Abstract
The present study is a corpus-assisted analysis of discourses on the Catalan movement of independence in the Catalan and Spanish press. A referendum for Catalan independence was held on 1 October 2017 but was not approved by the central Spanish government and was declared illegal. Spanish and Catalan newspapers differed considerably in their treatment and representation of the conflict, which also drew the attention of the international press to Catalonia. My analysis interconnects theoretical perspectives from linguistic critical discourse studies, cognitive linguistics, and media studies. The methodology is based on the analysis of two comparable corpora of journalistic texts (in Catalan, Spanish). My aim is to identify and compare different features of the semantic prosody of words describing the event of geopolitical separation and the different narratives on this sociopolitical conflict that these prosodies contribute to producing. More specifically, I focus on a selection of keywords, like independence, secession, and sovereignty, that seem to be indicative of different discursive strategies consciously, and possibly unconsciously, chosen by a selection of national Catalan and Spanish newspapers. By carrying out a contrastive analysis of the two corpora, I am able to describe certain similarities, like the similar frequent use of the word independence, and differences, like the more frequent use of the word secession in the Spanish corpus, that indicate different framings of the same event and can be related to different ideological positionings against or in favor of Catalonia’s independence from Spain.
Keywords
Introduction
The present study is part of an investigation I am carrying out into potential differences between the representations of national identities in journalistic sources in Catalan and Spanish. The project was initially motivated by the sociopolitical conflict in Catalonia sparked by the referendum held on 1 October 2017 that had been called by the Catalan government in order to decide the issue of Catalonia’s political independence from Spain. The referendum had not been authorized by the Spanish government and was declared illegal.
This study will focus on a number of linguistic elements that characterize the discourse surrounding independence in several Catalan and Spanish newspapers. By adopting a corpus-assisted approach to the analysis of linguistic discourse, my objective is to achieve an accurate description of the linguistic elements characteristic of the discursive patterns apparent in discussion of the independence question in the press and to compare these patterns in order to highlight differences and formulate explicative hypotheses about their causes.
Theoretical framework
The study adopts the theoretical framework of Discourse Studies from a critical and linguistic perspective (Otaola Olano, 2006). The approach is both qualitative and quantitative, inductive and descriptive. It is linguistic in the sense that it aims to highlight the textual and discursive functions that lexical items such as independence may have in relation to the evocation of certain semantic frames and the framing of newspaper narratives about geopolitical separation. It is critical because these narratives are often a consequence of explicit or implicit ideological positions or naturalized linguistic behaviors. The focus on lexicon is justified by the importance that the words used to describe events, people or objects may have in activating certain conceptual frames in the readership, as numerous studies in Cognitive Linguistics have shown (see Section 2.2). These frames are often constructed by highlighting certain linguistic features and toning down or erasing others. As a consequence of this perspectivization, different framings of the same event and different narratives about it may be produced. Morphosyntactic structures, phonetic features, and non-verbal elements such as colors, images, or a certain typographic style may also contribute to framing events in journalistic articles. Although important, these latter features are not taken into account here due to the limited scope of the study. Linguistic Discourse Studies (LDS) and Critical Discourse Studies (CDS) are inter-, multi-, and transdisciplines. This allows me to combine theoretical notions drawn from several fields of research, such as corpus linguistics, cognitive linguistics, and media studies. Despite the variety of theoretical perspectives, the notions adopted here are all closely interrelated.
Semantic prosody of a lexical item
The concept of a lexical item relates to ‘meanings that are different from residual meanings in that contextual features around a core are taken into account’ (Mahlberg, 2007: 194). Within the framework of CDS, the ultimate aim of which is to describe the relationship between social structures, power relationships and language, corpus-assisted linguistic research often focuses on lexical items. In order to identify the main linguistic features surrounding a lexical core, Sinclair (2004) introduces four categories: collocation, colligation, semantic preference, and semantic prosody. These categories are characterized by an increasing degree of generalization: the category of collocation takes into account the repeated co-occurrence of words in a text, colligation refers to the co-occurrence of grammatical phenomena, semantic preference refers to the relationship between a core and other words semantically related to it, and semantic prosody considers the pragmatic meanings of words appearing in a core’s immediate context. These four categories are interconnected since the observation of collocations or colligations may lead to the description of semantic preference and prosody. Semantic prosody is ‘the most variable and often most difficult component to describe’ (Mahlberg, 2007: 195) 1 but, as Sinclair (2004: 34) observes, it plays an important role since it indicates the approximate ‘function’ of the item in relation to the others and has, therefore, discursive relevance. The notion of semantic prosody was originally developed by Louw (1993) and defined as a ‘consistent aura of meaning with which a form is imbued by its collocates’ (p. 157). If a number of words frequently co-occur with a core, then their connotations may spread over to the core. Semantic prosody, also labeled ‘discourse prosody’ (Stewart, 2010: 161), is an important factor in the characterization and description of lexical units within a discourse. It has been frequently applied in connection with corpus linguistics since the study of concordances and collocates facilitates the detection of lexical nets that may produce certain types of semantic prosody (Baker et al., 2006; Tognini-Bonelli, 2001). 2
Semantic frames
When it comes to defining the relationships existing between entities both in discourse and in cognition, a second semantic notion imported from cognitive linguistics, namely that of semantic frames, turns out to be central to my research. Originally proposed by Tesnière (1959: 152) as a semantic counterpart to his dependency structures in syntax, it was later developed by computer scientists such as Minsky (1975), Õim and Saluveer (1985), and many other pioneers of Artificial Intelligence, and then adopted by Fillmore (1985) as a further theoretical refinement to his ‘Case Grammar’. It can be described as a tool for organizing experience, for making sense of it and for describing and explaining its lexical units (Fillmore, 1985: 232). Semantic frames are interrelated since a frame is ‘any system of concepts related in such a way that to understand any of them you have to understand the whole structure in which it fits’ (Fillmore, 1982: 111). Frames can be evoked by linguistic elements in the text but can also be invoked by readers through reference to the context (Fillmore, 1985: 232).
We can directly relate the idea of the semantic frame to the concept of framing, which was developed in the social sciences by Entman (1993) as an elaboration of Goffman’s (1974) concept of the frame and can be defined as a selection of textual elements that highlight and make salient some aspects of a perceived reality ‘in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation’ (p. 52). This notion has been successfully used, and problematized, in media studies (Cacciatore et al., 2016; Scheufele and Tewksbury, 2007). I use it here not only in relation to frames and semantic prosody but also because the focus on selection and salience in its definition provides a direct link to a corpus-assisted approach like mine (see, e.g. Czulo et al., 2020).
Methodology and corpus design
The general framework provided by CDS, the concept of semantic prosody and the methodologies drawn from corpus linguistics are closely interlinked (Koller and Mautner, 2004: 222). In most cases, research works with specialized rather than general corpora (as, e.g. in Baker, 2006; Partington, 2003) since the objective is to investigate the link between certain meanings circulating in society through specific sets of texts (e.g. journalistic texts, medical texts, or texts about specific themes). In my case, in order to identify recurring features characteristic of discourses about independence in several different newspapers, I adopted a corpus-assisted approach in order to achieve a blended reading 3 of certain textual and discursive phenomena related to the synchronic representations of independence in different cultural contexts.
The corpus has been analyzed using Sketch Engine (www.sketchengine.eu), a piece of software designed for use in text analysis and the text mining of large corpora in many languages. The main functions used, which we will describe during the analysis, are as follows:
Extraction of keywords and terms
Concordance analysis of Key Words in Context (KWIC)
Word Sketch, analysis of keyword collocates 4 as they occur in different grammatical contexts
Thesaurus
The research method consists of two closely interlinked stages: (a) an exploratory phase in which the analytical functions provided by the software are used systematically to gather quantitative data, interpret them and formulate hypotheses; (b) a targeted corpus-based analysis that aims at verifying initial hypotheses as well as those developed in the previous phase. The identification of two distinct stages implies a certain chronological order; however, in practice, they partly overlap and are not considered separately here, in order to avoid redundancies.
Corpus design
In the present study, I analyze two different corpora, in Catalan and Spanish respectively which I designed in accordance with the parameters of a German corpus (Giugliano, 2019), labeled CAT (see Table 1), adapted from the Sinner, Paasch-Kaiser, Neuhaus, and Gräsche Corpus of the University of Leipzig (Sinner, 2020). This decision was motivated by my intention to compare the German, Catalan, and Spanish corpora later in my future research. This operation is only possible if the three corpora are comparable, that is, if they are designed in accordance with similar parameters.
German corpus: CAT.
The majority of those corpus-assisted studies in CDS that use ad hoc corpora stress the importance of setting transparent criteria for the selection of the texts that are going to make up the corpus. Designing a corpus means making sure that it is representative and balanced (Bowker and Pearson, 2002). However, the parameters by which a corpus achieves such standards are not fixed since they depend on a wide number of features ultimately determined by the research objectives. In the case of the German corpus, representativeness and balance were achieved by assembling a homogeneous selection of articles concerning the Catalan referendum from four mainstream German newspapers and a broadcasting channel. The selection was limited to the period from 06.09.2017 to 10.10.2017, a time during which discussion of the question of Catalan independence was particularly lively in the German press due to the chronological immediacy of the events in question.
For the design of the Catalan and Spanish corpora, I used similar parameters. Both the Catalan corpus (CatRef) and the Spanish one (CatEs), summarized in Tables 2 and 3, are composed of two subcorpora corresponding to the articles included in the domestic politics section of two newspapers (El Punt/Avui and Ara for Catalan and El País and La Vanguardia for Spanish). As we can see from the tables, the Spanish and Catalan corpora are both larger than the German corpus. This is understandable if we consider that the number of articles devoted to the Catalan referendum in the Catalan and the Spanish press in 2017 was much higher than the coverage in the German press. The selection of the newspapers forming the subcorpora was determined by a number of considerations as well as practical constraints. El Punt/Avui and Ara are the most widely distributed newspapers in the Catalan language, if we consider both the print and digital editions. El Punt/Avui is more openly in favor of the process of independence but both newspapers can be considered moderate. 5 As for the Spanish corpus, I decided to select two of the most important mainstream Spanish newspapers whose political orientation is also moderate, that is center-right (La Vanguardia) and center-left (El País). 6 For this reason, I decided to exclude other mainstream right-wing-oriented newspapers, like ABC or La Razón. 7 Moreover, by selecting these four newspapers, I also achieved two corpora of comparable size, as can be seen in Tables 2 and 3.
Catalan corpus: CatRef.
Spanish corpus: CatEs.
Different instantiations of a frame of geopolitical separation and their discursive environments
Previous findings: The corpus-driven phase
In a previous study (Giugliano, 2019), I analyzed the coverage of the Catalan independence movement in the German press. My research allowed me to identify five key terms (Unabhängigkeit, Abspaltung, Loslösung, Separatismus, Sezession) often used by the newspapers in my German corpus to describe the geopolitical separation of Catalonia from the Spanish state. The study argued that each term evokes a conceptual frame that shares elements of a hyperframe labeled GEOPOLITICAL SEPARATION (see Figure 1) but at the same time focuses on different frame elements, offering a different perspective that may lead to a different framing of the events and, as a consequence, a different interpretation of them.
Features and structure of the hyperframe of GEOPOLITICAL SEPARATION (Giugliano, 2019).
Depending on the particular focus in question, it was hypothesized that some lexical items (Unabhängigkeit) were closer to the discursive environment in favor of Catalan independence, some were closer to the discursive environment against it (Separatismus and Sezession) and others occupied different positions in between (Loslösung, Abspaltung). It was also observed that the term Unabhängigkeit had the highest absolute frequency and the lowest degree of variation among the subcorpora (i.e. among the newspapers). This was seen as a possible indication of a generalized and ideologically neutral use of the word, since Unabhängigkeit can be considered a general label for the topic dealt with in the articles and was the literal translation of the Catalan word independència, which had international resonance in the media. Finally, it was established that, although the analysis may point to discursive patterns in the representation of the sociopolitical conflict in Catalonia, it is not possible to establish the ideological affinity of a newspaper through consideration of the discursive environment in which the five lexical items representing the various instantiations of the frame (Unabhängigkeit, Abspaltung, etc.) were located solely on the basis of the frequency of the term in a given subcorpus.
A comparative analysis of semantic prosodies of independència (CAT) and independencia (ES)
Drawing on the analysis of the German corpus, I shall focus in the present study on the detailed contrastive analysis of the semantic prosody that can be observed in relation to the core words independència/independencia in the Catalan and Spanish corpora respectively. I start the comparison by considering the frequency of both items, summarized in Table 4.
Frequencies of independence in the Catalan and Spanish corpora.
I have preferred to calculate the relative frequency per 100,000 words rather than the percentage in order to have results that can be read and compared more easily.
Despite the differences in absolute frequency, we observe that the relative frequency of the two words is similar. However, the Catalan word independència does not appear to have a high ranking in the automatically generated list of keywords (position 649). This fact seems discursively relevant, especially if we consider that in the German corpus the equivalent term, Unabhängigkeit, appears among the first 25 keywords. In the Spanish corpus independencia appears higher up in the ranking (position 193) than its Catalan equivalent, although it is not among the most frequently encountered items. As we know, keywords are generated by comparing the frequency of words in a focus corpus (i.e. the corpus we designed for the analysis) with the frequency of words in a non-specialized reference corpus (in the case of Catalan, caTenTen14). 8 If independència does not appear to be a keyword, it means that the lexical item is already widespread and relatively common in texts published in the Catalan Web in 2014. The resources currently at my disposal are insufficient to enable a comparison between my corpus and older Catalan corpora.
Collocations of independència in the Catalan corpus
For the collocation analysis, I use the software’s Word Sketch function, which proposes a list of collocations of independència in different syntactical relationships, such as, for example, prepositional phrases, phrases with independència as a subject or object and modifiers of independència.
The results include:
(a) lexical elements that suggest opposing positions in the conflict: those in favor of independence (as in partidari, a favor (in favor), or sí a la independència (yes to independence)) and those against it (as in contrari a la independència (against independence)).
(b) Verbs such as declarar and proclamar, which, however, seem to be a standard collocation in the general language, as a comparison of our focus corpus with the general reference corpus Catalan Web 2014 (caTenTen14) has confirmed.
(c) Modifiers of independència. In this case, however, there is lack of frequent modifiers. Independence is at times described as eventual (potential, 9 occurrences), hipotètic (hypothetical, 4), or catalana (9).
Finally, in order to identify and describe the semantic prosody of the term, I use the Thesaurus function that provides an automatically generated list of synonyms or words belonging to the same category. The list is based on the context in which the words appear in the corpus. The function draws on the theory of distributional semantics and considers that words appearing in the same context, namely words with similar collocations, are similar in meaning. The function works effectively with corpora of billions of words and with common terms. I did not apply this function in my analysis with the intention of identifying synonyms, which was unlikely due to the dimensions of the corpus, but because I was interested in potential similarities between the word independència and other words that share a broadly similar distribution. Figure 2 offers a visual rendering of the results as provided by Sketch Engine.
Words sharing context with independència.
Table 5 is a list of the lexical items visually represented in Figure 2, accompanied by their frequency and similarity scores. The similarity score shows the percentage of collocates the terms in the list have in common with the word independència. The words in the table are not synonyms of independència. We know, however, that they partly share the same distribution, so that we could hypothesize that they belong to a common discursive environment, share similar discursive strategies and help to characterize the semantic prosody of the lexical item. This hypothesis is partly validated by the analysis of the first terms in the list. The term referèndum, for example, is closely related to that of independència as both the frequency and score values seem to suggest. I also found the strong association established between independència and words such as sobirania [sovereignty] and república [republic] to be particularly noteworthy. The relevance of these terms, which have a relatively low frequency (94 and 282 occurrences respectively) compared to other collocates of independència, is confirmed by their similarity score. The score determines their significance in this context and allows us to assume that the discourse about sovereignty and the Republic uses verbal elements that are similar to those that orbit around independència, so that these words share features of semantic prosody. I shall focus here on the first three lexical items associated with independència (referèndum, sobirania, and república).
First 20 words sharing context with independència.
Referèndum
The word referèndum appears 3781 times in the corpus, that is, approximately three times more that the word independència. It can be found in collocation with independència in a relatively small number of cases (52) in the form of referèndum d’independència (30), referèndum sobre la independència (17) [referendum on independence] and referèndum per la independència [referendum for independence] (4). As mentioned above, however, what is relevant here is not the actual collocation of these two lexical items but their shared collocates. The most relevant collocations of referèndum seem to be those in which it appears as object, as in Table 6.
Verbs with referèndum as object.
The list of verbs with referèndum as object (of which Table 6 only presents the first 10) suggests the presence of a duality, namely two opposed positions in the sociopolitical conflict. The two positions are represented by verbs that can be grouped around two thematic domains: (a) against the referendum (as in prevent (impedir), stop (aturar) and, further down the list, avoid (evitar), suspend (suspender), suppress (reprimir) and ban (prohibir)) and (b) in favor of the referendum (as in promote (celebrar, impulsar), negotiate (pactar) and, again further down the list, protect (defensar, garantir) and organize/allow (organitzar, permetre)). A similar balance can be seen in the modifiers of the lexical item, which is described as illegal [il·legal] 30 times as well as legal (11), unilateral (10), and binding [vinculant] (5). Both positions have an individual focalization which is pro-Catalonia, so they do not indicate the presence of opposing discourses. This becomes clear if the immediate context is taken into account, in which verbs that indicate opposition to the referendum or modifiers that connote it negatively are mostly part of the reported speech of Spanish institutions or politicians.
The referendum is also described as the Catalan referendum [referèndum català] 33 times, which could be understood to suggest a point of view removed from the Catalan perspective and may refer to reported speeches. Finally, a very frequent collocation is referèndum d’autodeterminació [of self-determination].
Sobirania [sovereignty]
From a discursive viewpoint, the category ‘verbs with sobirania as object’ seems particularly relevant as it may indicate a range of related perceptions and attitudes, as can be seen in Figure 3.
Verbs with sobirania as object.
The notion of sovereignty may refer to both the Spanish state and to Catalonia. We need to take into account the context in which it is used in order to understand whether we can actually attribute the use of the word to one or the other of these opposing discursive fields. In both cases, in fact, sovereignty may be seen to be in danger, although discourse about the impending danger to Spanish sovereignty implies an ideological perspective diametrically opposed to discourse on the danger to Catalan sovereignty. The concordance analysis of sobirania as an object confirms that both parties use the term. When its use refers to the sovereignty of Spain the term is often, though not exclusively, accompanied by the adjective nacional. It is part of politicians’ speech (either in direct or indirect style) and is mostly presented in discourse against the separation of Catalonia from Spain, as an existing reality whose integrity is threatened by the independence movement, (of being fragmented (trossejar), shattered (esmicolar), despised (menysprear), eliminated (liquidar), or broken (trencar)). When the term is used to represent the sovereignty of Catalonia within pro-independence discourse, it is occasionally accompanied by the term nacional and is described as something that is in the process of being created (the expression cap a la plena
República
The term república appears 282 times in the corpus (relative frequency of 30.15). An analysis of its most important collocations does not show the presence of duality (duality of positioning in favor or against something, as in the case of referèndum, or duality of discursive use, as in the case of sobirania) and prompts questions regarding why this is so. Let us first observe the most relevant collocations, summarized in Table 7.
Collocations with república.
The analysis of modifiers in collocation with república shows that the institution is often qualified as Catalan (82 occurrences) and nova [new] (25). What is interesting about these first two collocations is that they together form a third collocation, nova república catalana (12 occurrences). From a discursive viewpoint, the phrases nova república and nova república catalana may suggest two coexisting concepts, even though they do not necessarily have the same cognitive prominence: on the one hand, the phrase points to the fact that the republican institution would be something that did not previously exist and was now to be newly created, as can be seen, for example, in the sentence Tot el món està pendent avui de Catalunya.
The absence of discursive oppositions seems to be suggested also by the verbs that have república as an object or subject. The word appears, in fact, like a target that needs to be achieved or a project to be realized and consolidated, as verbs like proclamar [proclaim], constituir [constitute], construir [build], néixer [be born], and voler [want] seem to suggest, as do reivendicar [reivendicate] and defensar [defend].
As for the interpretation of this absence of duality, there may be several reasons behind it. Firstly, for the pro-independence movement in Catalonia, the republic is envisaged as the only form of government after the separation from Spain and is conceptually in opposition to the Spanish monarchy. In discourse against the separation of Catalonia from Spain, conversely, independence is perceived as a threat to the unity of the Spanish State but not to the current form of the Spanish government (monarchy). Secondly, while it is possible to present sobirania from different points of view and be against or in favor of a referèndum, especially if it is considered illegal, the concept of republic is so strongly linked to positively connoted concepts such as democracy, for example, that it does not seem possible to be against it. The hypothesis that this observation prompts is that the lexical item is part of the semantic prosody about the independence of Catalonia in Catalan media but may not be found as a salient feature in discourse opposed to independence. Since the term cannot be explicitly opposed and presented in discourse in negative terms, it could be that it is simply omitted. The study of the Spanish corpus, in which positions opposed to Catalan independence are expected to be more prominent, is a first step toward the verification of this hypothesis.
Collocations of independencia (ES) and its comparison with Catalan
As in the previous section, I start the analysis by considering the word’s grammatical and collocational behavior and observe a number of similarities with the results of the Catalan corpus. Modifiers such as hipotética, unilateral, and eventual (Figure 4) are highly ranked in the corpus (i.e. they possibly have a high discursive relevance since their LogDice score measuring the strength of the collocation is high) although they do not necessarily have a high frequency.
Modifiers of independencia.
Another frequently encountered collocation is independencia catalana. In the Catalan corpus, the same collocation appears nine times with a score of 6.98. This comparison made me reflect on the discursive significance of these closely corresponding values in the two languages. The association of the notion of independence with the culture in which the sociopolitical conflict takes place, although superficially similar in both languages, may actually reflect differing ideological perspectives and discursive positions, depending on the language in question. My hypothesis was that in the Catalan corpus the collocation independència catalana contributes to the definition of a possibly positive Catalan self-image. If the perspective is external, then the collocation may imply a process of ‘othering’ in which a hetero-image of the Catalan culture is produced within the larger Spanish community. In order to avoid overinterpretation, I carried out a KWIC (Keyword in Context) analysis of the collocations in both the Catalan and the Spanish corpus and could observe that critical attitudes related to the collocation could be found in both corpora. However, while the examples of negative representation of Catalan independence in the Catalan corpus are fewer, relatively mild and found alongside positive comments (e.g. Fander Falconí publicava dimecres a El Telégrafo que ‘la independència catalana és un fet històric’ (Fander Falconí published on Wednesday in El Telégrafo that ‘Catalan independence is a historical fact’)), in the Spanish corpus the attitude is more openly, and more frequently, critical, as we can see in Examples 1 and 2.
Example 1
The English, not the Scots, have dug their own grave. The idea of
Another similarity between the two corpora is the frequent collocation independence of Catalonia (independencia de Cataluña) with 142 occurrences), which has a similar score in both (11.83 in the Spanish corpus and 11.05 in the Catalan). The phrases independencia catalana and independencia de Cataluña may seem synonymous, and certainly are to a certain extent. However, the significantly higher frequency of the second phrase in both the Catalan and the Spanish subcorpora suggests that its use is more common. This raises questions regarding the differences in meaning of the two phrases and leads to hypotheses related to a partly differing conceptualization of the event. It could be argued that, whereas the phrase independencia de Cataluña focuses more on the geopolitical event by mentioning the political region itself, the phrase independencia catalana may evoke identitarian issues. This hypothesis cannot be tested here due to the limited scope of the study.
Other frequent collocations are with the verbs declarer [declare] (124) and proclamar [proclaim] (41), and with declaración. As in Catalan, such collocations in Spanish can be considered frequent in the general language.
Finally, I analyze the results delivered by the Thesaurus function, summarized in Figures 5 (visual rendering) and 6 (list of the first 10 words).
Words sharing context with independencia.
List of words sharing context with independencia.
Secesión [secession] has the highest similarity score in the Thesaurus list, followed at a certain distance by referéndum, declaración, and ruptura [break]. The results are partly different from those obtained using the Catalan corpus. Lexical items like ‘sovereignty’, ‘republic’ and ‘liberty’ do appear in the Spanish thesaurus list as well, but with a much lower ranking and with lower similarity scores. Using the pattern of analysis adopted for the Catalan corpus, I will look at a selection of words (secesión, referéndum, declaración, and ruptura). During the analysis, I will also consider the contrastive perspective, using the results previously obtained from the Catalan corpus.
Secesión
My initial hypothesis was that secessió, the Catalan term that corresponds to secesión in Spanish, would not be found often in the Catalan corpus since it evokes a conceptual frame and an underlying ideology opposed to the separation of Catalonia from Spain (see also Giugliano, 2019), which does not conform to the ideological positions of the two Catalan newspapers represented in the corpus (Ara and El Punt/Avui). Conversely, I expect secesión to have, in comparison, a greater frequency in the Spanish corpus (using material from El País and La Vanguardia), since the political position of at least one of these newspapers (El País) was more openly against the separation of Catalonia from Spain and against the independence referendum itself.
In the Spanish corpus, secesión occurs 253 times. In terms of relative frequency, this corresponds to 22.7 occurrences per 100,000 words. This is approximately seven times more than the result from the Catalan corpus (3.4/100,000). I will consider here four types of collocation: (a) modifiers of secesión, (b) verbs with secesión as object, (c) verbs with secesión as subject, and (d) ‘de secesión’ the most frequent prepositional phrase that includes the word.
The most common modifier found in collocation with secesión (Figure 7) is the adjective unilateral, which so far had been found in collocation with referéndum in particular. What is atypical about this relatively frequent collocation is that it seems emphatic if not altogether pleonastic. According to the simple definition of the term in the RAE dictionary, secesión is an ‘Acción por la cual se separa de una nación una parte de su pueblo y de su territorio’ [Action by which a part of a nation’s people and territory is separated from it]. This action may theoretically be the result of mutual agreement; however, its common use in discourse usually refers to an act of separation that is unilaterally prompted and often unilaterally accomplished, meaning that it does not always enjoy immediate official recognition by the state. In the entry for ‘Secession’ in the International Encyclopedia of Political Science, Bartkus (2011: 2367) observes that:
[t]he decision to secede represents an instance of political disintegration [. . .]. Because of the state’s opposition and monopoly of coercive force, secessionist struggles frequently become violent and protracted. Thus, secession is disintegrative in the most fundamental sense: It involves not the overthrow of the existing government but rather its territorial dismemberment.
Modifiers of secession.
The word may, therefore, have negative connotations. 9 Other modifiers used to qualify the term (encountered less frequently) are eventual, improbable, hipotética, and posible. With regard to the verbs with secesión as object, we find similarities with the Catalan corpus, since the most frequent collocations are with declarar, avalar [endorse], apuntalar [fortify/unterpin], and proclamar. The most frequent collocate with the prepositional phrase de secesión is referendo/referéndum de secession, followed by several negatively connoted words that suggest danger or illegality, like riesgo [risk], guerra [war], proceso [process], absolutismo, vertigo, perjuicio [damage] in addition to more neutral terms such as caso [case], voluntad [will], favour and ventaja [advantage].
In the Catalan corpus, the term secessió appears 32 times, which, in terms of relative frequency, corresponds to 3.4 occurrences out of 100,000 words. Its low frequency, in itself discursively relevant, makes a quantitative study of its use in the corpus rather difficult. If, for example, the term appeared once only as object of the verbs avalar [endorse] and aconseguir [achieve], it would not really be possible to detect a discursive trend or identify discursive patterns. The same observation can be made when the term is found in other syntactic relations. A concordance analysis of all occurrences of the term provides additional contextual information about its use. I cannot trace any discursive patterns here, but I can observe that the word is often used in a context that suggests direct or reported speech by a person or institution. In the case of direct speech, this is evident from the use of inverted commas, as in ‘un fals dret a votar la secessió o, directament, al reconeixement de la independència’ [a false right to vote for secession or, directly, for the recognition of independence] whereas, in indirect speech, the concordance analysis of just a few lines of co-text seems to suggest this but is not always sufficient to identify it clearly.
To conclude, we can argue that the initial hypothesis seems confirmed. Secesió and the words derived from it are not part of the dominant discourse in the Catalan corpus. They tend to represent the speech of individual actors, either persons or institutions, whose standpoint is more distanced. As for the Spanish term secesión, its significantly higher frequency (seven times higher than in Catalan) and the types of collocations that we have identified suggest certain discursive patterns that stress the unilateralism of an act of secession, its illegality, the dangers it implies for the state, its improbability and unreality, making it an act, therefore, that cannot be acknowledged.
Referéndum
This term has a high frequency in both lists (3354 in the Spanish subcorpus vs 3781 in the Catalan one), but the score indicating how frequently it is found in collocation with independencia is higher than the score for the parallel terms in the Catalan list (0.206 compared to 0.187). An analysis of the modifiers of the term, and the contrastive perspective that is introduced, reveal differences that may be indicative of certain discursive patterns. As can be seen in Figure 8, some of the modifiers qualify the term negatively.
Modifiers of referendum.
Referéndum illegal is by far the most frequently encountered collocation (310 occurrences, with a similarity score of 13.23), both in comparison to other collocations in the category (the following term in the list, referéndum independentista, appears 55 times with a score of 10.29) and to the results from the Catalan corpus, in which the corresponding term appears 28 times. Since a referendum on the territorial independence of a region is not encompassed by the Spanish Constitution and the legislation relating to such a referendum approved by the local Catalan government was rescinded by the Spanish Constitutional court, it may appear that the modifier ilegal has a simple descriptive function. However, the difference in the frequency of occurrences in the two corpora may also suggest different discursive uses and two different contrasting ideological positions: one that aims at invalidating the referendum and another that aims to grant it legal substance. The opposite collocation, referéndum legal, only appears 27 times. As can be seen from the context, the expression (which is frequently found in the expanded form referéndum legal y pactado (agreed)) can be taken as an indicator of how few conciliatory voices there were calling for dialog and negotiation between the opposing parties. To sum up, a comparison of the semantic prosody of referendum in the two corpora shows an important difference: in the Catalan corpus anti-referendum positionings may not be shared, as the KWIC analysis of the various incidences has shown; however, voices both in favor of and against the referendum are made visible (the latter through direct or reported speech). In the Spanish corpus, conversely, the dominant voice is against the referendum and positions arguing in favor of its legality do not seem to be represented.
The referendum is also described as catalán (49 occurrences), which may trigger a process of ‘othering’ in relation to the Catalan community, as I have argued above.
With regard to the verbs in collocation with referéndum as object (Figure 9), we observe similarities with the Catalan corpus. The word celebrar [in the sense of ‘hold’, ‘conduct’] is more frequent in the Spanish corpus and has a higher score. However, verbs like celebrar and convocar are widespread collocates of the lexical item referéndum and, as a consequence, cannot be regarded as indicators of distinct discursive patterns in both corpora. In order to prove this point, I looked for collocations of the term in larger and older reference corpora. For Spanish, I checked the European Spanish Web 2011 (eseuTenTen11) with 2,021,633,644 words and, for Catalan, the Catalan Web 2014 (caTenTen14 v2) with 182,691,653 words. The results confirmed the collocates convocar and celebrar. As for the other lexical items in the list, they appear in both corpora with very similar scores, despite the different absolute frequency that may depend on the slightly different dimensions of the two corpora.
V words sharing context with independència. erbs with referéndum as object.
Declaración
This term occurs 658 times in the corpus and is found in frequent collocation with the modifier unilateral (217 occurrences) and with the noun independéncia (declaración de independéncia, 372 occurrences). If we add the total number of occurrences recorded for this latter collocation to the number recorded for the collocation declaración unilateral de independència (189 occurrences), the total (561) shows that declaración is almost exclusively used in reference to the sociopolitical event of separation. Other less frequent collocates suggest the precariousness of the event (posible, eventual, hipotético) or its institutional frame (institucional, formal, responsable).
The concordance analysis of referéndum and declaración and its comparison with the Catalan corpus show that these terms may be considered standard features of the semantic prosody of independència in both corpora but not distinctive traits of ideologically motivated discourses.
Ruptura
This term appears 210 times. Its discursive relevance lies not in its absolute frequency, or in the frequency of its individual collocations, which are never high. Rather, these become significant when considered as a whole. Its very presence in the Thesaurus list is noteworthy, since its Catalan equivalent (ruptura/trencament) does not appear at all in the corresponding list for Catalan. In order to appraise its function in discourse, I carried out a concordance analysis of all occurrences and could make the following observations:
The term is very frequently used in collocation with the word ley [law] (as in ley de ruptura, leyes de ruptura and norma de ruptura) and refers to the label used by the newspapers in our corpus to refer to the ley del referendum or ley de transitoriedad juridical y fundacional de la república [law on the juridical and foundational transitional nature of the republic], which were adopted by the Catalan parliament on 07.09.2017. A simple internet search reveals that its use was shared by other Spanish newspapers. Moreover, the terms ley/norma the rupture are almost exclusively found in El País subcorpus. Since this represents a difference from the Catalan corpus, the term is significant for the description of discourse. A qualitative assessment of larger fragments of context in which the term is used points to a depiction of these laws as illegal and non-constitutional.
The term is also used to indicate a disruption of legality, of Spanish and European institutions, of dialog and of communal living in society.
Finally, ruptura is used to refer to the possible geopolitical separation of Catalonia from Spain. In discourse, it is used occasionally as a synonym of secesión or independencia, as in the expressions proclamación/declaración de ruptura. In this context, it is also seen as illegal, dangerous, and unilateral.
Discussion
During my analysis, I combined methods of different granularity (from the broad overview of quantitative data via the concordance analysis of collocates to a more detailed study of larger segments of text) in order to identify the main lexical elements that compose the discursive prosodies of the word independence in two corpora of journalistic texts, in Catalan and Spanish respectively. A comparison between them reveals a number of linguistic similarities, but also some striking differences, and both of these aspects are relevant for a critical analysis of the discursive patterns that they may produce.
The frequency of the word independence in the Spanish corpus is basically similar to that of the corresponding word in the Catalan corpus. This similarity provides further support to the hypothesis, already mentioned in Giugliano (2019), according to which the notion of independence in an anti-independence discursive context becomes a general label for a sociopolitical event. Its use, however, is not discursively and ideologically neutral, as the analysis of the two semantic prosodies of the word in the two corpora shows. In Catalan, independència often co-occurs, appears in the immediate context of, or shares collocates with the words referèndum and sobirania. These words can present a duality of readings since they may be used for either fostering the empowerment of the pro-independence movement or discrediting it. In the case of the word referendum, this duality of positions seems to be equally represented in the Catalan corpus in terms of occurrences, though the anti-referendum voice is criticized in discourse. In the Spanish corpus, the pro-referendum position expressed is made invisible, as the very low frequency of positively connoted modifiers of the term seems to suggest.
The tendency to shun opposing voices in the Spanish corpus seems to be confirmed by the fact that the word sovereignty, which can be conceptualized in opposing terms (as sovereignty of Spain or of Catalonia) does not seem to characterize the prosody of independence in the Spanish corpus, while it does appear in its duality in the Catalan corpus. The presence of the word republic in both corpora is an additional example of difference and seems to confirm the discursive strategy detected in the Spanish corpus. The Thesaurus list in the Spanish corpus has shown that the term república appears in a much lower position in the ranking of collocates of the node word independence, and with a much lower similarity score. However, the frequency of the term in both corpora suggests a more complex phenomenon. In fact, in both the Catalan and the Spanish corpus, the term república has a very similar absolute and relative frequency (282 occurrences and 30 words per 100,000 for Catalan, 206 occurrences and 27 words per 100,000 for Spanish). These data make it possible to argue that, despite the similarities in frequency of the two equivalent words, the lexical unit in the Spanish corpus is less visible in discourse about independence than it is in the Catalan corpus. The word is present in the Spanish corpus, but absent in its discourse.
Finally, the discursive function of modifiers like Catalan or unilateral, whose score is similar in both subcorpora, suggests the hypothesis that these collocations may either produce a negative prosody, leading to a process of ‘othering’, or provide an inner perspective that consolidates discourse about one’s own self-image. This would be a consequence of recontextualization, according to which equivalent linguistic elements may bear different meanings when used in different linguistic and cultural contexts. This hypothesis was also prompted by the observation of other textual elements but needs to be verified with further research.
As for the linguistic features that clearly mark a difference between semantic prosodies and contribute to a critical reading of discourse about independence, the most relevant seems to be the use of the word secession in the Spanish corpus, whose occurrence in the Catalan corpus is very low. This lexical item is used as a linguistic alternative to independence in the Spanish corpus, occupies a similar distributional position and refers to the same geopolitical event of separation. The absence in Catalan becomes relevant as it highlights a specific discursive strategy in the media that form the Spanish corpus (especially the El País subcorpus, in which almost all of its occurrences are to be found) and a clear ideological positioning against Catalan independence. Secession is not a neutral term that stands for separation but rather a word evoking conceptual frames that focus on the negative consequences for Spain of any separation by perspectivization of the Spanish State’s point of view and stressing, with its semantic prosody, the danger, illegality and threat that the separation implies. As I observed during the analysis, it cannot be claimed that the general Spanish readership is aware of the framing produced by the term. Nonetheless, the claim seems confirmed by both the linguistic elements of its semantic prosody and its very low frequency in the Catalan corpus, in which the term, recontextualized, acquires different pragmatic meanings that shed negative light on the pro-independence social actors.
Another example of significant difference in the semantic prosody of independence is represented by the term ruptura, almost absent in the Catalan corpus and used in the Spanish corpus either to mention a number of laws approved by the Catalan Parliament whose official designation does not suit a certain anti-independence discourse or to refer to the geopolitical event of separation. In both cases, it reveals a discursive strategy (well-known in CDS and Cognitive Linguistics 10 ) that aims at framing the event in a negative light since the conceptual frame evoked by the word stresses its potential disruptive consequences for the Spanish state and for Spanish and Catalan society.
Conclusions
The corpus-assisted analysis and comparison of the semantic prosodies of the word independence in Catalan and Spanish has highlighted the importance for discourse description not only of differences between the two corpora but also of apparent similarities (e.g. of frequency and keyness). Similarities may be of different kinds and on different scales. We have observed similarities in the literal meaning of words in Catalan and Spanish, such as those between the Catalan word independència and the Spanish independencia, or secessió and secesión. These lexical items are semantically equivalent in the two languages, but their recontextualization, that is, their use in different discursive contexts, and characterization through different semantic prosodies lead to different, even opposed, pragmatic meanings that prompt different conceptual frames, produce different framings of the event and represent different ideological positions. If unquestioned, their use may unconsciously consolidate these cognitive frames in the discourses (and brains) of biconceptual (i.e. undecided) people and exacerbate discursive polarization. This aspect seems particularly relevant if we consider that the use of the word secession is found predominantly in the El País subcorpus and that El País is the leading Spanish newspaper in terms of distribution.
Other kinds of similarity concern frequency, such as, for example, that of independence in both corpora, whose semantic prosody, however, differs, or of republic, whose use does not have a similar discursive relevance in both corpora.
Differences have proved to be as relevant as similarities for a critical analysis of discourse. The different frequency and score of the word secession, the use of a negative label for a set of laws (leyes de ruptura) instead of their official designation, and the different characterizations of the words referendum and sovereignty in the corpora seem to confirm a discursive strategy, present in the Spanish corpus, that aims to highlight one political position (against independence) and silence the pro-independence voice. In the Catalan corpus, criticism is not spared for the anti-independence party, but both voices are reported.
Finally, the study confirms the validity of a corpus-assisted approach for the critical analysis of discourse since its results provide a solid quantitative foundation for further research about different representations of identities in the Catalan and the Spanish press and the risks of essentialization and polarization.
Footnotes
Acknowledgements
I would like to thank Dr. Héctor Hernández Arocha, Dr. Elia Hernández Socas, Prof. Yvette Bürki and Prof. Oliver Czulo for their helpful comments on this article.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.