A Constructionist Account of an Extragrammatical Process: Blending and Lexical Series in Italian

Extragrammatical Morphology lies outside the boundaries of morphological grammar and comprises, among others, metalinguistic and paralinguistic phenomena (Dressler 2000: 1). Blending, or “extra-grammatical compounding”, is sometimes referred to as a word creation process by which sophisticated and abbreviatory operations are performed (Dressler 2000: 2-4; Ronneberger-Sibold 2006). For Renner (2020: 9), blending serves two primary functions of word formation, that is, the transconceptual and the compacting functions. Both are displayed by the following examples of Italian blends (1a-b): (1) a. manzobab ‘a kebab made with beef’ < manzo ‘beef’ x kebab ‘kebab’ b. colf ‘domestic worker’ < collaboratrice ‘helper’ x familiare ‘family’ The corpus of 1136 Italian blends which is analysed in this study comprises two lists of items obtained adopting two different methodologies: the item in 1a pertains to a list of 820 lexical blends collected by manually checking all low-frequency words (absolute frequency ≤ 10) surrounded by single or double quotes (called metasignals in Svanlund 2018: 125) inside the Italian Timestamped JSI 2014-2022 (Bušta et al. 2017); the item in 1b pertains to a list of 316 lexical blends collected by past studies on blending in Italian. The items have been collected adopting a flexible approach to Italian blending, whereby blends are conceived as an often unexpected result of two formal operations: a combination of two (or more) input words; a moderate reduction of at least one input word (Micheli 2020: 172; Renner 2023: 2-3). In Italian, blending contributes to the emergence of new lexical series, i.e., a family of lexemes formed by the same splinter of an input word retained in the blend. For the inclusion of a lexical series, criteria similar to those adopted by Bauer et al. (2019: 59-60) and Micheli (2022: 3, note 8) have been followed, i.e., among 1136 items, if a splinter occurred at least twice and if it generated at least 15 different types in the Italian Timestamped JSI 2014-2022 corpus, such types have been considered as belonging to a lexical series. Below, 4 splinters that fall under this case are reported: Table 1: Lexical series generated by Italian blends Model blend Splinter Meaning Some example As it could be seen in Table 1, the length of the splinter retained can vary across types. In this study, such splinters are considered combining forms in view of the full lexical meaning conveyed (Iacobini 1999; but see Bauer et al. 2019: 62 in which they are compared to affixes). With time, splinters may undergo grammaticalization from “unan- alyzable simplexes” to “complex semi-idiosyncratic forms” (Brinton and Traugott 2005: 94). Moreover, the meaning of a splinter employed in lexical series acquires often a ’bound meaning’, similarly to what happens for compound constituents (Booij 2010: 66-71). Adopting a constructionist framework, if blends’ splinter reach the status of com- bining forms, lexical series can be schematized as constructional idioms, i.e., “schemas with partially prespecified constituents and corresponding meanings”, such the one below (Booij 2010: 74): (2) <[aperi- [X]Nj ]Nk ⇐⇒ [happy hour to celebrate SEMj]k > (3) <[pan(ta)-[X]Nj ]Nk ⇐⇒ [trousers with Relation R to SEMj]k > In sum, in this study, lexical series generated by blending in Italian are analyzed quantitatively from the perspective of Construction Morphology (Booij 2010) in order to look at lexical series patterns in Italian that have already undergone or are undergoing grammaticalization from a formal and semantic point of view. In particular, the aim is to analyse the types and tokens generated by each lexical series, to look at their type of concatenation (Bauer et al. 2019: 63), and to extract a constructional idiom from each of them. This particular issue could be described as well as a movement from the external, or extragrammatical, area of morphology to the periphery of it (Marginal Morphology) (Dressler 2000: 6-7).