Studies in Language. International Journal sponsored by the Foundation “Foundations of Language” - Volume 18, Issue 2, 1994

This paper suggest that in order to understand the cross-language distribution of the subjunctive mood one needs to understand the cross-grammar distribution of the irrealis modality, as well as have a general theory of modality, within which irrealis takes its rightful natural place. The subjunctive mood turns out to occupy two coherent sub regions within irrealis: (a) the subjunctive of lower certainty (within the epistemic sub-mode of irrealis), and (b) the subjunctive of weaker manipulation (within the deontic sub-mode of irrealis). A grammaticalized subjunctive may take the same form in both sub-modes (as in Romance), or two distinct forms (as in some Bantu languages). The paper also investigates the diachronic pathways of grammaticalization through which various subjunctive patterns arise, showing that analogical extension patterns exist in both directions — from deontic to epistemic and vice versa. The interaction between grammaticalization patterns and functional universals is also discussed.

A rigorous definition of 'suppletion' is proposed, based on prototypical cases and the following pivotal idea of the current notion: two linguistic signs are in relation of suppletion if the semantic difference between them is maximally regular (i.e., it is grammatical inflectional or derivational) while their formal difference is maximally irregular (i.e., it is not covered by any alternation). The paper offers a concise survey of major types of suppletion: derivational vs. inflectional suppletion; radical vs. affixal suppletion; suppletion of morphs vs. of megamorphs vs. of idioms; a few dozen examples are quoted. After an analysis of five difficult cases, suppletion is considered along the diachronic axis. As a conclusion, a related concept is introduced and delineated with respect to suppletion: pseudo-suppletion.

The present paper* deals with the important phenomena ergativity and embedding in the aboriginal Australian language Dyirbal.Embedding and ergativity are intimately connected in the syntax of Dyirbal. This motivates the composite theme of my paper.Ergativity, Dyirbal, and embedding are intensively discussed in current functionally oriented syntactic typology. However, despite the fact that these topics have been investigated in depth, the final integrative understanding seems to be still awaiting. This is so mainly because the original analyses of the phenomena in question have incurred the most diverse reinterpretations and even misrepresentations in mainstream theoretical discussions. Many theoreticians have either distorted facts and evidence so as to make these "data" fit their theories, or they have made superficial studies in which important details — which could not be accommodated in their theories — have been neglected. Functional Grammar (cf. Dik 1989; Siewierska 1991), though claiming typological adequacy, is nevertheless no better in this regard: without any evidence for the postulated scenario, Dyirbal ergativity is maintained to have developed out of the passive voice of a hypothesized earlier accusative stage of the language. An accusative analysis of the historical origin is indeed possible, but the evidence points in another direction than that speculated in Functional Grammar. The reconstruction of Dyirbal ergativity is one of the major points of this paper.The present understanding of the theme Dyirbal ergativity and embedding is presented, not only as valuable in itself, but also as a necessary revision of Functional Grammar in particular, as well as a contribution to the functionalistic paradigm in general.The outline of the presentation is:1. introduction — presentation of the theme;2. theoretical background, exemplification, and historical hyotheses concerning the theme;3. elaboration on the theme;4. the topic Dyirbal embedding in particular;5. theoretical implications of the present understanding of the theme;6. conclusion — moral;7. appendices.