Devos Maud, van der Wal Jenneke. Grammaticalization and (Inter)Subjectification in African langua... more Devos Maud, van der Wal Jenneke. Grammaticalization and (Inter)Subjectification in African languages. In: Africana Linguistica 18, 2012. p. 3
Determiner spreading, the phenomenon whereby adnominal modifiers carry an ‘additional’ determiner... more Determiner spreading, the phenomenon whereby adnominal modifiers carry an ‘additional’ determiner, has been studied extensively for a variety of languages, most notably Greek, Semitic, and Scandinavian languages. Interestingly, the same phenomenon occurs in the Bantu language Rukiga. We show how the Rukiga augment is parallel to the Greek determiner in the context of modification, and how it triggers a restrictive reading when present on a larger class of modifiers than familiar so far: relative clauses, adjectives, possessives, and certain quantifiers. Considering its morphosyntactic and interpretational properties, we propose that the variation in the presence versus absence of the augment on modifiers is due to different underlying structures, applying an analysis of determiner spreading in terms of a reduced relative clause structure.
In addition to the stem and noun class prefix, the structure of nouns in Bantu languages may cont... more In addition to the stem and noun class prefix, the structure of nouns in Bantu languages may contain an augment. This augment typically is a vowel, but some languages show a CV augment. Interestingly, the Bantu language Kinyakyusa shows nouns with a V as well as with a CV prefix, both of which have been analysed as augments (De Blois 1970). In this short paper we clarify the formal and functional properties of the ‘CV augment’ in Kinyakyusa. First we show that it does not behave like the V augment, but is a separate marker that is attached to the noun phrase. Second, we narrow down the previous analyses of the CV marker that describe it as ‘emphatic’ (De Blois 1970, Persohn 2020): On the basis of a range of focus tests, we argue that the CV marker functions as a marker of exhaustivity. This is remarkable, as exhaustive focus is in Bantu languages typically associated with marking in the clause and not on the noun itself.
This study sets out to investigate the insubordinated infinitive in the Bantu language Makhuwa-En... more This study sets out to investigate the insubordinated infinitive in the Bantu language Makhuwa-Enahara (P31, northern Mozambique), which is used with feeling predicates that have passive experiencers. The expression of bodily feelings and emotions in Makhuwa serves as a foundation, highlighting the unique formal and interpretational properties of the insubordinated infinitive within the domain of feelings.
The Bantu languages are in some sense remarkably uniform (subject, verb, order (SVO) basic word o... more The Bantu languages are in some sense remarkably uniform (subject, verb, order (SVO) basic word order, noun classes, verbal morphology), but this extensive language family also show a wealth of morphosyntactic variation. Two core areas in which such variation is attested are subject and object agreement. The book explores the variation in Bantu subject and object marking on the basis of data from 75 Bantu languages, discovering striking patterns (the Relation between Asymmetry and Non-Doubling Object Marking (RANDOM), and the Asymmetry Wants Single Object Marking (AWSOM) correlation), and providing a novel syntactic analysis. This analysis takes into account not just phi agreement, but also nominal licensing and information structure. A Person feature, associated with animacy, definiteness, or givenness, is shown to be responsible for differential object agreement, while at the same time accounting for doubling vs. non-doubling object marking—a hybrid solution to an age-old debate i...
This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For man... more This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For many people engaged in the field of African linguistics (and beyond), Maarten has been a teacher, an engaged colleague, and an inspiration. On the occasion of his 65th birthday, the present volume offers essays written by his former and current PhD students. The volume is divided into four sections: Language in use and contact, Morphosyntax, Number and numerals, and Phonology. It contains 25 articles and presents a variety of disciplinary and regional approaches to African linguistics.bookDescriptive and Comparative Linguistic
It has been claimed that in Bantu languages Agree and Move are related, i.e., that Agree has a mo... more It has been claimed that in Bantu languages Agree and Move are related, i.e., that Agree has a movement trigger (EPP feature). Exceptions to this generalisation are languages that have Agreeing Inversion, whereby the verb agrees with a postverbal subject. This paper reexamines the Agreement Parameter as proposed by Collins (2004) and its variants in Carstens (2005) and Baker (2003, 2008), suggesting that it can (at least partly) distinguish between two types of Agreeing Inversion: in the one type subject agreement not associated with a movement trigger/EPP feature (Matengo, Makwe, Matuumbi), and in the other type the agreeing subject is in fact moved, but there is remnant movement of the verbal complex to derive the inverted order (Makhuwa). This has many consequences, influencing word order (VSO or VOS), the form of the verb (conjoint or disjoint) and the interpretation (non-topic or also focus). * I would like to acknowledge and warmly thank a number of fellow linguists for sharin...
In this paper, we capture the crosslinguistic variation in Bantu nominal structure in a unified a... more In this paper, we capture the crosslinguistic variation in Bantu nominal structure in a unified analysis of gender on n (Kramer 2014, 2015). We demonstrate that this analysis accounts for the morphosyntactic properties of basic nouns as well as locative and diminutive derivations. Moreover, it allows us to capture intra- and inter-language morphosyntactic variation by reference to just three parameters – one strictly morphological and two structural. The presence of one or two n heads, and the size of the complement distinguish between different types of locatives (structural variation); the presence or absence of a spell-out rule of adjacent n heads differentiates “stacking” versus “non-stacking” prefixes in diminutive and augmentative derivations (morphological variation only).
A presentation at the 2020 Colloquium of African Languages and Linguistics, hosted at Leiden Univ... more A presentation at the 2020 Colloquium of African Languages and Linguistics, hosted at Leiden University.
L’otjiherero a un systeme de flexion tonale du nom appele ‘ cas tonals’. Ce systeme permet de dif... more L’otjiherero a un systeme de flexion tonale du nom appele ‘ cas tonals’. Ce systeme permet de differencier les noms selon le contexte syntaxique, notamment les cas qu’on appelle ‘ complement’ et ‘ par defaut’. Les noms marques comme cas ‘ complement’ ne se trouvent qu’immediatement apres le verbe et seulement dans un sous-ensemble d’aspects verbaux. Le groupe de noms qui peuvent prendre le cas ‘ complement’ inclut des objets directs et indirects, des noms adverbiaux, des sujets anticipes et des sujets postposes au verbe. On rencontre donc le cas ‘ complement’ avec differents noms, independamment de leur fonction grammaticale, du moment qu’ils sont places immediatement apres le verbe. Il existe deux arguments qui permettent de montrer que le systeme est (historiquement) lie a la structure informative. Tout d’abord, on ne trouve pas de cas ‘ complement’ dans les propositions relatives ou avec les noms deplaces par dislocation. Ensuite, les noms qui suivent un verbe au negatif factitif-habituel peuvent etre marques, soit comme cas ‘ complement’, soit comme cas ‘ par defaut’, selon qu’ils sont focalises ou non. A partir de la fonction et de la distribution des cas ‘ complement’, l’article suggere qu’il existe un parallelisme etroit entre le systeme de cas tonals et le systeme d’alternances conjoint-disjoint qu’on trouve en tswana. Les deux systemes impliquent un marquage prosodique, ne s’observent qu’a certains temps, marquent la relation entre un verbe et le nom qui le suit immediatement independamment de sa fonction grammaticale, sont limites a un domaine structurel comme la proposition et sont lies a la focalisation et a la structure informative. L’article ne propose pas de reconstruction detaillee, mais il suggere que les deux systemes, celui des cas tonals et celui de l’alternance conjoint-disjoint, sont des exemples de structure informative grammaticalisee. L’etroite similarite entre les deux systemes permet de penser qu’ils ont une fonction commune et peut-etre meme une origine historique commune.
This paper presents new data from Bantu languages, from which a hitherto unnoticed typological pa... more This paper presents new data from Bantu languages, from which a hitherto unnoticed typological pattern emerges: A) language-internally, causative, applicative and lexical ('give') ditransitives can differ with respect to symmetry; B) crosslinguistically, they are in an implicational relationship: if a language is symmetrical for one type of predicate, it is symmetrical for the predicate types to its right as well:<br> <br> causative > applicative > lexical ditransitive<br> <br> This can be accounted for if symmetry is due to low functional heads being flexible to license an argument in either their complement or their specifier \citep{HaddicanHolmberg2012,HaddicanHolmberg2015}. This flexibility is argued to be a sensitivity to topicality. The implicational relation can then be seen as a requirement for lower functional heads to have the same sensitivity: if Caus can license its specifier, then HAppl and LAppl should also be able to do so.
The Bantu language Rukiga (JE14, Uganda) shows tonal reduction on the verb in a subset of tenses,... more The Bantu language Rukiga (JE14, Uganda) shows tonal reduction on the verb in a subset of tenses, similar to the conjoint/disjoint alternation in Haya. Whereas in other languages the conjoint/disjoint alternation is usually marked by segmental morphology in at least one tense, Rukiga is unique in showing only tonal reduction. Nevertheless, our analysis shows that tonal reduction in Rukiga is not merely a phonological rule, but it encodes the conjoint/disjoint alternation. Furthermore, we show that tonal reduction in Rukiga is determined by constituent-finality, and there is no direct relation to focus
Les demonstratifs, ceux de distance en particulier, sont tres frequemment utilises dans les recit... more Les demonstratifs, ceux de distance en particulier, sont tres frequemment utilises dans les recits makhuwa. Meme si tous les types de demonstratifs peuvent etre utilises comme deictiques exophoriques, pour se referer a une entite intradiscursive, les demonstratifs de distance sont davantage utilises comme reference anaphorique. Ils interviennent dans les changements de topique, le lien queue-tete et aux frontieres d’episodes, renvoyant a des entites qui sont relativement moins accessibles a ce stade du recit (ce qui est conforme a la Theorie de l’Accessibilite). Les demonstratifs dans les narrations makhuwa ne sont pas seulement utilises anaphoriquement pour l’identification du referent, mais ils indiquent egalement l’importance d’un certain referent et attrait d’attention, et ils jouent un role important dans la structuration des textes.
Mauritian Creole displays an alternation between a short and a long form of the verb, which is re... more Mauritian Creole displays an alternation between a short and a long form of the verb, which is reminiscent of the conjoint–disjoint alternation found in some eastern Bantu languages. Based on comparison with other French-based creoles and socio-historical evidence, we conclude that the Bantu substrate must have had an impact on the grammatical system of Mauritian Creole. We compare the synchronic properties of the alternations in Mauritian Creole and the most likely substrate Bantu languages of northern Mozambique and examine two possible scenarios for the influence of Bantu on the Mauritian verbal alternation, concluding that probably only the (syntactic) basics of the Bantu alternation motivated the persistence of the alternation in Mauritian Creole.
Devos Maud, van der Wal Jenneke. Grammaticalization and (Inter)Subjectification in African langua... more Devos Maud, van der Wal Jenneke. Grammaticalization and (Inter)Subjectification in African languages. In: Africana Linguistica 18, 2012. p. 3
Determiner spreading, the phenomenon whereby adnominal modifiers carry an ‘additional’ determiner... more Determiner spreading, the phenomenon whereby adnominal modifiers carry an ‘additional’ determiner, has been studied extensively for a variety of languages, most notably Greek, Semitic, and Scandinavian languages. Interestingly, the same phenomenon occurs in the Bantu language Rukiga. We show how the Rukiga augment is parallel to the Greek determiner in the context of modification, and how it triggers a restrictive reading when present on a larger class of modifiers than familiar so far: relative clauses, adjectives, possessives, and certain quantifiers. Considering its morphosyntactic and interpretational properties, we propose that the variation in the presence versus absence of the augment on modifiers is due to different underlying structures, applying an analysis of determiner spreading in terms of a reduced relative clause structure.
In addition to the stem and noun class prefix, the structure of nouns in Bantu languages may cont... more In addition to the stem and noun class prefix, the structure of nouns in Bantu languages may contain an augment. This augment typically is a vowel, but some languages show a CV augment. Interestingly, the Bantu language Kinyakyusa shows nouns with a V as well as with a CV prefix, both of which have been analysed as augments (De Blois 1970). In this short paper we clarify the formal and functional properties of the ‘CV augment’ in Kinyakyusa. First we show that it does not behave like the V augment, but is a separate marker that is attached to the noun phrase. Second, we narrow down the previous analyses of the CV marker that describe it as ‘emphatic’ (De Blois 1970, Persohn 2020): On the basis of a range of focus tests, we argue that the CV marker functions as a marker of exhaustivity. This is remarkable, as exhaustive focus is in Bantu languages typically associated with marking in the clause and not on the noun itself.
This study sets out to investigate the insubordinated infinitive in the Bantu language Makhuwa-En... more This study sets out to investigate the insubordinated infinitive in the Bantu language Makhuwa-Enahara (P31, northern Mozambique), which is used with feeling predicates that have passive experiencers. The expression of bodily feelings and emotions in Makhuwa serves as a foundation, highlighting the unique formal and interpretational properties of the insubordinated infinitive within the domain of feelings.
The Bantu languages are in some sense remarkably uniform (subject, verb, order (SVO) basic word o... more The Bantu languages are in some sense remarkably uniform (subject, verb, order (SVO) basic word order, noun classes, verbal morphology), but this extensive language family also show a wealth of morphosyntactic variation. Two core areas in which such variation is attested are subject and object agreement. The book explores the variation in Bantu subject and object marking on the basis of data from 75 Bantu languages, discovering striking patterns (the Relation between Asymmetry and Non-Doubling Object Marking (RANDOM), and the Asymmetry Wants Single Object Marking (AWSOM) correlation), and providing a novel syntactic analysis. This analysis takes into account not just phi agreement, but also nominal licensing and information structure. A Person feature, associated with animacy, definiteness, or givenness, is shown to be responsible for differential object agreement, while at the same time accounting for doubling vs. non-doubling object marking—a hybrid solution to an age-old debate i...
This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For man... more This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For many people engaged in the field of African linguistics (and beyond), Maarten has been a teacher, an engaged colleague, and an inspiration. On the occasion of his 65th birthday, the present volume offers essays written by his former and current PhD students. The volume is divided into four sections: Language in use and contact, Morphosyntax, Number and numerals, and Phonology. It contains 25 articles and presents a variety of disciplinary and regional approaches to African linguistics.bookDescriptive and Comparative Linguistic
It has been claimed that in Bantu languages Agree and Move are related, i.e., that Agree has a mo... more It has been claimed that in Bantu languages Agree and Move are related, i.e., that Agree has a movement trigger (EPP feature). Exceptions to this generalisation are languages that have Agreeing Inversion, whereby the verb agrees with a postverbal subject. This paper reexamines the Agreement Parameter as proposed by Collins (2004) and its variants in Carstens (2005) and Baker (2003, 2008), suggesting that it can (at least partly) distinguish between two types of Agreeing Inversion: in the one type subject agreement not associated with a movement trigger/EPP feature (Matengo, Makwe, Matuumbi), and in the other type the agreeing subject is in fact moved, but there is remnant movement of the verbal complex to derive the inverted order (Makhuwa). This has many consequences, influencing word order (VSO or VOS), the form of the verb (conjoint or disjoint) and the interpretation (non-topic or also focus). * I would like to acknowledge and warmly thank a number of fellow linguists for sharin...
In this paper, we capture the crosslinguistic variation in Bantu nominal structure in a unified a... more In this paper, we capture the crosslinguistic variation in Bantu nominal structure in a unified analysis of gender on n (Kramer 2014, 2015). We demonstrate that this analysis accounts for the morphosyntactic properties of basic nouns as well as locative and diminutive derivations. Moreover, it allows us to capture intra- and inter-language morphosyntactic variation by reference to just three parameters – one strictly morphological and two structural. The presence of one or two n heads, and the size of the complement distinguish between different types of locatives (structural variation); the presence or absence of a spell-out rule of adjacent n heads differentiates “stacking” versus “non-stacking” prefixes in diminutive and augmentative derivations (morphological variation only).
A presentation at the 2020 Colloquium of African Languages and Linguistics, hosted at Leiden Univ... more A presentation at the 2020 Colloquium of African Languages and Linguistics, hosted at Leiden University.
L’otjiherero a un systeme de flexion tonale du nom appele ‘ cas tonals’. Ce systeme permet de dif... more L’otjiherero a un systeme de flexion tonale du nom appele ‘ cas tonals’. Ce systeme permet de differencier les noms selon le contexte syntaxique, notamment les cas qu’on appelle ‘ complement’ et ‘ par defaut’. Les noms marques comme cas ‘ complement’ ne se trouvent qu’immediatement apres le verbe et seulement dans un sous-ensemble d’aspects verbaux. Le groupe de noms qui peuvent prendre le cas ‘ complement’ inclut des objets directs et indirects, des noms adverbiaux, des sujets anticipes et des sujets postposes au verbe. On rencontre donc le cas ‘ complement’ avec differents noms, independamment de leur fonction grammaticale, du moment qu’ils sont places immediatement apres le verbe. Il existe deux arguments qui permettent de montrer que le systeme est (historiquement) lie a la structure informative. Tout d’abord, on ne trouve pas de cas ‘ complement’ dans les propositions relatives ou avec les noms deplaces par dislocation. Ensuite, les noms qui suivent un verbe au negatif factitif-habituel peuvent etre marques, soit comme cas ‘ complement’, soit comme cas ‘ par defaut’, selon qu’ils sont focalises ou non. A partir de la fonction et de la distribution des cas ‘ complement’, l’article suggere qu’il existe un parallelisme etroit entre le systeme de cas tonals et le systeme d’alternances conjoint-disjoint qu’on trouve en tswana. Les deux systemes impliquent un marquage prosodique, ne s’observent qu’a certains temps, marquent la relation entre un verbe et le nom qui le suit immediatement independamment de sa fonction grammaticale, sont limites a un domaine structurel comme la proposition et sont lies a la focalisation et a la structure informative. L’article ne propose pas de reconstruction detaillee, mais il suggere que les deux systemes, celui des cas tonals et celui de l’alternance conjoint-disjoint, sont des exemples de structure informative grammaticalisee. L’etroite similarite entre les deux systemes permet de penser qu’ils ont une fonction commune et peut-etre meme une origine historique commune.
This paper presents new data from Bantu languages, from which a hitherto unnoticed typological pa... more This paper presents new data from Bantu languages, from which a hitherto unnoticed typological pattern emerges: A) language-internally, causative, applicative and lexical ('give') ditransitives can differ with respect to symmetry; B) crosslinguistically, they are in an implicational relationship: if a language is symmetrical for one type of predicate, it is symmetrical for the predicate types to its right as well:<br> <br> causative > applicative > lexical ditransitive<br> <br> This can be accounted for if symmetry is due to low functional heads being flexible to license an argument in either their complement or their specifier \citep{HaddicanHolmberg2012,HaddicanHolmberg2015}. This flexibility is argued to be a sensitivity to topicality. The implicational relation can then be seen as a requirement for lower functional heads to have the same sensitivity: if Caus can license its specifier, then HAppl and LAppl should also be able to do so.
The Bantu language Rukiga (JE14, Uganda) shows tonal reduction on the verb in a subset of tenses,... more The Bantu language Rukiga (JE14, Uganda) shows tonal reduction on the verb in a subset of tenses, similar to the conjoint/disjoint alternation in Haya. Whereas in other languages the conjoint/disjoint alternation is usually marked by segmental morphology in at least one tense, Rukiga is unique in showing only tonal reduction. Nevertheless, our analysis shows that tonal reduction in Rukiga is not merely a phonological rule, but it encodes the conjoint/disjoint alternation. Furthermore, we show that tonal reduction in Rukiga is determined by constituent-finality, and there is no direct relation to focus
Les demonstratifs, ceux de distance en particulier, sont tres frequemment utilises dans les recit... more Les demonstratifs, ceux de distance en particulier, sont tres frequemment utilises dans les recits makhuwa. Meme si tous les types de demonstratifs peuvent etre utilises comme deictiques exophoriques, pour se referer a une entite intradiscursive, les demonstratifs de distance sont davantage utilises comme reference anaphorique. Ils interviennent dans les changements de topique, le lien queue-tete et aux frontieres d’episodes, renvoyant a des entites qui sont relativement moins accessibles a ce stade du recit (ce qui est conforme a la Theorie de l’Accessibilite). Les demonstratifs dans les narrations makhuwa ne sont pas seulement utilises anaphoriquement pour l’identification du referent, mais ils indiquent egalement l’importance d’un certain referent et attrait d’attention, et ils jouent un role important dans la structuration des textes.
Mauritian Creole displays an alternation between a short and a long form of the verb, which is re... more Mauritian Creole displays an alternation between a short and a long form of the verb, which is reminiscent of the conjoint–disjoint alternation found in some eastern Bantu languages. Based on comparison with other French-based creoles and socio-historical evidence, we conclude that the Bantu substrate must have had an impact on the grammatical system of Mauritian Creole. We compare the synchronic properties of the alternations in Mauritian Creole and the most likely substrate Bantu languages of northern Mozambique and examine two possible scenarios for the influence of Bantu on the Mauritian verbal alternation, concluding that probably only the (syntactic) basics of the Bantu alternation motivated the persistence of the alternation in Mauritian Creole.
Agreement in Bantu symmetric objects and subject inversion constructions can be understood as spl... more Agreement in Bantu symmetric objects and subject inversion constructions can be understood as split alignment systems based on information structure. [poster at GLOW 2018]
The starting point is the well-known observation that Bantu languages can show multiple object ma... more The starting point is the well-known observation that Bantu languages can show multiple object markers on multitransitive verb:
Luganda (Ssekiryango 2006: 67, 72) (1) a. Maama a-wa-dde taata ssente. 1.mother 1SM-give-PFV 1.father 10.money ‘Mother has given father money.’
b. Maama a-zi-mu-wa-dde. 1.mother 1SM-10OM-1OM-give-PFV ‘Mother has given it to him.’
This presents two puzzles for a theory of (multiple) Agree: Is this agreement? If so, how many probes are there?
Under a defective Goal approach to object marking (Roberts 2010, Iorio 2014, Van der Wal 2015), the answer to the first question is ‘yes’: each object marker involves an Agree relation, and object markers are due to the spell out of phi features on the probe.
The answer to the second question depends on the language, and I argue that there are three types of Bantu languages with multiple object markers. Type 1 has two phi probes as neighbour probes: one on v and one on Appl, which probe independently. Type 2 has two phi probes as twins on v, resulting in sensitivity to locality (Van der Wal to appear). Type 3 has only one probe on v, but this is a ‘phoenix probe’ that renews when all [Person[Participant[Author]]] features are valued (Béjar & Rezac 2009). A second object marker is thus only allowed when the first object marker is 1sg.
Each of these types thus produces multiple object markers, but each with its own characteristics: neighbour probes are independent and insensitive to locality (‘symmetric’), twin probes show locality effects, and phoenix probes are characterised by a restriction on the (Person) features of one of the objects.
A close look at two Bantu object marking parameters reveals a gap: there are no asymmetric langua... more A close look at two Bantu object marking parameters reveals a gap: there are no asymmetric languages with non-doubling object marking. I propose to account for this gap as follows: - doubling object marking is the presence of a Person feature in a separate projection on salient DPs; - symmetry is the sensitivity of low functional heads (Appl, Caus) to Person and Topic; - and languages must use Person/Topic in the clausal (Appl) and/or non-clausal domain (DP). This obligatory marking of salience explains the gap.
A second almost-gap occurs in another combination of parameters: overwhelmingly, languages with multiple object markers are symmetrical. The spread or clustering of ϕ probes on heads could be involved in this second gap.
The geographical distribution of the parameters settings suggests that doubling and multiple object markers are innovations.
A close look at three Bantu object marking parameters reveals two gaps: 1. there are no asymmetri... more A close look at three Bantu object marking parameters reveals two gaps: 1. there are no asymmetric languages with non-doubling object marking; 2 there are almost no asymmetric languages with multiple object marking. However, the three parameters cannot be reduced to two. Instead, I propose that obligatory marking of prominence in the clausal and/or non-clausal domain explains the first gap. The spread or clustering of ϕ probes on heads might be involved in the second gap. The geographical distribution of the parameters settings suggests that doubling and multiple object markers are innovations.
(Keynote talk at the Manchester Forum in Linguistics, 6-7 November 2015)
While we know that info... more (Keynote talk at the Manchester Forum in Linguistics, 6-7 November 2015)
While we know that information structure (topic, focus) influences the morphosyntax, we do not know how integrated the two really are. The two extreme approaches to this issue are not attractive; these are models that assume either that the two are completely independent (e.g. Chomsky 2008), or that all of information structure is represented in the syntax (e.g. Rizzi’s 1997 cartographic approach). In this talk I argue that before we can build an appropriate model of grammar, we need to look beyond the surface effects to establish which aspects of information structure are of direct influence in the grammar. I provide two case studies from Bantu languages of how this can be achieved. The first is Luganda, where I show that with a more elaborate toolbox we can specify a focus marker’s meaning as ‘exclusivity’ rather than an underspecified ‘focus’. The second is a typology of Bantu languages that on the surface have the same verbal focus alternation (conjoint/disjoint) but on closer examination turn out to be two classes: one directly determined by focus and the other only indirectly. These cases highlight the necessity to look beyond ‘what you see’ and the essential role of appropriate methodology in comparative research.
This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For man... more This book celebrates Maarten Mous, professor of African Linguistics at Leiden University. For many people engaged in the field of African linguistics (and beyond), Maarten has been a teacher, an engaged colleague, and an inspiration. On the occasion of his 65th birthday, the present volume offers essays written by his former and current PhD students. The volume is divided into four sections: Language in use and contact, Morphosyntax, Number and numerals, and Phonology. It contains 25 articles and presents a variety of disciplinary and regional approaches to African linguistics.
Uploads
Papers by Jenneke van der Wal
Luganda (Ssekiryango 2006: 67, 72)
(1) a. Maama a-wa-dde taata ssente.
1.mother 1SM-give-PFV 1.father 10.money
‘Mother has given father money.’
b. Maama a-zi-mu-wa-dde.
1.mother 1SM-10OM-1OM-give-PFV
‘Mother has given it to him.’
This presents two puzzles for a theory of (multiple) Agree:
Is this agreement?
If so, how many probes are there?
Under a defective Goal approach to object marking (Roberts 2010, Iorio 2014, Van der Wal 2015), the answer to the first question is ‘yes’: each object marker involves an Agree relation, and object markers are due to the spell out of phi features on the probe.
The answer to the second question depends on the language, and I argue that there are three types of Bantu languages with multiple object markers. Type 1 has two phi probes as neighbour probes: one on v and one on Appl, which probe independently. Type 2 has two phi probes as twins on v, resulting in sensitivity to locality (Van der Wal to appear). Type 3 has only one probe on v, but this is a ‘phoenix probe’ that renews when all [Person[Participant[Author]]] features are valued (Béjar & Rezac 2009). A second object marker is thus only allowed when the first object marker is 1sg.
Each of these types thus produces multiple object markers, but each with its own characteristics: neighbour probes are independent and insensitive to locality (‘symmetric’), twin probes show locality effects, and phoenix probes are characterised by a restriction on the (Person) features of one of the objects.
- doubling object marking is the presence of a Person feature in a separate projection on salient DPs;
- symmetry is the sensitivity of low functional heads (Appl, Caus) to Person and Topic;
- and languages must use Person/Topic in the clausal (Appl) and/or non-clausal domain (DP). This obligatory marking of salience explains the gap.
A second almost-gap occurs in another combination of parameters: overwhelmingly, languages with multiple object markers are symmetrical. The spread or clustering of ϕ probes on heads could be involved in this second gap.
The geographical distribution of the parameters settings suggests that doubling and multiple object markers are innovations.
While we know that information structure (topic, focus) influences the morphosyntax, we do not know how integrated the two really are. The two extreme approaches to this issue are not attractive; these are models that assume either that the two are completely independent (e.g. Chomsky 2008), or that all of information structure is represented in the syntax (e.g. Rizzi’s 1997 cartographic approach). In this talk I argue that before we can build an appropriate model of grammar, we need to look beyond the surface effects to establish which aspects of information structure are of direct influence in the grammar. I provide two case studies from Bantu languages of how this can be achieved. The first is Luganda, where I show that with a more elaborate toolbox we can specify a focus marker’s meaning as ‘exclusivity’ rather than an underspecified ‘focus’. The second is a typology of Bantu languages that on the surface have the same verbal focus alternation (conjoint/disjoint) but on closer examination turn out to be two classes: one directly determined by focus and the other only indirectly. These cases highlight the necessity to look beyond ‘what you see’ and the essential role of appropriate methodology in comparative research.