LINGUIST List 32.2439

Tue Jul 20 2021

Review: Afroasiatic; General Linguistics; Historical Linguistics; Typology: Frajzyngier, Shay (2020)

Editor for this issue: Jeremy Coburn <>

Date: 19-Jun-2021
From: Geoffrey Sampson <>
Subject: The Afroasiatic Languages
E-mail this message to a friend

Discuss this message

Book announced at

EDITOR: Zygmunt Frajzyngier
EDITOR: Erin Shay
TITLE: The Afroasiatic Languages
PUBLISHER: Cambridge University Press
YEAR: 2020

REVIEWER: Geoffrey Sampson, University of Sussex


The Afroasiatic language family (the name was coined by Joseph Greenberg in 1950, an earlier term was Hamito-Semitic) is the fourth or fifth largest in the world by number of current speakers, after Indo-European, Trans-Himalayan (or “Sino-Tibetan”), and Niger-Congo, and includes several hundred languages. (Estimated speaker numbers for Afroasiatic and Austronesian are too close to say for sure which is larger.) Afroasiatic languages are spoken throughout the Middle East and North Africa, as far south in Central Africa as northern Nigeria and southern Chad, in East Africa down through the Horn into northern Kenya (and also on Malta, which is politically part of Europe). The family includes languages of ancient high civilizations (Ancient Egyptian, Hebrew, more recently Arabic); a language like Hausa, lacking a long history of scholarship but currently with over twenty million native speakers and in which newspapers, books, and videos appear; and many languages which have never been described or for which only a single description is available, and which, as the editors put it, are “threatened with extinction because they are spoken by a small number of people in economically and politically unstable environments”. Many of the less-studied languages have no one standard name. I was startled to read of a language called Oromo with more than twenty million speakers in Ethiopia and Kenya, of which I had never heard, but then I noticed that Oromo is also called Galla (and indeed the editors’ introduction uses that name); I had heard of Galla.

This book offers a thorough survey of the family, emphasizing typological characteristics rather than historical developments (the recorded history of most Afroasiatic languages stretches back less than a hundred years). Contributors treat the family as having six top-level subfamilies: Berber, Chadic, Egyptian, Cushitic, Omotic, and Semitic. The only controversy there relates to Omotic, a group of 25 or so languages of southern and western Ethiopia, studied only since about 1940; some authorities believe that these languages, or some of them, are genetically unrelated to the Afroasiatic family, while others believe they should be seen as part of the Cushitic subfamily. Christopher Ehret, whose family-tree structure for Afroasiatic is displayed in the introductory chapter, believes that the first split in the Afroasiatic clade was between Omotic and the rest. (That would presumably imply that the family originated in Africa and spread into the Middle East, and this is what some believe, though others including Diamond and Bellwood (2003), referring to archaeological and botanical rather than linguistic evidence, have argued for a Middle Eastern origin.)

I am describing Afroasiatic as a language “family”, because that is the usual term for a maximally-inclusive grouping of languages sharing a common ancestor. These editors prefer the word “phylum”, on the grounds that relationships between the branches are more remote than is characteristic for better-known language families. They refer to a claim by Igor Diakonoff that the Chadic branch alone “is more typologically diverse than the entire Indo-European language family”. To my mind, borrowing a high-level clade term like “phylum” from the vocabulary of biological evolution is misleading, because it suggests that all languages of the world, like all living species, ultimately descend from a shared ancestor. Nothing in this book alludes to that idea, which few linguists take seriously.

The book comprises eight chapters: an introduction by the editors; one chapter on each of the six subfamilies listed above; and a closing chapter by the senior editor, “Typological outline of the Afroasiatic phylum”. The Chadic chapter was written by the editors; authors of the chapters on other subfamilies, many of them distinguished scholars, are: Maarten Kossmann (Berber); Antonio Loprieno and Matthias Müller (Ancient Egyptian and Coptic); Gene Gragg and Robert Hoberman (Semitic); Maarten Mous (Cushitic); and Azeb Amha (Omotic). In terms of institutional affiliation, four authors are American, three Dutch (Leiden), and two Swiss (Basel). It is worth noting that although the book has been offered for review in the year I am writing, 2021, it was published nine years earlier, with some contributions written a few years before that. Until I noticed the publication date I was puzzled by remarks on p. 8 suggesting that no satisfactory existing description of the Semitic subfamily was available, but they were written before Goldenberg (2013) appeared.

Many chapters cover the structures of the languages in the respective subfamilies at a level of detail, particularly with respect to morphology and syntax, that in a description of a single language would deserve to be seen as more than a “grammar sketch”; “concise grammar” would be a more suitable phrase.

One central concern of the introductory chapter is the evidence for seeing Afroasiatic as a single family or “phylum”. This is a particularly significant issue in view of the remoteness of the relationships between branches and great typological variation within some branches, and also because the editors and other contributors identify quite a number of features as likely to have entered some or all Afroasiatic languages through contact with neighbouring languages belonging to other families. (For instance, many though not all Afroasiatic languages are tone languages – the Omotic language Bench has as many as five level tones and a contour tone – but Azeb Amha points out that the most highly tonal Omotic languages are the ones in closest contact with Nilo-Saharan languages, also tonal.) The editors write that all Afroasiatic languages “have had extended contacts with … languages belonging to other families”, and “in most instances we are unable to state categorically which typological features are due to language contact”. Setting aside the debate about Omotic, it is evidently uncontroversial among the experts that the remaining languages do descend from a common ancestor – the editors say “the genetic unity of the phylum … is not in doubt” – but it is not obvious where this certainty comes from. Introductions to the Indo-European languages commonly make their family relationship look plausible by showing sets of visibly-cognate words for some basic concepts in different Indo-European languages, and perhaps by offering a few regular phoneme correspondences such as those which result from Grimm’s Law. We are given nothing like that here, and it may be that it would be difficult to assemble straightforward cognate sets when relationships between the branches of Afroasiatic are so distant. As someone familiar with only one Afroasiatic language, Biblical Hebrew, I certainly never noticed any cognate lexical roots in example sentences outside the Semitic chapter, though the examples often involved the kind of basic vocabulary which is sometimes stable over millennia.

Morphology was another matter, and from what the editors write, it seems as though the strongest evidence for the family relationship may be morphological. For instance, they identify as one particularly convincing piece of evidence the “internal ‘a’-plurals occurring in all Afroasiatic languages” – a Hebrew example would be the root /malk/, “king”, which in the plural becomes /malak-/, giving surface [mlåx-im] rather than regular *[malk-im]. (In the scholarship of some of the languages, though not of Hebrew, these are called “broken plurals”.) And in his closing chapter, Zygmunt Frajzyngier identifies as “the central characteristic of the phylum” the phonological fact that in addition to stops, continuants, and sometimes affricates, all Afroasiatic languages have one or more consonant series representing the types ejective, implosive, glottalized, or pharygealized. (When Biblical Hebrew was a living language, the letters teth, tsade, and qoph stood for ejective consonants, and Proto-Semitic had further ejectives which in Hebrew had merged with other sounds by the Biblical period, see Goldenberg 2013: 64–5, 71–2.) The impression that comes across is that genetic relationship is evidenced by a large number of individual facts most of which might, taken separately, prove very little but which cumulatively, for those knowledgeable enough to weigh them all up, make a case which feels decisive.

The final chapter is essentially a summary of what preceded, which could be read as a standalone introduction to the typological features of Afroasiatic languages, emphasizing those which are unusual on a world scale, and referring to individual chapters for those who wish to follow up particular points in more detail. I wondered about the statement on p. 620 that phonological coding of phrase-internal versus phrase-final position of words is found only in the Cushitic and Chadic branches. Is the Hebrew phenomenon of pre-pausal word forms (e.g. Ɂérec “land” becomes Ɂárec in pause) not similar?


This is an excellent survey of a large and under-studied group of languages. It contains a wealth of material that will be of interest to linguists generally, not just to specialists in its area. Some of the languages covered contain phenomena which seem unlike any in better-known languages, or involve unexpected twists on familiar categories or processes; I shall give a few examples.

Many Chadic languages encode a grammatical category “point of view of the subject”, which is “unrelated to the semantic and grammatical relations in the proposition”, but “Depending on the nature of the verb and its arguments, … may imply physical affectedness of the subject, beneficial or detrimental effects on the subject, the subject’s interest in the event, and other functions.” In some examples, use of this category produces a meaning that a European language might encode with a reflexive pronoun or middle voice, but for instance, in an example from the language Mina, the grammatical particle in question applied to a clause meaning “They have all been killed by God” implies that the speaker sympathizes; if “they” were enemies, a different particle would be used. (I have glossed the proposition in the passive rather than the authors’ active, because in the latter case the subject would be “God” rather than “they”.)

A contrast between alienable and inalienable possession is found in many languages, but in Gidar, another Chadic language, we find “X’s husband” coded as inalienable possession but “X’s wife” coded as alienable. (We are not told whether there is a correlation with marriage customs in this speech-community.)

Many languages have vowel-harmony, but a number of Omotic languages also have sibilant harmony: a root can contain palatal sibilants like š, č, or ž, or non-palatals like s or z, but may not mix palatals with non-palatals – and in some of these languages, the nature of sibilant(s) in a root determines which sibilant occurs in an affix. And vowel harmony sometimes operates in unexpected ways. In Gidar, a front or back high vowel causes a preceding /a/ to be raised to /e/ or /o/ respectively, but this is blocked by an intervening sonorant consonant whose place of articulation is similar to the relevant vowel: /n/ is made at the front of the mouth and therefore is a blocker for /a > e/. The word for “rat” is pronounced /máŋdwàŋ/ by some speakers but /mándwàn/ by others, so with the plural suffix /ɗé/ “rats” is [méŋdwèŋɗé] for one group but [mándwànɗé] for the other. This is a fascinating variation on sound harmony, even though Frajzyngier and Shay’s description as they word it does not quite work: they say the triggering vowel must be fully high, /i/ or /u/, but no such vowel occurs in their “rats” example. (They also mistakenly call /n/ a “high” rather than “front” consonant, but Frajzyngier rectifies that error when he gives another example of the same phenomenon on pp. 518–19 of his closing chapter.)

Many Afroasiatic languages have a two-gender system, but unlike European languages which assign nouns denoting inanimate objects to masculine or feminine gender apparently randomly, in some of these languages all inanimates are assigned to the same “default” gender, which is masculine for some languages and feminine for others. (Furthermore even nouns for animates are not always assigned to the “right” gender.) In some Cushitic languages, the two genders only derivatively relate to sex, with the fundamental semantic contrast being significance versus insignificance.

One problem which often besets scholars presenting their specialist subject to a wider audience is failure to appreciate that ideas which are very familiar to themselves may not mean anything to their intended readership. The longer-established the discipline, the greater the danger of this, so it is no surprise that the chapter where I found repeated difficulties of this kind was the one on Ancient Egyptian and Coptic. For instance, we are told that “In Ptolemaic and Roman times … an increasing consciousness of the symbolic potential inherent in the relation between hieroglyphic signs and semantic meanings led to the development of previously unknown phonetic values and also of so-called ‘cryptographic solutions’.” This is pretty mysterious to me, and what follows does not clarify it. We also read that a certain evolution in the usage of hieroglyphic signs “developed dramatically in Ptolemaic times, leading to a radical change in the laws regulating the use of hieroglyphics” – does this mean that hieroglyphic writing was controlled by actual laws in the sense of statutes, or is it just a picturesque way of referring to evolving scribal conventions? (And I wondered, on p. 114, in which varieties of English “aspiration appear[s] only in stressed syllables”.) In the Semitic chapter, the dual inflexion for Akkadian nouns is said to involve “a characteristic final ‘-n’ (nunation?)”. “Nun” is the name of the letter /n/ in the Hebrew alphabet, but I do not know what is implied by the abstract noun “nunation”, nor what issue is being raised by the question mark. In the Chadic chapter there are repeated unexplained references to “vehicular” languages – from the context I guess this may be a term for lingua francas, but if so it is new to me.

A term used by several contributors which I found opaque is “polarity” – some of the authors used “polarity-switching”, which is much clearer. One example is a feature of Hebrew that must have surprised generations of students: numerals agree with the gender of their head nouns “the wrong way round”. Before a masculine plural a numeral takes a form with the normally feminine ending -a, before a feminine plural it takes a form which looks like a masculine (e.g. šišša banim “six sons”, šeš banot “six daughters”). It turns out that other Afroasiatic languages, not in the Semitic branch, show related but distinct phenomena, and some of these give clues to how such an odd state of affairs may have arisen. Cushitic, according to Ehret’s family-tree diagram, is the most distant branch other than Omotic from Semitic, but we are told here that “Cushitic gender … is a property of the word and not of the lexeme … [e.g.] In Somali a large number of nouns have the opposite gender in singular reference—multiple reference pairs.”

The term “polarity” apparently originated with Carl Meinhof; for a detailed treatment Maarten Mous refers readers to Hetzron (1967). The latter is a paper I do not fully understand, but part of Hetzron’s explanation is that number was originally a grammatical category closely linked to, and subsidiary to, gender, with the plural of a masculine noun being thought of as a feminine singular – a collectivity rather than a set of separate entities. It was therefore natural for such a noun to take a feminine numeral form. Later, the situation was obscured in Hebrew because masculine plural nouns were reanalysed as genuinely plural and masculine and mostly given the distinctively masculine plural ending -im (though quite a few Hebrew masculine nouns do take a feminine suffix -ot in the plural, e.g. šem, šemot “name, names”); but numeral usage was not changed to match. (Hetzron seems to believe that using apparently-masculine numerals with feminine plurals was a natural corollary of using apparently-feminine numerals with masculine plurals.) I am not sure whether this theory is satisfactory (it does nothing to explain the parallel phenomenon in Hebrew 2nd-person sing. pronouns: “you” to a man is /atta/, which looks feminine, and to a woman is /att/, which looks masculine – surely this and the numeral usage are too similar and unusual to be a coincidence between unrelated phenomena?). The Cushitic language Bilin has a case of polarity-switching unrelated to gender: the verb for “have”, and two verbs for different senses of “be”, use what in other verbs is past-tense morphology to express the present, and vice versa – and the Hebrew phenomenon of “vav consecutiva” is often taught as a similar case of polarity-switching of tense, though it is discussed in the Semitic chapter here in a way that reduces the similarity. I do not know whether it could be reasonable to think of polarity-switching in general as somehow an inherent Afroasiatic characteristic. But it is certainly valuable to be able to compare Hebrew at this level of detail with African languages from which it must have parted company long before recorded history began, as this book enables us to do – Hebrew has traditionally been compared only with other Semitic languages.

A fact about the book which is very relevant to evaluation, though different linguists may perhaps set conflicting values on it, is that – as one would expect from Zygmunt Frajzyngier – it is concerned with presenting solid linguistic data, describing languages in their own terms, with no attempt to relate data to arcane theories about “syntactic universals” or the like.

The book has been splendidly produced – in over six hundred pages I noticed just four misprints, each too trivial to list here. (Also, on p. 273 there was a mysterious use of the word “quaternary” seemingly to refer to counting in fives, but perhaps I misunderstood the passage.) Typesetting must have been a challenging task in the case of the many examples, with lines of word-for-word glossing and idiomatic translation and often other complications, but these are almost always set out in a way that makes things easy for the reader (though there were problems I could not resolve with examples on pp. 121 and 133). I was particularly pleased to find that the editors had combined the various contributors’ lists of references into a single alphabetical list at the end of the book – too many editors take the lazy way out and leave these as separate lists for each chapter, forcing readers to do a lot of awkward flicking back and forth to look up a particular item. Altogether this is an admirable addition to the linguistic literature.


Diamond, J. and P. Bellwood. 2003. “Farmers and their languages: the first expansions”. Science 300.597–603.

Goldenberg, G. 2013. Semitic Languages: features, structures, relations, processes. Oxford University Press.

Hetzron, R. 1967. “Agaw numerals and incongruence in Semitic”. Journal of Semitic Studies 12.169–97.


Geoffrey Sampson graduated in Chinese Studies from Cambridge University, and his academic career was spent partly in Linguistics and partly in Informatics, with intervals in industrial research. After retiring as professor emeritus from Sussex University in 2009, he spent several years as Research Fellow at the University of South Africa. He has published contributions to most areas of Linguistics, as well as to other subjects. His latest book is ''Voices from Early China'' (2020).

Page Updated: 20-Jul-2021