Counter-Argumentation and Anaphoric Substitution

Introduction

When we communicate through discourse, utterances give us access to the mental representations that a speaker ostensively wants us to understand (Graesser et al., 1997; Sanders, 2005a: 105; cf. Portolés, 2007: 107; 1998/2001: 27–30)¹. Constructing these representations is an active cognitive process that requires interlocutors to integrate the upcoming information into a coherent whole with other mental representations previously obtained either from preceding discourse segments or from our world knowledge (cf. Knott/Sanders, 1998: 135; Blakemore, 2001; Sanders/Noordman, 2000: 38; Sanders et al., 1992: 1–2). In establishing this “connectedness of discourse” (Sanders/Spooren, 2012: 5), words with procedural meaning such as connectors (e.g. English: thus/however; Spanish: por tanto/sin embargo) and pronouns (e.g. English: he/she; this/it; Spanish: él/ella; eso/ello) play a fundamental role as they constrain the inferential processing of discourse by imposing their instruction over the different aspects of meaning (Escandell/Leonetti, 2011; §§ 1; 2). On the one hand, connectors constrain the recovery of implicatures and, in doing so, aid retrieving argumentative relations (e.g. causal, additive, counter-argumentative, among others) between the mental representations. On the other hand, pronouns act on the explicatures, directing readers’ processing toward a pronominal substitution that integrates into the mental representation a referent that has already been mentioned (anaphora) or is about to come in discourse (cataphora).

In Spanish, however, these two processes can be signaled by a single unit as is the case of the counter-argumentative connective a pesar de ello (despite this), which, due to its level of grammaticalization, explicitly signals a pronominal anaphoric substitution by means of the pronoun ello along with a counter-argumentative instruction (Dominguez, 2007; Martín Zorraquino, 2010: 163; Portolés, 1999). This feature separates a pesar de ello from other counter-argumentative connectives like sin embargo and no obstante (however), which are more advanced in their grammaticalization process and have already lost their anaphoric element. A pesar de ello, then, activates an inferential route by which the antecedent referred to by ello, which can vary in extension (1a-b), is to be considered the ineffective obstacle for a conclusion presented in the segment that follows the connective (cf. Montolío, 2014).

1. (1) a. John está muy estresado. A pesar de ello, es feliz.
   • (John is very stressed out. Despite this, he is happy.)
   • b. John está muy estresado. Tiene un trabajo muy demandante. A pesar de ello, es feliz.
   • (John is very stressed out. He has a very demanding job. Despite this, he is happy.)

Given these features, it is unclear whether or not a pesar de ello conditions the inferential routes during communication in the same way other less grammaticalized connectives do, or if its explicit anaphoric element triggers a different processing pattern. Moreover, its frequent use in cases of long antecedents raises the question about whether a longer antecedent might affect readers’ processing strategies as compared to short-antecedent utterances.

Previous theoretical and experimental research on discourse particles (i.e. connectors, focus particles, among others) has shown that these units have a facilitating role during the recovery of a communicated assumption as they activate certain inferential routes according to their morpho-syntactic, semantic, and pragmatic features (Cruz/Loureda, 2019; Martín Zorraquino/Portolés, 1999: 4057; Nadal, 2019; Loureda/Acín, 2010; Portolés, 1999: 773; Recio, 2020). In the case of Spanish, a number of eye-tracking reading experiments carried out by the research team Diskurspartikeln und Kognition at Heidelberg University have led to the proposal of some marking principles that describe how the inferential restrictions triggered by these units take place and how this phenomenon modifies the processing patterns of utterances by optimizing the creation and evaluation of the communicated assumption (Loureda et al., 2021b). In sum, although connectives suppose extra lexical material to process, they actually reduce the effort related to the integration of two adjacent discourse segments into a discourse relation since they reduce the semantic underdeterminancy of the utterance (Loureda et al., 2021b; Nadal, 2019; Narváez, 2019; Recio, 2020). In addition, these studies made it evident that the impact of connectives in the recovery of an intended assumption changes according to their position in the utterance and to whether their instruction recovers conflictive or non-conflictive mental representations. Finally, such impact slightly varies depending on the level of grammaticalization as seen in some causal connectives (see Cuello, 2021 and Recio et al., 2018 for the case of por ello/eso/tanto).

Adding to this experimental research, this book presents the results of two eye-tracking reading experiments on the processing patterns triggered by the less-grammaticalized Spanish counter-argumentative connective a pesar de ello. In particular, the study reported herein aimed at determining if:

1. a) there is a difference in the processing pattern of counter-argumentative utterances marked by the connective a pesar de ello and the processing pattern of analogous utterances linked in an implicit causal relation;
2. b) the processing pattern of the counter-argumentative relation signaled by the connective a pesar de ello is the same if this connective is placed in an unmarked initial position than if it is placed in a marked mid or final position;
3. c) the processing pattern of utterances formed by two discourse segments linked in a counter-argumentative relation is the same in cases where the assumption recovered by a pesar de ello is conflictive and in cases where such instruction is not conflictive; and
4. d) given its anaphoric value, a longer extension and higher complexity of a first discourse segment affects the processing patterns previously observed in counter-argumentative utterances marked by the connective a pesar de ello.

To provide experimental evidence to answer these questions, the study considered five variables for the design of the stimuli and for the evaluation of the results:

The analysis herein presented will show that the instruction of a pesar de ello conditions discourse processing in a similar fashion to what has been observed in other connectives. In addition, it will examine if any alteration in the expected pattern can be attributed to the explicit marking of the anaphoric pronominal substitution and to the extension and complexity of the antecedent.