Exploring the role of context models in memory-building: the completion of informative voids and the re-organisation of narratives in second-hand memories

Introduction

A major concern in post-conflict societies is the consensual construction of a shared memory regarding the past and often traumatic violence. Identities created during and polarised by the conflict usually last long after the official or factual end of visible violence (Webel and Galtung, 2007), leading to a 'discursive' version of the past conflict, in which the main battle shifts towards the sphere of contested narratives.

This is the case in the northern Spanish / southern French region of the Basque Country. October 2011 marked the end of ETA's (Euskadi Ta Askatasuna) armed attacks, the Basque separatist organisation dubbed as 'the last insurgency in the heart of Europe' (Doucet, 2017). A decade later, the first generation of local young adults who have no direct memories on the past violent conflict is coming of age, but a proper culture of memory has not been established yet. On the contrary, narratives about the Basque conflict are still a disputed topic in the political arena, as well as in formal education, leading to what many media outlets have called 'the battle of narratives' (Gara Editorial Board, 2017). These discrepancies have so far prevented the regional Parliament from designing textbooks or other pedagogic material aimed at transmitting memory (Velte, 2019).

It is in this complex context in which the first post-conflict generation is building its own narratives about the past, necessarily leaning on second-hand memories and external discourses. The lack of an "official" narrative in formal education and the conflict of interests in the political arena are not being balanced by a private transmission within families, as some parents do not know how to talk about the past. Others are afraid of transmitting part of past traumata to their children, or they often assume that knowledge about the past is socially shared information that does not need to be made explicit again. This leads among youngsters to a feeling of 'entering someone else's territory without permission' (their parents' experiences) or 'having to ask something we should know'. Transmission of memory is therefore often stalled, not only because of fear or grief caused by past experiences, but also because of embarrassment of one's own lack of knowledge on the side of the younger generation (Velte, 2020).

However, not having a planned public policy of memory does not mean that no transmission is happening at all. The younger generations are re-building the past in between informational voids, leaning on a variety of external sources in order to infer the missing information and construct coherent narratives that help them understand the past.

This paper aims to transcend the specific case analysed and provide a deeper understanding of the psycholinguistic mechanisms involved in collective memory-building and the concrete discursive strategies employed by post-conflict generations in disputed contexts, which can be applied to other situations in an international context.

Background

Transmission of collective memory between generations is often seen as a unidirectional process through which discourses and knowledge "flow" towards younger generations, who learn about past events by simply receiving and processing the information generated by older generations based solely on the latter's experience. However, interdisciplinary approaches to Memory Studies, combined with Cultural Studies and Discourse Analysis, agree more and more on the fact that the construction of a shared memory involves a series of actions by the transmitting group as much as by the receiving younger generation. Concepts such as 'technologies of memory' (Shahzad, 2011, p. 378) have been used to explain the changing nature of memorialistic discourses across generations: according to these, memory is not a fixed cluster of experiences or narratives that travel unchanged through time, but rather an inter-active process of meaning-making closely tied to the construction of discourse, and therefore highly dependent on the concrete context of the discursive event. Contextual information of the remembering person or community and their pre-established belief systems about this context are therefore crucial factors in memory-building.

In this paper, we explore some of the mechanisms that members of the first post-conflict generation in the Basque Country employ in order to fill in these gaps of information when remembering, and how they rely on contextual information in order to construct coherent narratives about the past violence.

On the theoretical level, we combine theories from Memory Studies and theoretical-methodological perspectives from Discourse Studies. On the one hand, we lean on Maurice Habwachs' concept of collective memory, which defines memory as a fundamentally socially constructed and dependant on metaphorical cadres sociaux (social frames), that is, previously established patterns of meaning which lead our process of interpretation and inferences of new information (Halbwachs, 2004). Memory as a mainly collective phenomenon is therefore strongly linked to the collective identity of the group in which the individual 're-members' (Halbwachs, 1968; Erll, 2017, p. 6).

This idea of relating shared knowledge or a common view of the world with individual meaning-making processes was also introduced by theories underlying Critical Discourse Analysis, namely Van Djik's sociocognitive theory, according to which mental models 'embody the interface between episodic, personal knowledge of events, on the one hand, and the socially shared beliefs of groups' (Van Dijk, 1997, p. 190). Or, as Ruth Wodak formulates it,

When updating information from [new discourses], mental representations necessarily must guide our understandings and mental models most probably support linking new information with stored information. Due to different belief and knowledge systems, news is experienced and stored depending on available cognitive frames. (Wodak, 2006, p. 182)

The transmission of memory is fundamentally a discursive action: memory is necessarily expressed through some kind of language (visual, verbal, or combined) or shared code. In order to infer meaning from external discourses, the receiving part needs a series of cultural knowledge, including contextual information. These pre-established cognitive frames enable individuals to fill in gaps when re-organising information about the past in the form of narratives.

Narratives are culturally canonical linguistic forms that modulate the organisation of experienced events. Narratives provide a sequential organisation that specifies the unfolding of an event along temporal lines, but even more so, narratives provide an explanatory and evaluative framework for understanding how and why events unfold as they do (Fivush, 2010, p. 89)

Apart from its strong collective component, remembering is primarily characterised by being a constructive and present-oriented process: versions about the past change with each act of remembering, according to the position of the speaker, their self-image and their collective identities within a changing present context. In this process, 'distortion and re-balancing, up to fiction' are often possible (Erll, 2017, p. 14).

Narratives about the past are also dependent on the perceptions speakers have about the context in which the communicative act develops. Van Dijk (1997, p. 194) defines context models as 'episodic, personal and hence subjective interpretations and experiences of the communicative event or context'. These models, understood as folders of information about one’s own position and about what is considered appropriate or expected in each situation may also guide the construction of discourses and the interpretation of new information.

Methods

On the methodological level, we lean on the Discourse-Historical Approach developed by Ruth Wodak, which 'centers on political issues and seeks to integrate as any of the genres of discourse referring to a particular issue as possible, as well as the historical dimension of that issue' (Van Leuwen and Wodak, 1999, p. 91). This approach focuses on the broader historical, socio-cultural and political background in which discursive events are embedded, and analyses how these change according to contextual factors. Speakers employ certain discursive strategies, defined as 'plans of action with various degrees of elaborateness, the realization of which can range from automatic to conscious and which are located on different levels of our mental organization' (Van Leeuwen and Wodak, 1999, p. 922).

In our analysis, we focus on strategies of transformation employed by our interviewees in order to re-construct narratives about the past and fill in gaps of information in their earlier constructed mental models. These include, according to Van Leeuwen and Wodak (1999, pp. 96-98), the deletion of information, the rearrangement of elements within discourse, the substitution of termini and metaphorisations and additions (including personal reactions).

Discourse Analysis is characterized by the importance of integrating contextual information with the linguistic analysis of concrete utterances. It takes into account four levels of analysis (Weiss and Wodak, 2003, p. 22):

1. The immediate, text internal co-text.
2. The intertextual and interdiscursive relationships involved.
3. The extralinguistic sociological variables and instutionalised frames of a specific 'context of situation'
4. The broader sociopolitical and historical contexts.

Following these steps, we analyse 9 text sequences extracted from 8 in-depth interviews conducted with young people aged 16 to 21, who were all born between 1997 and 2002, and therefore hold very few or no direct memories at all about the violence (that ended factually in 2009). The main topic of the interviews was the memory about the past and the construction of identities after the conflict: participants were asked about their personal and familiar experience, their feelings and attitudes towards concrete issues related to the conflict and the actors and historical events they considered most important.

These are very short fragments extracted from long interviews. Contextual information is a key factor in order to infer meaning out of these excerpts. On the one hand, the discourses of our young participants are full of ellipses; this is because many of the contextual information was considered to be shared knowledge between interviewer and interviewee. We have provided necessary explanations about this feature during our analysis.

On the other hand, some information is missing not because it is taken for granted, but because the speakers themselves lack it. We refer to these ellipses as informational voids within narratives.

In all the excerpts that are presented here, we could identify an unconscious search for information within the context in order to fill informational voids during the discursive action. These were identified through a) Markers of epistemic modality signalising uncertainty or doubt, as expressed by the speaker, and b) contradictions in the construction of meaning when contrasting the content of the utterance with contextual information.

Results

During our analysis, we identified four main strategies of rearrangement and completion employed by our participants in order to build discourses about the past in situations in which they encountered a void of information or narratives that resulted contradictory with their understanding of the context:

Rearrangement through logical association: 'I guess'

Example 1:

Q: If I ask you about Carrero Blanco, does it ring a bell?
A: I think it does. The guy in the car, right? Yes, yes. Well, I don't know much, but… that was in Gernika, right?
Q: What happened?
A: He had a bomb in the car and he went all the way up.
Q: And who was Carrero Blanco?
A: I suppose it was someone bad.

This excerpt refers to the attack against Luis Carrero Blanco, president of Francisco Franco’s government during the last years of hist dictatorship, who was killed in 1973 by a bomb placed by ETA in his private car, in what is probably the most well-known attack of the organization and the one with most popular support in its history, as it is widely interpreted that Carrero Blanco would have embodied the continuation of Francoism (Agirre, 1974). This shared cultural knowledge resonates in our participant's memory, when he refers to Carrero Blanco as ''the guy' in 'the car', remarking the particularity of the case through the use of singular articles. He draws on a shared imaginary when he describes a certain characteristic of the attack: that Carrero Blanco's car was propelled in the air and landed inside of a convent (Lopez Adan, 2021a). The slightly humorous tone of the discourse shows that the speaker feels eased or confident towards the content of the discussion: he does not perceive it as a sensitive topic.

However, although the speaker is able to re-build some aspects of the story in great detail, he expresses a difficulty to identify other elements, such as who was the actor in question. This is signalised by a series of modalisers that show uncertainty, such as epistemic verbs ('I think', 'I suppose'), and interpellations asking for positive feedback ("right?").

It is in these voids where the speaker draws upon the context in order to fill in the information that he lacks: on one hand, he introduces the idea of Gernika. Gernika is a Basque city internationally known because of the bombings it suffered by the Nazi German Luftwaffe and the Italian Aviazone Legionaria under Franco's guidelines in 1937; since then, and especially after Pablo Picasso's painting remembering the massacre, Gernika stands for the suffering of the local population under Francoism. However, it does not have any direct relation to Carrero Blanco's death, which happened in Madrid in 1973, 36 years after the bombing. A plausible reason for the introduction of this element is its symbolic importance in the history of the conflict; we can assume that the speaker does not know where Carrero Blanco's attack happened, and therefore associates the story to a geographical setting he can definitely relate to the conflict.

Secondly, he relies on logical reconstruction in order to describe the actor in question. When asked about who Carrero Blanco was, he does not provide information about the actor's characteristics or his role in history, but builds ad hoc a coherent narrative, which is based on pre-formed assumptions about 'good' and 'bad people'. This argumentative scheme could be summarized as follows:

• Premise 1: Good people do not deserve to be attacked, bad people do.
• Premise 2: Carrero Blanco was attacked.
• Conclusion: Carrero Blanco was “someone bad”.

This final evaluative statement is not the outcome of a thoroughly argued process, but rather of a spontaneously built narrative in the course of communicative action, nourished by the co-text of the discourse itself. A third premise based on an attitude regarding the historical context and the identity of the in-group (i.e. 'Francoism was mean to us', 'Carrero Blanco was a franquist') might also have influenced this logical reconstruction.

Example 2:

Q: Do you know why they closed Egunkaria?
A: I think it happened in 2001. That's because I was born in 2002, so I don't know if it was in 2001 or 2003. Well, so, I guess, er, because of their articles and such, right? Well, because of what they wrote, wasn't it? I guess.

This is a similar example of logical rearrangement. The speaker signalises uncertainty about the closure of the Basque-speaking newspaper Egunkaria, which was closed in 2003 because of alleged ties to ETA. The newspaper was accused of financing the organisation, not for its editorial line, which was never investigated by the police (Torrealdai, 2021). However, this fact is unknown to the speaker, who on the contrary relies on pre-established schemata about "closing newspapers", according to which, logically, newspapers are usually closed because of their content. She activates a frame of censorship, and then applies the contextual information from that frame to the topic that is being discussed in the communicative event.

Example 3:

Q: Do you know any prisoner?
A: Joseba Sarrionandia.
Q: Which is his story?
A: Eh – he was imprisoned, I don't know if he belonged to ETA or not, I know that he is a writer and a poet, and that he escaped inside of a baffle. During one of Imanol's concerts, I don't know, inside of a baffle. And afterwards he wrote a song for the band Kortatu.

The interviewee narrates the story of Joseba Sarrionandia, a well-known figure in Basque literature, who fled from prison in 1985 (in fact, inside of a baffle during a concert for prisoners), an event broadly remembered by the local society, among others, because of the popular song 'Sarri, Sarri', launched in 1985 by the punk-rock band Kortatu (which at the same time is a version of the song 'Chatty chatty', originally sung by the Jamaican band Toots & The Maytals). The lyrics of the song tell the story of Sarrionandia's escape, and are therefore a popular device of transmission of memory, but they were not written by Sarrionandia himself. The authorship is not very publicly known, but is attributed to close friends of the writer.

The speaker in Example 3 belongs to the social group that shares this memory and has had access to the device, and therefore knows many aspects of Sarrionandia's story. However, he re-organises the information differently: in his discourse, he attributes the writing of the song to the protagonist of the story. The reason might be a missing fact (the authorship of the song), which he compensates by associating the elements he could identify in previously built mental models. The logical reconstruction would be as follows:

• Premise 1: Poets write lyrics.
• Premise 2: Sarrionandia is a poet.
• Premise 3: The song 'Sarri, Sarri' is related to the story of Sarrionandia.
• Conclusion: Sarrionandia wrote the lyrics of the song.

This example, as well as Example 1, illustrates the influence that popular culture can have in the construction of mental models and meaning-making in the context of transmission of memory: films and songs that become popular often model the social perception of past events.

Rearrangement through addition or elaboration

Example 4:

Q: Which actions or attacks by ETA do you know?
A: The one [that happened] at the airport in Madrid.
Q: What happened?
A: I think two Ecuadorians died, right? And it was in the – where you need to pass through to get onto the plane. I think that – I don't know exactly where it was, but I think two Ecuadorians died, and many [were] wounded.

This participant describes the attack on Madrid's Barajas airport in December 2006. His discourse focuses on the elements of the story he can identify with, especially the nationality of the deadly victims, as the speaker himself is originally from a Latin American country. However, similarly to the speaker in Example 1, he tries to re-build the physical setting of the attack, but lacks this information. The bomb was actually placed in the parking lot, which was entirely evacuated with the exception of two men sleeping in their cars (Lopez Adan, 2021b). The speaker does obviously not know this, and therefore he introduces an element belonging to his pre-established scheme of "airport": the passenger boarding bridge, adding a "new" element to the story and distorting part of the original narrative.

Example 5:

Q: When you have heard your parents talk about ETA, what has been your perception?
A: Well, I remember once that I went home and I asked them, 'What is ETA', and – because I was told in class that they killed blonde and brunette women. I don't know why. So I said, 'Why did they kill blonde and brunette women?', and [my parents] told me: 'No, no, anyone'. As if they went house to house, or something like that. And that they killed whomever they wanted. I mean, not for being a woman or so, but for any reason.

This speaker, an 18-year-old boy from the outskirts of Bilbao, refers to a void of knowledge regarding ETA as an organisation: he does not know which were their aims or methods, and ignores the political background of the conflict. Instead, he elaborates a narrative of indiscriminate and unplanned violence on the basis of concepts he could associate beforehand with the idea of ETA. During this elaboration, the speaker introduces details that might seem a fictionalised or at least very hyperbolic description of actual events, for example, going "house to house" killing indiscriminately: it is known that ETA's modi operandi were either previously planned attacks on concrete individuals or bombings in the public sphere (Lopez Adan, 2012).

These freely introduced details do, however, play a role in the construction of coherence within this particular discourse: they provide some kind of evidentiality or logic (however far from actual facts) for an action the speaker evaluates as irrational or incomprehensible. This description remarks the perception of ETA as a group of cold-blooded assassins with no political or rational motivation. It functions therefore as a device of negative representation of the out-group.

Rearrangement through substitution

Example 6:

Speaker 1: (...) in the beginning, it was political against Franco, then they began to become military, and... I don't know. I mean, in an objective way? Eh, I don't know, at the beginning [ETA was] a student movement, a political one, against Franco or against repression. Then they took up arms, and from that moment on, well, little by little they left politics and began directly the armed struggle.

Speaker 2: We learnt that ETA had two groups: the armed group and the political group. The decision to use violence came because the armed group won inside of ETA. Which methods did they use? Well, most probably, in the last years, since they didn't get anything using violence, the political ETA probably acquired more sense, and therefore [they understood] that it would be better for the group to lay down arms and move on towards a rather political or peaceful manner.

In both these excerpts, the speakers are describing the evolution of one of the main actors involved in the past conflict, the organisation ETA. They use the adjectives 'military' and 'political' as synonyms for 'armed' and 'unarmed' (or violent and nonviolent) actions. What these speakers ignore is that their narratives are echoes or 'disguised quotations' (Reyes, 1996) of a terminology used by ETA itself during its history: the organisation split in 1970 into 'Political-Military ETA' (ETA pm) and "military ETA" (ETA m); the former would soon move towards institutional politics, and the latter would continue with its armed strategy until it decided to lay down arms in 2017.

Both our interviewees draw on this intertextual information through re-semantization of terminology. A clear influence of this cultural knowledge resonates in their speech, even though they do not seem to be aware of the actual development of facts: they reproduce terminology from a concrete moment in history and employ it with a new meaning in a new narrative. This could also be considered an example of colonization, linguistically understood as 'an "invasion" of concepts from one discursive sphere into another one' (Wodak, 2011, p. 169). This 'new' distinction between a 'military' and a 'political' ETA seems to function as a device of categorisation of actors: they represent the "mean" and "good" sides of the same actor, in the context of an inner conflict of interests in which eventually the 'good side' (the political one) wins.

This representation may be a strategy to cope with contradictory discourses about an actor the speaker (at least partly) identifies with. He acknowledges that the in-group did wrong in the past, but introduces the idea of a 'historical development towards peaceful or political methods' in what Rheindorf and Wodak call 'internal othering' (2017, p. 24). This would therefore function as a legitimation strategy of the in-group.

Example 7:

Q: Which of ETA's actions do you remember?
A: I remember this one, eh, I don't know, the train in Madrid... a train that they blew up with bombs, I don't know, eh? That's what I remember.

This excerpt refers to the March 11, 2004 attacks on several train stations in Madrid. Although they were initially attributed to ETA, this was soon denied, as Al-Qaeda claimed authorship of the attack. The speaker, who was six years old at that time, does not have this information. He can only remember some basic features of the event: an attack with bombs on trains and a setting (Madrid). The fact that he relates the attack to ETA might have two possible explanations when looking at the context: either he is drawing on an external discourse that was once expressed publicly and has been maintained in social knowledge, or he introduces the actor 'ETA' in this story because of a pre-established mental scheme, according to which ETA did similar attacks on civil society. The new narrative 'ETA bombed the train stations in Madrid' would therefore be a coherent narrative with his belief systems, even if the content is false. Moreover, similarly to Example 6, it would also play a role as a device of actor categorisation, but this time it would characterise ETA as an out-group, according to a series of negative characteristics that the speaker already attributes to the group.

Rearrangement through new contextualisation: an example of 'invented' memories

Example 8:

Q: Do you have any direct memories [about the conflict], about an attack or such?
A: Yes, I do. The [attack against the] ETB headquarters. I think we were at school, I was in my ikastola [Basque-speaking school] when we heard what had happened, I think they had placed a bomb or something, but we could hear a very loud noise.
Q: Did you hear the explosion?
A: Yes, we heard it from our school.

The attack on the Basque public television outlet EiTB, which the speaker assures to remember personally, happened on December 31st, 2008. As it was New Year's Eve, all schools were necessarily closed on that day: therefore, the situation that the speaker describes is physically impossible. However, the setting (the school) and the elements (the loud noise, the sound of an explosion) have become part of his perception of the past.

The reason for this might be an a posteriori re-construction of events, using different kinds of contextual information available in his mental models: information about the sound of an explosion, about the approximate year of the event (during his childhood) and a logical setting (he must have been at school). The resulting narrative serves firstly as a coherent organisation of events (although not accurate with reality) and secondly as a self-presentation strategy possibly refigured in accordance to the present context of situation (the speaker appears in this interview as somebody who can offer relevant information because he can provide direct memories). In both cases, pre-established context models greatly influence the outcome of the process of meaning making.

Results show that all our interviewees resorted to contextual information during the re-construction of narratives about the past, whenever they encountered a gap of information or contradictory discourses with their own beliefs or pre-established schemes. This often led to the creation of "new" narratives that were not always accurate with the actual development of historical events, but provided a plausible organisation of the available information, i.e. narratives that resulted coherent with the speaker's current interest and identity.

These processes were identified as transformation strategies (as defined by Van Leuwen and Wodak, 1999), specifically strategies of rearrangement. These strategies included logical re-organisation, additions or elaborations, substitutions and creating new contexts ad hoc. Most of them had a broader macro-function related to the position of the speaker within the context of the communicative event, i.e. they served as devices of self-presentation, legitimation of the in-group or a negative characterisation of the out-group. Elements from popular culture, as well as media discourses showed to have great influence in the formation of mental models of the youngsters participating in our interviews.

Some of the limitations of this study are related to the fact that the analysed data belonged to longer interviews whose main content was not included in this concrete paper; a deeper research of the immediate co-text would surely provide valuable information. Moreover, this study might open a pathway in order to continue analysing the discourses of future generations and be able to generate broader comparative studies.

Conclusions

The act of remembering is strongly related to the position and the social role that the speaker holds in the moment he re-constructs the past discursively, as well as to his identity within the community. This happens in two directions: on the one hand, the speaker's situation, his experiences and previously formed knowledge influence the choice of what is remembered and considered important, and guide their process of meaning-making when re-calling or processing new information. On the other hand, the role of the speaker in the exact moment of interaction can also influence the manners in which the past is re-interpreted or re-built in discourse, often distorting some of the original features of the story.

This idea is crucial in order to understand the changing nature of memory. Remembering is not a mere transportation of fixed knowledge in time through generations, but a completely new construction that takes place every time we remember, in the exact moment we create discourses about the past. During this process, speakers rely on several sources of information in order to build narratives that seem coherent within themselves and with their own pre-established belief systems, which are, at the same time, culturally and socially influenced. This would explain why memory changes from generation to generation: because speakers do not simply reproduce external discourses, but re-contextualise and adapt them to their own experience in each discursive act.

Further and more specifically focused research combining cognitive, linguistic and cultural studies could be helpful in order to understand which elements are more likely to be modified during these processes, and which ones are more likely to be maintained rather faithfully.

Acknowledgements

I warmly thank both anonymous reviewers for their feedback on this paper. Their comments were relevant and concrete, and therefore very helpful. The research project this paper is based on has received funding by the Basque regional government (Eusko Jaurlaritza: Predoc grant) and is sponsored by the research group NOR at the Department of Audiovisual Communication and Publicity of the University of the Basque Country.

About the author

Samara Velte is a predoctoral researcher and PhD student at the University of the Basque Country (UPV/EHU) in Bilbao, Spain. Her research focuses on the discursive construction of memory in the context of post-conflict societies. Previously, she has worked as a journalist and documentary filmmaker. She can be contacted at samara.velte@ehu.eus

How to cite this paper