Collocazioni lessicali in spagnolo: fra teoria linguistica e grammatica descrittiva

This study explores the phenomenon of lexical collocations in Spanish, aiming to provide a theoretical analysis of an intensely debated phenomenon which is still in need of a satisfactory definition. The research starts with the examination of a corpus of collocations extracted from the relevant literature and sheds light on the lack of homogeneity in the description of the combinations traditionally labeled “collocations”. Although this linguistic phenomenon has been approached from different perspectives in the literature, there is a fundamental question to be addressed concerning the definition of the linguistic nature of collocations: in fact, they cannot be considered the result of morphological processes because of their phrasal status; neither are they standard syntactic objects because of their fixity in use. Further, the frozen nature of collocations is not a consequence of their frequent use in common language, as argued in many traditional approaches (Firth 1968; Sinclair 1991; Halliday 2004). An appealing hypothesis, which has inspired the present research, is found in Bosque (2001), where it is argued that collocations are linguistic objects at the interface between lexicon and syntax. Their specificity lies in the peculiar selection restrictions that seem to involve complex metaphorical processes. The overall claim is that the collocational thread is stored in the lexicon along with most words and idioms. This idea finds support in Jackendoff (2002) who makes a clear distinction between storage and online computation. His conception of the lexicon is in contrast with Chomsky’s - whose lexicon is a list of irregularities - mainly because it allows the storage of items smaller or larger than words, from inflectional morphemes to whole texts. This heterogeneous set of lexical items can be viewed as part of a Lexicon-Syntax continuum of forms ranging from the most frozen (idioms) to the least ones (free combinations). The main feature of this pattern lies in scalarity - a property which is at odds with the classical view based on systems of categorization - which allows clarifying the proximity of collocations both to free combinations and to idioms, without forcing them into either one or the other class. Further, it is argued that the frozen nature of the collocational thread depends on the activation of some peculiar cognitive-semantic mechanisms. In particular, lexical decomposition processes can arguably shed light on the fact that the selection between the elements of a collocation (the base and the collocate) is orthogonal to the identification of the “formal” head of the construction. Specifically, concerning V-N and N-A (or A-N) collocations, the pivotal element corresponds to the more salient lexeme from a cognitive point of view (the base, namely N), which shows semantic autonomy and transparency but it does not automatically coincide with the syntactic head. On the contrary, the collocate (V or A) depends on the base for its semantic specification and it commonly receives a figurative interpretation. On these grounds, I will develop the hypothesis that it is the noun which selects its verb - and not viceversa - by activating subparts of its conceptual structure related to Pustejovsky’s qualia or to independent processes of metaphorical interpretation. Lexical 2 decomposition strategies also emphasize the regular activation of cognitive mechanisms in the selection processes in lexical collocations. Despite its semantic transparency, the base seems to undergo a conceptual representation which enables, through metaphorical mechanisms, the selection of a collocate belonging to a radically different conceptual domain. In particular, nouns belonging to abstract categories tend to be conceptualized as they were concrete objects. This kind of conceptualization (and the central role that spatial concepts play in it) triggers the selection of the collocate and its metaphorical interpretation. In destapar un secreto, for instance, the conceptualization of secreto ‘secret’ as a concrete, closed container which can be uncorked to let its content come out, triggers the semantic selection of a verb such as destapar ‘to uncork’, which is given a figurative sense. The internal cohesion of collocations is thus often the result of the metaphorical relationship between the base and its collocate. Given the high complexity of the interpretive mechanisms at the basis of the collocational thread, the most plausible option is its storage in long-term memory. To conclude, the corpus analysis has enabled us to propose a double categorization. Based on the taxonomy of metaphors proposed by Lakoff and Johnson (1980), the first categorization sorts collocations according to the type of metaphor detected on the collocate. The second one classifies different kinds of multi-word constructions on the basis of syntactic fixity. Further, the Lexicon-Syntax continuum model adopted in the present study highlights the scalar nature of fixity and idiomaticity, which are directly related in multiwords expressions: the more idiomatic a combination the more ‘fix’, hence ‘lexicalized’.