Gênese textual de um título: estudo sobre as redes associativas e ativação de memórias em alunos recém-alfabetizados

RESUMO Este estudo tem por objetivo compreender o processo criativo durante a gênese textual de títulos em histórias inventadas por alunos recém-alfabetizados (8 anos). A partir da perspectiva teórico-metodológica oferecida pela Genética Textual, dentro de uma abordagem linguístico-enunciativa, analisamos as relações associativas constituídas na recuperação de informações advindas das memórias semântica e episódica desses alunos. Tomamos como unidade de análise o diálogo face a face, estabelecido durante um processo de produção textual colaborativa, no contexto ecológico da sala de aula. Nossa análise, qualitativa e microgenética, permitiu mapear o modo como ideias e elementos linguísticos são articulados de forma imprevisível e inusitada, formando redes associativas, tanto por semelhanças semânticas, quanto por semelhanças sonoras. As redes associativas indicam que a criatividade se dá entre o reconhecimento do jogo de linguagem (homonímia, aliteração, sentidos) e, ao mesmo tempo, preservação das propriedades do gênero textual e da unidade de sentido do manuscrito escolar em construção.


Introduction
Writing fictional narratives is part of the didactic practices related to teaching and learning the Portuguese language. Among the different textual production proposals made by teachers in the early years of elementary school, they will certainly ask their students to write stories. In the field of Textual Genetics, researchers have highlighted the importance of the procedural aspect of writing, comparing different versions, copies or drafts of the same fictional narrative, highlighting the importance of erasure (FABRE, 1990). However, the interest in the creative dimension of invented stories is still little explored, especially if the study object is not only the school manuscript, but also the writing process in real-time-space of the classroom (CALIL, 1998), which generated the school manuscript.
We can consider the writing process as part of a "creative design" (SHARPLES, 1999), when the writer chooses multiple linguistic elements, which are articulated or excluded in order to produce unity of meaning. This process does not always involve conscious subjective values and judgments, but is strongly marked by the context, social uses and communicative functions of the written piece.
Studies with children on collaboratively produced writing processes of fictional narratives (DAIUTE, 1989;CALIL, 1998) have indicated how some of the choices of these novice writers are related to language games (alliterations, homonyms, senses), contributing to the creativity of written stories. These studies suggest that textual creativity is related to the recovery of knowledge from the memory system, rearranged in a singular and unpredictable way. The rearrangement of what is known in the production of something creative supposes, on the one hand, the occurrence of some form of association between different linguistic elements and, on the other, the writer's recognition of the effects of meaning produced by these associations, which are sometimes unusual, playful, and funny. Some initial studies on textual genesis and authorship in newly literate Brazilian students (CALIL, 1995;1998) have pointed out the importance of the role of language games in the construction of titles in fictional narratives as one of the central aspects for understanding the writing process of writers who are only 6 years old. Based on situations of collaborative writing in pairs, when these children are combining and linearizing All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 invented fictional narratives, one can identify the way in which creativity emerged from articulations between the linguistic and discursive properties of this textual genre, the cultural references and the coenunciatively established associations. The way in which these games took place were responsible for the creation of curious and original titles such as "Os três Todinhos e a Dona Sabor" [The three little Toddys and Madame Flavor] (CALIL, 1995;2000), "A família F atrapalhada" [The muddled F family] (CALIL, 1996;2003), These works showed how newly literate students take ownership of what is part of their cultural universes, putting together elements from both family and school literacy.
This appropriation was made from the dialogical and associative relationships established between what is recovered from the long-term memory and the semantic memory of each student, linked to the properties of the discursive genre and to the contextual elements reactivated in the here and now of spontaneous speech.
It is from this perspective, associated with genetic studies, that we will address and deepen the processes of creating titles among Portuguese students. Our main objective is not, at this moment, to establish comparisons between Brazilian studies already carried out, but to interpret the process of creating titles among Portuguese students, based on the theoretical-methodological framework proposed by the investigations with Brazilian students. Thus, this article proposes to advance the understanding of the textual genesis of titles for invented stories through the analysis of the function of linguistic-enunciation and its role in creative production. It aims at placing this issue in a more specific and interrogative manner: How is the different knowledge of writers, stored in their long-term memory, activated (recovered) and associated to produce language games and creative ideas for the composition of the title? To answer this question, we will analyze a real text production situation in the classroom, from the beginning of the dialogue between two students until the moment when they define the title for the invented story.

Writing Process and Memory
Written production involves a complex combination of cognitive components (motor coordination, memory, attention, reflection, executive system etc), linked to the linear and recursive character of the various linguistic elements that make up a textual piece, as well as the domains of referential and pragmatic knowledge. During its production, planning, formulation and revision processes are put into operation, when the cognitive components and knowledge domains interact, in a more or less automated way, according to the writer's cognitive abilities and the task demands.
In performing this task, an intricate memory system 1 plays an essential role. In general, within the scope of this study, we can say that the writer's knowledge is stored in his long-term memory. Long-term memory is, in turn, divided into "procedural memory" and "declarative memory." The latter is subdivided into "semantic memory" (processing facts, ideas, concepts and meanings; linguistic terms of common knowledge, book titles, stories, proper names, jokes etc) and "episodic memory" (autobiographical, linked to personal experiences, situations, contexts, events and so forth). These memories are recovered and articulated through working memory, responsible for storing and using information specifically linked to the performance of a task.
Those two forms of declarative memory (semantic and episodic) are constituted and fed, on the one hand, collectively, that is, by the cultural and socio-historical apparatus in which every speaker is inserted and, on the other, by each one's lived experience. Therefore, we hypothesize the fact that the linguistic elements verbalized by the writers, whatever they may be, when composing a title given to an invented story, would be related to the way in which these memories are brought forth and articulated on the textual surface, which may indicate the writer's greater or lesser creativity. 1 The memory system constituted in the human brain is extremely complex, with specialized literature describing different types of memory. As Izquierdo proposes, "It may be wise to reserve the use of the word 'Memory' to designate the general capacity of the brain and other systems to acquire, store and remember information; and use the word 'memories' to designate each one or each type of them." (2011, p.20

Creativity and Memory
Researchers Besançon and Lubart understand creativity as the "capacity to produce achievements that are, at the same time, new and adapted to the constraints of the situation in which they occur" (BESANÇON; LUBART, 2015, p.5). Among the different objects of study and the many questions that this notion covers, our interest focuses on understanding, from a linguistic-enunciative approach, the textual genesis of creative titles and the role of semantic and episodic memories in the formation of associative networks that led to its formulation and linearization on the sheet of paper.
Thus, we will deal with the process of creating an invented story, considering the cognitive and environmental factors, described in the multivariate approach proposed by Besançon and Lubart (2015) and Lubart et al. (2003). Cognitive factors refer to intellectual skills (knowledge and information processing) that facilitate and influence creative thinking. These skills involve the way "information will be sought in memory, whose knowledge will be integrated and applied." (LAUSTRIAT; BESANÇON, 2015, p.6) Environmental factors include the influences of each individual's family, school and cultural environment. These factors can "slow, stimulate or have a neutral effect on the development of creativity." (LAUSTRIAT; BESANÇON, 2015, p.11) They are directly related to the recovery of knowledge and information from long-term memory, in particular, semantic and episodic memory. To that end, creative thinking can be characterized by the formation of new combinations of ideas and linguistic elements or by the associations of ideas and elements distant from one another; the association between diverse and distant linguistic elements and ideas form the associative networks responsible for creation. In other words, these associative networks are "the basis of the ability to make associations of ideas that allow us to propose creative solutions to a problem." (BESANÇON; LUBART, 2015, p.8).
When a teacher asks his/her students to make up a story, he/she is posing a problem for them to solve. From a narrative point of view, they will need to create a title, define characters, their lines, establish time and setting (when and where the story takes place), the plot and the way the conflict will be resolved. The creative dimension involved in this task will depend on how the diverse knowledge they have, stored in their collective and personal memories, will be recovered and associated to solve the problem proposed All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 in the didactic context of a classroom. In view of the object of study chosen, that is, of the textual genesis registered through recorded film of the writing processes in pairs and the school manuscript generated (CALIL, 2016), we take the dialogue-text (DT) as our unit of analysis. Based on the established DT, we will identify the types of associations that have occurred, resulting in the title created, inscribed and linearized in the finished school manuscript.

Dialogue and Association
From a linguistic perspective, we characterize three types of associations, as defined by Suenaga (2005) in his re-reading of the notion of "associative relationship" presented by . Type 1 association is called "association by signifier and signified," characterized by the syntagmatic chain, attached to a morphological mechanism in which there is a community or group of morphemes. Thus, for example, the term "teaching" can raise associations between forms of conjugation (teach, we teach, teaching, will teach...), whose root is the linguistic element responsible for the formation of these associations. Type 2 association is called "association by meaning," in which there is a relation of semantic contiguity between the terms, by synonymy or part-whole relation. The term "teaching" can provoke associations between "educating," "learning," "school," "teacher," "difficulty" etc. Type 3 association, called "association by signifier," is established by phonic similarities, that is, associations of terms without morphosyntactic or semantic links. For example, the word "teaching" can recall words like "teacup," "tea cart," "thing," "ching" etc.
We note, however, that the description of these types of associations has a didactic character. Other types of associations, for example, associations between music (rhythm) and body movement or associations between images and feelings, such as those that occur when we appreciate a piece of work by Klimt or Pollock, indicate the limits of the linguistic associations described above.
Another important point to consider is the fact that in the functioning of speech, that is, in the linguistic-enunciative dimension of spontaneous speech in a face to face dialogue, "word for word" associations would not occur or would be merely the result of a single type of association. When verbalized, they are necessarily circumscribed by the All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 specific enunciative constraints of the here and now of enunciation (BENVENISTE, 1966), necessarily forming associative networks (CALIL, 2014a), which are anchored in the production of meaning. In other words, on the one hand, the very formulation of the syntagmatic chain supposes a type 1 association operation, insofar as the enunciation of a linguistic element in the syntagmatic chain restricts possible associations with other linguistic elements. On the other hand, the chain of the syntagmatic chain (type 1) will be structured and preserved by its articulation with associations of type 2 and 3. In the English language, for example, it is not expected that someone will spell out the word "teachity" or even the ungrammatical phrase "teaching was to hard, without whose every elephant can jump him." In this regard, the syntagmatic chain itself imposes the limits and possibilities of associations, insofar as each subsequent element is syntactically associated with the language restrictions imposed by the previous element and thus what will happen to it . This condition is necessarily associated with the socio-historical character of speech (VOLOŠINOV, 1973), 2 that is, the possibilities of association are delimited by what is already attested in the discourse, preventing, on the one hand, an association with what was never part of the language of the speaking subject and, on the other hand, preventing the production of meaning from completely breaking away from meanings already constituted in the memory of the speaking subject. Only under this double condition, syntagmatic and semantic, will the speaker establish type 1, 2 and 3 associations. Therefore, associations depend on the condition of "being a speaker," that is, of being subjected both to the stabilized and predictable functioning of the mother tongue, as well as to the socio-historical ballast of speech.
We understand, therefore, the face-to-face dialogue between teacher and students as a trigger for establishing these associative, socio-historical and contextually determined networks, although singular and unpredictable. These associative networks are formed from the recovery of ideas and linguistic elements arising from the semantic and episodic memory of the writers. The associations to be addressed here point to the multiple aspects that motivate different associations, depending on how different linguistic elements and ideas are recovered from the speakers' memories.
It is for this reason that mapping the formation of these associative networks can indicate the creative potential of the title, revealing how the newly literate writers articulate knowledge, information, and ideas that are "new and adapted to the restrictions of the situation in which they occur." (BESANÇON; LUBART, 2015, p.6).

Procedures for the Collection of Material and Choice of Material Analyzed
In order to foster investigations based on this form of film recording, the Laboratory of School Manuscript has offered a database since 1989, including data collected in classrooms in Brazil, France and Portugal. The data collected in these different countries performed the same didactic task: asking students, organized in pairs, to combine and write a single invented story together. This task was initially proposed by Calil (1998), whose study analyzed the textual genesis and the authorship of newly literate students.
Extracted from the material collected in a Portuguese school, in 2015, the material to be analyzed is part of the "EE" 3 Dossier, referring to a 2nd grade primary classroom, in the city of Aveiro (Portugal). In this classroom, for three months (January to March), six textual production tasks were recorded. In all tasks, applied by the teacher herself, the students were organized in pairs. Each pair should orally combine an invented story and then receive a sheet of paper to write this story together. The collection and postcollection method followed the current guidelines proposed in the description of the Ramos System (CALIL, 2020). According to the Ramos System, all the multimodal material collected was edited on film (synchronized film), through which one can have synchronous and simultaneous access to the dialogue between the pair of students and with the teacher, their facial expressions and gestures, the stroke of the pen on the sheet of paper and the classroom environment. Each shot video followed the same methodological protocol (CALIL, 2016), subdivided into six moments: 1. Classroom organization; 2. Presentation of the proposal by the teacher; 3. Combined student pairing for the story to be invented; 4. Inscription/Linearization of the story invented on the sheet 3 The authorization for this data collection was approved by CEP/CONEP (CAAE: 61488616.4.0000.5013).
All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 of paper; 5. Reading/Reviewing the written history; 6. Free drawing while the other pairs finish the task.
In three of the six textual production proposals, the teacher did not suggest a theme or characters (proposals with a free theme). In the other three proposals, the teacher suggested a theme or characters (proposals with a suggested theme). These proposals were made to the whole class on a weekly basis. Alternatively, in the 1st, 3rd and 5th proposals, the textual production had a free theme, and in the 2nd, 4th and 6th proposals, a suggested theme.
In this study, we selected the video recording of the textual production process of a dyad formed by students L (7:04) and M (7:06). As our interest is in understanding the process of creating an invented title, we selected the title of one of the three production proposals with a free theme. In these proposals there was less direct and deliberate interference from the teacher, since no titles, themes or characters were suggested. The pair produced three school manuscripts with the respective titles: "A princesa Sorridente" (The Smiling Princess) (story 1), "The Totónoffzil restaurant" (story 2) and "Lights, camera, animals!" (story 3). We will analyze the textual genesis of the title "The Totónoffzil restaurant," chosen from a public poll, 4 with the participation of 456 judges, as the most creative title.
After choosing the most creative invented story, we analyze its textual genesis.
The object of investigation of this analysis was the multimodal video recording (synchronized film), captured by the Ramos System. The unit of analysis will be the dialogue text (CALIL, 2016) established between the pair of students, through which we intend to identify the associative networks formed in the students' spontaneous speeches, that is, how the associative relationships between ideas and verbalized linguistic elements were established and semantic and episodic memories were recovered. This analytical path is characterized by an ethnographic, microgenetic and qualitative approach, with the 4 We suggest that readers participate in this poll, before continuing to read the article. To do this, just click on Poll 1 Creativity. We developed this survey based on a simple psychometric test using Likert Scale (LIKERT, 1932), establishing a scale from 1 (not creative) to 5 (extremely creative) items. No criteria were indicated for judging the creativity of these stories. We intended to have a choice conditioned on common sense, between independent and anonymous judges. Our survey received 456 responses from Brazilian participants, over 10 years old, literate and with internet access. Reading the three invented stories can help to better understand the content and the difference between them, already indicated in the respective titles.
All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 main objective of explaining how the students arrived at the final title, documented on the sheet of paper.

The Totónoffzil Restaurant and its Associative Networks
Mapping the formation of associative networks that generated the invented title "The Totónoffzil restaurant" requires following the dialog-text of the pair of students from the beginning, when the classroom was still being organized, identifying the main elements that were spontaneously delivered, which will generate different associations related to the title that will be invented. 1st associative network: From "Interwriting" to "cameras" At the very beginning of the video recording, we captured the following dialogue, still relatively distant from when the pair of students will begin to combine the invented story, but which will have an important effect in generating ideas for this fictional narrative. Whispering) Oh, no! He will install it alone! (Pointing to C) C will install it on its own… 5. D: (Looking at pair C and S) No, he isn't. 5 We highlight in red the first occurrence of words, expressions or key statements in the composition of the associative network.
The teacher is helping! 6. M: But C was putting the belt on by himself.
He could ruin the recorder! It's a danger, really bad, isn't it? Never play with recorders! Because they are from the Interwriting. 7. D: His is pink. 8. M: For a guy that likes... 9. D: (Interrupting) R's is purple.  13 and 15) interrupts this associative movement, establishing a relationship between the color "blue" and his football team (Porto! Porto!), whose uniform is "dark blue." M again mentions the name of the Interwriting project (turn 16), a fact that led D to make a new and unexpected association, which will be important in the process to create the title of this invented story: D associates the fact that the researchers of the Interwriting project are Brazilian to the fact he is also Brazilian and, therefore, will be able to "speak Brazilian." This unpredictable association changes the course of their dialogue, opening new associative relationships, between turns 17 to 24. M complains that he will not be able to understand what D says because "Brazilians speak very fast" and asks him to "speak... Portuguese!" (turn 22).
The confrontation between "being Brazilian," "speaking Brazilian" and "speaking Portuguese" leads D to associate this information with two other autobiographical pieces of information: his father and mother can only "speak Brazilian" and his mother's luck finds "work only of Brazilian women... Because then ... some will not understand (sic) what she is saying" (turns 25 to 29).
While some of this information is related to semantic memory (for example, the color of the football team, the common language that Brazilians and Portuguese speak, the differences in the way they speak the same language), the information about the work of D's mother is clearly recovered from his episodic, autobiographical memory. It is likely that his mother told him that.
M (turn 30), when asking what D's mother's work is, activates other associative relationships. D, answering where his mother works, retrieves the names of three restaurants ("Bella Pasta/Bella Pizza," "Pasta Grill," "Kebab") located in the food area of the Forum, an important shopping center located in the center of Aveiro, a touristic part of the city.
The two students share this personal information. Both have a lived experience (episodic memory) as they know and have visited the Forum. This is even more evident when they spatially place each of these restaurants in relation to the Bertrand bookstore (turns 31 to 40).
The reference to the Bertrand bookstore, frequented by the two students, evokes several and different associations: Garfield, editorial, castle, television...This last term, probably related to the large television set that decorated the bookstore environment, will reactivate associations related to the immediate context, using the term "fun" and the word "cameras," part of the electronic equipment used during the Interwriting project (turns 41 to 61) and which student D points to (turn 55).
In this first dialogue-text, which occurred before the teacher started the presentation of the task, while the researchers were now finishing installing the equipment in each student, we have the formation of different associative relationships. These relations, when coenunciatively interconnected, form the 1st associative network. In this network, we can observe that all established relationships emerged from the linguistic elements offered, arising both from the immediate context and from the semantic and episodic memory of each of the students.
Verbalized associations have a certain degree of expansiveness and interdependence, that is, an enunciation causes an explicit association in a new enunciation, which causes another association, which can cause another one, and another one, and another one, like a tangle of threads with knots, ends and bifurcations, unfolding or interrupting. Let us keep in mind, however, that the pair has not yet combined the story, as this dialogue took place when the classroom was being organized, before the teacher began to present the textual production proposal. Students are not yet talking about the story they are going to create. They are talking about what is happening in the classroom and their personal experiences like the work of D's mother, the food court at the Shopping Center Forum and the Bertrand bookstore, which both students know and have visited.
Although these spoken elements are prior to the teacher's presentation of the task, some will intercede and will ensue in the ideas that will lead to the created title, and also to a part of the content of the fictional narrative linearized in the school manuscript. The record of this dialogue reveals the unique and unpredictable dynamics of the interaction between students in a classroom. The same phenomenon happened in other writing processes of Brazilian students that we analyzed. In one of them, for example, Brazilian students Isabel and Nara invented the title "A rainha comilona" [The gluttonous queen] (CALIL, 2012a). This title was directly related to the Magali comic book (SOUSA, 1990) they had found under the desk and started reading before they began to combine the story and think of a title for it. As soon as the researchers leave the classroom, the teacher presents the task to her students: to write "a free text... you can invent (invented)" (turn 91). D reacts immediately to the proposal, choosing at this point of the dialogue text a textual object (CALIL, 2016), that is, the student recognizes the theme of the story to be invented as a problem to be solved. He proposes some common, predictable themes in students of this age group and inserted in this literate universe (CALIL, 2008b;2010): a text about joy or a text about the friends' club.
Other pairs, classmates from that same classroom, throughout the three-text production proposals with a free theme, gave titles with these same themes: "The friends," All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 "The two friends and the dog," "Best friends," "The club of funny friends." These two themes (joy and friends) proposed by D, respectively in turns 92 and 94, are predictable, not very creative and repetitive. The characteristics expressed in these titles depart from our criterion of creativity, defined as the association between diverse and distant ideas and linguistic elements, characterizing the uniqueness of the invented title. Both the first theme (joy) and the second (friends) one return to lexicons often found in texts read by students in their didactic and educational materials.
Another important aspect to highlight in the story creating process invented by students of this age group is the correlation and interdependence between theme, character and title. In all the writing processes analyzed in our previous works, there is an assessment of themes, plots, character(s), before deciding what the title of the story will be. However, the title will be related to the established narrative elements. This indicates that it is not really suitable to analyze the creation of the title without taking into account the creation of the narrated content itself, as we will see below.
In turn 99 the teacher resumes the presentation of the textual production proposal.
This speech will cause, however, an unexpected association. In the complementary guidance for carrying out the task of writing a "free" and "invented" text, she adds that they can write about what they "find funniest... like jokes," while also suggesting that they can choose characters and places.
The first unpredictable associative relationship, with greater creative potential, associating diverse and unexpected things, is done by M in turn 100, when she says she had an idea. By proposing "a restaurant where we eat materials... instead of food" as a The student also shows, right after proposing this theme, a metapragmatic concern when she says: "Do you think it will make the teacher laugh?" (turn 102). This concern characterizes the intention to respond to the task, as requested by the teacher. For D, even though this topic might be "funny," it would not have the effect of a joke. This interpretation is supported by the statement in turn 104: "How about telling a story of jokes ?!".
The terms "funny" and "joke," provided by the teacher, produced two different associations: for M, it would be a funny story to write about "a restaurant where we eat materials." For D, the teacher's suggestion made him propose to tell a "story of jokes." During the teacher's presentation of the task, her speech activates some associative relationships for the creation of the invented story. The main elements in the formation of this 2nd network of associations are the terms "funny" and "place," which generate associations with elements given in TD1 (equipment/materials -cameras, recorders, microphones-, Pasta Grill, Kebab, Bella Pizza): writing a (funny) story that takes place in a restaurant (site/place) where materials (equipment) are eaten. Until the end of this second DT, there is competition between writing a "funny story" and writing a "story of jokes," but the associations led by these themes have not yet been recombined.
This will happen shortly thereafter, in the next DT, when the recombination of these associations will generate the name for the restaurant, consolidating and finalizing the path of the textual genesis of the title that we find in the school manuscript. 3rd associative network: From "restaurant name" to "Totónoffzil"  A first attempt at a funny name is "Totóte." Why did M create that name? Why did she invent a word that is probably not part of her mental lexicon? From a semantic-lexical point of view, the term "totó" in the restaurant's name responds to the student's deliberate intention to create a funny name. On the one hand, the noun "totó" is also a proper name. "Totó" is easily recognized as a puppy's name or a person's nickname: Mr. Geek (Seu Totó). On the other hand, "totó" works as a qualifier, a qualification of the restaurant. In the informal Portuguese language of Portugal, the adjective "totó" means, according to the Priberam dictionary (2019), 8 "That or who is considered to lack skills, intelligence or resourcefulness (e.g., she is such a totó (geek) that she didn't even notice what happened; bunch of totós! (geeks!)." Or, according to the InFormal dictionary, 9 "In Portugal the term designates a person who is easy to deceive.
(e.g., Adriano is a totó (geek), even his daughter deceives him"). It is important to highlight that this second meaning ("totó" as an adjective) gains prominence in recent children's publications such as "Diary of a totó: stories of a little, or not, fantastic life" 10 (RUSSELL, 2010), and in publications for the adult audience, such as the (For geeks) "Para totós" collection by Porto publisher, 11 both on sale in Portuguese Bertrand bookshops. The sharp accent in "Totónoffzil," according to the name linearized in the school manuscript, also suggests the visual memory that students have of the word "totó." The reason for adding "te" to the term "totó" is less evident. Perhaps it was due to the homophony with the end of the term "restauran-te." There is no evidence this occurred consciously, but its enunciation suggests a type 3 association, a phonic one. This would help explain why the suggested name was not just "Totó restaurant." The first name suggestion for the restaurant (Totóte) is left out, but the funny idea for the theme remains: writing a story about a totó ( The collection Dork Diaries by the North American writer Rachel Renée Russel (https://en.wikipedia.org/wiki/Rachel_Renée_Russell) was released in 2009. Since its first edition in Portugal, in 2010, 13 volumes have already been released with the title "Diário de uma totó" (https://www.wook.pt/autor/rachel-renee-russell/1139777 ). In Brazil, this collection was translated as "Diário de Uma Garota Nada Popular" (https://pt.wikipedia.org/wiki/Diário_de_Uma_Garota_Nada_Popular). 11 Among the various volumes of this collection sold by Livraria Bertrand, there are: "Matemática para totós" (Math for geeks) , "Inglês para totós" (English for geeks), "Guitarra para totós" (Guitar for geeks), among other titles (https://www.bertrand.pt/pesquisa/+/coleccao/3886/Para+Totós). It is only in turn 117 that the name of food comes up: "stroganoff" ("strogonoff" in Portuguese). This turn, like the teacher's turn 99, will play a crucial role in redirecting the combination of the invented story. In addition to the name of the food, M explains the memory of the "stroganoff joke," managing to unexpectedly associate the themes: "restaurant where equipment instead of food is eaten" (TD2, M, turn 100) and "writing a Throughout the three TDs, new ideas and linguistic elements adapted to the constraints of the textual production situation with a free theme indicate an important language game, permeated by alliterations and homonyms: • "place," a term used by the teacher to guide textual production. • "restaurant," site (place/setting) where the invented narrative takes place. • "restaurant," the place where D's mother works. • "restaurant," with a menu and crazy (funny) foods. • "equipment," such as recorders, cameras, microphones in the classroom. • "electrical materials": English terms 'on' and 'off'. • "totó," masculine noun, proper name. • "totó," an adjective to qualify silly, goofy things or people. • "fun," "funny," how to make food with electrical materials. • "funny as a joke": joke about food. • "story about jokes" and "story about food." • "stroganoff joke (strog-on-off)"; what is the food that turns on and off?
The two students share this information, including the 'stroganoff joke', which is remembered by M as "there will be one of our jokes, the one about stroganoff." (TD3, M, turn 117). This joke, easily accessible on the Portuguese web page Omelhorsitede2011, is the first that appears among the many other jokes with food names that students may also be aware of. The joke retrieved from the semantic declarative memory shared by the students will compose the plot of the fictional narrative, whose unity will be guaranteed by the title, not yet formulated, of the invented story. The recombination of the idea of writing a "story invented about a totó (silly) restaurant, where materials (equipment) were eaten instead of food" with the idea of writing a "joke story about food" is done through the multiple associated linguistic elements, described above, to do something different, new, creative.
In the continuation of this DT, on the one hand, M continues to seek a name for the restaurant. However, D does not want to write a story about a restaurant (turn 120), insisting on another theme ("adventure food," turn 122) and plot ("the lady was singing... when you eat food you are happy," turn 124), both related to his working memory, what was said before, in TD2 ("text about joy," turn 92) and in TD3 ("story about food," turn 116).
In order to maintain the unity of meaning of the narrative that is being constructed, M reiterates another metatextual related concern, returning to the name to be given to the restaurant, when she verbalized that "it has to be a very... mega weird name." (turn 129).
After that turn, the language game between alliterations and homonyms is amplified, The richness of this intense process of creating the restaurant's name can be seen in the various rearrangements made with the terms "totó," "stroganoff," "on-off," "zil" and "pizza". These last two linguistic elements occur in two associations proposed by D (turn 132 "Totózil" and turn 146 "Totópizzazil"). This last term "pizza" retrieves what was said by the student in TD1, turn 31, when talking about the work the Brazilian mother got in the "Pasta Grill" restaurant, first enunciated by him as "Bella Pasta." At the Forum Shopping Center, as highlighted in a footnote, the correct name of the restaurant is "Bella Pizza." The "zil" ending is also enunciated for the first time D in "Totózil" (TD3, D, turn 132). Its occurrence strengthens our hypothesis about the importance of language games, particularly the phenomenon of homonymy, in the creative process of newly literate  (CALIL, 1995;2000;2003;2004;FELIPETO, 2013). This game is enriched by the composition of the syllable "zil" with the term "Totónoff," possibly motivated by the phonic association between the two terms mentioned above, which carry the same sound element: 1. The sound of the syllable "sil" in "Brazil" /bɾɐ.ˈziɫ/ ("Brazilian researchers," "Brazilian father and mother" "speak Brazilian").
The element "zil/il," in the name suggested by D ("Totózil," turn 131), will be incorporated by M, when proposing the name "Totónoffzil," enunciated for the first time in turn 139 and maintained in the last TD3 turn, when formulating the beginning of the story: "Once upon a time there was a restaurant, a space where they opened a restaurant called Totónoffzil." However, the entry of "zil" differs from intentional attempts, aiming at creating a "mega-weird" name for the restaurant. In other words, the combination of the terms "Totó," "totó," "stroganoff," "strog-on-off" and "on-off" appears to have been deliberate. The appearance of "zil" does not seem to be at the same level of deliberate intent. There is no evidence the students were consciously manipulating the "zil" element from their phonic relationship with "Brazil" and "Pasta Grill." This indicates the limit of Suenaga's proposal for the analysis of associative relations in the flow of the dialogue-text, since they do not operate "word for word" or "phonic similarity for phonic similarity." In the functioning of associative networks, the different types of associative relationships were operated in a simultaneous and interwoven way, in which it is difficult to isolate one type of association or another type of association without taking into account their effects on the syntagmatic chain, in the production of meaning and in textual unity. It is also not possible to analyze these relationships as if there was no speaking subject, submitted to their memories and knowledge. In addition to this aspect, it is necessary to point out that not all the effective associations operate at the same level of intentionality, as the entries of the elements "te" in "Totóte" and "zil" in "Totónoffzil" seem to indicate.
All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0

From "Interwriting" to "Totónoffzil": Final Relationships
In this article, we are interested in the textual genesis of an invented story title, highlighting the associative relationships and the role of semantic and episodic memories to understand its creative dimension. The established connections formed associative networks with functions that are closer to the multivariate approach, as proposed by Sternberg & Lubart (1995). This was observed in the combination of cognitive, environmental, conative (personality and motivation) and affective (emotional) factors in the creative potential of that title and invented story that accompanies it.
On the one hand, we observe the emergence of ideas and the retrieval of longterm memory information, both semantic memory (elements belonging to the collective culture) and episodic memory (elements belonging to the individual culture) of the speakers/writers. On the other hand, we saw that what is retrieved is associated with what is being said in the here and now of carrying out the task, both by students and by the teacher.
These two aspects alone do not guarantee the creative dimension of the invented story. The main component seems to be the way in which each student operates cognitively about what is spoken, interpreting the various elements spoken from language games (homonymy, alliteration, polysemy...). The playful dimension for these students, associated with the textual production task proposed by the teacher, as indicated by Daiute (1989), gains emphasis from his enunciation "what you find to be funniest ... like jokes" (TD2, turn 99). The way in which this enunciation subjectively affects each student is unique and unpredictable. The connections established between one idea and another, one memory and another, one linguistic element and another, created unusual and creative associations.
We show how the associative networks were being built by this pair of students, from the beginning of the face-to-face dialogue up to the moment when they reached the After deciding what the invented story would be like and what the title would be, students received the sheet of paper and student D wrote it, at minute 26: "The Totónoffzil restaurant." As indicated by Calil (2010), the nucleus of the syntactic structure of this title is massively recurrent in titles given to stories invented by newly literate students: Nominal Syntagma (NS = Det + Name). However, the 3rd element ("Totónoffzil") expands the NS, playing a double role. It is, at the same time, the name and qualifier of the restaurant, contributing to the creative (homonymous) dimension of the invented history. In its final linearized form, established after 10 different oral combinations, written with a capital letter, its proper name function is even more evident. In addition, the term "off" preserves the duplicated letter "F." This indicates that these Portuguese students, who are only 8 years old, have the knowledge (visual memory) of foreign words, both the name of the dish of Russian origin, "Stroganoff," and the English expression "On-Off." 12 This orthographic aspect, in addition to pointing to the importance of access to printed materials in the literate contexts (family and school) in which they live (DYSON, 2009;KRESS, 1997), strengthens our observation about the limits of the three types of associative relationships, given that the linguistic associations described by Suenaga (2005) do not consider the visual dimension of the written word.
It is also necessary to remember that in the field of "Titulology" (DUCHET, 1973), title and co-text (narrated content) maintain meta-textual and intratextual All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 relationships. According to Hoek (1981), the title functions as a "meta-text," anticipating the co-text, but keeping, at the same time, autonomy in relation to what it names.
Somehow, the meaning of the title will need to be associated with the meaning of the narrated story. Therefore, if the students had written "totonofzil," in the title, without a capital letter, without an acute accent and with only one 'F' letter, the produced meaning effects would not be the same.
Finally, the microgenetic analysis presented, despite being applied to a single case, reiterates a proposition already pointed out in other studies (CALIL, 2008a;2009;2012b;2014a;2016), when we describe the textual genesis of story titles invented by Brazilian students: the unpredictable and unusual associative networks that children build when inventing stories are formed by a recombination of diverse and distant linguistic elements and retrieved from their long-term memories (semantic and episodic), producing (recognizing) language games, as well as the preservation of the properties of textual genre and the unity of meaning of the school manuscript in progress.
It is precisely in this aspect that we find the common denominator between creativity in Brazilian students, who invented titles such as "The muddled F family" and "The three little Toddys and Madame Flavor" and Portuguese students, who invented titles such as "The Totónoffzil restaurant" and "Lights, cameras, animals!". It is this proposition that allows us to understand why an invented story is more creative (and therefore more authorial) than another.
It is likely that the recombination of diverse and distant linguistic elements and ideas in the school manuscript "The Totónoffzil restaurant" led the judges to choose this invented story as the most creative of the three stories invented by this pair, even without having access to the dialogue-text established during its textual genesis.
Similarly, the classroom teacher also only has access to the school manuscript (final product). It is not possible to follow what each of her students say, discuss, combine, or what are the metalinguistic and metatextual activities verbalized during the current school manuscript.
Thus, when discussing the relationships between creativity, formation of associative networks and memories established throughout the textual production process, we hope we are contributing to the understanding of the importance of didactic tasks that ask students to create stories invented by pairs of students. The students' All content of Bakhtiniana. Revista de Estudos do Discurso is licensed under a Creative Commons attribution-type CC-BY 4.0 creative potential is favored by face-to-face dialogues, through spontaneous and coannouncing speech, which could generate unpredictable and unusual associations.
We also suggest that students be given some space to combine the invented story before writing and linearizing it on a single sheet of paper, and that there are also textual production tasks with a free theme, with no suggestion of titles. These tasks can be of great value for developing metalinguistic and metatextual activities, as students, when building the title (metatext), will need to establish greater metatextual and intratextual relationships with the narrated content (co-text). That is, when inventing a story with a free theme, it will be necessary to link the title given to the invented characters, plot, conflict, scenario etc.