INTRODUCTION

This article presents the findings of research aimed at understanding the impact of competition between different epistemological foundations on the development of theoretical-methodological and pedagogical-curricular proposals for literacy in the country since the 1980s.

Given the significant influence of the constructivist perspective on literacy from this historical period onward, particularly in shaping the understanding of early writing acquisition, the research aimed to examine how this perspective was adopted for didactic and pedagogical purposes in official curricular reference documents.

The research examined the processes of maintenance, rejection, exclusion, (re)appropriation, and/or denial of themes, objects, instruments, and teaching and learning strategies. These processes are analyzed in the context of the relationships established between discourses on teaching and learning to read and write in the early stages, with constructivism serving as a central reference point.

At this stage, the research specifically examines how semantic processes related to language conceptions are addressed in proposals for teaching and learning early-stage writing, as produced in different historical contexts and conditions of discursive production. The analysis focused on the first and second versions of the Curricular Proposal for Teaching Portuguese Language — 1^st grade (^{São Paulo, 1988}; ¹⁹⁹¹).

Between editions, passages related to literacy propositions are maintained, excluded, or altered according to the social, political, and ideological conditions specific to each historical context. These changes reflect the interplay with competing epistemological foundations in the field of teaching and learning Portuguese, particularly in the area of writing acquisition.

The edition of the Curricular Proposal published in 1988 by the Coordination of Studies and Pedagogical Standards (Coordenadoria de Estudos e Normas Pedagógicas — CENP) of the São Paulo State Secretariat of Education was developed in a historical context marked by the struggle to expand access to quality public basic education and the country's redemocratization. The document emphasizes the collective and dialogical nature of its creation, resulting from a participatory process involving teachers from the education network and researchers. At that time, there was an urgent need to address the challenges posed by the influx of students from social groups that had previously lacked access to basic education. These students brought linguistic varieties and language practices into the classroom that differed significantly from those historically legitimized in the school environment.

To highlight what gained prominence in the production of curricular proposals during the 1980s, ^{Moreira (2000)} emphasizes:

The curriculum renewal movement of the 1980s primarily took place in the Southeast and South regions, following the election of governments opposed to the military regime. The primary goal was to improve the quality of education in public schools and reduce the high rates of repetition and dropout, which disproportionately affected children from lower-income backgrounds. Additionally, there was a focus on encouraging the participation of the school community in decision-making processes, aiming to overcome the authoritarianism of previous reforms, which were often imposed from the "top down". (^{Moreira, 2000}, p. 111, emphasis added)

In this context, research in sociolinguistics and socio-interactionist-based propositions provided crucial support for the development of pedagogical-curricular proposals (cf.: ^{Geraldi, Silva, and Fiad, 1996}). These efforts were further complemented by insights and propositions emerging from research on the psychogenesis of written language, whose findings began to gain broader dissemination across the country.

The reissue of the Curricular Proposal in 1991 occurred within the context of consolidating democratic foundations in Brazilian society, following the promulgation of the 1988 Constitution. During this historical period, the country began to implement neoliberal economic projects, which subsequently impacted the public education policies established at that time.

The literacy propositions in the 1988 document, while aligned with the constructivist conceptions that informed its development, also closely adhered to the social and interactionist principles that guide the general proposals for teaching Portuguese in the Curricular Proposal under analysis.

In transitioning between contexts and making changes from one version of the document to another, shifts in the epistemological bases supporting the theoretical-methodological propositions in the pedagogical-curricular proposal led to different conceptions of the relationships between language and the world. These shifts resulted in varying principles for understanding the semantic relationships underlying teaching and learning processes. The data analysis aimed to understand the effects of these recontextualization processes (^{Bernstein, 2000}) on conceptions of literacy practices, as they operate between academic-scientific and pedagogical-curricular discourses within the studied context.

THEORETICAL AND METHODOLOGICAL FOUNDATION

The conceptual bases of the study for data processing were employed to analyze the competition among discourses in the academic-scientific and curricular-pedagogical fields, considering their production by official bodies responsible for establishing public education policies. The study observes the development of theoretical-methodological conceptions and pedagogical-curricular propositions in relation to various epistemological perspectives and the discursive and ideological conditions characteristic of a particular historical context. These interrelationships are summarized in the Chart 1.

Considering the constitutive heterogeneity of discourses on literacy and the complexity of relationships between different fields, such as academic-scientific and curricular-pedagogical, as well as within each of these fields, the initial approach to the analysis involved referencing the production of data and their categorization based on two epistemological perspectives with their ideological affiliations historically evident in discourses on teaching early-stage writing: the classical and modern epistemes, which will be presented in more detail below.

The academic-scientific field, specifically within the area of literacy, refers to the domain where knowledge is produced regarding the teaching and learning of early-stage writing. This field includes various institutional contexts such as universities, research institutes, schools, and state bodies like Education Departments, etc. The pedagogical-curricular field, on the other hand, pertains to the spaces where regulatory principles are established to organize literacy concepts and practices within schools. This field includes various official bodies such as state and municipal Education Departments; the Ministry of Education; legislative chambers; and agents working in both school and higher education institutions, as well as private research institutions.

THE CLASSICAL EPISTEME

The so-called classical episteme guides propositions for language teaching and learning based on the conception that reality is an object of knowledge for the subject, who perceives it through the analytical nature of language. This perspective involves a temporal separation of representation from perception, presenting reality in an image, picture, or figure where data are gathered, planned, and organized systematically.

According to the principles of this episteme, speech is characterized by temporality, involving the arrangement of words in a linear, sequential order within a syntactic structure: temporal linearity serves as an analytical principle, intersecting with representation in its spatial context. This process involves selecting elements from the representation framework to be named and associating them in propositions organized around the verb to be. This results in attributing characteristics from one object to another (c.f.: ^{Foucault, 2007}).

The classical episteme is based on the notion that the relationship between words and things, or between entities in the world and language, is established through a referential relationship. In this view, the referred objects are perceived as completed, static, and empirically discrete, with language representing them as elements to be designated, akin to a labeling process.

This perspective belongs to a tradition rooted in principles of rationality, where the relationship between language and thought is seen as fundamental for making judgments, producing knowledge, and communicating it. Language is thought to develop in complexity to match the complexity of thought, as anticipated by the Port Royal grammar (^{Arnauld and Lancelot, [1660] 1992}), which will be discussed in more detail below. The logical foundations of a flawless language are believed to ensure accurate, fair thinking and effective communication. Consequently, this tradition involves not only understanding how languages are constituted and function but also improving their grammars to enhance their role as structured tools for concept formation, judgment, and reasoning.

The classical episteme informs theoretical propositions where syntactic construction holds a privileged position in understanding the relationships between language and thought, or between language and reality. In these propositions, the characteristic linearity of syntax is conceptualized in various ways: in some theoretical frameworks, it is seen as an expression of the representational nature of verbal language, marked by the sequential arrangement of elements in time. This linearity facilitates the categorization of judgments made by reason or facts of reality considered objects of knowledge. Thus, in this conception of syntax, linearity serves as a tool for organizing the representational process and defines the analytical nature of language.¹

In some theoretical frameworks, syntax is viewed as a surface phenomenon that is not directly linked to its represented counterpart. In this perspective, surface linearity does not imply that the relationships underlying it are inherently evident or capable of being directly "decomposed" or "segmented." These underlying relationships, which are temporally materialized in the linear ordering we perceive, reflect ways of organizing thought or meaning that do not necessarily correspond to the syntactic structure of a particular language. At its semantic core, conceived according to logical principles, knowledge is considered universal in its potential, while its linguistic representation is shaped by the idiosyncrasies of individual languages. According to these conceptions — such as those aligned with Chomskyan theory —, language does not linearly reflect thought or meaning, as each language has its own unique methods of expressing operations and categorizations.

The Chomskyan proposal is arguably the most prominent theoretical proposition associated with the classical episteme among recent developments. It is based on principles of heredity and innate predispositions for acquiring language, particularly its syntactic structure. According to Chomsky, this theory is indebted to the works of the Port Royal grammarians, developed in 17^th-century France. The Port Royal grammar (^{Arnauld and Lancelot, [1660] 1992}), rooted in Cartesian principles, was created to establish the logical foundations and proper use of languages, with the aim of identifying their universal bases. The intention was to develop a set of grammatical principles applicable to all languages, thereby aiding their specialization as tools for science and pedagogy. Noam Chomsky references Cartesianism and the principles that guided the creation of the Port Royal Grammar in its universalizing logical foundations to situate his linguistic theory within this philosophical tradition, proposing it as an explanation of human linguistic capacity based on cognitive and biological principles.

According to the author, within the framework of the Biolinguistic Program, which his explanatory theory of language acquisition aligns with, it is essential to consider the close relationship between available linguistic data and the selection procedures to make the specification of the language to be acquired feasible. Although the innate potential to develop grammar is universal, it must become specific to a particular language through the child's exposure to that language's occurrences in use. The theory must, therefore, account for "the vast gap between the available data and what the child comes to know."

It has been recognized since the early days of work in generative grammar that this Poverty of Stimulus problem is enormous, and subsequent investigations into what a very young child knows, along with the statistical study of the scarcity of available data, have revealed that the problem is even more severe than previously supposed. […]. (^{Chomsky, 2021}, p. 8)

In this context, the acquisition of language, particularly its grammatical foundations, occurs through the interaction between the subject's cognitive structures and the linguistic data that align with the syntactic organization of the language, based on principles of regularity and systematization. Since this process is grounded in logical principles, which are considered the foundation of thought across all languages, semantic processes hold a central position in this theory. This is evident when considering that language acquisition and development occur through exposure to linguistic data in use; additionally, the logical foundations of language are tied to thought processes, such as "conceiving (forming a concept), judging (making assertions about concepts), and reasoning (drawing conclusions based on established judgments)" (c.f.: ^{Cizescki, 2008}, p. 123). Furthermore, this relationship is also apparent when considering the prescriptive rules developed within the grammatical tradition, such as those outlined in the Port Royal Grammar, which aimed at refining language as a tool for thought and communication.

However, in Chomsky's theoretical framework, semantic foundations are repositioned outside the primary object of study. This shift occurs because Chomsky initially posits that reference is connected to elements external to language and, therefore, is not governed by its regularities or systematic structures. Instead, reference is subject to the free choices of speakers during use, placing it beyond the scope of naturalistic linguistics (^{Oliveira and Cruz, 2004}).

To uphold the principle that language is structured according to internal rules, Chomsky's theoretical propositions could not conceive of reference as a result of social convention, as such conventions would be processes external to the linguistic system itself. Additionally, the absence of an underlying rule within the semantic system that could be described and explained would undermine the fundamental objectives of the theory. Chomsky's proposed solution was to consider that reference is not established in relation to something in the world or to an agreed-upon meaning, but rather to mental entities formed during the process of lexical acquisition. This approach assumes that words are internally structured according to universal traits, and these traits constrain their use (^{Oliveira and Cruz, 2004}).

It is necessary to consider the function that semantics serves in reference theories for literacy, and how it is either appropriated or excluded according to their epistemological assumptions, in order to understand its effects on conceptual propositions about literacy and on the didactic processes involved in teaching writing in its initial phases. The application of the Chomskyan theoretical proposal in studies on the psychogenesis of written language reveals the underlying dynamics of reference conceptions in constructivist-based literacy approaches. In these studies, it is posited that language acquisition occurs based on the innate capabilities of the human subject, according to a Chomskyan perspective, from which the individual can then acquire a second modality of representing verbal language, writing:

Currently, we know that a child who arrives at school possesses a remarkable knowledge of their native language, a linguistic knowledge that they use "without knowing" (unconsciously) in their daily communication acts. Following Chomsky, it has become common in psycholinguistics to distinguish between competence and performance. This distinction warns us against the — markedly behaviorist — tendency to equate a subject's real knowledge of a particular domain with their actual performance in a specific situation. […]. It is unlikely that schools could have acknowledged this "linguistic knowledge" of children before psycholinguistics brought it to light; but can we now ignore these facts? Can we continue to act as if the child knows nothing about their own language? Can we continue to act in such a way that we force the child to disregard everything they know about their language to teach them, precisely, to transcribe this same language into a graphic code? (^{Ferreiro and Teberosky, 1985}, p. 24-25)

Thus, we observe the convergence of epistemologically divergent theoretical conceptions (c.f.: the "Debate between Chomsky and Piaget" in Piattelli-Palmarini, [1987]). In the Chomskyan perspective, reference is conceived as being established in relation to mental entities, as previously discussed. Conversely, in the Piagetian perspective, reference is constructed through the subject's interaction with the object of knowledge, in a process of referencing. This distinction will be discussed further to characterize what is referred to as the modern episteme.

THE MODERN EPISTEME

The second epistemological perspective guiding the production and organization of data, termed the modern episteme here, incorporates the subject as an active participant in its own processes of knowledge acquisition. Rather than merely reflecting on the world as a framework of representations to be described, the subject is repositioned to investigate the very possibility of knowing itself; this perspective focuses on understanding the processes through which cognition develops and the conditions under which this development occurs, treating these aspects as objects of knowledge.

The functioning of this episteme, which began at the end of the classical period while retaining some of its foundations, and which continues to have an impact today, is marked by the emergence of the subject as an element of knowledge (^{Foucault, 2007}). This shift allows for viewing time not merely as a principle of sequencing, analysis, or ordering through language, as was typical in the classical period, but as a constitutive element of both the object and the subject. Consequently, this perspective enables the consideration of knowledge as a developmental phenomenon.

This epistemological foundation is evident in various perspectives on methods and proposals for teaching reading and writing in its initial phases. These include cognitive approaches, such as the psychogenesis proposed by Jean Piaget; socio-historical approaches, such as Vygotsky's perspective on the dialectical relationship between language and thought; and socio-interactionist approaches, which examine historical and cultural interactions and the production and transformation of the social context through cooperative human work, including linguistic work.

The modern episteme integrates the subject as a key element in the pursuit of knowledge about its own constitutive processes. In this framework, the subject is placed in a position where the very possibility of knowing is questioned. The focus shifts to understanding the processes through which cognition develops and the conditions necessary for this development.

This episteme appears to be based not on a conception of reference but on one of referencing. It posits that objects are not pre-constructed in the processes of interaction between the subject and the world, or among subjects, but are constructed within the subjects themselves through interactive processes. From this perspective,

Knowledge objects are not given a priori as static referents but are constructed through interaction situations, in processes of referentiality resulting from our sociocognitive interaction with the physical, social, and cultural world. In this sense, "designated entities are seen as objects-of-discourse rather than objects-of-the-world". (^{Koch, 2002}, p. 79)

The conception that objects of discourse are constructed through the process of interaction appears epistemologically aligned with psychogenetic constructivist principles. However, it is important to note the differences in how these principles are applied in various research approaches. For instance, in the constructivism developed by Emília Ferreiro for studies on the acquisition of the alphabetic writing system, we observe the stages through which children progress as they construct hypotheses about the system's operation amid its inherent instability. This approach seems to resonate with the idea that the construction of discourse objects is a process of referencing. In this process, re-categorizations are made based on a presumed instability inherent to the categories:

Thus, it is contrasted with "the hypothesis of a referential power of language that is founded or legitimized by a direct (and true) link between words and things," but it is proposed that "instead of assuming a priori stability of entities in the world and in language, it is possible to reconsider the issue starting from the constitutive instability of categories, both cognitive and linguistic, as well as their processes of stabilization." (^{Mondada, 2002}, p. 118-119).

The assumption of referencing appears to support constructivist propositions about language, knowledge, and the processes of teaching and learning. As ^{Montoya (2006)} notes regarding the Piagetian perspective, and considering the advancements in Jean Piaget's research, the relationship between the learner and the reality to be known is constructed through internal mechanisms of thought formation, interacting with external processes of socialization and narrativization in which the subjects participate.

In contrast to empiricism and logical positivism, Piaget's formulation observes a true endogenous construction of concepts. As mentioned earlier, this construction does not occur solipsistically, isolated from interaction with other individuals. On the contrary, this interaction is a necessary, though not sufficient, condition. We have already observed the crucial role of narrative in the transition from sensory-motor schemas to concepts, as well as the collaborative advancement of linguistic organization and conceptual organization. (^{Montoya, 2006}, p. 126)

As ^{Montoya (2006}, p. 126) discusses, Jean Piaget's propositions in the study of human behavior development concerning intelligence and language are guided by the major thesis of "continuity with the reconstruction of previously acquired structures," from which one can infer, in contrast, the non-fixity of the reality to be known.

The conception of language based on the assumption of referencing appears to be replaced by the principle of reference when certain teaching strategies are applied to literacy work. It is important to consider the effects on the appropriation of the psychogenetic perspective when literacy is the learning process in focus, and writing is the object of learning as a secondary representation of language. In this context, various conceptions of writing in relation to speech, language in relation to syntactic and semantic processes, and language in relation to thought come into play.

According to ^{Dechauffour (2016)}, the Chomskyan proposal is linked to the philosophical aspect associated with realism, which assumes the pre-existence of general properties of objects that can be appropriated and described based on rationalist principles. ^{Dechauffour (2016)} refers to this tendency as rationalist fixism, characteristic of pre-Darwinian science. In this view, nature is conceived as "a fixed totality awaiting description" (^{Dechauffour, 2016}, p. 192). Consequently, the relationship between language and the world is characterized as a labeling of objects to be known, a perspective that ^{Mondada (2002)} critiques in her analysis of this conception of reference.

According to the author, the Piagetian perspective diverges from both realism and its nominalist opposition, as it does not presuppose the pre-existence of object properties or the primacy of the mind over the world. Instead, it suggests that the existence of objects arises from the process of classifying them. The Piagetian perspective is seen as a third position, distinct from the two more traditional ones, characterized by being neither ante re (realist) nor post rem (nominalist), but in re:

[…] a middle and new position, called conceptualism. Particularly developed by Abelard, it asserts that ideas are concepts of structural similarities observed in things. They are constructs that allow us to express things. They are the product of interaction with things, but this interaction is the occasion to experience the reality of the world. Concepts are products, but they are not arbitrary; the discovered connections are also as real as my interaction with the world. Thus, the idea is found in the thing or in the experience I have of the thing. It is possible to reconstruct the history of the discovery of these connections, but it is the history of discovering norms; it is not absolutely contingent. (^{Dechauffour, 2016}, p. 188)

According to the Piagetian perspective, the subject is seen as actively engaging with the reality to be known, with knowledge being a process of transforming reality to understand its development. This view critically distances itself from fixism.

However, when examining the conceptions of language specifically, both the Chomskyan and Piagetian perspectives assume that the stability of knowledge is ensured by the principle of reference. Even within the Piagetian framework — where knowledge is developed through the interaction between subject and object in a constructive process, with referencing initially emerging from this subject-language-world interaction, and the construction of knowledge leading to the transformation of the world —, language is viewed as a tool for classification, description, and ordering. According to ^{Dechauffour (2016}, p. 193), this perspective disregards, as does the Chomskyan proposal, the notion that language is fundamentally a matter of communication.

When dissociated from its communicative function and examined strictly on syntactic (syntagmatic) bases, as in the Chomskyan perspective, or as a classification instrument, on paradigmatic bases, in the Piagetian perspective, the conceptions of language in both frameworks would strip it of its semantic nature.

In this case, psychogenetic-based propositions are pedagogically appropriated in a manner that conceives objects of knowledge as fixed and pre-existing. This positions the learner as someone who apprehends these objects through a process of identification based on the stability of the reference relationship between language and what it denotes.

Considering the distinction between the conceptions of reference and referencing that underlie literacy propositions, it is possible to relate the ways constructivism is appropriated with pedagogical purposes to specific epistemological bases. Sometimes, these appropriations align more closely with the epistemological foundations of the classical episteme, while, other times, they align with those of the modern episteme. This process of competition influences how literacy proposals are conceived, distributed, and developed, both curricularly and pedagogically, within a school context.

In the initial stages of applying psychogenetic concepts to pedagogical practices for literacy in the 1980s (^{São Paulo, 1988}), it was argued, based on the proposals of Emília Ferreiro and collaborators, that learning conditions needed to be adapted to avoid adult-centered principles that overlooked the social and cultural backgrounds of students. These principles often prescribed teaching methods that conflicted with the child's developmental stage. The adoption of constructivist principles was seen as a revolutionary shift that repositioned teaching and learning subjects so that the learner's developmental time, rather than school time, dictated the acquisition of the alphabetic writing system. According to this perspective, early constructivist proposals focused on the child's interactions with learning objects — specifically, the alphabetic writing system — and the referencing processes established by the child.

The conceptions of the subject specific to each of the epistemes described above, in their relationship with the ways in which they establish the constitution of objects of knowledge and teaching and learning practices, generate distinct approaches to organizing theoretical-methodological conceptions and pedagogical-curricular propositions. These approaches vary depending on whether they emphasize external processes of learning control or internal processes of subject development. These distinctions align with two types of pedagogical practice as characterized by ^{Bernstein (1996)}: visible pedagogy is defined as one that seeks to produce explicit, comparable differences between learners in order to stratify them based on internal differences in potential. The focus is on evaluating and comparing the learner's output against a common external standard, thereby emphasizing transmission and performance. It highlights to the learner the expectations in terms of content and form of expression —hence the adjective visible. Invisible pedagogy, on the other hand, seeks to produce non-comparable external differences between learners, who recognize themselves through the uniqueness of how they express their shared competencies. This approach aims to develop learners’ internal procedures — cognitive, linguistic, affective, and motivational — thereby emphasizing acquisition and competence. In this case, teachers’ control and clarity over the criteria by which learners’ performances will be evaluated are limited — hence the adjective invisible.

What is considered visible or invisible are the contents deemed relevant and the legitimate forms of communicating what is learned — that is, in visible pedagogy, learners clearly understands what is important to learn and how they should express their knowledge; in invisible pedagogy, however, these elements are clear only to evaluators (teachers) and are based on criteria that are not understood by the learners or their families. In other words, in visible pedagogy, control of the pedagogical relationship is explicitly held by teachers, while in invisible pedagogy, it appears to be with the students — though, in reality, they are unaware of what truly matters, despite seeming to be highly valued in various aspects.

These differences in emphasis between visible and invisible pedagogies will clearly affect both the selection and the organization of what is to be acquired, that is, both the recontextualizing principle adopted to create and systematize the content to be acquired and the context in which it is acquired. (^{Bernstein, 1996}, p. 104-105)

The application of the two aforementioned epistemes in data analysis, while it may initially appear schematic, actually underscores the complexity and heterogeneity that are presumed to be inherent in the object of study, specifically, literacy as an object formed through the interplay between academic-scientific and pedagogical-curricular domains. In this context, elements of theoretical-methodological perspectives can simultaneously align with distinct, competing epistemes and, over time, selectively determine what is retained, rejected, reappropriated, or denied from the theoretical and methodological perspectives with which they establish interconstitutive relationships (^{Maingueneau, 1997}; ²⁰⁰⁵). This process is further shaped by the actions of agents situated in different recontextualizing fields, who operate with varying degrees of autonomy within specific sociopolitical frameworks. Within these fields, discourses circulate through various channels, supporting both continuities and ruptures that are linked to both classical and modern epistemes.

DATA ANALYSIS

Between the different versions of the Curricular Proposal for Teaching Portuguese Language — 1^st grade (^{São Paulo, 1988}; ¹⁹⁹¹), there is a noticeable shift toward more explicit didactic guidelines for conducting literacy processes in schools. The proposals for teaching and learning writing in its initial stages reflect a dual influence: they align with the socio-interactionist foundations that underpin the general section of the document on Portuguese language instruction, while also incorporating principles specific to the constructivist perspective, particularly in its more rigorous theoretical-methodological connections with studies on the psychogenesis of written language.

Thus, teaching time, which according to the initial constructivist-based proposals was meant to respect the learners’ pace, is later shown to be dictated by the school curriculum, which selects, organizes, and schedules what should be learned, how, and when. This shift is evident in the texts addressed to teachers, signed by the education secretaries in each edition of the document: in the 1988 edition, the curricular proposal is described as something that "should not be received as an instrument that restricts the teacher's performance, but rather as necessary support for the organicity of the pedagogical work that takes place in multiple school units" (^{São Paulo, 1988}, p. 5); by the 1991 edition, "the degree and limits of autonomy to be implemented in schools," stating that "while the organization of pedagogical processes, for example, should be the responsibility of the school, the basic curriculum will remain common to the entire network and determined by the Secretariat" (^{São Paulo, 1991}, p. 3). As a result, the responsibility of the teacher shifts to deciding how to teach, while the determination of what should be taught and when is prescribed by the Department of Education.

This shift reveals a transition from a didactic orientation predominantly shaped by the academic discourse of a socio-interactionist nature with emancipatory propositions, to one primarily influenced by the pedagogical-curricular discourse grounded in official prescriptive bases. This change in perspective impacts how semantic processes are conceived in the proposals for teaching and learning writing in its initial stages, as the relationships between conceptions of language and competing epistemes evolve. Consequently, assonances, tensions, conflicts, and/or accommodations emerge in the conceptions of the subject, language, and reference/referencing within the propositions for teaching and learning the Portuguese language.

Amid these changes, the later version of the curricular document under analysis includes passages that outline the characteristics attributed to the process of acquiring the alphabetic writing system and its relationship with school-based teaching and learning. This specification appears to signal a shift away from the socio-interactionist conceptions that originally guided the Portuguese language teaching proposals in the document. Instead, it leans toward literacy propositions associated with constructivist conceptions that have been reinterpreted within the framework of pedagogical-curricular discourse.

Thus, in the passage under analysis below, it has been stated since the 1988 edition that the formulation of hypotheses about children's writing occurs through "joint work," reflecting the conception of the classroom as a space for interaction among its participants. This perspective provides assumptions that can guide teachers in their work regarding what and how to teach:

Conceived, on the contrary, as a system of representation, writing becomes the appropriation of a new object of knowledge. From this perspective, the focus is primarily on the learning construction process, the hypotheses that the child develops about writing, and the ways in which, through collaborative work, the child transforms it into a new tool for their own experiences, for representing reality, and for communicating with others. By opting for this broader conception, it is, however, necessary to make explicit some assumptions that can guide the selection of content and the organization of teaching-learning strategies. (^{São Paulo, 1988}, p. 23, emphasis added)

In the 1991 edition, the section under observation includes a passage that appears to narrow learning to the child's individual work with their own writing. Here, interaction is framed more on cognitive grounds, focusing on writing as an object with which the child interacts. This is evident when it is stated that: "This process of interaction and construction of the child in relation to writing is what makes them learn to write and read, rebuilding their own path" (^{São Paulo, 1991}, p. 28):

Conceived, on the contrary, as a system of representation, writing becomes the appropriation of a socially and culturally constructed object of knowledge. From this perspective, the focus is primarily on the learning construction process, the hypotheses that the child formulates about writing, and the ways in which, through collaborative work, the child transforms it into a new tool for their own experiences, for representing reality, and for communicating with others. This process of interaction and construction in relation to writing is what enables the child to learn to write and read, reconstructing their own path, … this process should be understood as a ‘process of objectification,’ in which the subject continually constructs hypotheses and faces contradictions that force them to reformulate their hypotheses…*, in an attempt to understand how the written language system works and what it means. By opting for this broader conception, it is, however, necessary to make explicit some assumptions that can guide the selection of content and the organization of teaching-learning strategies. (^{São Paulo, 1991}, p. 27-28, emphasis added)

The circumscription of the individual's activity is further highlighted by the emphasis on the reconstruction of "their own path" and the objectification attributed to the process being thematized. These elements are then positioned as the reference for the anaphoric phrase "this more comprehensive conception," which replaces the socio-interactionist observations that constituted this referent in the previous version.

The shift from a social to an individual characterization of the writing acquisition process has significant implications for how meaning and reference and/or referencing are approached in the teaching and learning of writing. This shift reflects a movement away from viewing writing as a tool through which learners collectively transform their experiences, represent reality, and communicate with others. Instead, writing becomes positioned as the primary object to be learned, rather than as an instrument of learning and communication. As a result, writing is reconceptualized, moving away from its socio-interactionist foundation as a means of representing reality and communicating, to being understood as a system for representing the spoken modality of language.

Thus, the shift is from collaborative work in producing representations of lived experiences, where, from a socio-interactionist perspective, referencing relationships between language and reality are established based on the understanding that this reality is constructed and transformed by interacting subjects, to a perspective where the propositions, in later versions of the document, emphasize the subject's interaction with an object to be known. This object, in its stability, defines a terminal stage for the learning process (specifically the alphabetic principle of the Portuguese writing system). Here, a reference relationship between language and reality is assumed, wherein the analytical nature of the object itself teleologically determines the stages through which the subject progresses on their path of knowledge construction.

Thus, from one version of the analyzed document to the next, there is a discernible shift away from principles associated with the modern episteme, where subjects and objects are mutually transformed through their work within the space and time of their experiences, toward principles associated with the classical episteme, where the temporal ordering defined by the nature of the object to be known guides the cognitive process undertaken by the learning subject.

In this context, even though the Piagetian-based propositions underlying constructivist literacy approaches are initially aligned with principles of the modern episteme, where objects and actions are in continuous movement, even when operating with stable categories (c.f.: ^{Montoya, 2006}), these propositions undergo a shift when intersected with elements of pedagogical-curricular discourse. In this shift, they become guided by the assumption of reference, leading to a reliance on nominalism and aligning with principles of realistic fixism (c.f.: ^{Dechauffour, 2016}).

In the 1991 edition of the Curricular Proposal, the role of working with the child's name in literacy processes is prominently emphasized. This version of the document includes a passage listing texts to be used in the classroom that "present regularities, imply general rules of construction, and, therefore, offer information to children." It also provides an example of an activity "that has function and meaning for the child: writing their own name," reflecting justifications for working with the child's name as outlined by ^{Teberosky (1989)}.

The emphasis placed on the child's name may have influenced the revision of another passage in the document between versions: in the 1988 edition, the text begins by describing the initial phase in the process of acquiring writing as the moment when the child perceives the difference between writing and drawing, recognizing that writing involves using specific lines characterized by graphic or letter forms that they have learned to reproduce. It is observed, in this previous version of the document, that the child with whom the adult would be interacting is represented in a non-particularizing way, but as the subject of learning, in a given context — this, indeed, the object of considerations in the text under analysis, which is evident in the use of the present tense to represent in a generalized way the moment of the literacy process under consideration:

A first moment is characterized by the perception that writing is distinct from drawing. If we ask a child to draw a person, they might produce something like this — [reproduction of a drawing] —; however, if we ask them to "write," their initial attempts will likely consist of mere scribbles, wavy or jagged lines, repeated series of elements, or even sequences of letters they have encountered outside of school, though these bear no correspondence to the sounds they represent [reproductions of children's scribbles follow] […]. (^{São Paulo, 1988}, p. 24).

It should be noted, moreover, that the term child, to which the anaphoric them presumably refers, is not explicitly stated in the passage under observation, nor in the preceding paragraph:

From a cognitive-constructivist perspective, children's writing follows a surprisingly similar line of development across various cultural contexts, educational situations, and languages [note 3]. It progresses through different hypotheses about what writing is, successively understanding its properties and its relationship with speech. We will initially seek to understand this process in order to better guide our pedagogical approach. (^{São Paulo, 1988}, p. 24, emphasis added)

In this paragraph, there is ambiguity due to the imprecision of which preceding term the pronoun it refers to it progresses. It could refer to child, mentioned only in the preceding paragraph and subsumed under the term childhood; or to children's writing, which is closer to the pronoun and where childhood seems to imply a reference to the early stages of writing acquisition. Thus, although the paragraph begins with the mention of "the cognitive-constructivist perspective," it is the process of writing acquisition and its relationship with learning contexts, rather than the subject of learning, that appears to be the more specific focus of this version of the document.

The use of the present indicative, combined with the gerund, is also noted as a way to represent the observed process in a non-particularizing or non-individualizing manner. This approach aligns with the social foundations of studies on the psychogenesis of written language, explicitly referenced in the passage "in different cultural environments, in different educational situations, and in different languages." This passage includes a note referring to Ferreiro's publication (1986). However, it is also possible that this statement aligns with the socio-interactionist principles underpinning the curricular proposal under analysis.

In the corresponding passage reworked in the 1991 edition of the document, the referential ambiguity is resolved by using the term child in the phrase "The child is elaborating several hypotheses." Additionally, the object to which writing is related undergoes a change: while the 1988 edition describes literacy as understanding the relationship between writing and speech, the 1991 edition omits any mention of speech, instead, focusing on what writing represents:

From a cognitive-constructivist perspective, children's writing follows a surprisingly similar developmental trajectory across various cultural settings, educational situations, and languages [note]. The child gradually formulates different hypotheses about what writing represents and progressively constructs answers to two key questions: what does writing represent, and what is the structure of its mode of representation. (^{São Paulo, 1991}, p. 28, emphasis added).

There appears to be a shift in positioning the child at the center of the learning process, while simultaneously decoupling this learning from the communicative process, which was implied in the 1988 edition by the mention of the relationship with speech. Instead, the focus moves to the process of representation, specifically the alphabetic writing system. Although the child is now more explicitly identified as the learner, the emphasis is placed on the object to be learned rather than on the learning process or the context of interaction. This shift establishes the primacy of the principle of reference over that of referencing, emphasized by the focus on nominalization. To illustrate this phenomenon, the document presents an example of an episode, narrated in the perfect past tense, involving a specific child learning to write his mother's name:

At the beginning of the reconstruction process, the child assumes that writing is just another way of drawing things. For example, when asked to write the word "mommy," a child who had learned to write her mother's name (Dalva) but did not know the word "mommy," confidently wrote "Dalva." When questioned about the discrepancy, she was perplexed by the adult's inability to understand something so obvious: Dalva and mommy are the same person and, therefore, should have the same written form [note] (^{São Paulo, 1991}, p. 28, emphasis added).

In the example provided to illustrate the initial stage of writing acquisition, it is noted that "the child assumes that writing is another way of drawing things," with both the adult and the child operating under the assumption that the relationship between language and the world is established according to the principle of reference.

However, the issue at hand is not merely whether writing is considered a form of drawing, but rather how the process of representation is executed: for the child, the referent remains the same, even if the words used to refer to it (e.g., Dalva and mommy) may carry different meanings.

From the child's perspective, the issue at hand seems to extend beyond mere reference to encompass the process of referencing itself. This involves the selection of linguistic forms for constructing representations. The meaning of the forms used to refer to objects is guaranteed by the modes of presentation of that object — its meanings —, which can only be fully understood within the process of interaction. For the child, unlike the adult narrating the episode, this interactional context is presupposed, leading to the child's perplexity when the adult fails to grasp the meaning of words that were written as part of a dialogue constructed in that context, during their interaction.

The pedagogical-curricular discourse here appears to be the driving force behind the interpretation of learning processes, shifting the perception of these processes from collaborative work within a context to a relationship between the individual and the object to be learned. In the context of literacy, this shift confines the object of learning, in the proposition under analysis, to the writing representation system. This system is understood as a cohesive set of relationships, forming a referential unit to be comprehended by the learner. As a system, written representation is conceived as autonomous from speech, even though it is in speech that the alphabetic principle underlying this system is realized.

The shift from principles associated with the referencing process (typically involved in interactive processes — where the child has been continually shaped within the world of orality) to principles associated with the reference process also seems to guide the changes made between the 1988 and 1991 editions. This shift is evident in the deletion of the following passage (replaced by the one presented later):

Children are also unclear about the criteria by which different writings are related or distinguished. For example, when asked to provide other words similar to BABY in terms of spelling, a child learning to read responded with "baby bottle" and "pacifier." This indicates that the child was more focused on the meaning of the words and the relationships between them within the same semantic field, rather than looking for similarities in spelling or corresponding sounds. (^{São Paulo, 1988}, p. 25, emphasis added)

Attention to meaning in interactive processes underscores how children engage in these processes. Referencing emerges as a collaborative achievement between subjects interacting within a specific context. In such interactions, the forms used to construct discourse objects are not fixed but are continuously reconstructed through dialogue and recategorizations (c.f.: ^{Mondada, 2002}). In the narrated episode, the child uses writing to develop categories that stabilize meanings and interpretations within the ongoing interaction.

In this passage, the relationship between subjects and the context of meaning production is addressed, highlighting the semantic aspects associated with language production. This aligns with ^{Montoya's (2006)} consideration of the Piagetian perspective, which views objects and actions as being in continuous movement, even while operating with stable categories.

This perception is dispelled in the 1991 version, where the observed passage was removed and replaced with the following passage. Once again, the pedagogical-curricular discourse appears to influence the document review process, resulting in the circumscription of subjects and objects to a specific didactic relationship:

We term pre-syllabic the types of writing in which the child has not yet understood their properties or their relationship with speech. Such written manifestations have their own characteristics, and we could say that they follow a logic, subdividing into:

— undifferentiated writing. A series of identical graphemes, regardless of the stimulus; [followed by a reproduction of writing produced by a 5-year-old child]

— differentiated writing: writing displays a varied series of graphemes, responding to differences in stimuli. Differentiation is achieved through variation in repertoire, quantity, or position of graphemes. This variation may be influenced by the recall of some writing model (proper names, some learned words). [followed by a reproduction of writing — a text — produced by a 9-year-old child]. (^{São Paulo, 1991}, p. 28-29)

In this context, the relationship with speech appears to represent an update of the alphabetic principle of writing. The 1991 edition replaces the emphasis on semantic relationships established by the child in interaction with the adult, as seen in the 1988 edition, with a focus on the phases the child undergoes in the construction of the alphabetic writing system. The updated passage also continues to emphasize working with the child's name, now associated with a writing model. This shift seems to reaffirm the influence of pedagogical-curricular discourse on literacy conceptions and propositions for teaching and learning writing in its initial phases, steering them toward the principles of the classical episteme. Accordingly, the principles of this episteme prioritize the object of knowledge and the notion of reference when considering the relationships between language and reality.

FINAL CONSIDERATIONS

The analysis of the data, contrasting the different versions of the document, highlighted the effects of competing perspectives on the semantic foundations of pedagogical propositions for literacy. These perspectives influence the teaching and learning processes of writing in its initial stages.

From one version of the document to another, more explicit didactic guidelines for conducting literacy processes in schools are established. The propositions for teaching and learning writing in its initial phases must navigate both the socio-interactionist foundations referenced in the general part of the document on teaching Portuguese and the principles specific to the constructivist perspective, particularly its theoretical-methodological connections with studies on the psychogenesis of written language. Consequently, while earlier constructivist-based proposals emphasized respecting the learner's individual pace, the later version of the document presents learning time as dictated by the school curriculum, which determines what should be learned, how, and when.

The shift from a didactic orientation predominantly shaped by academic-scientific discourse to one primarily influenced by pedagogical-curricular discourse in the analyzed document affects how semantic processes are conceived in propositions for teaching and learning writing in its initial phases. This shift changes the relationships between conceptions of language and competing epistemes, leading to the emergence of assonances, tensions, conflicts, and/or accommodations in the conceptions of subject, language, reference, and teaching and learning.

The aforementioned shift can also be understood as a process of recontextualization, where discourse is removed from its original context of production — the academic-scientific field — and subsequently relocated into new contexts, such as curricular and/or teaching and learning proposals (^{Bernstein, 2000}). In this process, the discourse takes on new characteristics based on the choices made by agents working within recontextualizing fields: the official recontextualizing field (ORF) and the pedagogical recontextualizing field (PRF). According to ^{Bernstein (2000}, p. 33, our translation), ORF is "created and dominated by the State and its selected agents and ministers," while PRF is comprised of "pedagogues in schools and education departments (at universities), specialized newspapers, and private research foundations."

In the Brazilian context, where curricular prescriptions are produced at the federal level and then adapted at the state and municipal levels, the division into ORF and PRF takes on specific characteristics; for instance, in state and municipal education departments, agents include both those linked to the State and education professionals, such as university professors who act as advisors, as well as individuals associated with various foundations and civil organizations. Consequently, the degree of autonomy available to agents who translate discourses from one context to another varies — depending on the political scenario.

It is worth noting that, in the state of São Paulo, a process of administrative reorganization and curricular reforms had been underway since the 1970s, driven by the expansion of educational access to broader segments of the population. This scenario, characterized by increasing challenges related to the conditions of this expanded access, is likely connected to the concerns reflected in pedagogical-curricular measures. ^{Sampaio (1998}, p. 231) provides a description of the state of schools during this period:

The school, required to accommodate a large and disorganized urban population, relied on bureaucratic organization to respond to a chaotic situation. The teaching and learning process, around which all its work revolves, was constrained within possible limits, becoming reduced to mass education.

The author also states that the curricular proposals that began to be discussed in the state of São Paulo between 1984 and 1987 "did not find fertile ground for trials and reformulations that would make them advance and contribute to teachers’ practice" (^{Sampaio, 1998}, p. 227). She refers to the failure of the Education Departments to meet the needs of schools, as well as the absence of teacher training programs — in addition, of course, to the structural precariousness of the education network. It can be assumed, therefore, that the 1991 version, to some extent, adopted a tone that was less respectful of the autonomy of schools and teachers due to the difficulties imposed by the real conditions of implementing curricular proposals: if this makes sense, a more directive orientation of the teacher's work may represent, in the competition between epistemes highlighted here, the effect of an impoverished vision of what these professionals could do in a scenario in which the conditions for the development of their work with quality would not materialize.

It is worth thinking, therefore, about the way in which the discursive competitions highlighted in this article shape understandings and practices that defy any simplistic attempts at analysis.