Volume 28, Issue 4 -

The middle voice has received ample attention in the literature, yet the precise boundary between middle voice and other related constructions still remains elusive. For example, do sentences like Mary slept (*herself) well last night and Mary washed (herself) thoroughly and expertly belong to the middle voice or the reflexive domain, or are they simply intransitive one-participant structures? While ambiguity between reflexive and middles has been noted by (Kemmer, Suzanne. 1993. The middle voice . Amsterdam & Philadelphia: John Benjamins.), I show that the problem is more widespread. Depending on the marking patterns available in a given language, such ambiguities can occur between middles and prototypical intransitive one-participant events, or between middles and reflexives. Using data from two languages with distinct marking patterns – English (a language with a reflexive marker but no middle marker) and Romanian (a language with one marker of each type) – I discuss the possibility of distinguishing middles from other related constructions. In English, it is shown that the him/herself test can be used to distinguish direct middles from intransitive one-participant events. In Romanian, we see middle and reflexive markers used together in the same construction without contradiction, bringing together both middle and reflexive semantics (as well as marking). In agreement with (Maldonado, Ricardo. 2000. Spanish reflexives. In Zygmunt Frajzyngier & Traci Walker (eds.), Reflexives: Forms and functions , 153–185. Amsterdam & Philadelphia: John Benjamins.) and (Manney, Linda. 2001. Middle voice in Modern Greek: Meaning and function of a morphoyntactic category . Amsterdam &Philadelphia: John Benjamins.), the account given here supports the view of the middle voice as a unified phenomenon, and following (Maldonado, Ricardo. 2009. Middle as a basic voice system. In Lilian Guerrero, Ibáñez Sergio & Belloro Valeria (eds.), Studies in role and reference grammar . México: Instituto de Investigaciones Filológicas, UNAM.), the main function of the middle is to profile the core properties of events (but not necessarily to reduce the relative elaboration of participants, as proposed by Kemmer, Suzanne. 1993. The middle voice . Amsterdam &Philadelphia: John Benjamins.). So, unlike active and passive voice which concern the focusing of various participants (namely, Agents and Patients, respectively), the middle voice focuses the event itself.

This article explores the correlations between linguistic figurative features and their corresponding conceptual representations, by considering their respective continuities and discontinuities in language shift. I compare the figurative encoding of emotions in Kriol, a creole of northern Australia, with those of Dalabon, one of the languages replaced by this creole, with a particular focus on evidence from metaphorical gestures. The conclusions are three-fold. Firstly, the prominent figurative association between the body and the emotions observed in Dalabon is, overall, not matched in Kriol. Secondly, although this association is not prominent in Kriol, it is not entirely absent. It surfaces where speakers are less constrained by linguistic conventions: in non-conventionalized tropes, and gestures in particular. Indeed, some of the verbal emotion metaphors that have disappeared with language shift are preserved as gestural metaphors. Thus, Kriol speakers endorse the conceptual association between emotions and the body, in spite of the lower linguistic incidence of this association. The third conclusion is that therefore, in language shift, conceptual figurative representations and linguistic figurative representations are independent of each other. The former can persist when the latter largely disappear. Conversely, the fact that speakers endorse a certain type of conceptual representation does not entail that they will use corresponding linguistic forms in the new language. The transfer of linguistic figurative representations seems to depend, instead, upon purely linguistic parameters.

We advance theory formation in cognitive sociolinguistics by exploring the extent to which language users’ probabilistic grammar varies regionally. For this purpose, we investigate the effects of constraints that influence the choice between the two syntactic variants in the well-known dative alternation ( I give Mary a book vs. I give a book to Mary ) across nine post-colonial varieties of English. Using mixed-effects logistic regression and adopting a large-scale comparative perspective, we illustrate that on the one hand, stability in probabilistic grammars prevails across speakers of diverse regional and cultural backgrounds. On the other hand, traces of indigenization are found in those contexts where shifting usage frequencies in language-internal variation seem to have led to regional differences between users’ probabilistic grammar(s). Within a psycholinguistically grounded model of probabilistic grammar, we interpret these results from various explanatory perspectives, including language contact phenomena, second language acquisition, and semantic variation and change.

This article develops an analysis of event/result nominals and gerundives from a Cognitive Grammar perspective. By reviewing the previous research, we first point out that these phenomena are much more flexible than the extant research claims. Moreover, widely accepted generalizations concerning the phenomena are, at best, only partially true. We demonstrate that the said flexibility is ascribed to two different types of construals: [1] mass-like construal accompanied by reification and [2] zone-activation or metonymic shift. Event nominals arise, without respect to the types of the nominal affixes, when the relationship profiled by a verb takes an internal perspective. Result nominals arise through zone activation or metonymic shift in addition to the reification of the verbal base. Several grounding strategies apply to both event and result nominals, thereby yielding different realizations of instances such as (in)definite and possessive. We show that our analysis can be systematically extended to gerundives, which permit limited grounding methods. We also demonstrate that V-to-N converted event nominals are accounted for unproblematically in our analysis because the rise of event nominals does not rely on the nominalizing affixes.

In this paper, we assess the developmental trajectories by which children approach adult levels of complexity and informativeness in the linguistically and conceptually challenging domain of spatial language. To this end, we look at three types of spatial relations (localization, spontaneous and caused motion) in spontaneous German child speech (age 2;6 to 2;11 and 4;6 to 4;11), and in elicited Frog Story narratives from German child and adult speakers (3-, 5-, 9-year-olds, and adults. Children are generally sensitive to typological preferences. From early on, their productions reflect target-language-specific lexicalization patterns. Our analyses show that they still approach adult-like levels of information complexity and density only gradually. This concerns the local complexity (structural repertoire for the conceptual slots figure, verb, path/ground), as also established in previous research, but in particular the global complexity , as investigated in this study. Global complexity measures the structural integration of information, or the combinatorial complexity that surfaces at the utterance level. As predicted by usage-based theories, adult-like degrees of informativeness and information density are only reached gradually, although the component parts at the local level are available earlier in development.