Abstract
The morphological structure of poly-morphemic words (e.g. government) can affect processing, but it is unclear whether this effect is due to morphological structure or combined formal/orthographic and semantic effects. Setswana, a Bantu language, allows us to explore morphological, formal, and semantic effects: It has a noun derivation with an agglutinative agentive affix (Class-1, mo-rer-i “preacher”) and a noun derivation with vowel changes and little form-overlap (Class-9, ther-o “sermon”) that both apply to verbs (rer-a “preach”). In our masked-priming experiments (SOA = 60 ms), Class-9-forms were even more effective as primes for verbs than Class-1-forms, despite reduced formal overlap, suggesting that abstract morphological structure affects processing independently of formal or semantic relationships. Moreover, Class-1-targets showed reduced priming, indicating that unprimed morphological target material (the affix) reduces priming.
Acknowledgements
The experiments reported in this paper were carried out as part of Naledi Kgolo's doctoral dissertation project and are discussed in the dissertation in more detail (Kgolo, Citation2014). We are thankful for insightful comments and discussions from our anonymous reviewers and Bettina Landgraf, Nancy Kula, Ingrid Sonnenstuhl, Agnes Henson, the audiences of the Architectures and Mechanisms for Language Processing (AMLaP), the Bantu 5 conference, the Cambridge Morphology Meeting 2013, and the members of the Psycholinguistics Research Group at the University of Essex.
Funding
This work was supported by the University of Botswana PhD scolarship to Naledi Kgolo. The project described in the paper also benefited from a series of training workshops on fieldwork and psycholinguistics with our colleagues at Jawaharlal Nehru University (JNU), which have focused on “psycholinguistics in the field” (funded by the JNU-Essex Development Fund (JEDF) and a staff mobility grant by the UK-India Education and Research Initiative (UKIERI)).
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. The glosses used in this paper are as follows: FV final vowel, REC reciprocal, PERF perfect, CAUS causative, NCP noun-class prefix (e.g. NCP1 refers to noun-class prefix 1), SG singular, PL plural, NOM nominal suffix.
2. IPA equivalent/m`pʰó /
3. All materials are available upon request.
4. The results of the mixed-model analysis were in line with results of per-subject and per-item ANOVAs, which are available upon request.