| Popis |
Construction Grammar (CxG) is a cognitive approach to language that has gained considerable popularity in anglophone countries. It is therefore unsurprising that English is “the most widely analyzed language” in CxG (Boas 2010: 3) – at the expense of other languages, however. Boas, for example, has noted “a striking absence of cross-linguistic generalizations” (Boas 2010: 2) that is at odds with CxG’s “aspirations toward universal applicability” (Fried 2017: 249). Addressing this issue, I present a study on the English (1) and Spanish (2) comparative correlative (CC), a bi-clausal construction: (1) [The more I read,]C1 [the more I understand.]C2 (2) [Cuanto más leo,]C1 [tanto más entiendo.]C2 CUANTO more read:I TANTO more understand:I Each sub-clause of the CC consists of lexically fixed clause-initial elements (theC1/theC2 in English; cuantoC1/tantoC2 in Spanish), obligatory comparative elements (e.g. more; más) and optional clause slots (cf. lexically fixed idioms such as The more, the merrier). This complex structure encodes complex meaning, including asymmetric cause-effect (‘reading results in understanding’) as well as symmetric/parallel change-over-time (‘as more is being read, more is being learned’) (Hoffmann 2019; Horsch 2023).In line with Boas’ suggestion that English should serve as “basis” (2010: 14) for contrastive CxG-based investigations, I applied methodology from a study on the English CC (Hoffmann et al. 2019, 2020). I analyzed >8,000 CC tokens from English and Spanish corpora using covarying-collexeme analysis (Stefanowitsch and Gries 2005), testing for statistically significant cross-clausal associations between C1 and C2 regarding the variables FILLER TYPE in the comparative element slot and DELETION (whether the clause slot is realized or not), and their interactions. Assuming the usage-based view, i.e., constructions can be considered entrenched when they “occur with sufficientfrequency” (Goldberg 2006: 5), I posit these cross-clausal associations to be mesoconstructions that form taxonomic networks underlying the CC construction. Figures 1 (English) and 2 (Spanish) visualize the results of covarying-collexeme analysis in mental constructional networks, where the meso-constructional ‘nodes’ are interconnected based on formal similarities. The results show that despite language-specific idiosyncrasies, the ‘big picture’ reveals striking similarities: numerous meso-constructions, entrenched as a result of chunking and repetition (Bybee 2012), form elaborate constructional networks that are “baroque, involving massive redundancy and vastly rich detail” (Traugott and Trousdale 2013: 53). Furthermore, note that most cross-clausal associations are symmetric. In other words, what happens in C1 is mirrored in C2. This is explained by the principle of iconicity, i.e., the “(partial) motivation of a construction’s form by its meaning” (Hoffmann 2019: 12), in this case the CC’s symmetric/parallel change-over-time semantics (see above). Both observations confirm Goldberg’s Tenet #5, i.e., crosslinguistic generalizations can be attributed to domain-general cognitive processes (2003: 219) (chunking/repetition and iconicity) and prove that constructions are indeed “viable descriptive and analytical tools for cross-linguistic comparisons” (Boas 2010:15).
|