A Morphological Approach to Discourse in Spanish: A Theoretical and Experimental Review

The goal of this dissertation is twofold. First, this is an empirical attempt to investigate whether certain discourse contexts determine 1) word order and 2) phonology. Second, this project provides a theoretical account for these findings. The test case is the Spanish left periphery; namely clitic-doubled left dislocations (CLLD), focus fronting (FF), and their canonical counterparts. Starting from the canonical example in (i), FF involves fronting of a constituent (as in (ii)), while CLLD involves the same fronting plus the addition of a clitic-doubling pronoun (as in (iii)). (i) María ve [a Pedro] en el parque. Maria see.3rd DOM Pedro in the park ‘Maria sees Pedro in the Park.' (ii) [FF A Pedro] ve María en el parque. (iii) [CLLD A Pedro] María [clitic-doubling lo] ve en el parque. In general terms, FF contrasts some information given in previous sentences (It is Pedro that Maria sees, not Pablo), while CLLD links the constituent to a previous referent (Where do I see my friends? I see Pedro here and Pablo there). Further, FF marks its contrastivity with the addition of an emphatic phonology, which does not happen in CLLD. Despite the general claim presented above, previous accounts on the interpretation of CLLD and FF disagree on the details. First, Bianchi and Frascarelli (2010) argue that CLLD can fulfill multiple discourse contexts: Introduction of new information (What happened? I saw Pedro in the park), contrastive link (Where do I see my friends? I see Pedro here and Pablo there), and one-to-one link (Where do I see my friends? I see my friends in the park). Further, these discourse functions cannot be fulfilled by canonical utterances, and each interpretation corresponds to a unique intonation. However, previous experimental attempts to find such interpretation-intonation correlations found no significant results (Gupton, forthcoming; Stavropoulou & Spiliotopoulos, 2011; Pešková, 2015, 2018). In fact, discourse-phonology correlation seems to take a different form in the Spanish left-periphery, with contrastive-focus (see FF in example (ii)) being expressed via contrastive-stress (Hualde & Prieto, 2015; Ladd, 2008; Pešková, 2015, 2018). In a second approach to the problematic Spanish left periphery, López (2009a) argues for a unique interpretation of FF (as in (ii)) and CLLD (as in (iii)), which can also be fulfilled by canonical utterances. Third, Rubio Alcalá (2014) argues for a complete interpretational freedom of all canonical and non-canonical utterances. However, neither Bianchi and Frascarelli, López, nor Rubio Alcalá provide experimental evidence either. Despite the theoretical claims discussed in Chapter 3, it is the mismatches in empirical predictions and the lack of experimental evidence that opens up many avenues for experimental research. In order to fill-in some of the gaps presented in the previous paragraph, this dissertation presents three studies. Study 1 (Chapter 4) analyzes the Spanish section of the NOCANDO corpus (Brunetti, Bott, Costa, & Vallduví, 2011) as a first tentative approach to the issue. Given some limitations in Study 1, I propose two additional experimental studies with native speakers of Spanish from Madrid, Spain. In Study 2 (Chapter 5), participants provide their judgments on different word orders (CLLD vs. canonicals) in a series of discourse contexts. In Study 3 (Chapter 6), participants’ scripted production is analyzed in order to learn about the phonology of different word orders (CLLD vs. FF vs. canonicals). Results show that speakers produce (Study 1) and accept (Study 3) both CLLD and its canonical counterpart in multiple discourse contexts (contrary to López, 2009a and Bianchi & Frascarelli, 2010). However, CLLD does not work as an answer to a wh-word (contra Rubio Alcalá, 2014) Further, no specific intonation pattern is found for each discourse context (Studies 1 and 3; contra Bianchi & Frascarelli, 2010; aligning with Gupton, forthcoming; Stavropoulou & Spiliotopoulos, 2011; Pešková, 2015, 2018). Instead, there is a pattern for contrastive focus (FF and its canonical counterpart) that includes contrastive stress, but is different from topics (CLLD and its canonical counterpart) that do not include any sort of emphasis (aligning with Hualde & Prieto, 2015; Ladd, 2008; Pešková, 2015, 2018). In sum, the results of these three experiments do not fully confirm any of the previous models. My theoretical claim to explain such variability in word order and phonology can be summarized as follows: Discourse features are morphemes from numeration merged as functional heads into the derivation. Under this morphological approach to discourse, discourse features such as [anaphora] or [contrast] are merged to the relevant constituent (e.g. [anaphoric-XP [anaphora] [XP]], similar to a prepositional phrase). Further, an optional movement to the left periphery in CLLD is caused by an unvalued feature in C-head (e.g. [CP [[anaphoric-XP] [[unvalued-anaphora] [... [t(anaphoric-XP) ]]]]], similar to wh-movement). For example, a feature such as [anaphora] explains the linked interpretation as well as the clitic-doubling present in CLLD, while [contrast] explains the contrastive interpretation and the emphatic phonology present in FF. This model simplifies discourse by the addition of only a small number of morphemes to the basic architecture of human language (Chomsky, 1995; contra Bianchi & Frascarelli 2010, and López, 2009). A complete description of this model is provided in Chapter 7.