Discourse markers from processes of monologization: Two case studies Andrea Sansò 1. Introduction: the diachrony of discourse markers Diachronic research on discourse markers has repeatedly highlighted the chaotic nature of the changes leading to their emergence and the unpredictability of their semantic and pragmatic extensions over time (Brinton 1990: 49; 2001: 149; see also the recent surveys in Fedriani & Sansò 2017: 14ff., and Sansò 2020: 67-75). Such an impression of chaos has been variously explained as either the result of the particular conditions under which discourse markers acquire new functions in context, or by invoking some functional specificity of these items, which separates them from other grammatical elements. Brinton (2008: 129), for instance, has shown that identifying an ordered sequence of functional developments of the English reformulation marker I mean is a pointless enterprise. The subjective and intersubjective uses of I mean (e.g. as an emphatic marker or as a softener/mitigator) arise with no apparent order «as invited inferences in appositional structures, where a previous element in the discourse is restated or reformulated», and often coexist in a single context. Kaltenböck et al. 2011 (see also Heine 2013, Heine et al. 2017) have explained the special behavior of discourse markers in diachrony as intimately connected to the nature of the process leading to their emergence, which they call cooptation. This process is conceived of as an instantaneous operation consisting in setting off (syntactically and prosodically) an element from sentence grammar, which regulates constituency and dependency relationships among phrases, and deploying it as an element of discourse grammar (or thetical grammar). The instantaneous character of cooptation also explains why many of the concomitants of other types of diachronic development, such as morphological fusion, phonological reduction, and obligatorification are not present in such processes, while other phenomena such as layering (intended as the simultaneous presence of new functions alongside old ones) are, on the contrary, ubiquitous. The peculiarities of the diachronic developments of discourse markers, however, should not blind us as to the possibility of finding regular patterns in this domain. Such regularities include, for instance, the cyclical nature of changes (similar source meanings produce similar outputs in different languages, triggering the same or very similar inferences across different sociohistorical contexts; cf. Hansen 2018a; 2018b), the existence of politeness reversal phenomena following similar patterns across languages (cf. Ghezzi & Molinelli 2014; Mazzon 2017; Fedriani 2019), or the tendency to develop specialized functions depending on the specific periphery in which a given element appears (Degand 2014, among many others). The present article aims to contribute to the debate on the diachrony of discourse markers by discussing a type of change that will be labelled monologization. Monologization consists in the progressive emancipation of a given element from a dialogic structure to become an autonomous, monological marker/construction, which no longer requires the original dialogic structure in which it was embedded in order to be felicitously uttered. This type of change is not at all new: as will be discussed below, the literature on semantic and pragmatic change has described various instances of monologization using different labels (or no label at all). I will show that this type of change is pervasive as a source of discourse markers by virtue of their predominantly interactional nature. The two case studies that will be analyzed below involve two Italian discourse markers that have developed new functions in recent years: (e) niente (lit. ‘(and) nothing’) and che poi (lit. ‘that then/later’). Both these discourse markers have developed monological functions as topic orientation markers, «by which the speaker’s intentions concerning the immediate future topic of the discourse can be conveyed» (Fraser 2009: 893), starting from more dialogical functions. This development can be observed by comparing two corpora of spoken Italian collected in two different periods, with an