The article explores the middle and reflexive verbs in the Cushitic language group (with a focus on East Cushitic) and highlights the pervasiveness of the autobenefactive value in much of East Cushitic. After an overview of the values of the Middle, focused on Gawwada, the typology of the Middle will be discussed. It will be argued that to place the Middle along a transitivity scale (as proposed by Kemmer 1994) does not do justice to the manifold uses of this verbal extension in Cushitic, the autobenefactive in particular. It will be suggested that the Middle is a semantic and pragmatic function expressing the scope (interest and volitionality) of the subject, itself an aspect of subjectification (Traugott 1995) and reminiscent of the `affectedness of the subject' (Klaiman 1991) and `centrality of the body' (Mous and Qorro 2000) and `subject's experience' (Kawachi 2004).
Abstract This paper explores basic valency orientation in Semitic languages with a comparative diachronic approach. To do so, we analyse how Akkadian and Modern South Arabian languages ( MSAL ) formally encode causal and noncausal verb pairs. The study utilises a dataset of 24 verb meanings, distinguishing between animate and inanimate verbs, to identify patterns of augmentation, suppletion, and reduction in both language groups. Results indicate that Akkadian is strongly transitivizing, particularly for animate verbs, while MSAL exhibit a more mixed situation, with noticeable branch-internal variation. Our findings contribute to the understanding of diachronic change in Semitic verbal valency and raise questions regarding the structure of the Proto-Semitic system.
Abstract The ‘middle’ derivation in Somali (Lowland East Cushitic, Afroasiatic) has received various labels in the literature, and its exact semantic values are still unclear. Precise data comparing the lexical semantics of derived middles and their respective bases are still lacking. We aim at filling this gap on the basis of an elicitation work conducted with Somali native speakers. Following up on an intuition defended in Puglielli & Bruno (1988), and shared to some extent by Saeed (1995), we explore inhowfar the subject of middles can be viewed as systematically bearing an oblique, possessor-like semantic role. Our data largely confirm the semantics of this role, but they hint at an additional syntactic constraint: the respective role can never be assigned directly by the middle verb. It is always assigned through a (potentially implicit) co-argument.
Abstract This paper offers an analysis of a case of syncretism at the phonology-syntax interface, based primarily on empirical data from Kabyle, as well as other Berber languages. Specifically, while the imperfective aspect, the middle, and the passive are analyzed separately in the literature, it will be empirically argued that they all constitute a case of syncretism, as they are all realized by a common morphological exponent, /t s -/. Theoretically, it will be argued that this syncretism can be explained by the fact that all three functions are underlyingly identical and are realized under a single projection in the syntactic structure: AspectP. The theoretical contribution of this study is to show that this case of syncretism can be accounted for by an approach that links syntax directly to phonology, without any intermediate module, thereby contributing to the modeling of the syntax-phonology interface.
Abstract This paper provides a profile of the Hitpa’el template in Modern Hebrew ( MH ). In MH the template is extremely versatile, hosting a wide variety of verb types. Hitpa’el is traditionally taken to express middle voice, but it differs from Nif’al, the other middle voice template, in hosting productively formed agentive verbs. There is no uniform representation for all verbs in Hitpa’el, though it has one invariant property: verbs in this template never assign case. I focus on some less-discussed regular uses of Hitpa’el that pose challenges for previous analyses. Though morphologically Hitpa’el verbs are systematically derived from a base Pi’el verbal template, many Hitpa’el verbs are semantically derived from verbs in other templates or from nouns or adjectives. I discuss one case in which the Pi’el verb is clearly derived semantically from the Hitpa’el form. Hitpa’el verbs, then, offer an interesting case of divergence between morphological and semantic derivation.
Abstract Clausal connectives, such as complementizers or circumstantial markers, are highly grammaticalized functional elements. Nevertheless, part of their original meaning can persist after grammaticalization has occurred, so that a seemingly neutral clausal connective can be semantically or pragmatically marked. This article argues that such is the case for the highly polysemous Biblical Hebrew clausal connective כִּי kī . It shows that this clausal connective presupposes that the information content it introduces is in the Common Ground. Specifically, kī is used in three contexts: (a) to refer to discourse-old information, (b) to introduce information that the Addressee can accommodate, and (c) to present discourse-new information as part of the Common Ground, for discursive effect. The previously described wide variety of uses of kī (introducing object and subject clauses as well as causal, temporal, conditional, adversative, concessive, and resultative circumstantial clauses) can then be derived from contextual clues given the general function of marking Common Ground. Furthermore, I argue that this function of kī is related to its origin as a [+distal] deictic lexeme, which provides a parallel with similar complementizers in unrelated languages. The analysis thus adds to a growing body of evidence for the possibility of employing referential features in the left periphery to express relations between interlocutors on the one hand and information content on the other.
Abstract “Middle” t , with functions ranging from autobenefactive to passive and beyond, is one of the best known cross-Afroasiatic cognates within derivational morphology, along with causative s . In some cases, its functions appear to extend beyond diathesis into the domain of aspect; whereas causative s has remained aloof from TAM , middle t has been argued to have developed imperfective functions in at least three branches of Afroasiatic: West Rift Cushitic, Akkadian (Semitic), and Berber. Yet voice morphology is not a typical source for aspectual morphology cross-linguistically, and grammaticalisation pathways deriving imperfective from middle, specifically, appear poorly attested and difficult to motivate conceptually. This paper seeks to eliminate this apparent paradox, arguing that West Rift and Akkadian developments reflect fusion with earlier imperfective markers, and putting forward a novel hypothesis in which imperfective t in Berber is derived from another morpheme with reflexes in both Berber and Semitic: nominalizing t .
Abstract The Afroasiatic t -morpheme is found in Akkadian (East-Semitic) with three distinct functions: a syncretic ‘Middle’, a pluractional, and a progression-marker. As a Middle, the morpheme may denote passive, reflexive/reciprocal, or noncausal functions. As a pluractional, it denotes iterative, frequentative, habitual, continuous, and distributive meanings. As a progression-marker, it typically links the marked event to a chain of perfective -marked events. No consensus has been reached thus far as to how this syncretic system came to be. This paper develops a semantic diachronic reconstruction, arguing that Middles of unergatives, which regularly derive atelic predicates, led to a first process of reanalysis wherein the t -morpheme became associated with general atelicity. From there on, by attaching to the imperfective , the morpheme developed from atelic > pluractional and by attaching to the perfective from atelic > progression-marker.
Abstract This paper explores potential connections between the Afroasiatic verbal morpheme *t and specific features of Ancient Egyptian verbal morphology. It focuses on three key cases: the passivizing morpheme t(ỉ/w), the suffix -t in the sḏm-t=f form, and the -t ending found in certain infinitives. Although all three have been previously linked to Afroasiatic *t, their functions and historical development remain uncertain. The study reexamines these proposals in light of recent scholarship. Particular attention is devoted to the infinitival -t, whose distribution is reassessed through an analysis of its associations with verbs across different semantic domains known to interact with *t in other Afroasiatic languages. Given the ambiguity of the Egyptian writing system and the limitations of the available data, the conclusions remain tentative. Overall, the paper emphasizes both the plausibility of these connections and the methodological challenges involved.
Soqotri is a Semitic language belonging to the Modern South Arabian branch and spoken by the approximately 100,000 inhabitants of the island of Soqotra. As elsewhere in Semitic, Soqotri verbal derivation involves a complex system of the so-called verbal stems. The present contribution deals with morphology and semantics of the stem formally corresponding to Arabic Stem x and labeled accordingly in the studies of the Russian-Soqotri research team (the same stem is sometimes referred to as causativereflexive, e.g., in Simeone-Senelle 2011:1092). Data are drawn from the two published volumes of our Corpus of Soqotri Oral Literature (CS OL I-II) as well as from the authors' fieldwork.
This article revisits the system of gender exponence of a class of Omani Mehri plural nouns initially analyzed in Fathi (2017a,b), the faSowal/faSoyal class. In that class, it is argued, gender is exponed canonically on one site and non-canonically on another site. Concomitantly, another class, not discussed in Fathi (2017a,b), is introduced here, the faSol/faSel plurals. That second class also displays canonical and non-canonical gender but contrary to the first class, both are exponed on the same site. An account is offered in this article whereby both classes are shown to be different realizations of the same underlying morphosyntactic structure.
This paper provides a unifying, Optimality-Theoretic analysis of subtractive morphology and morphological gemination in Moroccan Amazigh (Berber) and Arabic. I start with a Morpheme Realization and Anti-faithfulness account of subtraction and a floating mora affixation account of gemination. Then, I provide a preliminary, unifying analysis of both phenomena based on the floating mora approach of Generalized Mora Affixation (GMA), which extends naturally to facts of other Afro-asiatic languages. I also show how the analysis explains an intricate case involving subtraction and gemination occurring in tandem. Unifying though the GMA approach is, residual issues in the morphology of Amazigh are a serious contender. The proponents of GMA defend "pivot affixation" and criticize "phonological affix dislocation" for its allowing variable infixation. The gemination vs. tt-prefixation facet of the Amazigh intensive aorist allomorphy is a clear case of variable infixation which seems to require a weaker version of the "pivot affixation" approach.
This article investigates the adaptation of French nasal vowels in Moroccan Arabic. It departs from the traditional analysis, which adopts nasal vowel unpacking and, hence, default VN adaptation. I argue that borrowers do not unpack nasal vowels; instead, they employ their L1 phonology to perceive the L2 acoustic cues of such vowels as VN or V. I assume that adjustments occur during perception; the latter is, however, conditioned by the borrowers' exposure to various forms of L2 word forms and exploitation of all grammatical information (e.g. acoustic, morphological). Adopting the Bi-Phon model of Optimality Theory (Boersma & Hamann 2009), I show how the ranking of cue and structural constraints generate all the seemingly irregular adaptations. The findings will contribute in the understanding of the literature on perception, loanword adaptation, Moroccan Arabic phonology as well as French phonetics.
This paper provides a unifying, Optimality-Theoretic analysis of subtractive morphology and morphological gemination in Moroccan Amazigh (Berber) and Arabic. I start with a Morpheme Realization and Anti-faithfulness account of subtraction and a floating mora affixation account of gemination. Then, I provide a preliminary, unifying analysis of both phenomena based on the floating mora approach of Generalized Mora Affixation ( GMA ), which extends naturally to facts of other Afro-asiatic languages. I also show how the analysis explains an intricate case involving subtraction and gemination occurring in tandem. Unifying though the GMA approach is, residual issues in the morphology of Amazigh are a serious contender. The proponents of GMA defend “pivot affixation” and criticize “phonological affix dislocation” for its allowing variable infixation. The gemination vs. tt-prefixation facet of the Amazigh intensive aorist allomorphy is a clear case of variable infixation which seems to require a weaker version of the “pivot affixation” approach.
Soqotri is a Semitic language belonging to the Modern South Arabian branch and spoken by the approximately 100,000 inhabitants of the island of Soqotra. As elsewhere in Semitic, Soqotri verbal derivation involves a complex system of the so-called verbal stems. The present contribution deals with morphology and semantics of the stem formally corresponding to Arabic Stem X and labeled accordingly in the studies of the Russian-Soqotri research team (the same stem is sometimes referred to as causative-reflexive, e.g., in Simeone-Senelle 2011:1092). Data are drawn from the two published volumes of our Corpus of Soqotri Oral Literature ( CSOL I – II ) as well as from the authors’ fieldwork.
This article investigates the adaptation of French nasal vowels in Moroccan Arabic. It departs from the traditional analysis, which adopts nasal vowel unpacking and, hence, default VN adaptation. I argue that borrowers do not unpack nasal vowels; instead, they employ their L1 phonology to perceive the L2 acoustic cues of such vowels as VN or V. I assume that adjustments occur during perception; the latter is, however, conditioned by the borrowers’ exposure to various forms of L2 word forms and exploitation of all grammatical information (e.g. acoustic, morphological). Adopting the Bi-Phon model of Optimality Theory (Boersma & Hamann 2009), I show how the ranking of cue and structural constraints generate all the seemingly irregular adaptations. The findings will contribute in the understanding of the literature on perception, loanword adaptation, Moroccan Arabic phonology as well as French phonetics.
This article revisits the system of gender exponence of a class of Omani Mehri plural nouns initially analyzed in Fathi (2017a,b), the fəʕṓwəl/fəʕṓyəl class. In that class, it is argued, gender is exponed canonically on one site and non-canonically on another site. Concomitantly, another class, not discussed in Fathi (2017a,b), is introduced here, the fəʕṓl/fəʕḗl plurals. That second class also displays canonical and non-canonical gender but contrary to the first class, both are exponed on the same site . An account is offered in this article whereby both classes are shown to be different realizations of the same underlying morphosyntactic structure.
This paper analyses the function and historical development of the N-stem and Gt-stem, two derivational templates for the middle voice in Ugaritic. Based on a comprehensive sample of 61 N-stem and 63 Gt-stem cases, the study shows that these templates have distinct roles. The N-stem typically promotes a patientive, often inanimate, subject, creating unaccusative, anticausative, and medio-passive verbs. In contrast, the Gt-stem aligns an affected, animate Agent as the subject, forming autobenefactive, reflexive, and reciprocal verbs. The distribution of these usages varies between Ugaritic poetry and prose. The N-stem is more common in prose, where it is frequently used for medio-passive functions, a trend also seen in later Canaanite languages. The Gt-stem is more prominent in poetry, reflecting an earlier stage of development that parallels the trajectory of Aramaic, where this stem became the main marker for middle and passive voice. The study concludes that the Ugaritic corpus, with its linguistic variations between poetic and prosaic registers, provides a crucial historical context for understanding the divergent paths of middle voice development in the Canaanite and Aramaic branches of Northwest Semitic languages.
