Compound worlds and metaphor landscapes: Affixoids, allostructions, and higher-order generalizations

Abstract

This paper investigates the use of German -landschaft ‘landscape’ and -welt ‘world’ as compound constituents. Both occur frequently in metaphorical uses such as Korpuslandschaft ‘corpus landscape’ or Arbeitswelt ‘labor world’. The high productivity of both compound types raises the question of whether [N-landschaft] and [N-welt] form constructions in their own right, both with a collectivizing meaning, and if so, how they relate to their respective higher-order schemas. More specifically, the question arises how the metaphorical-collectivizing uses relate to non-metaphorical ones. As both metaphorical and non-metaphorical reading variants coexist, these patterns show interesting similarities to those semantically bleached compound constituents that have been discussed under the label of “affixoids” in the morphological literature and in recent constructionist approaches. Drawing on synchronic data from the DWDS reference corpus of the 20thcentury and from the webcorpus DECOW16B as well as diachronic data from the German Text Archive, I argue that they form part of a family of collectivizing constructions. On a more theoretical note, I discuss how the higher-order generalizations connecting this family of constructions can be conceptualized.

Constructionist approaches: Past, present, future

“…Traditionally, scholars have largely relied on theoretical argumentation to motivate the existence of constructions at a certain degree of abstraction. In particular, researchers have increasingly posited lower-level schemas at intermediate levels of abstraction rather than highly abstract constructions (e.g., Boas 2003;Dąbrowska 2008;Hartmann 2019;Hilpert 2015). Recently, attempts have also been made to base such modeling decisions on quantifiable factors: Schmid (2020: 234), for example, suggests that the likelihood with which speakers form a schematic construction depends on the frequency and similarity of its instances, as well as the (syntagmatic) size of the pattern and its paradigmatic range.…”

“…Similarity, meanwhile, is more difficult to measure, but researchers have uncovered a number of corpus-based and experimental methods for the purpose. On the corpus side, collostructional analysis (see Stefanowitsch 2013 for an overview) has been used to compare the typical lexemes that combine with two constructions and thus obtain at least a rough impression of their similarity (Gries 2011;Hartmann 2019). In addition, distributional semantic methods such as semantic vector space analysis (see Lenci 2018 for an overview) yield quantitative measures of the semantic similarity between lexemes -or, if averaging over those lexemes, of the abstract constructions in which they occur -based on their collocational profiles (Hilpert & Perek 2022;Percillier 2020).…”

Constructionist Approaches

“…In particular, researchers have increasingly posited lower-level schemas at intermediate levels of abstraction rather than highly abstract constructions (e.g. Boas 2003;Dąbrowska 2008;Hartmann 2019;Hilpert 2015). Recently, attempts have also been made to base such modeling decisions on quantifiable factors: Schmid (2020: 234), for example, suggests that the likelihood of speakers forming a schematic construction depends on the frequency and similarity of its instances, as well as the (syntagmatic) size of the pattern and its paradigmatic range.…”

“…Similarity, meanwhile, is more difficult to measure, but relevant evidence could come from a number of corpus-based and experimental methods. On the corpus side, collostructional analysis (see Section 3.3) has been used to compare the typical lexemes that combine with two constructions and thus obtain at least a rough impression of their similarity (Gries 2011;Hartmann 2019). In addition, distributional semantic methods such as semantic vector space analysis (see Lenci 2018 for an overview) yield quantitative measures of the semantic similarity between lexemesor, if averaging over those lexemes, of the abstract constructions in which they occurbased on their collocational profiles (Hilpert & Perek 2022;Percillier 2020).…”

’n Gebruiksgebaseerde beskrywing van Afrikaanse prefiksoïede

“…Teenoor hierdie gebruike staan dood+ in komposita soos dood+s•kis en dod+e•mars, waar dood met die letterlike betekenis daarvan gebruik word. Daar is al 'n gevestigde literatuur vir die bestudering van affiksoïede in onder andere Nederlands en Duits, wat ouer bronne soos Schmidt (1987), Motsch (1996) en Van den Toorn (1983) asook onlangse bronne soos Hartmann (2019) en Stratton (2020) insluit. Benewens die onlangse beskrywing van affiksoïede in die konteks van Afrikaans morfologiese evalueringskonstruksies in Trollip (2022) ontbreek daar egter nog beskrywende en eksperimentele werk oor Afrikaanse affiksoïede.…”