. Verb-final vowels also appear to be synchronically or diachronically grammatical markers (for a different view, see Schuh, 2017).
Infixation of the vowel a after the first consonant of the root is still a productive process in many Chadic languages, coding verbal plurality. This process has been noted in Semitic and Berber languages and reconstructed for Proto-Afroasiatic by Greenberg (1955).
In several East Chadic languages, the verbal inflection codes tense and aspectual distinctions, mainly through vocalic suffixes, which then trigger vocalic changes within the stem (Frajzyngier, 1981).
Suffixation to the verb includes subject markers coding the person, number, and gender of the subject, in languages that have a gender distinction. In some languages, for example Gidar and East Dangla, there are suffixes coding only the plurality of the subject. In all languages, there are suffixes coding the object. The coding of the object has, however, functions other than the mere indication of the presence of an object in the clause. While first- and second-person objects are marked if they are a part of the proposition, the third-person object is often unmarked. In some languages, for example in Gidar, the coding of the gender and number of the third-person object on the verb is a function within the domain of reference, indicating that the object is determined or is a proper name. In some languages, the third-person object marker is added only when it is unpredictable from the inherent properties of the verb.
A characteristic of many West and Central Chadic languages is the presence of suffixes to the verb, called “verbal extensions,” which code a variety of functions, including point of view; repetition of the event (similar to the function of the English prefix ‘re-’); partial affectedness of the object; presence of an additional argument; presence of an associative argument; and a host of spatial relations, such as movement toward a place, movement into a place, movement up, movement down, and movement away.
In some Central Chadic languages, the verb is repeated to code certain aspectual distinctions, in particular imperfective, completive, and backgrounding. In some of these aspects, verbal subject pronouns, object pronouns, and extensions are inserted between the reduplicated forms. In other aspects, subject pronouns are suffixed after the last reduplicant (Frajzyngier & Shay, 2002; Frajzyngier, 2012b).
In many West and Central Chadic languages, verbs may have possessive suffixes coding the person, gender, and number of the subject. These markers, called “intransitive copy pronouns” in Newman (1971) and “possessive subject pronouns” in Frajzyngier (2011), indeed occur with intransitive verbs. The function of these pronouns is to code the point of view of the subject rather than intransitive diathesis.
6.6 Morphology of Complementizers, Associative Markers, and Conjunctions
In Gidar (Central Chadic), the associative marker may have a suffix coding the gender and number of the second member of the associative construction. In East Dangla (East Chadic), complementizers can be combined with markers of the subject of the following clause.
6.7 Linear Orders
The majority of Chadic languages have the subject in clause-initial position, followed by the verb. Other noun phrases or prepositional phrases follow the verb. Linear order thus codes the grammatical relations of subject and object. Some Chadic languages, mainly from the Central branch, are verb-initial, with the subject following the verb. Some of these languages, for example Hdi, are also predicate-initial, where the nominal or adjectival predicate precedes the subject in the equational clause. In Frajzyngier (1983) it is proposed that in Proto-Chadic verb was in clause-initial position.
In some languages, clause-initial position is the topicalization position. In some languages, for example, Pero (West Chadic), clause-final position codes the focus on subject (Frajzyngier, 1989).
In Mupun (West Chadic), lexical adverbs in post-verbal position have an adverbial function, but in clause-initial position they function as tense markers (Frajzyngier, 1993).
In Lele (East Chadic), the third-person plural subject pronoun suffixed to the verb has an anaphoric or deictic function. The same pronoun codes the unspecified human subject when it precedes the verb.
6.8 Functions Coded in Grammatical Systems of Chadic Languages
6.8.1 Noun and Nominal Modification
Most Chadic languages code the distinction between singular and plural, with singular being the unmarked category. There are a few languages, however, in which there is no coding of number on the noun and there are no traces of such coding, for example Gwandara and Pero (Frajzyngier, 1989). In some languages, the nominal plural marker is coded syntactically, by deployment of the third-person plural pronoun at the end of the noun phrase and not at the end of the noun. Given the fact that the majority of Chadic languages have nominal plural markers, such a function should be reconstructed for Proto-Chadic.
Even in languages that have a gender distinction, nouns are not overtly marked for gender. However, gender is overtly marked in the singular subject, object, or independent pronouns. There is no gender distinction in the plural in any Chadic language. In the singular pronouns, gender is marked most often in the third-person singular, and less often in the second-person singular. Demonstratives and relative clause markers also code gender in a language with gender distinction.
Modification of one noun by another is coded by the order head modifier, with a marker of relationship between the nouns (recall that there is no case system in Chadic).
6.8.2 Relationship Between Predicate and Noun Phrases
It appears that the fundamental distinction in Chadic languages is the distinction between the point of view of the affected subject and the point of view of the goal. In West and Central Chadic languages, some verbs inherently represent the event from the point of view of the affected subject and some verbs from that of the goal. The distinction between the two points of view cuts across the distinction between transitive and intransitive verbs. For many verbs that in Indo-European languages are considered to be inherently transitive, for example ‘break,’ the inherent point of view is that of the subject undergoing a change rather than the controlling subject.
6.9 Further Readings
For a brief typological overview of the Chadic family, see Frajzyngier and Shay (2012). For a discussion of selected problems in a few West and Central Chadic languages, see Schuh (2017). About fifty descriptions of languages from all branches of Chadic have been published since 1970. A reader who would like to have a more nuanced opinion about Chadic languages is encouraged to consult at least one grammar from each branch, as Chadic languages differ significantly in their grammatical systems.
The following account of Berber is mainly based on Galand (2010) and Kossmann (2012), with important information provided by Amina Mettouchi (personal communication) and occasional information drawn from other sources. First presented are some formal means of coding available in the languages, the phonological, morphological, and linear order means. This is followed by some functions encoded in the grammatical systems.
7.1 Areas Where Berber Languages Are Spoken
The term “Berber” refers to a family of languages that are not always mutually intelligible. These languages are spoken in Mauritania (Zenaga), Morocco (Tashelhiyt, Tamazight, Riffian, Figuig), Tunisia (Djerba), Algeria (Kabyle, Mzab, Chaouia, Ourgla, Gourara), Mali (Adagh of the Ifoghas), Libya (Ghadames, Nefusa, Audjilah), Burkina Faso (Oudalan) Niger, and Egypt (Siwa). Tuareg is spoken in northern Mali, southern Algeria, and Niger (Ayer, Iwellemmeden). Some Berber languages have become extinct. The Berber languages are thus spread over a large expanse of North and West Africa. Some have undergone many changes as a result of contact with Arabic, and the lexicon of some has undergone further changes through contact with French.
The consonantal systems of Berber languages differ with respect to places of articulation and to secondary features, specifically pharyngealization, labialization, and palatalization. Common characteristics of all Berber languages include the presence of voiced and voiceless plosives and fricatives and the presence, in some languages, of pharyngealized stops and fricatives. In addition, a number of languages display a distinction between simple and long (geminated) consonants. Some scholars analyze long consonants as consisting of two phonemes, others as consisting of one phoneme. In languages that have long consonants, there exist sequences of a long consonant followed by simple consonant.
A characteristic of most contemporary Berber languages is pharyngeal consonants and pharyngealization, which phonologically correspond to the so-called emphatic consonants of Semitic languages. In several languages, the presence of a pharyngeal consonant triggers pharyngealization of the surrounding consonants and vowels.
Some Berber languages (Tashelhiyt) have three vowels, while others may have up to a ten-vowel system (Tuareg). Schwa is often inserted to form syllables, but some authors also postulate the existence of underlying schwa.
There is no overall description of syllable structure for all Berber languages, and it appears that the same speakers can syllabify the same underlying structure in different ways (cf. Galand, 2010, pp. 78–79). Berber languages have no tone.
Berber languages have the following lexical categories: nouns, verbs, independent pronouns, adverbs, prepositions, and a few postpositions. The lexical category adjective has been postulated for Zenaga by Catherine Tain-Cheikh (Amina Mettouchi, personal communication). Property concepts have often been lexicalized as nouns or stative verbs. Most underlying forms of verbs, called roots, consist of consonants only, although there are some underlying forms for which vowels must be postulated. Vocalic segments and gemination are added to erstwhile verbal forms to code a variety of functions including mood, aspect, and verbal plurality.
There are various verbal morphological means. The prefix ss derives verbs of sound from onomatopoeic nouns (Mettouchi, personal communication) and transitive verbs from intransitive verbs, as well as signaling causative function (make somebody do something). The prefix ttw, and its variants in different languages, codes the passive. For most languages, the passive does not allow agentive phrases, but it does imply the presence of an agent. In Figuig, however, there are two varieties of passive, one that implies the presence of an agent in the event and one that does not. The prefix mm, in conjunction with the verbal plural marker a, codes the reciprocal function in Moroccan Berber and codes the passive function in those varieties of Eastern Berber that have lost the prefix ttw (Kossmann, 2012, p. 37). These prefixes can be combined.
Other morphological processes applying to verbs include reduplication, which codes distributivity and verbal plurality; prefixes, suffixes, and vocalic patterns coding mood, polarity, and aspects; prefixes and suffixes coding person, gender, and number of the subject; and suffixes or clitics coding the object. Verbal nouns can be derived from verbs through a variety of derivational processes, the most common of which is the addition of nominal prefixes.
Nominal inflection consists of prefixes and suffixes marking the gender and number of the noun and marking its state, whether absolute or annexed. A large number of borrowed nouns do not have state or gender marked.
The function of the marking of state on the noun remains a controversial issue. One widespread approach is that the states are a type of case marking, albeit limited to just two cases, and that the absolutive marks the object and the annexed the subject; a second approach is that the states associate the noun in the annexed state with the coding on the verb through subject clitic, and a third is that the noun in the annexed state has only one transparent function, that of providing the variable for a grammaticalized function borne by a constituent to its left (see Mettouchi & Frajzyngier, 2013 for Kabyle). For a typological overview of the contexts in which the opposition absolute- annexed plays a role in several Berber languages, see Mettouchi (2014).
7.4 Linear Orders and Prepositional Phrases
In pragmatically unmarked verbal clauses, that is, clauses without topicalization and focus, the verb occurs in clause-initial position. The verb may be preceded by a subject clitic. In clauses with non-verbal predicates, the subject precedes the predicate. The contrasts between these default linear orders, combined with prosodic means, are deployed to code a variety of pragmatic functions, including focus and topicalization. The order of nominal arguments after the verb also indicates a pragmatic function.
7.5 Functions Coded in the Grammatical System
Modality. The unmarked clause indicates assertive modality. Negation is marked by pre-verbal negative markers in some languages and by a pre-verbal and a post-verbal negative marker in others. Polar questions can be marked by intonation alone or by a polar question particle placed in clause-initial position. Content questions are marked by the question word m, followed in some languages by markers identical with prepositions. Content questions may well be derived from constructions identical with cleft sentences.
The verb has a distinct imperative form. Other deontic modalities include wishes with respect to the first person, marked by suffixes added to the imperative form of the verb (Kossmann, 2012, p. 91). Mettouchi (2009) postulates a major division in mood in Berber between realis and irrealis. She claims that the semantic value of the perfective form in Kabyle is that of coding realis. She further claims that the traditionally postulated aspectual distinction between complete and incomplete aspect is actually between the modal categories factual and counterfactual.
Aspect. There is great variation among Berber languages with respect to the number and types of aspectual distinctions. Kossmann (2012, p. 78) considers the system in Figuig to be conservative and postulates the following aspects: aorist, which codes the same aspect as the preceding one; perfective, which codes state or a dynamic event in the past; and imperfective, which codes a variety of semantic relations ranging from an ongoing event to a habitual event. Galand (2010, pp. 208–209) postulates that the completive aspect in Berber represents the event as a whole, while the incompletive does not attribute boundaries to an event.
Grammatical relations. Within the pronominal system, Berber makes a distinction between subject and object by means of different positions and forms of the pronominal clitics. Nominal arguments are not marked for their grammatical roles. The role of the noun phrase that follows the verb in the clause is mediated through its association with pronominal clitics attached to the verb. The association with the pronominal clitics is in turn marked by the annexed state. If the noun precedes the verb, it is in the absolutive state. Mettouchi postulates for Kabyle the categories “affecting subject,” inferred from the presence of the causative marker on the verb, and “affected subject,” inferred from the presence of the passive marker on the verb.
Complex sentence structure. Berber has a clausal conjunction corresponding to ‘or’ but, like the Chadic languages, does not have a clausal coordinating conjunction corresponding to ‘and.’
7.6 Further Readings
There exists a large body of scholarship on individual Berber languages as well as on Berber in general. Recent general works on Berber include Kossmann (2012) and Galand (2010), and papers in Mettouchi (ed.) (2011). Reference to individual Berber languages can be found in Kossmann (2012), Galand (2010), and Mettouchi (2011).
The following sketch of Cushitic is based mainly on Mous (2012). Although some scholars consider Cushitic and Omotic to be one family, in the present article, as in Frajzyngier and Shay (2012) and Amha (2012), the Cushitic and Omotic families are treated separately. Nevertheless, the two families share many typological similarities, hence some information is repeated in the sections on Cushitic and Omotic.
8.1 The Areas Where Languages Are Spoken
There are currently more than 30 Cushitic languages spoken in East Africa, from Egypt to Tanzania. The northernmost Cushitic language, Beja, spoken in Egypt, Sudan, and Eritrea, constitutes a subfamily of its own. Some linguists postulate that Beja should be classified as a branch of Afroasiatic rather than as a single-language subfamily within the Cushitic family. The remaining Cushitic languages are divided into four subgroups: Agaw, Highland East-Cushitic, Lowland East Cushitic, and Southern Cushitic.
The coding means within the phonological system of Cushitic languages include consonants, vowels, and tones or accents. Unlike in tone languages, not every word has tonal distinctions. However, tones play an important role in the morphology and syntax of Cushitic languages.
Some Cushitic languages have three series of plosives, voiceless, voiced, and ejectives, paralleling the three series in other Afroasiatic languages. Some Cushitic languages have a ten-vowel system, with five short and five long vowels. Some languages have no contrast in vowel length. Somali has advanced long root vowel harmony.
8.3 Verbal and Nominal Morphology
Cushitic languages have the following lexical categories: verbs, nouns, postpositions, conjunctions, and ideophones. The existence of the category adjective is a controversial issue. In some languages property concepts are lexicalized as stative verbs and in other languages as nouns. Morphological processes include suffixation, prefixation, infixation, vocalic changes in the stem, reduplication, and tonal marking. There are derivational processes to derive nouns from verbs.
Morphological process affecting nouns include number marking and case marking. Both types of marking deserve special attention in Cushitic and Omotic languages. The unmarked form of a noun has no number value. Most languages can derive plural nouns and singular nouns through morphological means. The case system in Cushitic languages includes an absolutive case (morphologically unmarked), a nominative case (morphologically marked), a genitive case, a dative case, and a locative case. The nominative case is used with subjects of tensed clauses. In some languages, the nominative case is restricted to certain classes of nouns, for example masculine nouns in Saho, Afar, and a few other languages. The absolutive case is the citation form of the noun.
Cushitic languages have rich systems of verbal inflection coding tense, aspect, mood, evidentiality, dependent or independent status of the clause, polarity, and the person of the subject. For most verbs, the person of the subject is coded by suffixes, but some verbs in various languages code the person by prefixes. Coding by prefixes is considered to be the trace of older conjugation (Mous, 2012, pp. 391ff). A characteristic feature of many Cushitic languages is the identity of forms for the second-person feminine singular and the third-person masculine singular. Another feature, which is attributed to Cushitic origin and which spread to some Ethio-Semitic and Omotic languages of Ethiopia, is the use of the verb ‘say’ as an auxiliary following another lexical item, often an ideophone.
Another interesting characteristic of Cushitic languages is morphemes or groups of morphemes called “selectors,” “indicator particles,” or “focus markers” (Oromo). In most languages, selectors code various functions in the domain of information structure, with focus being the most important function. In most languages, selectors also mark the subject and object.
8.4 Linear Orders
Cushitic languages differ in the linear order of the head and modifier, in that in some languages the modifier precedes the head and in others it follows the head. These word orders do not correlate with the order of predicates and arguments in the verbal clause.
In verbal clauses, the verb is in clause-final position. However, an argument can be placed after the verb in clause-final position to code focus.
8.5 Functions Coded in Cushitic Languages
Number in the nominal system distinguishes between singular reference and multiple references. Bayso also has paucal number. The coding of number in Cushitic languages interacts in important way with the coding of gender, and in some languages the distinction between the two domains is blurred (Mous, 2008).
Modality. All Cushitic languages distinguish between imperative, assertive, interrogative, and negative modalities. In a number of languages, negative paradigms are not based on affirmative paradigms.
Aspect and tense. Most Cushitic languages have a distinction between perfective and imperfective marked by distinct verbal paradigms. Cushitic languages have also a distinction between past tense and present tense.
Relationships between the predicate and noun phrases. Mous (2012, p. 412) states that “[g]rammatical relations tend not to be the most central organizational principle in Cushitic syntax.” In languages that have the accusative case, this case marks the object. But not all languages have the accusative case. In a number of languages, the nominal object is negatively defined as the argument which is not the subject. While cliticized pronominal arguments distinguish between subject and object, independent pronouns do not.
The semantic relations of the subject are marked by the causative suffix s-, which introduces an argument that causes the event but does not necessarily participate in it. Cushitic languages have a suffix referred to as “middle” or “reflexive,” marked by the suffix d, which codes indirect affectedness of the subject, whether benefactive or malefactive. Some Cushitic languages have a passive marker that indicates that the subject is affected and there is an external agent. This agent is not, however, overtly marked in the clause. Reduplication of the verb is one of the means to code plurality of the event, which may imply plurality of the subject of the intransitive verb or plurality of the object of the transitive verb. Postpositions code genitive-locative, benefactive, instrumental, directional, and ablative relationships.
The present account of Omotic is based on Amha (2012), with important corrections and updates based on personal communication with Azeb Amha.
9.1 Geographic Location and Classification of Omotic
Omotic languages, currently about 28 in Fleming’s 1976 classification, are mainly spoken in South Ethiopia, with Shinasha (Boro) spoken in North Ethiopia and Hadza in southeast Sudan. The genetic status of the Omotic languages (the term was proposed by Harold Fleming) remains quite controversial. Some scholars postulate that the Omotic family is really a branch of Cushitic; some propose that it is a family within Afroasiatic on a par with other families; and some claim that it is not even a part of the Afroasiatic family, but rather a part of the Nilo-Saharan family. The differences in genetic classification mainly stem from what kind of weight is given to various data and the choice of data. In the present article, Omotic is taken to be a separate family, although one cannot avoid noticing great typological similarity between Cushitic and Omotic languages. This similarity, in turn, may well be a product of extensive language contact. The internal classification of Omotic languages as per Fleming 1976 involves major division into Eastern and Western branches. For an extensive discussion of the history of classification of Omotic see Amha (2012, pp. 425–428), and for modern subclassification within Omotic see Amha (2017).
Proto-Omotic, according to Bender (2003), had three series of stops, voiced, voiceless and glottalized, and three series of affricates, voiced, voiceless, and ejective. Fricatives had only one series, voiceless. The places of articulation were bilabial, alveolar, palatal, velar, and glottal. Contemporary Omotic languages have much richer systems of consonants. Most Omotic languages have five phonemic vowels, with short and long counterparts. There is considerable variation in tone: two-tone languages (some with a pitch-accent system because of the restriction on the distribution of the high tone, as in Ometo languages); three-tone languages (Yem, Kafa, Dizi); a four-tone language (Sheko); and a six-tone language (Benchnon) (Azeb Amha, personal communication).
Both open and closed syllables are common. In several Omotic languages there is sibilant harmony (Hayward, 1988), whereby sibilants within a word must agree with respect to the value of the palatal feature; that is, they must all be [+palatal] or [−palatal]. The scope of sibilant harmony varies across languages and may involve only the stem or both stem and affixes.
9.3 Lexical Categories and Their Morphology
Omotic languages distinguish lexical categories in nouns and verbs. Property concepts are realized as either nouns or verbs. The status of postpositions is controversial in that some analyze them as case markers.
A characteristic of Omotic languages, shared to a certain degree with Chadic languages, is that some word-final vowels in nouns are deleted when a suffix is added, while other vowels are retained. In most languages, these vowels appear not to have a morphological function. Many Omotic languages mark gender on nouns. Gender is also marked on verbs. Nouns also code case. In addition to nominative, accusative, and genitive, some languages code dative, locative, and vocative. Most languages have a two-number distinction, singular and plural. One language, Dizi, also has dual number.
Postpositions can follow case markers such as genitive and accusative, resulting, in a way, in one noun having several markers of its semantic relation to the verb. Thus, a dative marker may follow the accusative marker in Dizi.
Inflectional markers on the verb are largely suffixes, although prefixes are attested in languages from the Aroid and Dizoid groups. Root-internal changes are attested only in a few languages (in contrast with Cushitic and Semitic), where they code distinctions between aspects (perfective or imperfective) and modalities. Inflectional suffixes code tense, aspectual, and modal distinctions, as well as the person, gender, and number of the subject. Suffixes called derivational code the semantic role of the subject. In a few languages, markers coding the person, gender, and number of the subject can occur on categories other than the verb, for example adverbs of time and question words.
Omotic languages have the category converbs. These verbs are predicates of clauses preceding the main clause in a complex sentence. The use of converbs may indicate whether the subject of the preceding clause is the same as or different from the subject of the main clause.
9.4 Linear Order
In verbal clauses, the verb occurs in clause-final position. In existential, equational, and attributive predications in languages that have a copula, the copula occurs in clause-final position. In Maale and Wolaita, the linear order in verbal clauses is OSV and is used to code the focus on the subject.
In noun phrases, the head does not have to be in phrase-final position. Different word orders apply to different types of modifiers.
9.5 Functions Coded in Omotic Languages
Within the domain of modality, some Omotic languages make a distinction between affirmative and negative polarity, in that the number of tenses is greater in the affirmative (present, past, future) than in the negative (present/future and past). Within deontic modality, in most languages there is a distinction between optative (third person) and imperative (second person).
Omotic languages have a relatively rich case-marking system, including nominative, accusative, dative, ablative, elative, instrumental, and locative. In many languages, the nominative is unmarked within the clause and the accusative is marked, although often on definite nouns. In some languages, the accusative is unmarked and the nominative is marked. The dative case, when present, codes the indirectly affected argument. Within the noun phrase, the genitive case codes the possessor-possessed relationship.
In the nominal system, there is a distinction between singular and plural marking for nouns. Pronouns have singular and plural marking. Dizi and Northern Mao also code dual in the pronominal system.
10. Common Characteristics of the Phylum: Formal Means
Within the consonantal inventory, all Afroasiatic languages have at least three series of stops: voiced, voiceless, and a third series sometimes characterized as glottalized, sometimes as ejectives, and sometimes as pharyngealized.
The number of underlying vowels varies between two and seven; the choice may depend on approaches to the analysis of phonology. Some linguists postulate the existence of palatal and labial prosodies, which decreases the number of underlying vowels. For many Afroasiatic languages, linguists postulate schwa, sometimes a mid central and sometimes a high central vowel. In many languages schwa is epenthetic, serving to break down disallowed consonant clusters, or prothetic, serving to override word- and syllable-structure constraints. In some languages, however, schwa is analyzed as underlying rather than epenthetic. It is important to note that in some languages, for example in Wandala (Central Chadic), all vowels, with the possible exception of a, can be epenthetic (Frajzyngier, 2012b).
Some Afroasiatic languages display harmony with respect to some vocalic features, such as back and front or low and high, or consonantal features such as [+pharyngeal] or [+palatal]. Functions of these harmonies have yet to be described. One of the potential functions of vowel harmonies is to signal the formal entity ‘word.’
Three families within the Afroasiatic phylum, Chadic, Cushitic, and Omotic, have tones as a coding means. The role of tone in Afroasiatic languages is mainly to code grammatical functions rather than lexical distinctions. While in Chadic every word has a distinct tonal pattern, in Cushitic languages some words do not have to have a tone. In the Ometo group of Omotic, every word has to have at least one high tone. It appears that tones have developed independently since the split of the Afroasiatic phylum.
Syllables in Cushitic, Omotic, and Egyptian have to have a consonantal onset. This is also the case in most Semitic languages, but significantly not all. In Tigrinya and Amharic, syllables may have vocalic onsets. In some Chadic languages, syllables have to have consonantal onsets, but in others they can have vocalic onsets. In Cushitic and Chadic languages, a prothetic glottal stop or glottal continuant is inserted to satisfy syllable-structure conditions if the underlying form of a word begins with a vowel.
Some Cushitic, Omotic, and Chadic languages have rules of final vowel deletion. These rules operate in different ways in different languages, but one of the common characteristics is vowel deletion before a suffix is added. In some Chadic languages, final vowel retention is a mark of phrasal boundary, while deletion of the final underlying vowel is a mark of phrase-internal position. In some languages, vowel reduction has word-internal vowels in its scope.
10.2.1 Lexical Categories
All Afroasiatic languages have the lexical categories verb, noun, and adposition. The category adjective exists in some languages, but it may contain relatively few lexical items. Many property concepts are lexicalized as either nouns or verbs. In some languages, there is no morphological distinction between adjectives and adverbs.
A large number of Afroasiatic languages have the category ideophones. The number of known ideophones varies significantly from a few dozen to several hundred. In some languages, ideophones have an adverbial function, modifying a clause. In Omotic, ideophones can be combined with auxiliary verbs and serve as predicates.
Across the Afroasiatic phylum, there is a phonological distinction between verbs and other lexical categories. In Semitic and Egyptian, verbs historically had only consonants in their underlying representation, while other lexical categories could have vowels. In many Chadic languages, verbs cannot begin with an underlying vowel, while other lexical categories can.
Afroasiatic languages have prepositions and postpositions, and some have both categories (Ghadames and Awdjilah, both Berber languages). Postpositions occur in Southern Cushitic and Omotic languages. Whether a language has postpositions or prepositions depends on the lexical source of those markers. In addition to prepositions, some Chadic languages have grammaticalized a class of spatial specifiers that code spatial relations of the subject or object with respect to the locative complement. Prepositions in such constructions mark the noun phrase simply as a locative complement, without any indication of the nature of spatial relations.
10.2.2 Morphological Processes
The following morphological processes code a variety of related and unrelated functions across the languages of the phylum. The functions coded are listed later in the present section.
Prefixation is attested in Berber, Egyptian, Semitic, Cushitic, and Omotic languages. In Chadic languages, prefixation is limited to a few derivational morphemes found in nouns.
Suffixation is attested in all Afroasiatic languages in nouns, verbs, and adjectives. In some languages, even prepositions (Gidar, Frajzyngier, 2008) and complementizers (Nigerian Arabic, Owens, 1993) can have suffixes added.
Infixation, or the insertion of a morpheme between the underlying segments of another morpheme, is attested in the Berber, Semitic, Cushitic, and Chadic families. The formation of various forms of verbs through the addition of vowel patterns is a type of infixation.
Discontinuous morphemes exist in some Chadic, Berber, and Semitic languages.
Gemination of consonants is a common morphological device in all families of the phylum.
Vowel lengthening as a morphological means is attested in all languages of the phylum.
Reduplication of two or more segments, not necessarily syllables, in a sequence is a morphological process attested in all languages of the family.
Tonal changes as morphological means are attested in Chadic, Cushitic, and Omotic languages. Tonal changes are deployed to code a large variety of functions, including modality, tense, aspect, number, and semantic relations between the predicate and noun phrases.
10.2.3 Morphology of the Noun
Number. Most Afroasiatic languages code number on the noun. The most frequent distinction is between singular and plural, although in some languages there is a tripartite distinction between singular, plural, and dual. An important characteristic of nominal number marking in some Afroasiatic languages has to do with when number is actually marked. In some Chadic and Cushitic languages, number is not marked on the noun if number is marked on the verb or if the noun is preceded by a numeral larger than one. In some Cushitic languages, for example in Somali (Saeed, 2007) and Gawwada (Tosco, 2007), the basic form of the noun is unmarked, and the noun can be marked for either singulative or plural.
Case. In some languages of the Semitic, Cushitic, and Omotic families, nouns can be marked for case. In Arabic and Akkadian (Semitic), there are three cases, nominative, accusative, and genitive. In some Cushitic and Omotic languages, the number of cases is much larger and includes absolute, accusative (for definite object), genitive, dative, comitative, locative, and ablative (Appleyard, 2007, p. 487).
State. The term “state” is used in two unrelated functions. One function, observed in Semitic languages, involves a distinction between absolute state and construct state. Construct state refers to the overt coding of the head noun in the head-modifier construction. Such coding usually involves phonological reduction. The other use of the term involves some Berber languages, where the distinction is between absolutive state and annexed state. The absolutive state is the unmarked form of the noun, which does not indicate any specific function, and the annexed state marks the noun as providing the value for the variable of the function borne by the immediately preceding constituent (verb or noun).
10.2.4 Morphology of Adjectives
In Berber, Egyptian, Semitic, Cushitic, and Omotic, adjectives can be marked for gender, number, and case. In some languages, adjectives and adverbs can be morphologically derived from other lexical categories.
10.2.5 Inflectional and Derivational Categories of Verbs
The notion of root. In Semitic and Egyptian, the semantic structure of the verb is represented by the consonantal skeleton, which contains between two and three consonants. This is also the case for some verbs in various Chadic languages, where the consonantal root may consist of just one consonant or of two or three. But unlike in Egyptian and Semitic languages, verbal root in Chadic may also include vowels. The addition of vowels in Semitic provides information about grammatical functions, such as mood and the semantic role of the subject. Vocalic changes in Chadic provide information about the semantic role of the noun phrase following the verb.
Gemination of a consonant. Gemination of consonants is a formal means of coding found in all families. Languages differ with respect to which of the consonants is geminated: penultimate, first, or last. The functions of gemination differ across languages. Gemination may mark plurality of the event, the relationship between the verb and the subject (causative), and other functions.
Reduplication of a sequence of segments. In some languages, reduplication codes the frequentative or intensive character of the event. In others, it codes the plurality of the verb, and in some languages it codes aspect. Both perfective and progressive aspect can be coded by reduplication. In some languages, reduplication of one sequence of segments codes aspect and reduplication of another sequence of segments codes plurality (Wandala and Hdi, Central Chadic, Frajzyngier, 2012b; Frajzyngier & Shay, 2002).
Verbal affixes in Afroasiatic languages code grammatical relations between the verb and noun phrases; point of view; semantic relations between the verb and arguments through causative, passive, and middle markers; definiteness of the object (Tigrinya [Ethiosemitic] and Gidar [Central Chadic]); tense, aspect, and mood; and presence of an additional argument in a proposition, though not necessarily in the clause (West and Central Chadic languages).
All Afroasiatic languages have formal means to derive nouns from verbs.
10.2.6 Inflectional Categories of Prepositions and Complementizers
In some Chadic languages and in the Nigerian variety of Arabic, inflectional markers may be added to prepositions and complementizers. In Gidar (Central Chadic), a suffix coding the gender and number of the second participant may be added to the associative preposition. In East Dangla (East Chadic) and in Nigerian Arabic, a marker of the subject of the embedded clause may be added to the complementizer.
10.3 Linear Orders and Other Syntactic Coding Means
The linear order of the major components of the verbal clause varies significantly across Afroasiatic languages, within individual families, and even within individual languages. In Old Egyptian the verb was in clause-initial position. Late Egyptian had SVO word order. In some Semitic languages, such as Classical Arabic, Biblical Hebrew, and Ge’ez, the verb is in clause-initial position. In other Semitic languages, the verb follows the subject in clause-initial position (all contemporary dialects of Arabic, Modern Hebrew). In Akkadian (East Semitic) and in modern Ethio-Semitic languages (Amharic and Tigrinya), the verb is in clause-final position. The latter is attributed to contact with Cushitic and Omotic languages. Most Chadic languages have SV order, but some languages in the Central Branch are verb-initial and even predicate-initial, when the predicate is nominal. All languages of the Cushitic and Omotic branch are verb-final. The linear order in verbless clauses is subject-predicate in many languages but predicate-subject in some. In languages that have default verb-initial or verb-final order, the position of the noun before the verb or after the verb in verb-final languages is a means of coding pragmatic functions.
Both head-modifier and modifier-head orders have been observed, and they do not depend on the position of the predicate in verbal clauses.
In some Chadic languages (Gidar, Frajzyngier, 2008) the repetition of the verb, or the verb with its complement in clause-initial position, is a coding means for the progressive aspect.
Cushitic and Omotic languages have the category converbs, deployed to form complex sentences. In some languages converbs do not code the category subject, and in others they do.
Cushitic languages have a category called selectors; these are complexes of grammatical markers that may include modality, case, and subject.
In Cushitic, Omotic, and Ethio-Semitic languages, equivalents of the verbs ‘say’ and ‘do’ are deployed to form predicates for semantic notions that have been lexicalized by categories other than the verb. The verb ‘say’ is used to code intransitive predications, and the verb ‘do’ to code transitive predications.
In at least one Chadic language, Mina, the auxiliary which occurs in clause-final position coding the end of event is identical with the verb ‘say’ (Frajzyngier, Johnston, & Edwards, 2005).
11. Functional Commonalities and Differences Across the Afroasiatic Phylum
11.1 Number in Nouns and Verbs
In some languages, the unmarked form of a countable noun is inherently singular, and plurality is overtly marked. This is the case in Semitic languages. In some languages, the unmarked form of a countable noun is semantically unmarked for number, and the language may have overt markers of singulative and plural number. This is the case in some Cushitic languages (see Mous, in press). In some languages, there is no number distinction for nouns, but there is a number distinction for pronouns and for verbs, as in Pero (Frajzyngier, 1989). In some Cushitic and Chadic languages, if the number is implied by a numeral larger than one, or if the verb is coded for plurality, the plural marker on the noun does not have to appear.
In languages that have verbal plurality, the unmarked form does not necessarily indicate a singular event. The form marked for plural (labeled “pluractional” by Newman) codes plurality of the intransitive event or plurality of the object in a clause with a transitive verb. The form does not imply plurality of a transitive event as performed by many agents (Frajzyngier, 1985). Verbal plurality is coded in Afroasiatic languages by a variety of means, including the infix a, a retention from Proto-Afroasiatic; gemination and reduplication; and number of other markers.
The assertive modality is the unmarked epistemic modality in all Afroasiatic languages. Most of the languages have grammatical means to code hedging on assertion. One of such means is the use of the verb ‘say.’ At least one Cushitic language, Awngi, has the category of evidentiality. In Tigrinya, imperfectives formed with different auxiliaries convey different degrees of doubt in truth (Leslau, 1941, p. 90).
Negation as a function is expected to exist in all languages. In Afroasiatic languages, the ways of coding negation are varied. In Berber and Cushitic languages, there exists a distinction between affirmative and negative verbal forms. While in Berber the negative form of the verb co-occurs with the negative particle, in some Cushitic languages the negative verbal form alone can code negation. Afroasiatic languages also differ with respect to the number of negative markers in each language.
Within deontic modality, most Afroasiatic languages make a formal distinction between the imperative and wishes expressed to persons other than the second person, variously called jussive, optative, and subjunctive. In some languages, for example Lele (East Chadic, Frajzyngier, 2001), the same verbal form is used to express wishes to all persons. Some Berber languages have a three-way distinction between orders to the listener, wishes with respect to the speaker, and wishes with respect to the third person.
In Afroasiatic languages the category tense, the grammaticalization of time, most often does not display as rich a system of distinctions as does the grammaticalization of aspect, described in the next section. Some Omotic and Cushitic languages, such as Somali, have a three-way distinction between past, present, and future. Some Ethio-Semitic languages have a distinction between non-past and past tense in the imperfective and between present and past tense in the perfective (Meyer, 2016).
The most frequent tense distinction is between unmarked tense, with no specific time reference, and future, which is overtly marked. Some Chadic languages, for example Hdi, have a category of specific past tense, which indicates that the event happened at some specific time in the past. That specific time in the past may have been determined in previous discourse or may be expressed within the same clause. The reference to specific past contrasts with the unmarked tense, which may have in its scope an event in the present or in the past (Frajzyngier & Shay, 2002).
Unlike tense, aspect is a rich category in Afroasiatic languages. In most English descriptions, authors talk about the perfective-imperfective distinction, and in many French descriptions authors talk about the accompli-inaccompli (completed versus incompletive) distinction. The difference between these two approaches lies not only in the terms used but sometimes also in the functions described. Most descriptions of aspect are confined to the traditional approach, where aspect is a category describing the internal constituency of the event. The internal constituency of the aspect referred to as imperfective most often refers to an event that is unbounded and not completed. On the other hand, the event referred to by the aspect called perfective in some languages may involve an event that is bounded, and by implication completed, while in other languages the perfective may involve only a completed event, which does not have to be bounded.
In addition to the perfective-imperfective distinction, which in Semitic languages is encoded in the verbal stem, Afroasiatic languages also code other aspectual distinctions, including prospective, habitual (Egyptian), progressive, stative (Hdi, Central Chadic), and backgrounding (Wandala, Central Chadic).
One issue relating to aspects is their role in discourse. The backgrounding aspect in Wandala, which most often implies a completed event, actually indicates that the event supplies background information for the subsequent proposition (Frajzyngier, 2012b). Mettouchi (2015) discusses the role of aspect in discourse in Kabyle.
An important characteristic of Egyptian, Chadic, and some Cushitic languages is the presence of two tense and aspectual systems. There are two perfective, two imperfective, and two progressive forms, with identical aspectual and temporal properties. One system is used to code pragmatically independent clauses—that is, clauses that can be interpreted on their own—and the other system is used to code pragmatically dependent clauses, that is, clauses that must interpreted in connection with some other, previously mentioned, proposition or in connection with some event. Thus, content questions are pragmatically dependent, because they assume the truth of the rest of the proposition. A comment-on-topic construction is pragmatically dependent, because the clause must be interpreted in connection with the topicalized noun phrase. The use of any form from the pragmatically dependent set forces the interpretation of the clause with some other proposition or some event (Jungraithmayr, 1994; Frajzyngier, 2004).
11.5 Semantic Relations in the Clause
Semantic relations in Proto-Afroasiatic have yet to be reconstructed. For all Afroasiatic languages, linguists postulate the existence of the familiar categories subject and object. Moreover, many languages have nominative/accusative alignment. There are, however, important differences. Although in modern Ethio-Semitic, Cushitic, and Omotic languages there is the category case, the accusative is used only to mark definite objects. Some Cushitic languages have a system whereby the nominative case is marked and the absolutive is unmarked (König, 2006, 2008). Semitic, Egyptian, Berber, and Cushitic languages have the category passive, but some of these languages have more than one type of passive. For example, Berber has two passive forms, both of which indicate that the subject is affected; one implies presence of an agent in the event, and the other makes no such implication.
In addition to the well-known distinction between transitive and intransitive predications, with its assumption of the prototypical properties of the categories subject and object (Lazard, 2001), Afroasiatic languages display another distinction in which the prototypical properties play no role. This distinction may well be a remnant of the Proto-Afroasiatic verbal system, as postulated by Cohen (2014). Here are the details.
In Berber there is a large class of verbs that are characterized as inherently stative (Galand, 2010). In Wandala, most verbs of a type that in Indo-European languages are considered to describe prototypical transitive events (Lazard, 2001 and his more recent writings)—that is, events that are supposed to have controlling participants and an affected object, such as verbs referring to breaking and smashing—represent the event from the point of view of the affected argument. If one wants to add a controller to such verbs, a morphological marker must be added (Frajzyngier, 2012b; Frajzyngier & Shay, 2016). Cohen (2014, p. 15) states, and he probably is not the first one, that the basic form of the verb was stative. Further on, on p. 17, Cohen elaborates on this notion, stating: ‘Il est proposé de voir l’origine du système chamito-semitique dans une opposition apophonique entre des forms qui seront provisoirement définies comme médio-statif intransitif et processif transitif’ [‘I propose that the source of the Hamito-Semitic system is the apophonic opposition between the forms that one can, for the time being, label as middle-stative and intransitive, and the processual transitive’; translation by Z.F.]. The term “apophonic” refers to the vowel alternation a versus schwa. The fundamental issue here is which verbs fall into which class. Most verbs of a sort that in traditional Indo-European linguistics are considered as prototypically transitive, that involve a controlling subject and an affected object, in Chadic languages inherently indicate the point of view of the affected subject (Frajzyngier & Munkaila, 2004; Frajzyngier, 2012b; called “unaccusative” in Schuh, 2017). The same is true of stative verbs in Berber. Deployment of these verbs with two arguments, one of which is controlling, requires an additional morphological marker. It would appear that Proto-Afroasiatic had a basic distinction between two verbal predications: point of view of the affected argument and point of view of the goal.
11.6 Areas for Future Investigation
A number of functional areas have not yet been properly explored in a sufficient number of individual languages and therefore cannot yet be subject to any generalizations across the Afroasiatic phylum. Here are some of these areas:
Reference systems. Linguists have postulated the category of definiteness in Semitic, Cushitic, and Omotic languages. In Chadic languages, some linguists postulate the category of previous mention (Newman, 2000; Frajzyngier, 2012b), and for some languages authors postulate the category definiteness. The category indefinite is postulated much less often. Information about other functions within the system of reference in Afroasiatic languages is lacking. For an attempt to discover such functions, see Frajzyngier and Mettouchi (2015).
Relations within noun phrases. In addition to the general category of modification of a noun, which can be assumed to exist in any language, Afroasiatic languages have grammaticalized more fine-grained functions, including alienable relations, inalienable relations (often called inalienable possession), and kinship relations.
Locative predication. Chadic languages have grammaticalized the category of locative predication, distinct from all other predications in the language. As a result, nouns and verbs are either inherently [+locative] or [−locative]. In locative predications, if the noun is [+locative], no preposition is used to mark the locative complement, and if the predicate is [+locative], no other markers are required to code locative predication. If the verb is [−locative], a locative predicator must be used to code the locative predication. We do not have sufficient information about the relevant issues in other Afroasiatic languages.
Relations between clauses in complex sentences. Relations between clauses in complex sentences have for the most part been analyzed within the traditional framework of conjoining and embedding. The situation in Afroasiatic merits a comprehensive study. In Egyptian, there were no markers of relationships between clauses in a complex sentence. In Berber, there is a disjoint conjunction corresponding to ‘or.’ In neither Berber nor Chadic languages is there a clausal coordinating conjunction corresponding to ‘and.’ An in-depth analysis of functions coded within the domain of the complex sentence indicates that the types and number of functions coded are quite different from what has been proposed in traditional linguistics, based on Indo-European languages (Frajzyngier, 1996). Studies of complex sentences in other Afroasiatic families have yet to be made.
For recent overviews of Afroasiatic, see Frajzyngier (2012a, Hayward (2000), and Diakonoff (1988). For phonology see Diakonoff, Militarev, and Stolbova (1992); Belova et al. (1993); and Ehret (1995). For a collection of recent papers on Afroasiatic languages, see Mettouchi and Lonnet (2006) and Lonnet and Mettouchi (2005). Frajzyngier and Shay (2012), Mettouchi and Lonnet (2006), and Lonnet and Mettouchi (2005) contain extensive bibliographies of works related to Afroasiatic languages. For an older survey, see Hodge (1971).
- Allen, J. P. (2014). Middle Egyptian: An introduction to the language and culture of hieroglyphs. Cambridge, UK: Cambridge University Press.
- Amha, A. (2012). Omotic. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 423–504). Cambridge, UK: Cambridge University Press.
- Amha, A. (2017). The Omotic language family. In R. M. W. Dixon & A. Aikhenvald (Eds.), The Cambridge handbook of linguistic typology (pp. 782–814). Cambridge, UK: Cambridge University Press.
- Appleyard, D. L. (2007). Bilin morphology. In A. S. Kaye (Ed.), Morphologies of Asia and Africa (Vol. 1, pp. 481–504).Winona Lake, IN: Eisenbrauns.
- Belova, A., Diakonoff, I., Militarev, A., Porkhomovsky, V., & Stolbova, O. (1993). On the principles of Afrasian phonological reconstruction. St. Petersburg Journal of African Studies, 1, 7–15.
- Belova, A., Diakonoff, I., Militarev, A., Porkhomovsky, V., & Stolbova, O. (1994–1997). Historical-comparative vocabulary of Afrasian. St. Petersburg Journal of African Studies, 2, 5–28; 3, 5–26; 4, 7–38; 5, 4–32; 6, 12–35.
- Bender, M. L. (2003). Omotic lexicon and phonology. Carbondale: Southern Illinois University.
- Bubenik, V. (2017). Development of tense/aspect in Semitic in the context of Afro-Asiatic languages. Amsterdam: Benjamins.
- Cohen, D. (2014). L’hypothèse d’une laryngale dans la morphogenèse du système verbal des langues chamito-sémitiques. Quaderni di Studi Arabi, 9, 11–20.
- Cohen, M. (1947). Essai comparatif sur le vocabulaire et la phonétique du chamito-sémitique. Paris: Champion.
- Cohen, M. (1953). Sémitique, égyptien, libyco-berbère, couchitique et méthode comparative. Bibliotheca Orientalis, 10(3/4), 88–90.
- Diakonoff, I. M. (1965). Semito-Hamitic languages: An essay in classification. Moscow: Academy of Sciences of the USSR.
- Diakonoff, I. M. (1975). Lingvisticheskiye dannyye k istorii drevneyshikh nositeley afraziyskikh yazykov. Africana: Afrikanskiy etnograficheskiy sbornik, 10, 117–130.
- Diakonoff, I. M. (1980). Afro-Asiatic languages. In Encyclopedia Britannica (Vol. 8, pp. 589–598). Chicago: Encyclopaedia Britannica.
- Diakonoff, I. M. (1988). Afrasian languages. Moscow: Nauka.
- Diakonoff, I. M., Stolbova, O., & Militarev, A. (1992). Proto-Afrasian and Old Akkadian: A Study in Historical Phonetics. Princeton: Institute of Semitic Studies.
- Diakonoff, I. M. (1998). The earliest Semitic society: Linguistic data. Journal of Semitic Studies, 43(2), 209–219.
- Diamond, J., & Bellwood, P. (2003). Farmers and their languages: The first expansions. Science, 300(5619), 597–603.
- Ehret, C. (1995). Reconstructing Proto-Afroasiatic (Proto-Afrasian): Vowels, tone, consonants, and vocabulary. University of California Publications in Linguistics 126. Berkeley: University of California Press.
- Ehret, C., Keita, S. O. Y., & Newman, P. (2004). The origins of Afroasiatic [letter to the editor]. Science, 3, 1680–1681.
- Fleming, H. (1976). Omotic overview. In L. M. Bender (Ed.), The Non-Semitic Languages of Ethiopia (pp. 299–323). East Lansing, MI: Michigan State University.
- Frajzyngier, Z. (1981). Some rules concerning vowels in Chadic. Bulletin of the School of Oriental and African Studies, 44(2), 334–348.
- Frajzyngier, Z. (1983). Marking Syntactic Relations in Proto-Chadic. In E. Wolff & H. Meyer-Bahlburg (Eds.), Studies in Chadic and Afro-Asiatic Linguistics (pp. 115–138). Hamburg: Buske.
- Frajzyngier, Z. (1985). Ergativity, number, and agreement. In M. Niepokuj, M. VanClay, V. Nikiforidou, & D. Feder (Eds.), Proceedings of the eleventh annual meeting of the Berkeley Linguistic Society (pp. 96–106). Berkeley: Berkeley Linguistics Society.
- Frajzyngier, Z. (1989). A grammar of Pero. Berlin: Reimer.
- Frajzyngier, Z. (1993). A grammar of Mupun. Berlin: Reimer.
- Frajzyngier, Z. (1996). Grammaticalization of the complex sentence: A case study in Chadic. Amsterdam: Benjamins.
- Frajzyngier, Z. (2001). A Grammar of Lele. Stanford Monographs in African Linguistics. Stanford: CSLI.
- Frajzyngier, Z. (2004). Tense and aspect as coding means. Journal of West African Languages, 30(2), 53–67.
- Frajzyngier, Z. (2008). A grammar of Gidar. Frankfurt: Peter Lang.
- Frajzyngier, Z. (2011). Possessive subject pronouns in Wandala: Point of view of the subject. In A. Storch, G. G. Atindogbé, & R. M. Blench (Eds.), Copy pronouns: Case studies from African languages (pp. 47–64). Cologne: Koeppe.
- Frajzyngier, Z. (2012a). Coding relations between the verb and noun phrases in Afroasiatic: A sketch of typological explanations. In M. Brenzinger & A.-M. Fenn (Eds.), Proceedings of the 6th World Congress of African Linguistics, Cologne, 17–21 August 2009 (pp. 57–68). Cologne: Koeppe.
- Frajzyngier, Z. (2012b). A grammar of Wandala. Berlin: Mouton de Gruyter.
- Frajzyngier, Z. (2012c). Typological outline of Afroasiatic languages. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 505–624). Cambridge, UK: Cambridge University Press.
- Frajzyngier, Z. (2016). Inflectional markers of sentential parsing. Lingua, 183, 1–33.
- Frajzyngier, Z., Johnston E., with Edwards, A. (2005). A Grammar of Mina. Berlin/New York: Mouton de Gruyter.
- Frajzyngier, Z., & Mettouchi, A. (2015). Functional domains and cross-linguistic comparability. In A. Mettouchi, M. Vanhove, & D. Caubet (Eds.), Corpus-based studies of lesser-described languages: The CorpAfroAs corpus of spoken AfroAsiatic languages (pp. 257–278). Amsterdam: Benjamins.
- Frajzyngier, Z., & Munkaila, M. (2004). Grammatical and semantic relations in Hausa: ‘Point of view’ ‘goal’ and ‘affected object.’ Cologne: Koeppe.
- Frajzyngier Z., & Shay, E. (2002). A grammar of Hdi. Berlin: Mouton de Gruyter.
- Frajzyngier, Z., & Shay, E. (2012). Chadic languages. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 236–341). Cambridge, UK: Cambridge University Press.
- Frajzyngier, Z., & Shay, E. (Eds.). (2012). The Afroasiatic languages. Cambridge, UK: Cambridge University Press.
- Frajzyngier, Z. with Shay, E. (2016). The role of functions in syntax: a unified approach to language theory, description, and typology. Benjamins: Amsterdam.
- Galand, L. (2010). Regards sur le berbère. Milan: Centro Studi Camito-Semitici.
- Goldenberg, G. (2013). Semitic languages. Oxford: Oxford University Press.
- Gragg, G., & Hoberman, R. (2012). Semitic. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 145–235). Cambridge, UK: Cambridge University Press.
- Greenberg, J. H. (1950). Studies in African linguistic classification: IV. Hamito-Semitic. Southwestern Journal of Anthropology, 6, 47–63.
- Greenberg, J. H. (1955). Internal a-plurals in Afroasiatic (Hamito-Semitic). In J. Lukas (Ed.), Afrikanistische Studien (pp. 198–204). Berlin: Akademie Verlag.
- Greenberg, J. H. (1963). The languages of Africa. Bloomington: Indiana University Research Center in Anthropology, Folklore and Linguistics.
- Greenberg, J. H. (1966). The Languages of Africa (2nd ed. with additions and corrections). Bloomington: Indiana University.
- Hayward, R. J. (1988). Remarks on the Omotic sibilants. In M. Bechhaus-Gerst & F. Serzisko (Eds.), Cushitic and Omotic: Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, January 6–9, 1986 (pp. 263–299). Hamburg: Helmut Buske.
- Hayward, R. (2000). Afroasiatic. In B. Heine & D. Nurse (Eds.), African languages: An introduction (pp. 74–98). Cambridge, UK: Cambridge University Press.
- Hayward, R. (2003). Omotic: The “empty quarter” of Afroasiatic linguistics. In J. Lecarme (Ed.), Research of Afroasiatic grammar II: Selected papers from the Fifth Conference on Afroasiatic Languages, Paris, 2000. Amsterdam: Benjamins.
- Heath, J. (2005). A grammar of Tamashek (Tuareg of Mali). Berlin: Mouton de Gruyter.
- Heine, B., & Nurse, D. (Eds.). (2000). African languages: An introduction. Cambridge, UK: Cambridge University Press.
- Hetzron, R. (1974). La division des langues sémitiques. In A. Caquot & D. Cohen (Eds.), Actes du premier congrès international de linguistique sémitique et chamito-sémitique, Paris 16–19 juillet, 1969 (pp. 181–194). The Hague: Mouton.
- Hetzron, R. (1987). Afroasiatic languages. In B. Comrie (Ed.), The world's major languages (pp. 645–653). New York: Oxford University Press.
- Hodge, C. T. (1970). The linguistic cycle. Language Sciences, 13, 1–7.
- Hodge, C. T. (Ed.). (1971). Afroasiatic: A survey. Janua Linguarum Series Practica 163. The Hague: Mouton.
- Hodge, C. T. (1975). Lisramic II. Anthropological Linguistics, 17(5), 237–272.
- Hodge, C. T. (2004 ). Lisramic (Afroasiatic): An overview. In S. Noegel & A. S. Kaye (Eds.), Afroasiatic linguistics, Semitics, and Egyptology: Selected writings of Carleton T. Hodge (pp. 75–100). Bethesda, MD: CDL Press.
- Izre'el, S. (Ed.). (2002). Semitic linguistics: The state of the art at the turn of the twenty-first century. Winona Lake, IN: Eisenbrauns.
- Joüon, P. (1947). Grammaire de l’hébreu biblique. Rome: Pontifical Institute.
- Joüon, P., & Muraoka, T. (2006). A grammar of Biblical Hebrew (2nd ed.). Rome: Pontificio Istituto Biblico.
- Junge, Friedrich. (2001). Late Egyptian Grammar. An Introduction. Translated from the German by David Warburton. Oxford: Griffith Institute.
- Jungraithmayr, H. (1978). Les langues tchadiques et le proto-tchadique: documentation, analyse et problèmes. In J.-P. Caprile & H. Jungraithmayr (Eds.), Préalables à la reconstruction du proto-tchadique (pp. 17–30). Paris: SELAF.
- Jungraithmayr, H. (1994). “Zweite Tempora” in afrikanischen Sprachen—Ägyptisch-Tschadische Gemeinsemkeiten? In M. Bietak, J. Holaubek, H. Mukarowsky, & H. Satzinger (Eds.), Zwischen den beiden Ewigkeiten: Festschrift Gertrud Thausing (pp. 101–122). Vienna: Institute für Ägyptologie der Universität Wien.
- Jungraithmayr, H., & Ibriszimov, D. (1994). Chadic lexical roots. Berlin: Reimer.
- Kaye, A. S. (Ed.). (2007). Morphologies of Asia and Africa (Vol. 1).Winona Lake, IN: Eisenbrauns.
- König, C. (2006). Marked nominative in Africa. Studies in Language, 30(4), 655–732.
- König, C. (2008). Case in Africa. Oxford: Oxford University Press.
- Kossmann, M. (1997). Grammaire du parler berbère de Figuig (Maroc oriental). Paris: Peeters.
- Kossmann, M. (1999). Essai sur la phonologie du proto-berbère. Cologne: Rüdiger Köppe.
- Kossmann, M. (2000). Esquisse grammaticale du rifain oriental. Paris: Peeters.
- Kossmann, M. (2012). Berber. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 18–101). Cambridge, UK: Cambridge University Press.
- Kuryłowicz, J. (1973). Studies in Semitic metrics and grammar. London: Curzon.
- Layton, B. (2011). A Coptic grammar (3rd ed.). Wiesbaden: Harrassowitz.
- Lambdin, T. O. (1983). Introduction to Sahidic Coptic. Macon: Mercer University Press.
- Lazard, G. (2001). La typologie actantielle. In G. Lazard (Ed.), Études de linguistique générale: Typologie grammaticale (pp. 65–77). Leuven: Peeters.
- Lecarme, J. (Ed.). (2003). Research in Afroasiatic grammar II. Amsterdam: Benjamins.
- Lecarme, J., Lowenstamm, J., & Shlonsky, U. (Eds.). (2000). Research in Afroasiatic grammar. Amsterdam: Benjamins.
- Lepsius, R. (1981 ). Nubische Grammatik, mit einer Einleitung über die Völker und Sprachen Afrika’s. Wiesbaden: LTR-Verlag.
- Leslau, W. (1941). Documents tigrigna (éthiopien septentrional): Grammaire et textes. Paris: C. Klincksieck.
- Lipiński, E. (2000). Semitic languages: Outline of a comparative grammar (2nd ed.). Leuven: Peeters.
- Lonnet, A., & Mettouchi, A. (Eds.). (2005). Les langues chamito-sémitiques (afro-asiatiques) (Vol. 1). Paris: Ophrys.
- Loprieno, A. (1995). Ancient Egyptian: A linguistic introduction. Cambridge, UK: Cambridge University Press.
- Loprieno, A., & Müller, M. (2012). Ancient Egyptian and Coptic. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 102–144). Cambridge, UK: Cambridge University Press.
- Lottner, C. (1860–1861). On sisterfamilies of languages, especially those connected with the Semitic family. Transactions of the Philological Society, 20–27 and 112–132.
- Mettouchi, A. (2009). Mood and modality in Berber. In B. Hansen, F. de Haan, & J. van der Auwera (Eds.), Modals in the languages of Europe (pp. 431–456). Berlin: Mouton de Gruyter.
- Mettouchi, A. (Ed.). (2011). Parcours berbères: Mélanges offerts à Paulette Galand-Pernet et Lionel Galand pour leur 90e anniversaire. Berber Studies 33. Cologne: Rüdiger Köppe.
- Mettouchi, A. (2014). Foundations for a typology of the annexed/absolute state systems in Berber. STUF, 67(1), 47–61.
- Mettouchi, A. (2015). Aspect-Mood and discourse in Kabyle (Berber) spoken narratives. In D. Payne & S. Shirtz (Eds.), Beyond Aspect: The expression of discourse functions in African languages (pp. 117–144). John Benjamins: Amsterdam-Philadelphia.
- Mettouchi, A., & Frajzyngier, Z. (2013). A previously unrecognized typological category: The state distinction in Kabyle (Berber). Linguistic Typology, 17(1), 1–30.
- Mettouchi, A., & Lonnet, A. (Eds.). (2006). Les langues chamito-sémitiques: Afro-asiatiques (Vol. 2). Faits de langues 27. Paris: Ophrys.
- Meyer, R. (2016). Aspect and tense in Ethiosemitic languages. In L. Edzard (Ed.), The morpho-syntactic and lexical encoding of tense and aspect in Semitic (pp. 159–239). Wiesbaden: Harrassowitz.
- Militarev, A., & Shnirelman, V. (1984). Towards the problem of locating the early Afrasian speakers, preprinted abstracts and papers of the 1st national conference “Language Reconstruction and Prehistory of the Orient” (pp. 35–53, part 2) (in Russian). Moscow.
- Moscati, S. (Ed.). (1964). An introduction to the comparative grammar of the Semitic languages. Wiesbaden: Harrassowitz.
- Mous, M. (2008). Number as an exponent of gender in Cushitic. In Z. Frajzyngier & E. Shay (Eds.), Interaction of morphology and syntax: Case studies in Afroasiatic (pp. 137–160). Typological Studies in Language 75. Amsterdam: John Benjamins.
- Mous, M. (2012). Cushitic. In Z. Frajzyngier & E. Shay (Eds.), The Afroasiatic languages (pp. 342–422). Cambridge, UK: Cambridge University Press.
- Mous, M. (in press). Number in Cushitic. In P. Cabredo Hofherr & J. Doetjes (Eds.), Oxford handbook of grammatical number. Oxford: Oxford University Press.
- Müller, F. (1876). Grundriss der Sprachwissenschaft. Vienna.
- Newman, P. (1971). Transitive and intransitive in Chadic languages. In V. Six, N. Cyffer, E. Wolff, L. Gerhardt, & H. Meyer-Bahburg (Eds.), Afrikanische Sprachen und Kulturen—Ein Querschnitt: Festschrift J. Lukas (pp. 188–200). Hamburger Beiträge zur Afrika-Kunde 14. Hamburg: Deutsches Institut für Afrikaforschung.
- Newman, P. (1977a). Chadic classification and reconstructions. Afroasiatic Linguistics, 5(1), 1–42.
- Newman, P. (1980). The classification of Chadic within Afro-Asiatic. Leiden: Universitaire Pers Leiden.
- Newman, P. (1990). Nominal and Verbal Plurality in Chadic (Publications in African Languages and Linguistics, 12). Dordrecht: Foris Publications.
- Newman, P. (2000). The Hausa Language: An Encyclopedic Reference Grammar. New Haven: Yale University Press.
- Orel, V. E., & Stolbova, O. V. (1995). Hamito-Semitic etymological dictionary: Materials for a reconstruction. Leiden: Brill.
- Owens, J. (1993). A grammar of Nigerian Arabic. Wiesbaden: Harrassowitz.
- Owens, J. (1998). Case and Proto-Arabic. Bulletin of the School of Oriental and African Studies, 61, 51–73 and 217–27.
- Saeed, J. I. (2007). Somali morphology. In A. S. Kaye (Ed.), Morphologies of Asia and Africa (Vol. 1, pp. 547–586). Winona Lake, IN: Eisenbrauns.
- Schuh, R. G. (2017). A Chadic cornucopia. Ed. Paul Newman. eScholarship California Digital Library Oakland, California.
- Tosco, M. (2000). Cushitic overview. Journal of Ethiopian Studies, 33(2), 87–121.
- Tosco, M. (2003). Cushitic and Omotic overview. In M. L. Bender (Ed.), Omotic lexicon and phonology (pp. 87–92). Carbondale: Southern Illinois University.
- Tosco, M. (2007). Gawwada morphology. In A. S. Kaye (Ed.), Morphologies of Asia and Africa (Vol. 1, pp. 505–528). Winona Lake, IN: Eisenbrauns.
- Tucker, A. N. (1967). Erythraic elements and patternings: some Eat African findings. African Language Review, 6, 17–25.
1. Michael Avina and Antonio Loprieno provided very useful comments on Egyptian. Amina Mettouchi has kindly commented on a variety of issues and corrected my mistakes in the discussion of Berber. Maarten Mous has commented on Cushitic languages, and Azeb Amha has commented on Omotic languages. Erin Shay has read the whole text and made numerous substantial and editorial comments. I alone am responsible for any errors and infelicities contained in this article.