Volume 59, Issue 2 - Special Issue: Left and right peripheries in discourse; Guest Editors: Matthias Klumm and Augustin Speyer

This paper provides an introduction to a collection of papers focusing on the left and right peripheries in discourse from both theoretical and empirical perspectives as well as across various languages. After outlining the main aims and questions underlying research on the left and right peripheries of discourse units, this paper presents an overview of the current state of the art by summarizing some of the key topics that have recently received increasing scholarly attention. These include (i) the question of granularity, i.e. the nature of the unit in relation to which the left and right peripheries can be defined, (ii) the question of how the left and right peripheries are internally structured and linguistically realized, and (iii) the question of where to draw the boundary between the core and the peripheries of discourse units. This introduction concludes with short summaries of the six papers included in this special issue.

In recent discourse approaches to the peripheries, sentence adverbs have received less attention than categories such as pragmatic markers or discourse connectors. This study compares the development of sentence adverbs and discourse connectors via a case study of evidential apparently , evaluative sadly , and discourse-connector instead . These expressions occur in present-day English in all three ‘peripheral slots’. The study shows that the three expressions evolved into peripheral stance expressions via similar mechanisms, the developments being favoured by occurrence in constructions which appear to provide ‘bridging contexts’. The expressions are viewed in the light of two models of the development of such peripherals: Grammaticalization theory and Diachronic Construction Grammar. The findings point to some compatibility with a discourse-prominence-based approach to grammaticality and some compatibility with a model of productivity within a higher-level modification schema. Apparent regularities in the diachronic development, together with parallels between sentence modifiers and constituent modifiers, and synchronic relations between the constituent source items and the emergent peripheral items, suggest that the ‘boundary’ between core and periphery is both permeable and gradient rather than clear-cut.

This paper traces the formal and functional development of evidential extra-clausal fragments of the form ( the ) last I/we  + VERB in American English. Using data from the Corpus of Historical American English and the Corpus of Contemporary American English , the study shows that these fragments are first attested in the late 19th century as a truncation of specificational structures (e.g. The last thing I remember is …). In the course of the 20th century, however, they exhibit a shift in allegiance by becoming increasingly associated with a temporal adjuncts template (e.g. The last time I checked …). Functionally, these evidential fragments are used as downtoners with additional functions being added in the course of the 20th century, viz. booster function and ironic use. The processes identified for their development are cooptation and subsequent grammaticalization, together with semantic specialization and constructional contamination. More generally, the fragmentary form ( the ) last I/we  + VERB is shown to specialize in evidential use, with non-evidential verbs becoming increasingly rare in this string, while the full form remains firmly non-evidential. Fragmentary form thus has become an important marker for evidential (and, by extension, epistemic) function.

The N- is construction has evolved from main clause to parenthetical, and bottom line has even become a mobile marker. The spoken section of COCA shows that left-peripheral bottom line can be surrounded by other extra-clausal expressions compatible with its pragmatic functions (projection, summation/(sub)topic shift, emphasis, epistemic endorsement and objectification). The aim of this paper is to establish whether these categories exhibit ordering preferences. The study reveals that the left periphery can be divided into nine optional slots. The first four slots usually express disjunction with the previous utterance; they include turn-initiators, then discourse connectors, then markers of disalignment, and then attention-getters/(sub)topic shifters. The second part of the left periphery (Slots 5–9) is more argumentative. The fifth slot comprises subjectifiers: expressions of viewpoint meant to attenuate the peremptory tone of bottom line , which occupies the sixth slot. The seventh slot is filled by the stance-marking discourse marker though , and the eighth slot is typically composed of expressions which de-subjectify the upcoming assertion. Finally, the ninth slot is occupied by addresses, which can be used strategically to maintain the recipient in an inferior interactional position. Together, these expressions form powerful rhetorical devices meant to convince interlocutors while preserving their faces.

Discourse particles have been often associated with the sentence peripheries. However, considering for instance optative particles, we observe that their distribution can vary cross-linguistically, since some languages use middle-field particles, while other display left-peripheral particles. In addition, we observe further differences between these two groups, for instance, the grammaticalization process of optative left-peripheral particles is different from that of middle-field particles. Furthermore, the combination with the subjunctive mood does not seem to be sufficient in the case of middle-field particles. Based on these and other properties, we propose a syntactic analysis adopting current views assuming an Agree operation between the optative particles and the left periphery of the clause, in particular with two distinct heads encoding clause type and illocutionary force.

The article describes the distribution of fronted adverbial constituents in Danish in the Prefield and the Extra-left (‘left dislocation’) positions: i morgen skal vi i skole versus i morgen der skal vi i skole ‘tomorrow we go to school’. The data, 635 sentences with a fronted adverbial, are extracted from seven sociolinguistic interviews. The article shows that some types of adverbials are more liable to occur in the Extra-left position than others. Conditional clauses and når -clauses (‘when’) occur almost always in the Extra-left position. Attitude and manner adverbials like  heldigvis ‘luckily’ and traditionelt ‘traditionally’ occur infrequently in Extra-left position. Contrary to what has been suggested in the literature, a principle of lexical weight does not have much explanatory power. Instead, grammatical principles and principles related to informational backgrounding provide explanations for the distribution.

The right clause edge links the utterance to the larger discourse and supports interlocutors in navigating the communicative situation, making such linguistic alternatives prime candidates to be exploited for register distinction. We investigate whether right-peripheral subjects behave similar with respect to registers in the verb-final languages German and Persian, using the multi-lingual Lang*Reg corpus (Adli, Aria, Elisabeth Verhoeven, Nico Lehmann, Vahid Mortezapour & Jozina Vander Klok. 2023. Lang*Reg: A multi-lingual corpus of intra-speaker variation across situations. Version 0.1.0. Zenodo ), which includes intra-speaker variation for six communicative situations differentiated by interactivity (monologue vs. dialog), social distance (close vs. distant), social hierarchy (equal vs. unequal) and mode (spoken vs. written). We identified four formal types of right-peripheral subjects: extraposition, right dislocation, afterthought + pronoun correlate and afterthought + NP correlate. Our analysis of the functions of these types across languages revealed that Persian extraposition and German right dislocation behave functionally similar whereas German extraposition is highly restricted and no Persian right dislocation could be found. The corpus analysis reveals that Persian and German speakers used the right periphery more often in interactive in contrast to monological communicative events.