FOUR ANALYSES ON THE OFFICIAL LETTERS OF RIO GRANDE DO NORTE (BRAZIL)

A BSTRACT Between variation and change, as we know, not only language, but also the texts, with their structures, undergo changes over time. This article aims to analyze textual changes observed in letters of the 18 th , 19 th and 20 th centuries. These documents are part of the corpus called Cartas oficiais norte-rio-grandenses (Rio Grande do Norte official letters) combined to the regional project called História do Português Brasileiro no Rio Grande do Norte ( PHPB-RN ) and the national project História do Português Brasileiro ( PHPB ). Four analyses were made from these letters, built upon the ideas of Coseriu and on the studies about discourse tradition (Koch 1997, Kabatek 2006), according to which texts are shaped to follow their own tradition (Coseriu 2007). After that, we focus on the letter and the letter of registration genres and their macro-structures as well as on the clausula (valediction) and the application expressions micro-structures. This analysis shows some processes of variation and change which these official letters underwent, besides how textual dynamics can indicate social change.

We have chosen a fragment of one of Camoens' sonnet to illustrate our study.The word change (mudam-se, muda-se, mudança), with different meanings in Camonian text, seems to be a very interesting term for us to outline an essay that intends to point out the linguistic features of handwritten texts from another time -precisely from the first half of the 18 th Century to the first one of the 20 th -besides presenting the material available to the researchers whose studies are concentrated in the domain of Diachronic Studies.The texts referred to are part of the regional project called História do Português Brasileiro no Rio Grande do Norte -PHPB-RN.
In this article, four analyses are offered to the reader -two of them previously unpublished, all of them reviewed and developed-from a larger work, Melo's dissertation (2012).We also follow another major national project named Para a História do Português Brasileiro -PHPB, coordinated at this moment by Professor Dr. Ataliba de Castilho, from the Universidade de São Paulo (USP).The two projects were joined together owing to the main goals of each one: establishing a minimum common corpus that could be useful for linguistic research elaborated in Rio Grande do Norte (a state in the Northeast of Brazil) as much as for the construction of a national database (formed by data from each region in Brazil).
Therefore, we created a corpus to represent the material based on which we would carry out the studies about the process of tradition-innovation, which is especially noticeable from a diachronic point of view, i.e. in the history of the language.
The arrival of a new movement of Diachronic Studies in Brazil took place in the 1980s and more intensively in the mid-1990s, its most important event possibly being the foundation of PHPB.The project originally covered states from the South and the Southeast -as well as Bahia, with its own tradition of very productive studies on Diachrony-and these have been the ones that have backed the project the most.As Castilho wisely points it, we are still far from achieving a whole history of Brazilian Portuguese, because «parece mais factível consagrar-se ao conhecimento diacrônico das variedades regionais do português brasileiro para, num segundo momento, obter-se a visão do conjunto» (Castilho 1998: 15).Therefore, we trust that by means of that challenge to which we have put ourselves -namely, to select and transcribe the official letters of Rio Grande do Norte-we can somehow give our contribution to writing process of some of the regional history of Brazilian Portuguese.
The PHPB decided to put together a database, discriminated according to two types of corpora that are common to all teams affiliated to the PHPB project: the «corpus mínimo comum -manuscritos» and the «corpus mínimo comum -impressos».To the first group belong the following clusters of texts: 1) wills; 2) criminal cases; 3) council minutes; 4) private letters; 5) private administration letters; 6) official letters.In the second group are: 1) letters from editors / editorials; 2) letters from readers; 3) advertisements.Besides these two corpora, there is still a common differential corpus, «a ser usado dependendo da disponibilidade regional e do tema» (Castilho 2010: 8), which may comprise: 1) inventories; 2) memoirs / historical reports and travel journals; 3) interludes and other drama; 4) oral inquiries (NURC) (vid. Hora and Silva 2010: 423).
Our research deals with one of the genres of the manuscript corpus, the official letters, and goes through the time frame specified by PHPB, ranging from the 18 th to the 20 th century, each analyzed in its first and second halves separately.We were guided by the demands of PHPB, aiming at obtaining 5.000 words for each half-century, from the first half of the 18 th century until the second half of the 20 th .However, since we have only found typed texts from the second half of the 20 th century and we had limited ourselves to manuscripts only, we decided to abandon the 1951-2000 texts.Our new time frame encompasses thus the first half of the 18 th up to the first half of the 20 th , totaling five sets of 50 years and at least 25.000 words.
The corpus is composed of official letters, a textual category 1 1 We discriminate genres and textual categories because there are some PHPB demands that are not configured as genre, as the private administration letters and official letters, the latter concerning the letters that circulated in public administration.Unlike other requests -such as wills, criminal cases, letters from readers and inventories-which are traditionally classified as genres, we consider private administration letters and official letters as textual categories, because these designations act as umbrella terms, for the reason that each represents not one, but a certain set of genres.Under the first term, for example, there can be the circular genre (circular) or the memorandum (memorando), while in the second textual category we can find an administrative letter (ofício) and a letter (carta).
Once the letters were described, transcribed and edited, we carried a central search: describing, investigating and analyzing the macro-structure of genres and microstructures such as the formulaic expressions both synchronically and diachronically, since we would operate in the temporal thread that generates the dynamics between maintenance (tradition), variation and change.The latter (change) is reflected on the temporal course of official letters and also reveals changes in the social conjuncture of Brazil, significantly felt in formulaic expressions.

ON THE CONSTITUTION OF THE OFFICIAL LETTERS OF RIO GRANDE DO NORTE
Once the doors are open to working with texts from past language stages in order to constitute a single corpus, the first step to be taken is to find sources from which to get the data to be selected.When browsing through texts from the public administration, the sources are supposed to be the public bodies such as chambers, institutes etc.Therefore, the first repository was the Instituto Histórico e Geográfico do Rio Grande do Norte (IHGRN), one of the most important collections of historical documents in the State, where we began our search, concentrated initially on the 19 th century.But resorting to more data became necessary.We discriminate, below, the material that makes up the corpus and the means used to obtain them: a) Texts from the 18 th century were transcribed from the photographs of some manuscripts relating to the Captaincy of Rio Grande found in the Arquivo Ultramarino de Lisboa documents, catalogued by Lopes (2000), professor at the UFRN History Department, as part of the Projeto Barão do Rio Branco; b) Texts from the 19 th century: again we use the corpus organized by Lopes (2000).The other texts, both from the first half as well as all from the second were administrative letters (ofícios) located in the IHGRN, preserved in eight filing boxes, each one with sub-folders organized by date.The first administrative letter of our corpus dates from 1812, and the latest from 1889.These documents had not been photographed, as had those from the 18 th century.Consequently, field work was needed in order to view the collection, select the letters and photograph them so as to do the proper transcription; c) Texts from the 20 th century: transcribed from documents photographed in the Administration Department of the Municipal Archives of Currais Novos (one of the cities of Rio Grande do Norte), during a field trip made by the PHPB-RN team to the city.
Initially, we selected documents and subsequently transcribed them.In the case of texts from the IHGRN, we photographed more letters than what was required, so we had to make choices.As for the letters from the 18 th and 20 th centuries, we had access to a vast database which had been already photographed, so our task was only selecting them according to criteria of readability.
The official letters, once edited, totaled 107 documents.Each letter is represented by C next to its number, ranging from C 1, the first letter from the 18 th century, dated 1713; to C 107, the last letter from the first half of the 20 th , dated 1931.Each half is symbolized by the century in Roman numbers followed by a full stop and either the number 1, indicating the first half, or 2, indicating the second one.The letters are thus distributed as follows: For the transcription, we followed the standards proposed by the PHPB.The transcription model is available in Mattos e Silva (2001) and it is entitled Normas para transcrição de documentos manuscritos para a história do português brasileiro.In 2010, they were passed on to all regional project teams, among which to the PHPB-RN branch, with some updates, entitled Normas de transcrição de documentos manuscritos e impressos.The model offers guidelines towards a semi-diplomatic editing of documents, what Spina (1977: 79) also called interpretive-diplomatic, since vai mais longe na interpretação do original, pois já representa uma tentativa de melhoramento do texto, com a divisão das palavras, o desdobramento das abreviaturas (trazendo as letras, que não figuram no original, colocadas entre parênteses) e às vezes até com pontuação (Spina 1977: 79).
In the case of the criteria proposed by the PHPB, there is no inclusion of punctuation and the words that are joined because of some linguistic reason, such as the clitic nature of some of the terms, are kept together.However, those joined due to the unique morphology of the manuscript letters (owing to their usually elongated lines, they reach the next word by their ornaments) must be separated.The transcription of the corpus was an exercise of significant importance, especially because, by opening a sort of Pandora's box, we could discovery manuscripts, corroded by dust and by the time that shed light on the roads that language and texts traveled from the 18 th to the 20 th century.

AN ANALYTICAL MICROFILM: THE DISCOURSE TRADITION
The main activity that allowed us to profile the official letters of Rio Grande do Norte, regardless of the route taken for this description, was a careful reading of each of them.During this inevitable and fundamental task, we enhanced our perception to understand the letters by two different means: that of the genre (a more generalized viewpoint, because first we had to perceive the genres, and, to a further extent, on a deeper internal level, the macro-structure of those genres) and that of the micro-structures, especially in terms of formulaic expressions.Both, genres with their macro-structures, as well as the micro-structural level, were seen as discourse tradition (henceforth to be named DT), and therefore a brief presentation of this theoretical model is necessary before the moving on to the analytical microfilm.
To Kabatek (2006), the defining feature of DT is the repetition.However, not every repetition is configured as DT, because a DT must possess sign value, i.e. it must represent something that is external to itself, symbolizing it.To develop the notion of repetition, Kabatek (2006)  We see, in the schema, that a situation 1 (which is the discourse constellation) evokes the production of a text 1.Using Kabatek's example (2006: 511), the situation of an encounter on the street evokes a greeting (the text that is a DT).Whenever this discourse constellation occurs again, that is, when this constellation updates itself in another historical moment (a position 2, 3, 4 ...), the same DT will be evoked.We consider important to note that, from the point of view of representation, both conditions (1 and 2) as much as both texts (1 and 2) must be considered the same.They are, however, from the viewpoint of realization, different, since each situation will occur in a chronotopia (that is to say an inseparable junction of time and space, for that reason, unique and unrepeatable) and each text will be enunciated in that chronotopia.That is why it seems better or clearer to us the use of the term update than of repetition.
Probably the concept of DT most often referred to in Brazil is that of Kabatek.The author (Kabatek 2006: 512) defines the DT as «a repetição de um texto ou de uma forma textual ou de uma maneira particular de escrever ou falar que adquire valor de signo próprio».This shows the broad domain of the DT studies: from a speech act (of a question, for example), through a genre (an application, e.g.) to a more abstract level, as the style (the affectionate style, e.g.).
Following Koch's ideas (1997), it is important to note that a DT is part of cultural traditions and, consequently, never grows out of nothing -«nunca surgem ex nihilo» (1997: 17)-and are always in a dynamic relation between conservatism and innovation.Below we show a representative plan that Koch, quoting from Strube (1989apud Koch 1997: 16), uses to comment on these other attributes of DT: The schema shows that the ABCD tradition bequeathed three of its traits to the next one, BCD.That is, BCDE underwent a process of innovation, bringing a new unknown element for its earlier tradition, E, but kept other three that reveal its conservatism.With the passage of time (t), we arrive at an EFGH tradition that, after several successive innovations, preserves not a single element that was found in its original tradition, ABCD.We realize, besides, that a tradition is not born from nothing, but always comes from a previous one.In the case of the first expression, if it did not come from another discourse tradition, it emerged certainly from some pre-existent cultural tradition.

Criteria for the Identification of Genres and their Macro-Structures
In order to operate the identification of genres that shape the official letters, we elected -from the three dimensions (the thematic content, style and compositional construction) identified by Bakhtin (2010: 261-262), by which the genre reflects its purpose-the compositional construction, i.e., an inner macro-structural organization.We added another criterion to this formal one, which is its function/communicative intention, usually taken as definitive principle to distinguish one genre from another.
In this sense, we attached communicative intention and structure, attributing to them extra-linguistic features, namely, the nature of the sender and the addressee, and on a secondary level, the relationship between them is established.This was done also because we decided to preserve certain limits that are presupposed in the field of textual genres and Diplomatics.In practical terms, the effects of such choice can be explained in the following situation.Basically, the letter, the administrative letter (ofício) and royal charter genres have the same communicative intention, which is to obtain facts -in the case of our official letters, facts relating to Rio Grande do Norte-that need to be known to the public administration.Nonetheless their functional specificities (ordering, sending, asking, etc.), this more general objective to report facts to other public bodies possibly can be considered the common denominator of these three genres.However, the fact of being addressed to the king, being written between members of the government other than the royal authority and being sent by the king, is what makes these documents what they are, respectively: letter, administrative letter and royal charter.
Despite the fact that there is no record of the term royal charter (carta régia) in any document from our corpus, and instead an expression akin to letter of His Majesty is always used, we wanted to keep the distinctive label of royal charter because it is a Diplomatics category (vid. Belloto 2002: 54), and even to follow a symmetry in discriminating letters, in line with the distinction letter (carta) and administrative letter (ofício): nature of the sender and/or addressee.
As for the structure, we were based on the study of the form made in Diplomatics, a science which has as its object «a estrutura formal do documento» (Belloto 2002: 18) and presents, thus, among other functions, the function of classifying «classes ou grupos documentais, baseada na relação da entidade produtora e estruturas externa e interna do documento» (Berwanger, Leal 2008: 27).
As for the macro-structure of the official texts, Spina (1977: 53) summarizes it as two parts that intrinsically organize a document: an internal one, called text, that constitutes the body of the document and contains the fact registered; and an exterior part, the protocol, formed by an exordium (opening), also called protocol, and a peroration (conclusion), the eschatocol.Belloto (2002) introduces the same idea, systematized, however, in a three-part classification: the initial protocol or just protocol (Spina's exordium), the text, and the eschatocol or final protocol (Spina's peroration).For the analysis of macro-structures of the genres of our official letters, we chose the word protocol for this pre-textual section; text to the main part of the document; and eschatocol for the post-textual fragment.
Each of these parts has a number of components.In a complete model there would be invocatio, intitulatio, inscriptio and salutatio in the protocol; prologus or exordium, notificatio or promulgatio, narratio, dispositio, sanctio or minatio and valoratio or corroboratio in the text section; and subscriptio, datatio and apprecatio in the eschatocol3 .Not all these components appear in official letters of Rio Grande do Norte and they do not include the valediction, which we will also name in Latin as clausula.It is important to remember that, for better visualization, tables were created in Melo ( 2012), that record all possibilities of realization of each of these three analytical partitions of the document.Below, we reproduce one of these tables, where the possibilities of arrangement of one of the parts of the official letters in study, the protocol, are registered, without taking into account the nature of inscriptio and subscriptio or the order in which they have their elements:

THE FIRST ANALYTICAL SAMPLING: THE CASE OF THE GENRE LETTER
The tradition of the letter is formally very similar to that of the administrative letter (ofício), but it is evoked by another discourse constellation, which is the need to communicate to the King any matter relating to public administration, and therefore they all have as inscriptio only the appellation Senhor.Such an appellation, as stated by Fonseca (2003: 150), is addressed exclusively to royal authority.Due to the fixity of its address, the King, it is always an official letter directed upwards, that is to say, written by someone hierarchically inferior to the person to whom the document is addressed.The letter is responsible for 9,3% of the corpus 4 The macro-structure presented in 90% of letters (9 out of 10 letters) is illustrated below by the C 3 model.and, like the administrative letter (ofício), it presents a plurality of communicative functions represented, in the documents in question, by communicating any subject (C 1), committing to the fulfillment of any royal order (C 3), submitting a document (C 5), arguing in the defense of someone and proposing an attitude (C 15), and requesting (C 16).Thus, their common denominator lies in socio-functional elements: the context of production and the nature of the person to whom the document is addressed.Fonseca's work (2003) to describe our corpus, consisting of nine textual genres, namely: 10 letters (9,3% of the corpus), one royal charter (0,9%), 2 certificates («certificado») (1,8%), 3 consultations (3,8%) , 5 letters of registration (4,6%), 9 applications («requerimento») (8,4%), 1 provision («provisão») (0,9%), 75 administrative letters («ofício») (70%) and 1 warning («aviso») (0,9%).In this analytical sampling, «The case of the letter», we treat that particular genre (quite similar to the royal charter and administrative letter [«ofício»] from the viewpoint of its structural organization, but different, however, because of the role of its partners: the royal charter is drawn up by the king, the letter is addressed to the king and the «ofício» does not involve the king as addresser nor addressee). 5 The angles indicate insertions by another hand.In this case of C 3, that insertion was written in the upper left margin of the document.This convention, like all other notations used throughout this work, is according to the rules of transcription mentioned above (cfr.Mattos e Silva 2001).The other letter C 15 (the remaining 10%), diverges from this model because it presents, in the text section, a prologus before the narratio, and its datatio is composed only by the topical date.Unlike the most common models of the administrative letter (ofício), the letter (carta) does not reveal an inscriptio in eschatocol, certainly because of the important character of the King, who is one and unique.Neither Manual (Mendes e Forster 2002) nor Medeiros (2006) point out the letter as a genre of official writing anymore.Medeiros (2006: 215) defines it only by the terms found in Brazilian dictionaries, that is to say, this species has become generally conceptualized, losing thereby the particularity it had in the 18 th century, the only period when it occurs in the corpus.

THE SECOND ANALYTICAL SAMPLING: THE CASE OF THE GENRE LETTER OF REGISTRATION
The letters of registration (cartas de registro) represent 4,6% of the official letters of Rio Grande do Norte, and are restrict, in our corpus, to the first half of the 18 th century.They are evoked neither by a communicative situation, as happened to administrative letters (ofícios) and letters (cartas), nor by co-text, as will be the case of the formulaic expressions, but by the existence of another official letter, of which the letter of registration works as a kind of an attachment to summarize the subject of the main document.We can say, then, that the letter of registration is a DT evoked by the existence of another DT.To exemplify the situation, we bring out at first the letter C 5, written by the Ombudsman of the Royal Treasury of Rio Grande do Norte, José Barbosa Leal, to the King João v.In this letter, Barbosa tells the King about the books needed for the collection of the Royal Treasury.Evoked by the existence of this letter C 5, the letter of registration C 6 appears, which we take as an example, since it characterizes the macrostructure of the genre in all five letters of registration that our corpus holds: Table 3. Model of the macro-structure of the genre «letter of registration» according to C 6.
We notice, by the example, that this letter of registration C 6 announces the content of the letter C 5. Although strongly dependent on another text, this DT (the genre letter of registration) showed us certain formal independence malleability, as it was written on a separate page and it had its own datatio, which is the same as the letter to which it binds, and it also brought the indication of the author of the main letter, always preceded by the preposition de.We decided, by metonymy logic, to treat this component under the name subscriptio.The gist of the text (C 6), its dispositio, eventually is the account of the main letter (C 5) content.
If, at first, the letter of registration looks somehow unimportant, actually it gives us a clue to understand the origin of one of the components of three genres of the official writing that the Manual da Presidência da República (Mendes e Forster 2002: 11) called official pattern (padrão oficial): the warning (aviso), the administrative letter (ofício) and the memorandum.This same communicative intention of recording or announcing the content of a more important document, which was fulfilled with «certain formal independence malleability» in the letters of registration, loses its flexibility, stiffens and becomes one of the sections (the subject section) that the warning, the administrative letter and the memorandum must contain, located in the pre-textual part of these genres, i.e., in the protocol.In this way, we can see, over the course of time, the change from a macrostructural TD to a new tradition, this time a microstructural one, now turned into a component that integrates the macrostructures of the three aforementioned document species.

CRITERIA FOR IDENTIFICATION OF MICRO-STRUCTURES SUCH AS «TEXTUAL FORMULAE»
For the analysis of textual formulae -which are the micro-structures dealt with by Melo (2012)-, we departed from the same starting point of the study of textual genres: the reading of the texts.As we were reading them, we listed the textual components that seemed to possess characteristics of a new sign, that is, expressions that received a new stratum over the one signifier -one meaning denotative layer, a significant meaning to a new layer in which the signifier started to gain a new meaning; in other words, we listed the possible DT.The following examples illustrate the procedure: in the first of the letters that were read, C 1, an expression appeared, shortly after the inscriptio: «Foi VMagestade servido ordernarme por carta | de coatro de fevereiro deste anno, de [inint.]a rezão | que tinha»; and in C 77, starting the narratio, there is «Inteirado pelo officio de V. Ex. a de | hoje datado».Both the expressions «Foi VMagestade servido ordernarme» and «Inteirado pelo officio de V. Ex. a de | hoje datado» looked like possible DT to us.In the first case, we thought of a possible textual tradition evoked as a mention of the King's prior order starting the narratio and, in the second case, as a strategy to start the narratio with a participle, from which the information of dispositio derives.
Meanwhile, we added other forms -identical, similar or different, as long as responsible for the expression of a particular meaning (or, in other words, according to the terms of the DT, as long as evoked by the same discourse constellation)-to those expressions already selected.This happened with the expression «Foi VMagestade servido ordernarme», which reappears in the same way in C 5; and by the «V.Magestade foi servido ordenarme» and «foy V. Magestade servido determi= | nar» in C 3 and C 19 respectively.Some impressions, however, were not matched to others from our official letters and were therefore excluded from the chart, such as the use of the participle of C 77.We also counted as DT only the elements that were repeated (i.e., which appeared in more than one letter), for they were the only ones allowed -empirically (not speculative) and based on an examination limited to our corpus (without resorting to other texts to check the likelihood of repetition by the evocation of the same discourse constellation)to prove the value of a new sign.

THE THIRD ANALYTICAL SAMPLING: SOME APPLICATION EXPRESSIONS
The application is also one of the documentary species that has formulaic expressions as its characterizing stamps.We reserve this section only for this genre since there are two formulas that we will analyze as well as due to the development of the analysis.
The first of them is symbolized by the following notation verbum dicendi + identification of the applicant + narratio.Usually, together with the identification of the supplicant was their intitulatio.Some examples of how this expression has been updated in the corpus are «Diz João deTeyve Barreto eMenezes fidalgo da caza de VMa | gestade, q VMagestade foy servido», in C 17; «Diz Francisco Paulo Moreira, que o Governador, eCappitam Ge-| neral», in C 25; and «Disem Saboya Albuquerque e Ciª con-| tractantes da construcçaõ do açude "Garga-| lheiras" no Municipio do Acary, d'este Estado, | que tendo de transportar» in C 87.This DT is evoked by the beginning of the protocol or after inscriptio when it is presented.
The majority of the occurrences, in 87,5% of the updates (7 of 8 cases) of this expression, is syntactically integrated to the narratio by means of the subordinating conjunction que.Only in C 27, the update breaks the syntactic unit, as showed below: (1) A Vossa Magestade se queixa Manoel Francizco hom ẽ branco esoldado | condestavel da Fortaleza desta cidade do Rio grande do norte, deJoaquim | Felix e Lima Capitam mor, e Governador desta mesma, comarca dada Para= | hýba donorte, e a rezaõ desua queixa aespoein pelos itenis seguintes.
There is not a verbum dicendi in this case (if there were it would be queixar-se) and the speech is retextualized by «e a rezaõ desua queixa aespoein pelos itenis seguintes», followed by a colon to introduce a large appositive, which was developed in four paragraphs with the function of narratio.
The second characteristic application expression is evoked within the eschatocol, representing either the only element in the eschatocol (in 66,6% of the applications) or the first component of it (in 100% of cases): ERM and its variants as a kind of clausula, signing off the letter.Flexor (2008), in her dictionary of abbreviations, only records «E receberá mercê» as the possible meaning of the formula.We find, however, more appropriate, considering the nature of the official letter (an application), the development of the acronym given by Acioli (1994) and Belloto (2002): «Espera receber mercê».
This expression, in general, has a high degree of fixity, which means it is quite crystallized.«Espera» and «receber» were written as acronym in 88,8% of cases, representing 8 of 9 requirements.The minimum change that happened in these 88,8% focuses only on «mercê», which follows the technique of acronym in C 18 and C 25; which is abbreviated as syncope, keeping the final letter «e» of the word in C 17 and C 27; and also by syncope, but preserving the last syllable «ce» in C 28 , C 30 , C 54 and C 74.
Stating that this expression was, in these 8 requirements, updated by a sequence of ABCD traits, being A «espera», B «receber», C «mercê» and D an alleged trait indicating their syntactical independence and formal rigidity, now, we examine the only case that differs from that use, the 11,1% which C 87 is responsible for: «os requerintes esperam que V. Excelen | cia se digne deferir a sua petiçaõ», snippet highlighted in italics below: This C 87 is the only application that is part of the first half of the 20 th century, composed by mostly administrative letters.On the track of ABC traits that gave body to the ERM expression, in the course of the 18 th and 19 th centuries, we could see in this 1915 update the permanence only of the verb «esperar», that is, of the trait A, to which other new traits were added, like the mention of the target of the application, «que V. Excelen | cia», trait E, associated to the main ideas of «deferir» and «petição», that we could consider alleged traits F and G. Apart from it, the meaning expressed is not drawn anymore as a rigid expression (such as ERM), but, on the contrary, it is syntactically integrated to the end of the dispositio.Thus, it loses trace D, generating an AEFG expression.This mechanism reveals the interaction between tradition and innovation, a potential process that any DT is subject to and that was specifically seen in this particular case.It is true that, by the way it is presented in C 87, it has more reasons to be categorized as something different from a formulaic expression than not, but we decided to preserve this classification to highlight this transmutation according to the same perspective we are working with.
Observing today's application model proposed by Medeiros (2006), we notice that the first of the two textual formulas studied in this topic (verbum dicendi + identification of the applicant + narratio) is partly dissolved since verbum dicendi stops being the heading, yet the allusion to the applicant (who is the sender of the letter) remains as a third person, as we can verify in this model given by Medeiros (2006: 323): «Carlos Alberto, que atualmente ocupa o cargo de Servente, nível 4, com exercício no Departamento de Ensino de 2.º Grau, requer a V. Sa. se digne conceder-lhe Auxílio-Doença, nos termos».As for the second formula, the closing ERM, the current expressions recommended are «Nestes termos pede deferimento», «Espera deferimento», «Aguarda deferimento», «Pede deferimento» or «Termos em que pede deferimento» and it has to be written with capital letters, «PEDE DEFERIMENTO», for instance (Medeiros 2006: 322).
One of the possibilities given by Medeiros (2006), «Espera deferimento», reuses exactly the same trait A that arose in 1734 C 17 from our corpus, espera, and adds a new trait, deferimento, derivative from F, in which the same sense of acceptance remains.Nevertheless, it is not in fact a brand new element, since in the course of our official letters, in 1915 C 87, the noun deferimento had already appeared.Therefore what we seem to face in today's recommended formulas is the result of a mix of traits that were used in different historical moments, but that were evoked by the same discourse constellation, which is, in this case, the opening of the eschatocol of an application.
The option «Aguarda deferimento» expresses the same operation, but with the verb aguardar, which, yet modified in form, closes up semantically to esperar.The other possibilities do not keep trace of primal traits from ERM, but «pede deferimento» remembers the noun petição in C 87 and most of them repeat the word deferimento which, as just aforementioned, was also used in this crucial 1915 letter.All these interconnections shown make us see it as a mathematical curious chess, which seems to remind us that change (innovation) is closely attached to the tradition and that tradition (permanence), in turn, only survives due to the change.

THE FORTH ANALYTICAL SAMPLING: THE CASE OF CLAUSULA 6
One of the most recurrent formulaic expressions in the letters is represented by the notation «Deus guarde» («muitos anos») + the beneficiary's name (+ «para amparo dos vassalos» or a similar idea, in some letters from the 18 th century).It is evoked by the opening of the final protocol of the letters and acts as clausula.This pattern occurs in 57 (65,5%) of 87 letters that have clausula or 53,2% of the total of the official letters.The update of that fixed expression changes with the passage of time.The first time it occurs is in C 1 as «A Real Pessoade | VMagestade guarde Deos como todos seus vassallos avemos mister», which is updated in C 3 by the recurrence of the exact same construction.The arrangement of the elements of this DT varies, sometimes the verb coming before, as in «Guarde Deos aVMagestade | muitos annos», in C 5, or by adding vows to God's care, as in C 15: «Deus a VMagestade guarde pormuitos efelises | annos, para bem, econsolaçaõ de seus vassa-| los fieis».
In these instances, as noted, they are letters (cartas) because they are being sent to the King.That same DT, however, is also updated in administrative letters (ofícios), for which this DT works almost as a badge throughout the 19 th century.It still appears as clausula in the only warning (aviso) of the corpus, always evoked by the discourse constellation of completion, marked diplomatically by the eschatocol section.The almost Baroque strategies of the first occurrences will give way to simpler variants, up to culminating in the laconic «Deus guarde aVossa Excelência», the variant that most often updates the TD in question, representing 43 of the 57 appearances (75,4%).This formula that had already been performed by four elements -«Deus Guarde», as element A; «muitos anos», B; the name of the benefitted person, C; and «para amparo dos vassalos» or a similar expression as D-generating therefore an ABCD discourse tradition, it is updated throughout the 19 th century as concisely as AC.
There is yet another strategy, in the 18 th century, that can act as clausula, expressed by the notation reference to God + verb + benefitted person recipient of divine grace + good wishes + for subject's and, in particular, sender's protection.It shows in the disposition of C 32, as «o alticimo comceda aV.Ex. a ejuntamente acom | panhada de todas as felicidades dignas de sua | bem conhecida ver tude p. a meu emparo e detodos | os que vivem debaixo da portecam deV.Ex. a », with the clausula filled by «aVEx.a q' Deus guarde muitos annos».The letter C 26, however, does not require this valediction, so that we can understand «DEUS comceda a V.Ex.ª aquelas felecidades que todos | os seus humildes subditos lhedesejaõ, e[cu] muito emparticular p. a meu amparo» as a new closing strategy, especially for its location, the typical place of this introductory component of the eschatocol: at the end of the letter, immediately before datatio.
Hence we see how, even inside a synchronic stage, a discourse constellation can evoke different DTs or, to be more exact, a DT with different expressions, formally different, but functionally equivalent; and how the same expression can accumulate signifying layers in a given text, representing a DT, as it happens in C 26, or being merely an expression of its denotative sense, as in C 32.
Mainly from the first half of the 20 th century, the same discourse constellation that evoked «Deus guarde a» and its variants, besides of «DEUS comceda a V.Ex.ª aquelas felicidades» in C 26, it starts to evoke other TDs.One of them is the textual formula «Saúde e Fraternidade», representing 5,7% of the clausulae of official letters and 4,6% of the total of these letters; the other is noted by (modifier +) + the expression «saudações», recurrent in 18,3% of the clausulae and present in 14,9% of the corpus.With these new DT, the two remaining elements, AC, disappear and there are new features, conforming a new DT: an E, «Saúde e Fraternidade», representing a DT; an F, the possibility of a modifier prefixing «Saudações», and G, «Saudações» itself, resulting, for example, in a tradition of the sort FG, like «Cordeaes saudações» in C 88, or just G, «Saudações», in C 89.This no longer reveals the dynamic variation through which the DT goes, but the process of change.
Two considerations can be drawn from that extract.The first concerns a fortunate find in the last letters of the corpus.We relied on chance, because our corpus was randomly composed from the point of view of its content, since it was readability that conducted the selection of documents.For the DT we represented as (modifier +) the expression «Saudações», emerged, as the clausula of C 102, the form «Respeitosas saudações» and C 107, the official letter that finishes our corpus, updates the same DT by the formula «Attenciosas saudações».It is inevitable to compare these two clausulae appearing in 1930 and 1931, respectively, with the recommended valedictions «Respeitosamente» and «Atenciosamente» established by the Manual de Redação da Presidência da República.The element F, as we claim above, i.e. the possibility of a modifier prefixing «Saudações», can be seen as a source-feature, from which the present famous and familiar «respeitosamente» e «atenciosamente» emanated.We could even suppose the process of this change: «respeitosas» and «atenciosiosas» lost their dependence as adjectival satellites of the clausula, acquired adverbial status on a certain moment in the 20 th century and became autonomous fixed formulae, emancipated in official letters until they were valued in a specific Brazilian social-political situation as the ideal forms to fill out the closing component in the letters of public administration.
The second consideration concerns society and it is an overview of all the other clausulae.We have observed that the change from the DT «Deus guarde a» to «saudações» or «saúde e fraternidade» has a parallel development with the history of Brazil, even if strictly referential.This change happens in the corpus for the first time in 1891 C 84 with «Saúde e Fraternidade».This is the first official letter after 1889, since its antecessor, C 83, dates from 1887.This means such change occured immediately after the proclamation of the Republic, when a secular state is imposed, to the detriment of the monarchy, a form of government which had affinities with Catholicism across Europe and therefore with Portugal, from which Brazil inherits it.And the reduction of possibilities suggested (by the Manual de Redação da Presidência da República) harmonizes with an increasingly objective and impartial culture, especially, at least in theory, on a governmental level.These changes let us see another facet of DT, which relates to its connection with cultural traditions.According to Koch (1997), a discourse tradition is never born ex nihilo, always arising from another DT or from a cultural tradition.In this case, we have new DTs, «Saudações» and «Saúde e Fraternidade», that derived not only from a discourse constellation of a textual nature, namely the need for a clausula introducing the eschatocol, but also from new social and political traditions that had been emerging in Brazil.
Other changes in traditional ways of writing letters in the public administration are evidenced by the abolition of certain expressions in official letters from today, like «cumpre-me», used in 9,3% of the letters, or «tenho a honra de», written in 8,4% of them.Both forms are evoked to introduce dispositio, just after the prologus, if it occurs.These expressions represent one of the diplomatic components, notificatio or promulgatio, which were obliterated in the analytic partition we proceed to.The Manual da Presidência, on those terms, states the following: «Evite o uso das formas: "Tenho a honra de", "Tenho o prazer de", "Cumpre-me informar que", empregue a forma direta» (Brasil 2002: 12).

FINAL CONSIDERATIONS
Look in thy glass, and tell the face thou viewest, Now is the time that face should form another (Shakespeare 1970(Shakespeare : 1043)).
At the end of our presentation, we believe we have demonstrated the dynamic interaction between tradition and innovation, verified by studying the preservationsvariations -changes that some genres and some formulaic expressions, all of them regarded as DT in this paper, went through in the 18 th , 19 th and 20 th centuries.We briefly review some of them below:  We see the texts (genres) and their textual forms (the micro-structures, especially the formulaic expressions), or better yet, we see the DTs that make up the official letters of Rio Grande do Norte as the common denominator for the texts of our corpus, because these TDs weave a large network of structural, formal possibilities, whose updates, repeated over time, end up functioning as a defining characteristic of the letters.The existence -although, in some cases, potential, given, for example, the formulae do not appear in all genres-of letters, administrative letters (ofícios), clausulae such as «Deus salve a» or «Saudações» and other structures, all this represents linguistic/textual criteria, which, along with extra-linguistic criteria, allow for the categorization of such heterogeneous texts under a single term: official letters.Moreover, by this reasoning we find it less risky to think about a form of bureaucratic language or, more accurately speaking, a virtual set of forms of bureaucratic language.

Letter of registration
The textual changes may reflect or allow us to see changes which society goes through.If this was evident in the change of valediction from «Deus guarde a» to «Saúde e Fraternidade» in the post-Proclamation of the Republic documents, we suspect that perhaps changes beyond the lexical level, such as those in the macro-structure of genres, might also reveal, under closer scrutiny, that social influence.
Our analyses and the constitution of our corpus -now belonging to the PHPB-RNcan contribute to the study of a category (official documents) about which Silveira (2007) says that studies are almost inexistent.The author states: «esse tipo de produção escrita tem sido marginalizado tanto na escola básica como na universidade» (2007: 1) and only mentions two academic papers dealing with administrative genres, one of which, she claims, does exactly deal with these genres, but only discusses the bureaucratic slang.
However, it should be remembered that, for the history of language, these texts played a crucial role, since they marked, especially after the 70's, a new way of doing Diachrony based no more on literary texts, but on notarial writings, providing us with data that if, on the one hand, sin by routinization of some formulas, on the other hand have a great advantage: they are free from the language stylization.Nevertheless, it is precisely in the «sin» of the routinization of certain expressions that part of the importance of a work like ours is: identifying which structures are routinized (in this case, as DTs) allows us to separate what is in the domain of the text from what is in the domain of the language.A great teacher of History of Language in Mexico, dealing with graphic trends, gave us the following equation: orthography = language + tradition.To approach the «reality of the language», he said, the mathematical inversion: language = orthography -tradition.The same logic could be applied to texts: text = language + tradition.Thus, language would be reached much more truthfully if we deduct texts from their traditions.When we recognize the elements that are textual traditions (Diplomatic has a strong potential to be a very important ally in search for such distinction), or, in other words, when we know the history of the texts, we will have a great filter in hand to do, more accurately, studies on history of language.
We hope our work, with its textual character and from a diachronic perspective, contributes to this not-so-prolific investigative field in Academia, namely, the study of a) XVIII.1 2 b) XVIII.2 -15 letters, from C 20 to C 34; -19 letters, from C 1 to C 19; c) XIX.1 -23 letters, from C 35 to C 57; d) XIX.2 -29 letters, from C 58 to C 86; e) XX.1 -21 letters, from C 87 to C 107.
Summary of some of the innovation/tradition processes examined along this paper

Table 1 .
Different arrangements of the protocol in the official letters of Rio Grande do Norte