Next Article in Journal
Tamil Speakers in Switzerland: An Intergenerational and Typological Perspective
Next Article in Special Issue
Different Degrees of Analyzability—The Case of the Spanish Verbal Periphrasis [Tardar en + Infinitive]
Previous Article in Journal
The Geography of Meaning: Investigating Semantic Differences Across German Dialects
Previous Article in Special Issue
Psych Light Verb Constructions in Old Catalan: Patterns and Contrasts with Present-Day Catalan
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Back-and-Forth of assim que in the History of Portuguese

by
Aroldo Leal de Andrade
* and
Glayson Martins Oliveira
Faculdade de Letras, Federal University of Minas Gerais (UFMG), Belo Horizonte 31270-901, MG, Brazil
*
Author to whom correspondence should be addressed.
Languages 2026, 11(3), 57; https://doi.org/10.3390/languages11030057
Submission received: 26 November 2025 / Revised: 27 February 2026 / Accepted: 10 March 2026 / Published: 16 March 2026

Abstract

This paper investigates the diachronic development of the sequence assim que (lit. ‘such that’) in the history of Portuguese, with a comparative perspective on the parallel construction así que in Spanish. A corpus-based approach was employed, analyzing approximately 1800 tokens from the Corpus do Português: Historical Genres, spanning eight centuries of written European Portuguese. The results show that assim que remained highly analyzable until the end of the Old Portuguese period, with the adverb assim often followed by a complement or result clause. The grammaticalization of assim que appears to have evolved partly independently from standalone assim. While Portuguese and Spanish share many uses of the construction, modern European Portuguese has diverged, with assim que losing its status as a discourse marker. This change is best explained by the frequent use of cleft constructions (e.g., foi assim que), which reanalyzed que as a subordinating connector, undoing the earlier single-unit interpretation. These findings suggest that even deeply entrenched grammaticalization processes may undergo retraction when the semantic analyzability of component elements allows it.

1. Introduction

The aim of this paper is to describe the diachronic developments associated with the complex connector assim que (also attested as assi que and other orthographic variants in older texts) in the history of Portuguese, which has followed a different path from that of Spanish así que. Additionally, the paper discusses how the analyzability of complex items may be maintained even in the context of highly advanced grammaticalization processes.
As can be seen in examples (1) and (2), whereas in modern Portuguese the temporal meaning is most typically associated with this connector, in modern Spanish the consecutive use is more frequent.
(1)
É provável que o criminoso mude o e-mail e número de telefone associados a uma conta assim que consegue o acesso. (Público, 23 February 2023)
‘It is likely that the criminal will change the email and phone number associated with an account as soon as they gain access.’
(2)
No había aire acondicionado, así que había que buscarse la vida.
(El País, 14 August 2025)
‘There was no air conditioning, so you had to make do on your own.’
This analysis contributes to current research on the grammaticalization of connectors in Portuguese. Previous studies have either examined the word assim in isolation or compared the expression assim que with other adverbial locutions (e.g., Souza, 2009; Santos & Silva, 2019), without focusing on the diachrony of assim que and its development into modern European Portuguese. In Spanish, there are few studies addressing the diachrony of así que; however, those that do examine this structure have been highly enlightening for our purposes.
Due to space constraints, we focus primarily on the trajectory observed in the development of Portuguese, supplemented by reflections on the corresponding change in Spanish based on previously published studies. To this end, we conducted a corpus-based investigation, selecting representative texts from written European Portuguese spanning the 13th to the 20th centuries. Our study also examines cleft constructions involving sequences such as é assim que and assim é que, which occur throughout the history of Portuguese but have generally not been considered in diachronic studies of the grammaticalization of assim.
The main aim of the paper is to show that, as these constructions became increasingly frequent, they contributed to both semantic change (encoding temporal immediacy rather than consequence) and, indirectly, to structural change (leading to the loss of its use as a discourse marker) in assim que. This development is attributed to the increased prominence of the anaphoric relation associated with the lexical entry of assim, which ceased to be merely inferential and became a central, profiled component of the construction’s semantic representation.
For the analysis, we adopt a primarily descriptive approach, using Rhetorical Structure Theory (Mann & Thompson, 1988) as a framework for data classification. The study is grounded in the constructional approach to grammaticalization (Traugott & Trousdale, 2013), within a broader cognitive framework in which linguistic meaning is modeled in terms of conceptual structure and construal.
The paper is organized as follows: Section 2 reviews the literature on assim (que) and así (que) in historical Portuguese and Spanish. Section 3 outlines the theoretical framework relating grammaticalization to analyzability. Section 4 describes the methodology for data retrieval and classification. Section 5 presents the results. Section 6 discusses issues arising from the results, and Section 7 provides the concluding remarks.

2. Previous Studies

2.1. Assim/Assim que in Portuguese

According to Longhin-Thomazi (2006), assim originated from the Latin expression ad sīc (‘to/for this’, ‘in this way’). It is an early form rooted in Vulgar Latin adsi (example from Leges et Consuetudines, yr. 1188–1230, as cited in Machado, 1967, p. 335).
(3)Et si adsinon fecerit,perhicadat…
andifthusnothas.donebythesemay.fall.3sg
‘And if he does not do so, let him fall because of these…’
Following the developments described in Williams (1962), the change from ad sīc to assi involves regressive assimilation of /d/ to /s/ before /s/ (/ds/ > /ss/), followed by the loss of the final consonant /k/, yielding assi. In Classical Portuguese, the form underwent final nasalization, resulting in assim.
In Old Portuguese, assi primarily expressed manner deixis (‘in this way’), referring to a specific event or circumstance in the discourse or situational context, as in this example from Cantigas de Santa Maria (Cantiga 45):
(4)
mais los ángeos chegaron/dizendo: “Estad’, estade!/ca non quér Santa María/que a vós assí levedes.” (Afonso X, 13th century/2021, p. 102)
‘But the angels arrived/saying: “Wait, wait!/because Saint Mary does not wish/you to take her (the soul) away in this way.”’
Over time, however, the form developed increasingly abstract relational functions, acquiring notable multifunctionality. It came to function as an anaphor (5), a cataphor (6), and a non-prototypical coordinating conjunction characterized by both retroactive (anaphoric-like) and projective (cataphoric-like) orientation, yielding a conclusive interpretation (7) (Lopes-Damasio, 2008, p. 110). The following examples are cited in Longhin-Thomazi (2006):1
(5)
e porẽ lhe rogava que lhe outorgasse de lhe chamarẽ rey de Portugal, e que lhe assi o chamasse ẽ suas cartas e privilégios. (Crónica Geral de Espanha, 1344, p. 1776)
‘And for that reason, he asked them to grant him permission to be called King of Portugal, and that he might be addressed as such in their letters and privileges.’
(6)
chorava e nom sabia porquê, disse assi: —Senhora Santa Maria, madre de piedade, socorre-me e nom me leixe ainda morrer. (Demanda do Santo Graal, 1210–1240, p. 1776)
‘He wept and knew not why, and said thus: —Lady Saint Mary, mother of mercy, help me, and do not let me die yet.’
(7)
E quando amanhece as mais das vezes está o ceo todo cuberto de nuuẽs, & assi as mais das manhaãs choue nestas partes… (Historia da Provincia de Santa Cruz, Gandavo, 1576, p. 1777)
‘And when day breaks, most of the time the sky is completely covered with clouds, and therefore it rains on most mornings in these regions…’
Note that the coordinating conjunction e (spelled as “&”) was found until the 16th century and disappeared thereafter. Lopes-Damasio (2008) considers that, at this stage, assim becomes a bona fide coordinating conjunction.
One of the grammaticalization paths followed by assi(m) is: cataphoric adverb → modifier → discourse marker. Accordingly, in Middle Portuguese, assim comes to assume a dual function: anaphoric, specifying a preceding referent, and cataphoric, as in (8), where assi exhibits a focalizing function.2 In modern Brazilian Portuguese, assim may function as a discourse marker that both points to the immediately preceding noun phrase and signals the insertion of new information specifying its content, as in (9), or serves as an indicator of expressive content, as in (10) (examples cited in Lopes-Damasio, 2008):3
(8)
começou o cavallo bullir rijamente coma as maãos e com os pees; e permeando assi rijamente, açertou o canello da ferradura da maão ho teçido dhuma fivalla… (Crónica de Dom Fernando, Lopes, 1436–1443, p. 155)
‘The horse began to thrash its legs and feet vigorously; and, kicking so violently, it struck the shin against a metal buckle…’
(9)
—André, e não tem nenhuma outra história assim que você ficou sabendo que você possa me contar? (Iboruna Corpus, 21st century, p. 192)
‘—André, isn’t there any other story like that one that you heard and could tell me?’
(10)
—ai não de amigo ahh… é muito triste… porque assim o amor pode não durar pra sempre mas pra mim a amizade… é pra sempre… porque você conhece a pessoa… (Iboruna Corpus, 21st century, p. 193)
‘—Oh no, not from a friend, uhm… that’s really sad… because, like, love might not last forever, but for me friendship… it’s forever… because you really get to know the person…’
This development of assi(m) can be explained following Traugott’s (1995) proposal, as it exhibits a range of meanings—from a mitigating element to a syntagmatic-level intensifier—ultimately extending to interactional and subjective levels.
Another grammaticalization path observed for assi(m) is: anaphoric adverb → coordinating conjunction with a basic consecutive meaning, adapting the proposal by Longhin-Thomazi (2006). In the initial stage, the non-prototypical nature of the conjunction assi is corroborated by its co-occurrence with e (‘and’), as exemplified in (7). A stable interpretation as a coordinating conjunction emerges in contexts with a resultative meaning, either purely sequential or indicating cause–consequence, as in (11) and (12), respectively (examples cited in Lopes-Damasio, 2008).
(11)
Fôram os mortos no lugar da briga 25 olandeses, & os mais dos uiuos firidos da espada, & frecha, fogindo com tanto desacordo, que largando os mosquetes, não puxauaõ das espadas. Assi se recolheraõ os nossos carregados dos despojos das armas do inimigo. (Jornada dos Vassalos, Guerreiro, 1625, p. 114)
‘Twenty-five Dutchmen were killed at the site of the fight, and the survivors, wounded by swords and arrows, fled in such disorder that they abandoned their muskets and did not even draw their swords. Thus, our men withdrew, laden with the enemy’s spoils.’
(12)
Quem duuida disso? que como de tres tirando huu, não fica ternario, assi tirando de hua substancia o que é proprio da substancia da cousa, já não fica a mesma cousa. (Desengano dos Perdidos, Leão, 1573, p. 115)
‘Who doubts this? Just as, when one is subtracted from three, it is no longer a triad, so when what is proper to a substance is removed from it, it is no longer the same thing.’
Notwithstanding these long and complex grammaticalization paths of assi(m), it came to participate in a second grammaticalized unit with que, eventually forming a non-compositional construction already in the medieval period, as in (13).4
(13)
Per aazo do cuydado que ei porque som pastor das almas, o meu coraçõ padece os negocios dos homẽs segraes […] E, assy que entendo aquello que padeço e ẽtendo aquello que perdy, faze-sse a mĩ mays graue aquello que soffro. (O Orto do Esposo, 14th–15th c., as cited in Santos & Silva, 2019, p. 62)
‘Because of the care I bear, as I am a shepherd of souls, my heart suffers from the affairs of worldly men […] And thus, as I understand what I endure and realize what I have lost, what I suffer becomes all the more grievous to me.’
In fact, the connector assi(m) que was predominantly employed as a resultative linker throughout the Old, Middle, and Classical Portuguese stages, and it was only in the transition from Classical to Modern Portuguese that the temporal meaning of subsequent action (‘as soon as’) emerged:
(14)
Não foi preciso que os sucessos verificassem aquele vaticínio, porque Vossa Majestade assim que veio ao mundo, só com se mostrar, disse o que havia de ser. (Reflexões sobre a vaidade dos homens, Aires, 18th century, as cited in Santos & Silva, 2019, pp. 63–64)
‘It was not necessary for events to confirm that prophecy, because Your Majesty, as soon as you came into the world, merely by appearing, revealed what you were destined to be.’
Few studies have analyzed the evolution of assi(m) que in Portuguese. Santos and Silva (2019) examine it as a subconstruction of the general [X que] connector, in the same vein as já que (‘given that’) and ainda que (‘although’), among others.5 However, several questions remain open, among which why assi(m) que changed in Modern Portuguese in such a way that it lost its resultative meaning.
An analysis of the parallel construction in Spanish, así que, which followed a very different trajectory, may provide a more complete picture of what Portuguese might have become had a break in its grammaticalization not taken place.

2.2. Así/Así que in Spanish

Unlike in Portuguese, the Spanish literature shows broader interest in the diachrony of así que, whereas the usage of standalone así is more often discussed in terms of its synchronic functions. According to Rodríguez Molina (2014), the basic and earliest use of así is as a deictic element, from which the anaphoric and cataphoric uses subsequently developed, followed by its grammaticalization as an epistemic adverb.6
Martín Zorraquino and Portolés Lázaro (1999), in a study on the use of así as a discourse marker, classify it as a consecutive connector, noting that some of its functions are not yet fully grammaticalized. They distinguish two main usages of así: in the first, the discourse segment containing así echoes or exemplifies a preceding one:
(15)
Ese es uno de los problemas, pero hay otros. Así, el progresivo deterioro de la capacidad lingüística de los españoles. (El dardo en la palabra, Lázaro Carreter, 1997, as cited in Martín Zorraquino & Portolés Lázaro, 1999, p. 4105)
‘That is one of the problems, but there are others. For instance, the progressive deterioration of the linguistic ability of Spaniards.’
In another usage, the connector así, by virtue of its anaphoric capacity as a modal adverb, expresses a circumstance that enables or licenses a given consequence. The specific nature of this circumstance depends on the context and may convey, for instance, a purposive or conditional meaning, the latter illustrated in (16).
(16)
—Padre, si me condenaste a la soledad, ¿por qué no me arrancaste los ojos? Así no lo hubiera visto. (La hija de Rappaccini, Paz, 1983, as cited in Martín Zorraquino & Portolés Lázaro, 1999, p. 4106)
‘—Father, if you condemned me to solitude, why didn’t you tear out my eyes? That way I would not have seen it.’
Regarding así que, the first diachronic analyses devoted to this construction in Spanish were conducted by Narbona Jiménez (1978) and Fuentes Rodríguez (1985). The former proposed the following grammaticalization pathway:
(17)
así used cataphorically to introduce a subordinate clause → así…que with a modal/consecutive or final/consecutive meaning → fully grammaticalized así que (in its continuous form).
Fuentes Rodríguez (1985) offered a more detailed account, addressing the relationship between the connective expressions así que (continuous) and así es que (discontinuous), and outlining a slightly different evolutionary trajectory:
(18)
discontinuous consecutive connector → continuous consecutive connector → discursive connector → continuative (sequential) marker → verbal filler.
It is important to note that both authors agree on the possibility that the sequence así es que may have served as the source of the grammaticalized form así que. Their main justification for this hypothesis is that the former predates the latter. However, this alone does not provide a sufficient explanation, since other complex conjunctions formed by Adverb + que already existed during the Vulgar Latin period (Herman, 1967/2000, p. 92, among others).7
According to Bartol Hernández (1986), another foundational study on the topic, in the 13th century así que exhibited three main connective uses: an illative use (with the verb in the indicative), a final use (with the verb in the subjunctive), and a consecutive use (with the verb in either the subjunctive or the indicative). The distinction between the first and the last lies in their level of operation: the illative use functions at the textual level, as it expresses a logical or inferential conclusion, whereas the consecutive use operates at the clausal level, involving a certain degree of speaker subjectivity. These are illustrated in examples (19) in Old Aragonese and (20) in Old Spanish (examples cited in Guzmán Herrera, 2014, p. 121).
(19)
Encara es otorgado mas a los infançones. que pueden escaliar… Enpero assi que por aquesta compra no desfagan el cabo maso… (Fueros de Aragón, 1247)
‘It is still granted more to the noblemen, who can climb… But so that, with this purchase, they must not undo the end of the hammer…’
(20)
Et aun a esta piedra otra uertud que da muy grand alegria al coraçon assi que el que la cata de mannana andara alegre todo aquel dia (Lapidario, Afonso X, 1253)
‘And moreover, this stone possesses another virtue: it brings great joy to the heart, so that whoever looks at it in the morning will be cheerful throughout the day.’
The emergence of a temporal-succession meaning for así que occurred in the 17th century, similar to what has been documented in the literature on Portuguese, as illustrated in examples (21)–(22), cited in Guzmán Herrera (2014):
(21)
Diós sea conmigo, y cayó muerto en el suelo, y así que cayó lo arrebataron sus enemigos. (Fragmentos de una historia de Nueva Galicia, Tello, 1650, p. 126)
‘God be with me—and he fell dead to the ground, and as soon as he fell, his enemies seized him.’
(22)
Así que calmó un poco su primera cólera, subió a ver a su hija. (Suprema Ley, Gamboa, 1896, p. 217)
As soon as his first anger subsided a little, he went upstairs to see his daughter.’
In addition to the functions previously identified, she observed a progression from textual to clausal and then to discourse-level uses in the Peninsular data—a sequence that is the reverse of the order found in the Mexican samples. Based on synchronic data from the Corpus del Español, the Corpus Sociolingüístico de México, and Google, compared with diachronic data from the first corpus and CREA, she also noted that así que developed its discourse-marker status with a dialogical function in the second half of the twentieth century. This usage signals the existence of shared awareness between speaker and hearer, eliminating the need for an explicit linguistic antecedent (example proposed by Guzmán Herrera, 2014, p. 137):
(23)
Así que utilizas las bolsas de plástico para la basura.
‘—So you use plastic bags for the trash!’
In modern Spanish, así que may function as a connector introducing result, temporal, causal, or even concessive clauses (cf. RAE/ASALE, 2009), although its temporal interpretation became virtually obsolete in spoken data by the 20th century (cf. Guzmán Herrera, 2014).
Hennemann and Schlaak (2015) acknowledge the status of así que as a discourse marker and propose that this use accounts for approximately 10% of its occurrences in the 20th century, according to data from the Corpus del Español. According to these authors, it may fulfill a summarizing function or signal a shift in sentential topic:
(24)
—Estás hecho todo un hombre, no se puede dudar. Ahora tenés un coche muy lindo. Así que el departamento... ¿tenés el departamento en el centro para una oficina? (El Habla Culta de Buenos Aires, 20th century, as cited in Hennemann & Schlaak, 2015, p. 13)
‘You have really become quite the man, there is no doubt about it. Now you have got a very nice car. So, the apartment... you’ve got the apartment downtown for an office?’
Another piece of evidence for new usages related to the pragmatics of así que is provided by Bravo Cladera (2005). In her analysis of an oral corpus of monolingual and bilingual students in Sweden, she observes that así que may occur either within a single turn or at the boundary between turns. The latter pattern, however, is more frequent among bilingual speakers.
A particularly noteworthy use arises when así que appears at the end of a turn, in which case the resulting inference is left for the interlocutor to draw (p. 159).
(25)
[Context: The interlocutor questions the speaker about how he feels about his coming of age]
—Claro, pero no se siente nada especial, ¿no? Es un año no más... Es un año solamente entre diecisiete y dieciocho... Así que...
‘—Sure, but you don’t feel anything special, right? It’s just a year... only a year between seventeen and eighteen... so...’
The results for assi(m)/así que in Portuguese and Spanish reveal a clear contrast: in Spanish, several innovative pragmatic uses have emerged, including the use of así que as a discourse marker with (inter)subjective interpretation, while such developments have likely not occurred to the same extent in Portuguese. This asymmetry appears to stem from the preservation of the consecutive reading in the latter, which remained highly productive in both languages until the 16th century. Their typological and areal proximity naturally explains this shared diachronic background.
We put forward the following hypothesis to explain why this change occurred in Portuguese: clauses containing the sequence assim que began to occur with high frequency in cleft constructions (e.g., foi assim que ‘it was thus that’). This syntactic development, we argue, prompted a semantic reorganization within the system of adverbial connectors, ultimately restricting the functional expansion of assim que in Portuguese and preventing the emergence of the pragmatic values later observed in Spanish.
Section 3 outlines the theoretical framework underpinning our analysis and introduces the key concepts that guide the study.

3. Grammaticalization, Pragmaticalization and Analyzability

As is well established in the literature, grammaticalization refers to the diachronic process whereby lexical items or relatively independent expressions develop into grammatical elements, such as affixes, clitics, or function words. In their seminal account, Hopper and Traugott (2003) characterize grammaticalization as a predominantly unidirectional process involving semantic bleaching, phonetic erosion, and increased syntactic dependency, underscoring the shift from lexical to grammatical meaning.
Within a constructional perspective, however, Traugott and Trousdale (2013) reconceptualize grammaticalization as a type of constructional change, arguing that change affects not merely individual lexical items, but constructions understood as conventionalized form–meaning pairings. From this standpoint, grammaticalization constitutes one subtype of constructionalization within a broader network of related constructions, with particular emphasis on contextual modulation, increasing schematicity, and the gradual reorganization of the language’s constructional inventory.
Unlike more traditional accounts, Traugott (1995) acknowledges the possibility of semantic enrichment, that is, the emergence of new meaning components that run counter to the predictions of semantic bleaching, and therefore merit analytical attention. In this respect, it is crucial to distinguish between conventionalized pragmatic meaning and contextually inferred meaning, arising from the specification of conversational implicatures. Grammaticalization can thus be viewed as the mechanism through which such implicatures become conventionalized, typically via metaphorical and/or metonymic processes.
Wischer (2000) proposes a distinction between two principal types of grammaticalization. Type I involves a shift toward morphological expression, whereas Type II concerns developments in the domain of discourse organization. The latter trajectory is frequently labeled pragmaticalization. Within functional–cognitive frameworks (e.g., Givón, 1979, 1995; Langacker, 1987, 1991), pragmaticalization is conceived as a gradual evolution from discourse-anchored structures toward more grammaticalized and less compositionally transparent forms, driven by factors such as frequency effects, conventionalization, and processing economy.
On the other hand, subjectification, as defined by Traugott (1995) and others, is the semantic–pragmatic process through which expressions acquire meanings that encode the speaker’s beliefs, attitudes, or evaluations regarding the message. The link between pragmaticalization and subjectification resides in their shared directionality: as forms shift from propositional or textual meanings to speaker-oriented, interpersonal meanings, they undergo both grammaticalization (or pragmaticalization) and subjectification. In this sense, pragmaticalization represents a domain where subjectification is particularly salient—for example, when adverbs such as actually, well, or so evolve from describing states or manners to marking discourse stance or speaker attitude.8
A concrete illustration can be seen in the verb go: in He goes to work every morning versus The year 1920 went by quickly, the verb’s meaning, initially grounded in physical motion, is metaphorically extended to express temporal progression, reflecting a shift from an objective interpretation to a more subjectively construed perspective.
Focusing specifically on connectors, Traugott and König (1991) describe their evolution as moving from meanings anchored in extralinguistic situations—which are more concrete, objective, and impersonal—toward meanings serving textual organization, such as anaphoric functions, and ultimately to meanings expressing the speaker’s attitude or belief toward the proposition. This final stage is therefore less concrete, more subjective, and consequently more personal. Martelotta (2004) further distinguishes between argumentative operators, which occupy relatively fixed positions within the sentence and assume prototypically grammatical roles, and discourse markers, which occur in freer positions and primarily function to adapt discourse to its production context.
From this discussion, one can identify three principal categorizations for assi(m)/así que in diachronic texts:
  • Compositional meaning: assi(m)/así functions as an adverb and que as a connector.
  • Non-compositional meaning I: assi(m)/así que functions as a conjunction.
  • Non-compositional meaning II: assi(m)/así que functions as a discourse marker.
Accordingly, we consider the classification proposed by Martín Zorraquino and Portolés Lázaro (1999) to be of limited utility from a diachronic perspective, highlighting the need to focus particularly on the distinction between connectors and discourse markers. Following earlier research, Hennemann and Schlaak (2015) propose two structural criteria for identifying discourse markers:
  • They are often followed by a pause.
  • They tend to occur in sentence-initial position.
Of these two criteria, Hennemann and Schlaak (2015) argue that only the second—sentence-initial position—is reliably applicable to written corpora, since pauses must be interpreted prosodically and are not necessarily indicated by punctuation. In older texts, where punctuation practices were highly variable, both criteria prove ultimately unreliable. Consequently, it becomes necessary to consider textual and discursive criteria as well—criteria that were also discussed by the authors—although these are more difficult to apply. For clarity, the criteria adopted in this study to differentiate between conjunctions and discourse markers are outlined in Section 4.2.
For Langacker (2008), analyzability refers to the extent to which a linguistic unit can be mentally decomposed into its constituent parts—that is, how transparently a speaker can perceive its internal structure. By this measure, maker is more analyzable than lipstick because it is formed via the highly productive derivational suffix -er (‘one who makes’), rendering its meaning immediately transparent. In contrast, lipstick is a lexicalized compound whose meaning cannot be fully predicted from its components and does not instantiate a productive morphological pattern. Even less analyzable items are those that have undergone substantial phonetic or semantic change, such as husband, from Old Norse hús + bóndi (‘master of the house’). In composed forms like lipstick maker, the salience of individual morphemes is further reduced, as both elements become backgrounded.
Traugott and Trousdale (2013) link the loss of analyzability to the process of grammaticalization, which typically involves semantic bleaching, decategorialization, and fusion, marking the shift from a compositional structure to a new grammatical construction. They argue that this loss of analyzability often co-occurs with increased schematicity and entrenchment, as speakers store the construction as a single cognitive unit.
In relation to our study, standalone assi(m)/así is not readily analyzable, having begun its grammaticalization process centuries ago, already during the Latin period. In contrast, assi(m)/así que remains more analyzable, since its immediate components—assi(m)/así and que—still function independently as connectors in both languages. This difference in analyzability suggests that the reanalysis or retraction to a non-compositional usage of assi(m) que may be facilitated by the recognizability of its constituent parts.

4. Materials and Methods

In the following subsections, we present and justify the methodological guidelines adopted for this research.

4.1. Corpus and Data Retrieval

Data were extracted from the Corpus do Português: Historical Genres (Davies & Ferreira, 2006). Two concordance queries were constructed, spanning texts from the 13th to the 19th century. The first query retrieved both assim que and é assim que, whereas the second query specifically targeted occurrences of assi(m) é que, with ‘|’ representing the conjunction or:
  • assy|assi|assim + que
  • assy|assi|assim + é|he|foi|foy|sendo
These queries allowed us to capture both the general use of assim que and its variants in copular constructions, facilitating a diachronic analysis of compositional and non-compositional usages.
After running these queries, all occurrences were displayed in Key Word in Context (KWIC) format, with each instance centered and accompanied by several words of surrounding context. This layout facilitated the identification of recurring syntactic and semantic patterns. Once the expanded context span was established, the concordance results were saved within the corpus interface and exported to an Excel file. This exported KWIC file was subsequently used for qualitative coding and quantitative analysis, allowing a systematic comparison of how assi(m) (é) que functions across different historical periods. To ensure comparability of textual genres across centuries, for the 20th century we included only European Portuguese data from fictional works.
Already in the corpus interface, we excluded instances not relevant to the research objectives. First, Brazilian authors from the 18th and 19th centuries were manually removed, using the worksheet provided by the CDP to identify textual sources.9 Additionally, the publication dates of two texts were adjusted to reflect their original release rather than the date of the edition consulted.10
  • Tratado da Sciencia Cabala (‘Treatise on the Science of Kabbalah’), by Francisco Manuel de Melo, considered as 17th-century rather than 18th-century material;
  • O Livro da Montaria (‘Book of the Hunt’), by D. João I, considered as 15th-century, rather than 18th-century material.
Second, in the Excel file we excluded other tokens that could potentially bias the results, such as:
  •  
    1.
    Instances of independent assim + que, where que functions as an expletive conjunction expressing an optative or injunctive meaning, since these cases do not contribute to the grammaticalization of the consecutive meaning of assim que, as in (26) (cf. a list of the full text titles in Appendix A).11
(26)
e nam no fazendo assim que eses lauradores nam ssejam costrangidos pera o pagarem a djnheiro (14:Manuel:1498)
‘and by not doing so, those plowmen will not be forced to pay it in money’
2.
Instances in which assim is preceded by a modifier, such as tanto (‘so much’), likely represent compositional adverbial structures that did not undergo the syntactic reanalysis leading to the formation of the connector assim que. The presence of a modifier indicates that assim retains its lexical and syntactic autonomy in these contexts, suggesting that such constructions fall outside the grammaticalization pathway that produced the complex conjunction. Accordingly, we analyzed tanto assim in (27) as a complex adverb followed by a consecutive clause introduced by que:
(27)
Tanto assim que Santo Inácio disse que, se alguém jejuasse aos domingos ou sábados, excepto os da Semana Santa, esse tal era matador de Cristo. (16:Bernardes:Floresta)
So much so that Saint Ignatius said that if anyone fasted on Sundays or Saturdays, except during those of Holy Week, that person was a slayer of Christ.’
In sum, following the first round of data exclusion, we collected 1917 tokens from the query output, of which 31 were excluded, leaving 1886 tokens for analysis.

4.2. Data Classification: General Aspects

The next step was data classification, which involved coding the dependent variable corresponding to the category of assim—as adverb, conjunction, or discourse marker—thereby capturing the compositional or non-compositional meaning of assi(m) que. This procedure was followed by classification of the morphosemantic meaning of each unit, or of the two items individually, as appropriate.
If assim functions as an adverb:
  • assim was classified as an anaphor or cataphor;
  • que was classified according to the type of clause it introduced, e.g., a complement, relative, adverbial, or embedded cleft clause—either canonical (‘ser’ + assim + que) or reverse (assim + é + que).
If assim combines with que to form a conjunction or discourse marker, we classified:
  • the semantic/pragmatic link connecting the clause introduced by assim que with the related unit (clause or discourse segment), and
  • the rhetorical relation connecting the clause introduced by assim que with the related unit.
Adopting the well-established framework of Rhetorical Structure Theory allows us to systematically classify rhetorical relations, which can then be applied to the categorization of assim que. This procedure is detailed in Section 4.3.
In the non-compositional reading of assim que, we needed to decide whether to classify it as a conjunction or as a discourse marker. This decision was based on whether assim signaled an intra-clausal or an inter-clausal relation, respectively. In this framework, assim que is considered a discourse marker only when a clear discourse break can be identified between its clause and the preceding one, signaled by one of the following criteria:
  • a change in the discourse topic;
  • a shift in the polyphonic source;
  • the beginning of a new stage in the narrative, marked by a different communicative intention.
Because the distinction between these two types of connectors can be subtle, we present two examples from the 16th century in which the semantic link is identical (Cause/Consequence):
(28)
Logo à tarde acodirão trinta pessoas nobres... [e baptizou-os]. [Assim que neste dia baptizou duzentos.] (15:Frois:Japam)
‘Later in the afternoon, thirty noble people arrived... [and he baptized them]. [Thus, on that day he baptized two hundred (people).]’
(29)
[quando eu livre minha alma de vossa mão, não na podereis livrar da presença do Senhor, porque ninguem pode fugir de sua vista nem de sua mão escapar vivo ou morto, que seu dominio e sobelos vivos e mortos.] [Assim que milhor sera que animosamente moura, e deixe ao meu povo exemplo de constancia...] (15:Usque:Israel)
‘[when I have freed my soul from your hand, you will not be able to deliver it from the presence of the Lord, for no one can flee from His sight nor escape His hand, living or dead, since His dominion extends over both the living and the dead.] [Therefore, it will be better that I die courageously and leave my people an example of constancy...]’
Assim que functions as an intra-clausal connector in (28), signaling a logical or temporal progression between two events. There is no significant shift in topic or discourse plane, as the narrative remains within the same episode, continuing the account of the baptisms. Moreover, both the clause containing assim que and the preceding clause are Elementary Discourse Units, each consisting of a single clause. Although the clause introduced by assim que appears as an independent sentence, it could equally be preceded by a comma rather than a period, reflecting its syntactic integration at the event level. By contrast, in (29), assim que signals a textual or rhetorical relation between discourse units. Here, we observe a shift from reasoning to conclusion or resolution, involving Complex Discourse Units (each comprising more than one syntactic clause), marked by brackets. These examples also illustrate that punctuation alone is insufficient to distinguish between discourse markers and conjunctions.

4.3. Data Classification: Rhetorical Structure Theory

Regarding the second type of classification, we employed Mann and Thompson’s (1988) version of Rhetorical Structure Theory (RST) to enable systematic quantification of the data.
RST is a framework for analyzing the coherence and functional organization of texts. It segments a text into minimal discourse units, typically clauses or sentences, called Elementary Discourse Units, and describes the relationships between them through rhetorical relations such as Elaboration, Contrast, Cause, or Condition. Each relation generally involves a nucleus, representing the central or essential information, and a satellite, which provides supporting, clarifying, or modifying content. These relations are hierarchical, allowing complex texts to be represented recursively in a tree-like structure. Some multinuclear relations exist as well, in which the connected discourse units carry equal importance and are both marked as nuclei.
RST emphasizes the functional role of discourse segments rather than their syntactic form. By modeling how information is connected and why certain segments support others, it provides a systematic approach to studying text coherence, argumentation, and communication strategies. Its hierarchical structure makes it a foundational tool for understanding both written and spoken discourse.
To systematically capture both the functional and interpersonal roles of assim que in discourse, we applied a combination of rhetorical and pragmatic analyses. In this study, RST was applied to identify the rhetorical relation between the discourse unit containing assim que and the immediately preceding unit, as listed in the second column of Table 1.12 In addition, we expanded the RST list to include a subjective/dialogical nexus, with more specific categories adapted from studies such as Bravo Cladera (2005) and Hennemann and Schlaak (2015). The relations identified in the data are underlined, and an example in English is provided in the third column, with the nucleus (N) and satellite (S) marked, to aid reader comprehension. Beyond this fine-grained classification, similar relations were grouped into broader semantic or pragmatic nexa, which are presented in the first column of Table 1.
It is important to note that the texts were analyzed based on our intuitions as speakers of a contemporary variety of Portuguese. To ensure the reliability of the analysis, both authors independently classified the data and subsequently discussed any examples on which they disagreed.
This system predicts a predominance of Cause/Consequence links, not only because these relations encompass a larger number of rhetorical categories, but also because the basic non-compositional meaning of assim que is associated with many of these notions. Conversely, given that the texts are predominantly monological (except for some plays), relatively few examples of the subjective/dialogical nexus are expected. It should be noted that, apart from the subjective/dialogical link—which is specific to discourse markers—all other semantic and pragmatic links expressed by assi(m) que can occur both when it functions as a conjunction and when it functions as a discourse marker.

5. Results: The Back-and-Forth of assi(m) que

Our data are distributed across the centuries as shown in Figure 1. The distribution is highly uneven, with relatively low counts in the early centuries and particularly high values in the 15th and 16th centuries, producing a skewed pattern compared to the 13th and 18th centuries. The median is 161.5, and the overall total is 1886 sentences.
In the following subsections, we present the detailed results in two parts: the first covering the 13th to the 16th centuries, and the second covering the 17th to the 20th centuries, followed by a third analytical section. This division is significant because it aligns with the most widely recognized periodization of Portuguese, separating Classical from Modern Portuguese, and facilitates a more organized presentation of the results.
The two periods established for analysis closely align with Castro’s (2006) proposal for the historical periodization of Portuguese, which was originally based on an unpublished paper by Lindley Cintra. The first period encompasses Old (1175–1420) and Middle Portuguese (1420–1550), marked by intense linguistic expansion, structural diversification, and the consolidation of written norms. The second period, spanning the 17th to the 20th centuries, roughly aligns with Classical (1550–1800) and Modern Portuguese (1800–present), characterized by the progressive stabilization of syntactic structures, the codification of stylistic and orthographic conventions, and the emergence of a standardized written variety.

5.1. Assi(m) que in Old and Middle Portuguese

The results from the earliest centuries represented in our database indicate a period during which assim retained its adverbial use (ADV), followed by que interpreted either as a complementizer or as a relative pronoun. Nevertheless, as early as the 13th century, we find evidence of the non-compositional meaning of assi(m) que, functioning either as a conjunction (CONJ) or as a discourse marker (DM), as illustrated in Figure 2.
From Figure 2, we can observe that the proportion of adverbs remains relatively stable from the 13th to the 15th century. This pattern changes markedly between the 15th and 16th centuries, when there is a sharp decline in the proportion of adverbs—approximately 20%. Conversely, the proportion of discourse markers (DMs) increases steadily, rising from 3.6% to 19% by the 16th century.
We also identified several instances of assim with a cataphoric function, particularly when preceded by verbs such as avir/acaescer (‘to happen’) or dizer (‘to say’), which typically have a presentational use. In these cases, que introduces a subjective complement clause:
(30)
 
a.
Andados seis ãnos do reynado deste rey Vuallya—que foy na era de quatrocentos e sesenta e dous ãnos—, aveo assim que Costancio patricio evyou dizer per suas cartas a rey Vuallia que lhe rogava que fosse ajudar os d’Espanha contra os barboros que os assenhoreavam mallamente. (14:CIPM:Demanda)
‘Six years having passed in the reign of this King Vualia, which corresponded to the year 462, it happened thus that Constantius, the patrician, sent word by his letters to King Vualia, asking him to go help the people of Spain against the barbarians, who were oppressing them severely.’
b.
Et sse uosso senhor outra uez acó enviar seu messageyro, çerto seia que nuca lle aló tornará, Et a uós, conde, digo assý que uos gardedes de nõ estar mais en Salamjna… (14:CIPM:Demanda)
‘And if your lord sends his messenger here again, let him be certain that he will never return from there. And to you, count, I say thus: take care not to remain any longer in Salamina…’
Furthermore, a few cases show assim functioning as an anaphor, followed by que as a connector introducing an adverbial clause, typically conveying a resultative meaning:
(31)
E os da outra parte não se atreverõ a viir empos elles ne de os cometer de nem uma guysa mas ficarom em seu outeyro que avyam tomado e teveronno. E steveron assim que nõ lydarom aquelles dous dias. (13:CIPM:CGEsp)
‘And those on the other side did not dare to come against them nor attack them in any way, but remained on the hill they had taken and held. And they stayed so, such that they did not fight for those two days.’
From the structure in (31), the non-compositional values of assim que as a connector appear to have developed to indicate either a Cause/Consequence or an Interpretation relation, illustrated respectively in (32) and (33).
(32)
e da outra parte passava huu barco de gente e seendo em meo do rio com grande arebatamento da agua começou se torvar o barco, assi que o que regia o barco desesperou de poderem sair em terra (15:FlosSanct)
‘And on the other side, a boat carrying people was crossing, and when it reached the middle of the river, the strong current began to overturn it, so that the person steering the boat despaired of their being able to reach the shore’
(33)
empero suas proprias entenções nunca son senom en servirem bem a seus senhores e fazerem-lhes prazer, en tal guisa que, pollo prazer e serviço que lhes assi fazem, possam vir a aver delles bem. Assi que as entençõys non som iguaaes em bondade… (14/17:DJoao:Montaria)
‘yet their own intentions are never anything but to serve their lords well and to please them, in such a way that, through the favor and service they thus render, they may come to receive some good from them. Therefore, the intentions are not equal in goodness…’
We considered the impossibility of replacing assim with an adverbial phrase such as de tal maneira/dessa maneira (‘in such a way’/’in this way’) as the primary criterion for identifying a non-compositional instance of assim que. The second diagnostic concerns whether assim can take scope over the preceding verb phrase, and a third is the presence of a punctuation mark before assim. Example (31) fails all three diagnostics; (32) meets the first and third, and (33) meets all of them.
The use of non-compositional assim que during Old and Middle Portuguese exhibits increasingly diverse patterns. As shown in Figure 3, in the 13th century its use was almost exclusively associated with the Cause/Consequence meaning, with only minimal instances of Sequence linking. Over time, the types of nexa became progressively more varied, reaching six distinct categories by the 16th century, including subjective linking. (cf. a more detailed classification in Appendix B).
We analyze these meanings in the order in which they emerge. In the 13th century, the construction with assim que, which originally marked Result, extended to express Sequence through a metonymic shift along the event chain. Because a result naturally leads into the subsequent event, the boundary between causal outcome and simple temporal succession becomes blurred. Consequently, the “result” stage of an event sequence comes to represent the progression itself, as illustrated in (34).
(34)
E assi andado ja ouve per muitas terras, assi que ouve chegado a Terena, u a Virgen fez muito miragr’ onrrado… (12:Mettman:CantigasSM3)
‘And so, having already traveled through many lands, he thus arrived in Terena, where the Virgin performed a miracle for him, one of great honor…’
In the 14th century, assi(m) que begins to mark a different type of nexus between clauses: Elaboration. As the temporal link weakens, the expression loses its event-advancing force and shifts from signaling what happens next to indicating what is said next. This discursive weakening enables assim que to introduce not consequences or subsequent events, but clarifying details, specifications, or exemplifications of the same situation. In this way, a former result/sequence marker is reinterpreted as an elaborative connector—one that expands the preceding idea by unpacking its content rather than advancing the narrative, as illustrated in (35).
(35)
E a ssegunda maneira he como quãdo aiutam dous moesteyros ou duas eygreias en um de guisa que nõ he sometuda umã a outra mais que som yguaes assim que os que som mõges ou uelogos da umma som esse mesmo da outra & esse mesmo da outra & todalas cousas que am som comunaes tanbe aos ums como aos outros (13:Afonso:Partida1)
‘And the second manner is when two monasteries or two churches assist one another in such a way that neither is subordinate to the other, but they are equals. Thus, those who are monks or clerics of one belong also to the other, and vice versa; and all the things they possess are common to both groups’
In the 15th century, assim que becomes available for higher-level discourse functions, including Evaluation and Summary. In the evaluative use, the connector no longer signals a consequence but instead introduces the speaker’s interpretive stance toward the preceding material, as shown in (36). In the summarizing use, the connector serves to recapitulate or wrap up prior discourse units, functioning like ‘thus’ or ‘in sum’ in English, and signaling that what follows condenses or restates the essential point of the preceding discussion, as illustrated in (37).
(36)
E se a virgem se calar nõ pecca: empero os taes aueran tribullaçom da carne Mas eu os perdo. assim que ysto digo Irmaãos ho tempo he breue (14:SantaMaria:Evangelhos)
‘And if the virgin remains silent, she does not sin; however, such people will have trouble in the flesh. But I pardon them. In sum, I say to you, brothers: the time is short’
(37)
porque hera muito estimado & amado de todolos bõos da corte… se foram allgus fidallgos & gemtis homes da corte, os quaes rrequererão liçemça a ell rrey pera ho hir servir aquella çidade. assim que amtre os que foram do rregno & os que la estavã heram na çidade CC de cavallo. (14:CondeMeneses)
‘because he was greatly esteemed and loved by all the honorable men of the court… some noblemen and gentlemen of the court went to him and requested the king’s permission to go serve him in that city. Thus, among those who came from the kingdom and those who were already there, there were two hundred horsemen in the city.’
Finally, in the 16th century, assim que develops into a general discourse-boundary marker—that is, a device speakers use to manage the flow of discourse through progressive abstraction. In (38), an excerpt from the popular play by Gil Vicente, Farsa dos Almocreves (‘The Farce of the Mule-Drivers’), the discourse marker signals the speaker’s decision to pivot the discourse, indicating a shift to a new sentence topic: in this case, a silver saltcellar offered by a goldsmith to a nobleman, which had been mentioned previously.
(38)
Fidalgo: —Rijamente m’apertais e fazeis-me mentiroso que eu gabei-vos doutro jeito e s’eu tornar ao desfeito nam será proveito vosso.
Ourives: —Assi que o meu saleiro peito?
Fidalgo: —Ele é dos mais maus saleiros que eu em minha vida comprei.
(15:Vicente:Obra1)
‘Nobleman: —You press me too hard and make me out to be a liar, for I praised you differently before; and if I take back what I said, it will not be to your advantage.
Goldsmith: —So (should I) offer my saltcellar?
Nobleman: —It is one of the worst saltcellars I have ever bought in my life.’
The final dimension of change concerns the possibility of separating the elements of the sequence assi(m) que through a cleft construction. Two constructions are relevant, as already mentioned: the canonical cleft and the reverse cleft. In both cases, assim can only display its basic adverbial (anaphoric) meaning.
(39)
E sse este amor & temor tẽdes bem ante vossos olhos sem duuida vos sooes no caminho que dereytamente vos leuara ao lugar onde vos chamamos. scilicet aas virtudes. Hora së duuida he assy que toda pesoa que bem ama deos. o deue mostrar per obras segundo elle mesmo diz no euãgelho… (15:Pisan:Cristina)
‘And if you keep this love and fear clearly before your eyes, you will undoubtedly be on the path that will directly lead you to the place to which we call you, namely, to the virtues. Now, undoubtedly, it is thus that every person who truly loves God must show it by deeds, as He Himself says in the Gospel…’
(40)
E todo he muy graue mal e sinal de mayor. Oo, Senhor Jhesu Christo, leuãta-te e esperta-te, porque dormes e te esqueeces, Senhor, das nossas pobrezas e da nossa tribulaçom! Leuãta-te, Senhor, e nõ nos dessenpares ataa a fim! E, pois assy he que ẽnos tenpos dagora he theudo por sandeu aquelle que leixa muy pouca fazenda por seruir a nosso Senhor… (14:Maler:Orto)
‘And all this is a very grave evil and a sign of something even worse. O Lord Jesus Christ, rise and awaken, for why do You sleep and forget, Lord, our poverty and our tribulation? Rise, Lord, and do not abandon us until the end! And since it is thus, in these times, anyone who leaves even a small portion of his possessions to serve our Lord is considered a madman…’
Figure 4 shows the proportion of these constructions within the adverb dataset (comparing clefts with other examples in which assim is an adverb, i.e., with a compositional meaning) and within the full dataset (i.e., comparing clefts with all examples with assim…que) for each century. The numbers show that the reverse cleft is always more frequent than the canonical one. In addition, the number of clefts becomes more relevant in the 15th and 16th centuries, reaching about 75% relative to the number of adverbs. However, within the full dataset, these constructions show a clear increase in frequency only in the 15th century (about 17%).
On the one hand, the increased use of cleft constructions during the Middle Portuguese period reinforces the compositional character of assi(m) que. At this stage, clefts become the primary locus of the adverbial meaning of assi(m), as the earlier uses illustrated in (30)–(31) disappear with the stabilization of the non-compositional meaning of the sequence. However, given the predominance of the consecutive value of assi(m) que, together with the semantic extensions discussed in (34)–(38), the growing frequency of clefts did not immediately trigger a systemic grammatical change. Nevertheless, this configuration of the data did not remain stable in the subsequent centuries.

5.2. Assi(m) que in Classical and Modern Portuguese

Between the 17th and 20th centuries, assim expanded its adverbial use (ADV), although in virtually all cases it occurs as part of a cleft construction. In instances where assi(m) que exhibits a non-compositional meaning, these centuries show increased use as a conjunction (CONJ) and progressively lower percentages as a discourse marker (DM), as illustrated in Figure 5.
From Figure 5, we can observe that the proportion of adverbs drops sharply in the 18th century but doubles from the 17th to the 19th century, remaining high thereafter—a pattern clearly linked to the expansion of cleft constructions, as we will demonstrate. Conversely, the use of assim que as a discourse marker (DM) declines progressively, with a corresponding increase in its categorization as a conjunction. As illustrated in Figure 6, this shift is closely associated with the emergence of its immediate temporal value.
By the 17th century, the immediate temporal sense of assim que is already attested, likely developing from the sequential value documented in Old and Middle Portuguese. Notably, the discourse unit introduced by assim que establishes its rhetorical linkage prospectively, anchoring to the subsequent clause rather than the preceding one, as becomes evident when comparing (34) with (41).
(41)
& como taes, sempre ficão com o sabor da primeira agoa, que recebèrão, ainda que depois lhe lancem outra mais, ou menos saborosa. [Assim que V. R. A. chegar ao tempo da adolescencia, & nelle industriado na Doutrina Christãa; sem cujos documentos, & comprehensão não està idoneo para os empregos de seus espirituaes affectos;] [lembrese, que o amor de Deos he o primeiro mandamento da Ley Divina]… (16:Prazeres:Abedario)
‘and as such, they always retain the taste of the first water they received, even if afterwards other water—more or less flavorful—is poured into them. [As soon as V. R. A. reaches adolescence, and is instructed in Christian Doctrine—without whose teachings and understanding you will not be fit for the duties of its spiritual affections—] [remember that the love of God is the first commandment of the Divine Law]…’
A significant semantic shift has occurred in the use of the clause introduced by assim que (clause A), even though its position relative to the clause it relates to (clause B) may vary. Compare (41) with (42), in which the clauses are indicated by square brackets.13
(42)
e assim o fez ao sermao, acabado o qual foi António Vaz queixar-se a el-Rei contra o Figueiredo, [que lhe disse] [assim que ele se pôs de joelhos]: —Folguei muito de vos ouvir. (17:Morais:Apotegmas)
‘And so he did at the sermon, and after it ended, António Vaz went to complain to the King against Figueiredo, [who said to him] [as soon as he knelt down]:—I was very pleased to hear you.’
In Old and Middle Portuguese, clause A functioned as the main point (nucleus), while clause B acted as the subordinate element (satellite), describing a subsequent event. In Classical and Modern Portuguese, this hierarchy is reversed: clause A now serves as the satellite, indicating an event that precedes and sets the temporal context for the nuclear event in clause B—observe that the natural order is not shown in (42), where assim que appears in an intercalated clause. This shift coincides with the retraction of assim que from its use as a discourse marker and is consistent with its use as a conjunction.
The use of the new immediate temporal construction increased sharply from 1% to 73% between the 17th and 18th centuries, continuing to rise until it became virtually categorical by the 20th century. During this shift, other peripheral meanings that had emerged during the widespread consecutive use of assim que—such as those conveying Conduction, Elaboration, and Interpretation—gradually disappeared.
The motivation for this change appears to be linked to the high analyzability of assim que, combined with the significant increase in cleft constructions that interrupt the assim…que sequence, as shown in Figure 7. (Note that the absence of corresponding bars indicates no cleft data were reported for the 18th century.)
The data in Figure 7 reveal several key differences compared to Figure 4, which presents results from earlier periods. First, the number of canonical cleft constructions consistently exceeds that of reverse clefts across all centuries analyzed in this period. Second, cleft constructions are nearly categorical among instances of assim with an adverbial value. Third, when combining canonical and reverse clefts, the total proportion of cleft constructions in the full dataset exceeds 50% in the 19th and 20th centuries, reaching 64.8% and 57.2%, respectively.

5.3. Analysis of the Diachronic Change Involving assi(m) que

Here, we aim to highlight more explicitly the rise in cleft constructions in Portuguese as a driving factor behind the semantic change involving assim que and its relation to analyzability.
As illustrated in the following examples, the function of assim in cleft constructions—both in the canonical order (é assim que) and in the reverse order (assim é que)—remains systematically anaphoric or deictic throughout the entire period under study:
(43)
o que o seu coração nos poderia contar d’interessante, d’original, d’inédito, provàvelmente o pudor lhe proibiu que o revelasse, e é assim que muitas vezes esta hipocrisia imposta dá aos livros de mulher a possibilidade de parecerem ter sido escritos por um homem. (18:Almeida:Gatos5)
‘What your heart might have revealed to us—what is interesting, original, or previously unknown—was probably restrained by modesty; and it is in this way that such imposed hypocrisy often allows a woman’s book to appear as if it were written by a man’
(44)
Maria: —E bem sabeis, querida mãe, que eu ando há tanto tempo para ir àquele convento para conhecer a tia D. Joana.
Jorge: —Sóror Joana: assim é que se chama agora. (18:Garrett:Sousa)
‘Maria: —And you well know, dear mother, that I have been wanting for so long to go to that convent to meet Aunt D. Joana.
‘Jorge: —Sister Joana: that is what she is called now.’
In (43), assim refers to the preceding proposition. Notably, the entire cleft construction appears to relate to the previous discourse unit via Interpretation (concerning the quality of books written by women), a nuance no longer conveyed by non-compositional uses of assim que. In (44), assim refers to an immediately preceding referent, Sóror Joana (‘Sister Joana’), which corrects the noun phrase a tia D. Joana (‘Aunt D. Joana’) in the interlocutor’s turn.
Given this, the first step in our analysis is to examine whether the increased use of cleft constructions reinforced the adverbial use of assim. Figure 8, which combines data from Figure 2 and Figure 5, illustrates this pattern: assim as an adverb never exceeded 35% before the 19th century, whereas in the 19th and 20th centuries this proportion nearly doubled. Therefore, cleft constructions somehow reactivate the compositional structure in which assim functions as a genuine (anaphoric) adverb and que as a connector. Remember that the high analyzability of assim que is a prerequisite for this process.
Second, the reactivation of the compositional structure of assim que may be linked to the increase in its immediate temporal meaning, as the anaphoric and deictic values of assim become more salient. Simultaneously, this change disfavors cataphoric readings, which were more frequent in earlier uses of assim que as a resultative complex conjunction. This process ultimately led to the conventionalization of assim que as an immediate temporal connector, since this meaning entails an anaphoric link to another discourse unit that expresses the nucleus in the rhetorical relation.
Third, the former functions of assim que came to be expressed by other connectors marking Cause/Consequence, such as porque, já que, and visto que.14
We now aim to describe this chain of events using the theoretical framework of Cognitive Grammar as a guiding perspective. From this perspective, assim profiles manner while remaining highly context-dependent, with deictic, anaphoric, and cataphoric potential. Regarding assim que, the resultative-sequential schema was its prototypical interpretation, a meaning that gradually extended to encompass other relational notions such as Interpretation and Conduction, as we condense in the following representations.
  • Assim (relational dependent schema):
    Semantic nucleusProfiling of Manner with contextual dependence
    Extension potentialDeictic, Anaphoric or Cataphoric
  • Assim que1 (resultative schema)
    Relation Event A → Result B
    OrientationCataphoric
    ProcessingProspective scanning
    Inferential loadStrong causal inference
As the construction specialized toward immediate temporal sequencing, the resultative-sequential schema—which relied on prospective anchoring—lost functional prominence. Consequently, cataphoric interpretations declined, and assim que became conventionalized as a temporal connector.
  • Assim que2 (immediate temporal schema)
    RelationEvent A → Result B
    OrientationAnaphoric with deictic persistence
    ProcessingSequential scanning
    Inferential loadLow causal inference
  • É assim que/Assim é que (cleft frames)
    StructureCopula (variable position), Assim, Que-clause
    o 
    Assim (adverb)
    Semantic nucleusProfiling of Manner, [+Anaphoric], [+Deictic]
    FunctionHighlighting how an action or event occurred in relation to the previous context.
    o 
    Que-clause
    FunctionSpecifying the event or action described.
The third step—namely, the redistribution of causal-resultative functions to more specialized connectors (porque, já que, visto que)—may reflect systemic paradigmatic pressures within the language. From a diachronic perspective, however, we interpret this development as a case of retraction, following the term proposed by Haspelmath (2004, p. 33).15 The author defines this phenomenon as a type of change in which more grammaticalized usages become marginalized or entirely obsolete, while less grammatical usages increase in frequency. This precisely characterizes the evolution of assim que in Portuguese: its highly grammaticalized use as a discourse marker indicating Cause/Consequence, Interpretation, and related notions became entirely obsolete, whereas its function as a temporal conjunction of immediacy was conventionalized.
Interestingly, the expression of resultative meanings can still be conveyed today—either by assim alone, by other consecutive discourse markers such as portanto, or by cleft constructions, owing to their discursive flexibility (cf. Andrade & Antonio, 2024 for a study on Portuguese wh-clefts illustrating this issue). However, this is pragmatic information that is not semantically encoded in the assim…que sequence itself.
We further discuss the implications of this analysis in Section 6.

6. Discussion

In this section, we expand upon three issues: (i) why this change does not constitute a case of degrammaticalization; (ii) its relevance of this change for the periodization of Portuguese and its relation to the diachronic development of cleft constructions in the language, and (iii) the extent to which this proposal may account for the syntactic divergence between Portuguese and Spanish.
First, why should the change in assim que not be classified as an instance of degrammaticalization? According to Norde (2009, p. 120), degrammaticalization is a change “whereby a gram in a specific context gains in autonomy or substance on one or more linguistic levels (semantics, morphology, syntax, and phonology).” A crucial diagnostic for identifying such cases is movement in the reverse direction along established grammaticalization clines. In the present case, however, no such reversal can be observed. Moreover, the development does not fit any of the three types of degrammaticalization proposed by Norde—namely, degrammation, deinflectionalization, and debonding.
Rather than representing a shift from a more grammatical to a less grammatical status, the development of assim que is more adequately analyzed as a semantic reconfiguration within the domain of conjunctions, accompanied by the retraction of other meanings, as discussed in Section 5. Crucially, the change does not involve a shift in syntactic category: in both stages, assim que functions as a clause-linking device. What changes is the internal semantic profile of the relation, moving from consequence-oriented sequencing to temporally bounded succession. The development can therefore be understood as semantic narrowing and reweighting within the connector domain.
More specifically, the change involves:
  • the maintenance of the subordinating function of assim que;
  • increased semantic specialization;
  • the weakening or loss of causal inference; and
  • the strengthening of temporal profiling.
All these developments are consistent with constructional change rather than with a reversal in grammaticalization status.
Turning to our second point, we would like to consider the relevance of this change for the history of Portuguese. The period in which it is implemented coincides with the onset of Modern European Portuguese (18th century). This temporal convergence is significant: it confirms that during the classical period Portuguese and Spanish shared a substantial number of grammatical structures, whereas they grew more distinct with the emergence of Modern Portuguese. Some typical developments of that grammar include the consolidation of the prepositional infinitive (estar + a + Infinitive) with progressive interpretation; the stabilization of the periphrastic perfect (ter + Past Participle) with iterative value; and an increase in syntactic rigidity in word order.
The latter development is likely directly related to the emergence and expansion of cleft constructions. According to Lahousse and Lamiroy (2012), increased rigidity in word order correlates with greater reliance on syntactic constructions capable of highlighting discourse-salient constituents. Cleft constructions provide precisely such a mechanism: they allow for the explicit marking of focus, thereby compensating for the reduction in syntactic flexibility which, in earlier stages, could encode informational distinctions without the need for additional morphosyntactic material.
Within this framework, the expansion of the cleft inventory can be interpreted as part of a broader reorganization of the discourse grammar of Portuguese. Longhin (1999, pp. 92–100), based on an extensive corpus of notarial and literary documents, identifies the possible early emergence of cleft constructions involving assim as early as the 13th century. Her analysis relies on a set of diagnostic tests applied to paraphrases of the attested sentences. She ultimately dismisses the medieval structures as fully developed clefts, classifying them instead as “incipient cleft constructions,” since the pattern had not yet achieved syntactic stabilization. This assessment is grounded in the absence of parallel clefting strategies involving other constituents at the time.
However, given the presence of emphasis associated with focus marking, these structures can reasonably be analyzed as bona fide cleft types. They appear to serve a dual function: on the one hand, they encode focal prominence; on the other, they correlate with specific semantic specializations. In the examples under discussion, canonical clefts tend to signal Elaboration, whereas reverse clefts are associated with Result:16
(45)
 
a.
Porque o finamento d’elrrei fora feito muito simprezmente e não suas exequias como deveram, hordenou a rrainha de mandar chamar todollos senhores e fidallgos do rregno que vehessem ao saimento do mes, pera sse fazer o mais honrradamente que podesse; e foi assim que o fezerom ho melhor que pôde ser, como compria à honrra d’elrrei; (14:Lopes:Fernando)
‘Because the king’s burial had been conducted very simply and not with the funeral rites that were due, the queen ordered all the lords and noblemen of the kingdom to be summoned to attend the month-later procession,17 so that it could be conducted as honorably as possible. And so they did it in the best possible way, as was fitting for the king’s honor.’
b.
[Almançor] amava outrossy os seus, ẽtanto que todos se trabalhavã em fazer serviço. E assi foy que, des o ryo de Doyro, que entom era en comarca antre os mouros e os cristãos, todo correrom e destruyrõ ataa o ryo que chamã Escola. (13:CIPM:CGEsp)
‘He [Almanzor] loved his own people (i.e., the Moors) so much that all worked hard to serve him. Thus, from the Douro River—which at the time marked the border between the Moors and the Christians—they advanced, destroying everything along the way up to the river now called the Escola.’
We may now ask whether the early presence of clefts with assim is in any way related to their later role in the change discussed here. The answer is at least partially affirmative. Cleft constructions require a clefted constituent—that is, the element occupying the main clause position alongside the copula—with strong potential for discourse identification. This requirement is readily satisfied by assim, given its anaphoric and relational nature. Indeed, the earliest cleft constructions in Portuguese involved assim precisely because its discourse-dependent profile made it structurally and pragmatically compatible with identificational focus. As a context-sensitive modifier of manner, assim already encoded discourse linkage, which facilitated its integration into the cleft schema.
Another crucial factor concerns the diachronic distribution of clefts. According to Longhin (1999), their overall frequency increased substantially after the Old Portuguese period: until the 16th century, absolute counts remained below ten tokens; by the 17th century, the total reached twenty; and by the 19th century, it had risen to over two hundred. Similarly, Silveira (2014), drawing on corpus-based research using the Tycho Brahe Parsed Corpus of Historical Portuguese, observes that the frequency of both clefts and wh-clefts increases markedly from the 18th century onward. She attributes this shift to the decline of the relatively relaxed V2 grammar characteristic of Classical Portuguese (cf. Galves, 2020, among others).
In contemporary Portuguese, constructions belonging to the cleft family are highly productive (Braga et al., 2015), and there remains ongoing debate regarding the chronological order of their emergence. Concerning the initial stages of this trajectory, Longhin (1999) and Kato and Ribeiro (2009) concur that canonical wh-clefts emerged first, followed by reverse wh-clefts and reverse clefts (the latter developing from the use of ser in existential clauses), and only subsequently canonical clefts. This developmental sequence accounts for the prominence of reverse clefts in our medieval data.18
Finally, let us consider the expectations our study generates regarding Spanish—specifically, why this language not only preserved the conservative semantic uses of así que, but also extended them to pragmatic nuances that appear to have been curtailed in Portuguese. Despite the relative scarcity of diachronic research on cleft constructions in Spanish, comparative work based on parallel corpora (e.g., Dufter, 2009) indicates that clefts are considerably more frequent in Portuguese than in Spanish. This asymmetry is highly relevant: in the absence of a comparable expansion of a construction forcing a compositional reading of así in Spanish, such as clefts, even a highly analyzable sequence such as así que could remain structurally and functionally stable.
In present-day Spanish, así que displays a broad functional range, from conjunction to discourse marker, with context-dependent interpretations. Nevertheless, the sequential or consequential meaning remains central. This pattern suggests a greater degree of diachronic continuity between medieval and modern Spanish than between medieval and modern Portuguese.

7. Final Remarks

This study has shown that assi(m) que functioned as a dynamic conjunction and discourse marker during Old and Middle Portuguese, alongside compositional uses as an adverb followed by a subordinate clause. A period of parallel grammaticalization with Spanish was interrupted by the rise in cleft constructions, which became increasingly frequent in late Classical and Modern Portuguese. This development reanalyzed assim as an anaphoric adverb and que as the connector introducing the embedded clause that forms the cleft construction. The only grammaticalized meaning that remained categorical for assim que was temporal succession.
The retraction of assim que to its use as a conjunction demonstrates that it is possible to interrupt a long-standing process of change in the domain of connectors. In addition, this case suggests that the trajectory of assim que is not isolated, but rather illustrative of how residual analyzability can redirect grammaticalization pathways.
The present study opens the door to extending this proposal to other instances of change driven by renewed analyzability in different languages, particularly within the domain of connectors, where the boundary between lexical transparency and functional opacity is often historically fluid. Such comparative analyses may provide a deeper understanding of the conditions under which grammaticalization can undergo retraction.

Author Contributions

Conceptualization, A.L.d.A.; methodology, A.L.d.A. and G.M.O.; software: A.L.d.A.; validation, A.L.d.A. and G.M.O.; formal analysis, A.L.d.A.; investigation, A.L.d.A. and G.M.O.; resources, A.L.d.A.; data curation, A.L.d.A.; writing—original draft preparation, A.L.d.A. and G.M.O.; writing—review and editing, A.L.d.A.; visualization, A.L.d.A.; supervision, A.L.d.A.; project administration, A.L.d.A. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data presented in this study are available in the Center for Open Science (OSF) at https://osf.io/q8hpu/overview?view_only=7d9636465e754ff1bdb27a5e9d5671ae (accessed on 26 November 2025), reference number 10.17605/OSF.IO/Q8HPU. These data were derived from the following resources in the public domain: Corpus do Português: Historical Genres: https://www.corpusdoportugues.org/hist-gen/ (accessed on 26 November 2025).

Acknowledgments

We thank the anonymous reviewers for their feedback and questions and the audience at Semana do Conhecimento UFMG 2025, where a portion of this material was presented. We also acknowledge the editorial assistance of Ariana Zhou, Josep Alba-Salas and Manuel Delicado Cantero. The customary disclaimers apply.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A. List of Text Abbreviations

Table A1. Text abbreviations used in Corpus do Português.
Table A1. Text abbreviations used in Corpus do Português.
AbbreviationTextSource
12:Mettman:CantigasSM3Cantigas de Santa Maria, vol. 3, Afonso X [Edited by Walter Mettman]Scanned text
13:Afonso:Partida1Primeira Partida, Afonso XScanned text
13:CIPM:CGEspCrónica Geral de Espanha de 1344 (Ms. L)CIPM *
14:CIPM:DemandaA Demanda do Santo Graal, 15th-century CIPM *
14:CondeMenesesCrónica do Conde D. Pedro de Meneses, ZuraraCIPM *
14/17:DJoao:MontariaLivro da Montaria, Dom João IUniversity of Aveiro
14:Lopes:FernandoCrónica de Dom Fernando, LopesScanned text
14:Maler:OrtoOrto do Esposo [Edited by Bertil Maler]CIPM *
14:Manuel:1498Forais Manuelinos, 1498 FCT **
14:SantaMaria:EvangelhosEuangelhos e epistolas con suas exposições en romãce, Santa MariaUniversity of Minho
15:FlosSanctFlos SanctorumScanned text
15:Frois:Japam1Historia de Japam, Part 1, FróisUniversity of Aveiro
15:Lucena:SFXavierHistoria da vida do Padre S. Francisco Xavier, LucenaUniversity of Aveiro
15:Pisan:CristinaO espelho de Cristina, PisanUniversity of Minho
15:Usque:IsraelConsolaçam às Tribulações de Israel, UsqueScanned text
15:Vicente:Obra1Obra Completa, Part 1, Vicente,University of Lisbon
16:Bernardes:FlorestaNova Floresta, BernardesCTB ***
16:Prazeres:AbedarioAbedario Real e Regia Instrucçam de Principes Lusitanos, PrazeresUniversity of Aveiro
17:Aires:VaidadeReflexões sobre a Vaidade dos Homens, AiresCTB ***
17:Morais:ApotegmasColecção política de apotegmas, MoraisScanned text
18:Almeida:Gatos5Os Gatos, Volume 5, AlmeidaUniversity of Aveiro
18:Garrett:SousaFrei Luís de Sousa, GarrettUniversity of Aveiro
* CIPM = Computerized Corpus of Medieval Portuguese; ** FCT = Foundation for Science and Technology, Portugal; *** CTB = Tycho Brahe Parsed Corpus of Historical Portuguese.

Appendix B. Complementary Tables with Data Classification

Table A2. Rhetorical relations connecting assim que in Old and Middle Portuguese.
Table A2. Rhetorical relations connecting assim que in Old and Middle Portuguese.
LinkRelation13th Century14th Century15th Century16th Century
Cause/
consequence
Cause10.3%
Condition10.2%
Enablement10.8%102.8%183.3%
Evidence51.4%122.2%
Justify21.6%51.4%61.1%
Purpose34.7%72.0%101.9%
Result6093.7%10283.6%25171.7%39273.1%
ConductionPreparation20.6%
Summary20.6%152.8%
Restatement10.3%40.7%
ElaborationBackground41.1%20.4%
Elaboration21.6%277.7%203.7%
InterpretationEvaluation236.5%224.1%
Motivation61.7%30.5%
SequenceSequence11.6%10.8%61.7%305.6%
SubjectiveTopic-shift10.2%
Total64122350536
Table A3. Rhetorical relations connecting assim que in Classical and Modern Portuguese.
Table A3. Rhetorical relations connecting assim que in Classical and Modern Portuguese.
LinkRelation17th Century18th Century19th Century20th Century
Cause/
consequence
Cause
Condition
Enablement11.1%
Evidence66.5%
Justify22.2%12.4%
Purpose
Result5862.4%921.9%24.1%11.1%
ConductionPreparation
Summary1212.9%12.4%
Restatement
ContrastAntithesis11.1%
ElaborationBackground 24.1%
Elaboration11.1%
InterpretationEvaluation66.5%
Motivation55.4%
SequenceTemporal11.6%3073.2%4591.8%8998.9%
Total93414990

Notes

1
Other intermediary, mostly short-lived, uses of assim are reported in the cited works, including its use as a correlative conjunction.
2
According to Lopes-Damasio (2008), assi could also be used in optative and deprecatory formulas, also referred to as volitive expressions. However, in (i) its function is cataphoric, as it points to the motive for the help:
(i)
Senhor, eu sou Galvam, o vosso amigo, que vós amávades tanto. E, assi Deus me ajude, pêsa-me desta maa-andança tanto, como se me aviesse com alguu de meus irmaãos. (Demanda do Santo Graal, 1210–1240, as cited in Lopes-Damasio, 2008, p. 107).
‘My Lord, I am Galvam, your friend, whom you loved so much. And so may God help me; I so deeply regret this misfortune, as if I were dealing with one of my brethren.’
3
The following abbreviations are used for the corpora cited: CDP = Corpus do Português (Davies & Ferreira, 2006); Iboruna Corpus = Electronic excerpts from the Portuguese of São Paulo State (Gonçalves, 2007), representing Southeastern Brazilian Portuguese.
4
See also Lopes-Damasio (2011) for other another constructional view, which analyzes assim is combination with other words.
5
A parallel study showing the productivity of adverb + que in Spanish, as well as other sequences involving que, is Barra Jover (2002), conducted within a formal framework. I thank an anonymous reviewer for this observation.
6
Rodríguez Molina (2015) discusses the various historical forms of así in Spanish, which could vary and appear as , asín, ansí, ansina, asina, among others.
7
We thank an anonymous reviewer for drawing our attention to this fact.
8
While subjectification is a mechanism through which meanings become more speaker-centered, intersubjectification shifts them toward the hearer, both in epistemic and social terms (cf. Traugott, 1995). It is important to emphasize that intersubjectification—as the explicit representation of attention to the hearer’s attitudes—arises from and crucially depends on subjectification; that is, it functions as a subsequent stage.
9
Nor did we include authors who were born in Portugal and spent long periods in Brazil, such as António José da Silva.
10
This decision was made primarily to ensure that the sentences in our database correspond to the remaining data from the century in which they were originally written. Unfortunately, this greatly reduced the amount of data available for the 18th century.
11
The corpus data are accompanied by this reference pattern, with optional elements in brackets:
  • century of publication: (edition source or author name): abbreviated title.
Centuries are indicated using the first two digits of the year (14 = 14xx, i.e., 15th century).
12
For Cause and Result, we did not consider the distinction between volitional and non-volitional, as proposed in Mann and Thompson (1988). Another relevant adaptation is that we disregarded the originally proposed relation Joint, since it did not appear in our data, and treated Temporal Succession as a separate rhetorical relation distinct from Sequence. The difference between the two is explained in Section 5.
13
In Modern Portuguese, a clause introduced by assim que can also be parenthetically intercalated within a sentence, often appearing alongside other adjunct phrases.
(i)
porque Vossa Majestade [assim que veio ao mundo], [só com se mostrar], disse o que havia de ser. (17:Aires:Vaidade)
‘because Your Majesty, [as soon as you came into the world], [merely by showing yourself], revealed what you were destined to be.’
14
Pons Bordería (2014, p. 124) also proposes that paradigmatic pressure was one of the causes for the pragmaticalization of Spanish o sea as a discourse marker. We thank an anonymous reviewer for suggesting the volume which contains that chapter.
15
We thank an anonymous reviewer for suggesting a deeper analysis of this issue.
16
The few occurrences of sendo assim que may be analyzed as ser + cataphor assim + insubordinate que. Interestingly, they are concentrated in a few texts from the 16th century:
(i)
[Nossas perguntas são] tão conformes à pouquidade e limitação sua delas, que não perguntamos de alguma coisa que é, senão limitando-lhe o ser [...] Sendo assim que do infinito ser, menos impróprio fora perguntar que não é que o que não é que o que é... (15:Lucena:SFXavier)
‘[Our questions are] so aligned with our own scantiness and limitations that we do not inquire something that exists, except by constraining its being… It being so: regarding to the infinite being, it would be less improper to ask what it is not than to ask what it is…’
17
To be more precise, the term saimento do mes referred to the mass held a month after someone’s death.
18
The chronology proposed in the latter study, however, has been challenged by subsequent research, such as Silveira (2014).

References

  1. Afonso X. (2021). Cantigas de Santa Maria: Texto crítico completo (J.-M. Montero Santalha, Ed.). (Original work published 13th century). Available online: https://www.academia.edu/50870774/Cantigas_de_Santa_Maria_Texto_crítico_completo_2021_ (accessed on 10 July 2025).
  2. Andrade, A. L. d., & Antonio, J. D. (2024). Brazilian portuguese wh-clefts in a multilevel analytic perspective. Pragmatics, 35, 475–503. [Google Scholar] [CrossRef]
  3. Barra Jover, M. (2002). Propiedades léxicas y evolución sintáctica. El desarrollo de los mecanismos de subordinación en español. Toxosoutos. [Google Scholar]
  4. Bartol Hernández, J. A. (1986). Oraciones consecutivas y concessivas en las siete partidas. Ediciones Universidad de Salamanca. [Google Scholar]
  5. Braga, M. L., Kato, M. A., & Mioto, C. (2015). As construções-Q no português brasileiro falado: Relativas, clivadas e interrogativas. In M. A. Kato, & M. do Nascimento (Eds.), Gramática do português falado no Brasil, Vol. 2: A construção da sentença (pp. 187–227). Contexto. [Google Scholar]
  6. Bravo Cladera, N. (2005). Y, pero, así que y es que: Un estudio de su uso en las interacciones del español de jóvenes bilingües y unilingües [Studia Romanica Upsaliensa] (Vol. 70). Acta Universitatis Upsaliensis. Available online: https://uu.diva-portal.org/smash/get/diva2:164818/FULLTEXT01.pdf (accessed on 25 September 2025).
  7. Castro, I. (2006). Introdução à história do português (2nd ed.). Colibri. [Google Scholar]
  8. Davies, M., & Ferreira, M. (2006). Corpus do Português: Historical genres. Available online: http://www.corpusdoportugues.org/hist-gen/ (accessed on 10 October 2025).
  9. Dufter, A. (2009). Clefting and discourse organization: Comparing Germanic and Romance. In A. Dufter, & D. Jacob (Eds.), Focus and background in Romance languages (pp. 83–121). John Benjamins. [Google Scholar]
  10. Fuentes Rodríguez, C. (1985). Sintaxis oracional: Las oraciones consecutivas en español. Ediciones Alfar. [Google Scholar]
  11. Galves, C. (2020). Relaxed V2 in Classical Portuguese. In R. Woods, & S. Wolfe (Eds.), Rethinking verb second (pp. 368–396). Oxford University Press. [Google Scholar] [CrossRef]
  12. Givón, T. (1979). On understanding grammar. Academic Press. [Google Scholar]
  13. Givón, T. (1995). Functionalism and grammar. John Benjamins. [Google Scholar]
  14. Gonçalves, S. C. L. (2007). Banco de dados Iboruna: Amostras eletrônicas do português falado no interior paulista. Available online: http://www.alip.ibilce.unesp.br (accessed on 25 September 2025).
  15. Guzmán Herrera, A. R. (2014). Luego, así que y apenas: Marcadores pragmáticos, de inmediatez y de consecución [Doctoral dissertation, Universidad Nacional Autónoma de México]. [Google Scholar]
  16. Haspelmath, M. (2004). On directionality in language change with particular reference to grammaticalization. In O. Fischer, M. Norde, & H. Perridon (Eds.), Up and down the cline: The nature of grammaticalization (pp. 17–44). John Benjamins. [Google Scholar] [CrossRef]
  17. Hennemann, A., & Schlaak, C. (2015). Así que als diskursmarker. Ein fall von adfunktionalisierung. Universität Potsdam. Available online: http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-75412 (accessed on 10 March 2025).
  18. Herman, J. (2000). Vulgar Latin (R. Wright, Trans.; 2nd ed.). The Pennsylvania State University Press. (Original work published 1967). [Google Scholar]
  19. Hopper, P. J., & Traugott, E. C. (2003). Grammaticalization (2nd ed.). Cambridge University Press. [Google Scholar]
  20. Kato, M. A., & Ribeiro, I. (2009). Cleft sentences from old Portuguese to modern Portuguese. In A. Dufter, & D. Jacob (Eds.), Focus and background in Romance languages (pp. 123–154). John Benjamins. [Google Scholar]
  21. Lahousse, K., & Lamiroy, B. (2012). Word order in French, Spanish and Italian: A grammaticalization account. Folia Linguistica, 46, 387–416. [Google Scholar] [CrossRef]
  22. Langacker, R. W. (1987). Foundations of cognitive grammar (Vol. 1). Stanford University Press. [Google Scholar]
  23. Langacker, R. W. (1991). Foundations of cognitive grammar (Vol. 2). Stanford University Press. [Google Scholar]
  24. Langacker, R. W. (2008). Cognitive grammar: A basic introduction. Oxford University Press. [Google Scholar]
  25. Longhin, S. R. (1999). As construções clivadas: Uma abordagem diacrônica [Master’s thesis, Universidade Estadual de Campinas]. [Google Scholar]
  26. Longhin-Thomazi, S. R. (2006). Gramaticalização, (inter)subjetivização e modalidade epistêmica: O caso de assim. Estudos Linguísticos, XXXV, 1772–1779. [Google Scholar]
  27. Lopes-Damasio, L. R. (2008). A emergência do marcador discursivo assim sob a ótica da gramaticalização [Master’s thesis, Universidade Estadual Paulista]. [Google Scholar]
  28. Lopes-Damasio, L. R. (2011). Diacronia dos processos constitutivos do texto relativos a assim: Um novo enfoque da gramaticalização [PhD dissertation, Universidade Estadual Paulista]. [Google Scholar]
  29. Machado, J. P. (1967). Dicionário etimológico da língua portuguesa (2nd ed., Vol. 1). Editorial Confluência. [Google Scholar]
  30. Mann, W. C., & Thompson, S. A. (1988). Rhetorical structure theory: Toward a functional theory of text organization. Text, 8, 243–281. [Google Scholar] [CrossRef]
  31. Martelotta, M. E. (2004). Operadores argumentativos e marcadores discursivos. In M. E. Martelotta, S. J. Votre, & M. M. Cezário (Eds.), Gramaticalização (pp. 325–344). Faculdade de Letras da UFRJ. [Google Scholar]
  32. Martín Zorraquino, M. A., & Portolés Lázaro, J. (1999). Los marcadores del discurso. In I. Bosque, & V. Demonte (Eds.), Gramática descriptiva de la lengua española (Vol. 3, pp. 4051–4213). Espasa. [Google Scholar]
  33. Narbona Jiménez, A. (1978). Las proposiciones consecutivas en español medieval. Publicaciones de la Universidad de Granada. [Google Scholar]
  34. Norde, M. (2009). Degrammaticalization. Oxford University Press. [Google Scholar]
  35. Pons Bordería, S. (2014). Paths of grammaticalization in Spanish o sea. In C. Ghezzi, & P. Molinelli (Eds.), Discourse and pragmatic markers from Latin to the Romance languages (pp. 109–136). Oxford University Press. [Google Scholar] [CrossRef]
  36. RAE/ASALE. (2009). Nueva gramática de la lengua española. Espasa-Calpe. [Google Scholar]
  37. Rodríguez Molina, J. (2014). Adverbios y locuciones adverbiales de manera. In C. Company Company (Ed.), Sintaxis histórica de la lengua española (Vol. 1, pp. 733–937). Universidad Nacional Autónoma de México/Fondo de Cultura Económica. [Google Scholar]
  38. Rodríguez Molina, J. (2015). El adverbio así en español medieval: Variantes morfofonéticas. In J. M. García Martín, T. Bastardín Candón, & M. Rivas Zancarrón (Eds.), Actas del IX congreso internacional de historia de la lengua Española (pp. 1049–1064). Iberoamericana/Vervuert. [Google Scholar]
  39. Santos, M. P. K. d., & Silva, T. d. S. (2019). Conectores advindos da construção [XQUE]conect no português: Uma visão construcional de mudança. LaborHistórico, 5, 53–69. [Google Scholar] [CrossRef]
  40. Silveira, D. M. (2014). Clivadas e pseudo-clivadas na história do português: Uma análise diacrônica das estruturas de foco e implicações da gramática V2 [Master’s thesis, Universidade Estadual de Campinas]. [Google Scholar]
  41. Souza, E. R. F. d. (2009). Gramaticalização dos itens linguísticos assim, já e no português brasileiro [Doctoral dissertation, Universidade Estadual de Campinas]. [Google Scholar]
  42. Stede, M., Taboada, M., & Das, D. (2017). Annotation guidelines for rhetorical structure. University of Potsdam and Simon Fraser University. Available online: https://www.sfu.ca/%7Emtaboada/docs/research/RST_Annotation_Guidelines.pdf (accessed on 20 March 2025).
  43. Traugott, E. C. (1995). Subjectification in grammaticalisation. In D. Stein, & S. Wright (Eds.), Subjectivity and subjectivisation (pp. 31–54). Cambridge University Press. [Google Scholar]
  44. Traugott, E. C., & König, E. (1991). The semantics-pragmatics of grammaticalization revisited. In E. C. Traugott, & B. Heine (Eds.), Approaches to grammaticalization (Vol. 1, pp. 189–218). John Benjamins. [Google Scholar]
  45. Traugott, E. C., & Trousdale, G. (2013). Constructionalization and constructional changes. Oxford University Press. [Google Scholar]
  46. Williams, E. B. (1962). From Latin to Portuguese: Historical phonology and morphology of the Portuguese language (2nd ed.). Philadelphia University Press. [Google Scholar]
  47. Wischer, I. (2000). Grammaticalization versus lexicalization. In O. Fischer, A. Rosenbach, & D. Stein (Eds.), Pathways of change: Grammaticalization in English (pp. 355–370). John Benjamins. [Google Scholar] [CrossRef]
Figure 1. Distribution of sentences forming the database per century.
Figure 1. Distribution of sentences forming the database per century.
Languages 11 00057 g001
Figure 2. Proportion of values of assim as part of assi(m)… que as an adverb, conjunction or discourse marker during Old and Middle Portuguese.
Figure 2. Proportion of values of assim as part of assi(m)… que as an adverb, conjunction or discourse marker during Old and Middle Portuguese.
Languages 11 00057 g002
Figure 3. Proportion of semantic/pragmatic categories expressed by assim que during Old and Middle Portuguese.
Figure 3. Proportion of semantic/pragmatic categories expressed by assim que during Old and Middle Portuguese.
Languages 11 00057 g003
Figure 4. Proportion of clefts with assim… que during Old and Middle Portuguese.
Figure 4. Proportion of clefts with assim… que during Old and Middle Portuguese.
Languages 11 00057 g004
Figure 5. Proportion of values of assim as part of assi(m)… que as an adverb, conjunction or discourse marker during Classical and Modern Portuguese.
Figure 5. Proportion of values of assim as part of assi(m)… que as an adverb, conjunction or discourse marker during Classical and Modern Portuguese.
Languages 11 00057 g005
Figure 6. Proportion of semantic/pragmatic categories expressed by assim que during Classical and Modern Portuguese.
Figure 6. Proportion of semantic/pragmatic categories expressed by assim que during Classical and Modern Portuguese.
Languages 11 00057 g006
Figure 7. Proportion of clefts with assim… que during Classical and Modern Portuguese.
Figure 7. Proportion of clefts with assim… que during Classical and Modern Portuguese.
Languages 11 00057 g007
Figure 8. Proportion of values of assim as part of assi(m)…que as an adverb, conjunction or discourse marker in the history of Portuguese.
Figure 8. Proportion of values of assim as part of assi(m)…que as an adverb, conjunction or discourse marker in the history of Portuguese.
Languages 11 00057 g008
Table 1. Classification into semantic/pragmatic links (relations and nexa).
Table 1. Classification into semantic/pragmatic links (relations and nexa).
Nexus RelationExample in English
Cause/consequenceCause[Mayer Jochen Wagner reacted with surprise, too.]N [After all, just on Monday the community council had agreed to expand the village of Diepensee.]S
Condition[If the sanitary facilities are still not available in the coming season,]S [Radewege is in danger of losing the competition for attracting boats to the campground.]N
Enablement[Replace the spark plugs.]N [A square key wrench can be found right under the cap.]S
Evidence[And now even our state government seems determined to remove this apparent equality between the two subjects.]N [Stolpe, Reiche and others do say Yes to the possible compromise offered by the Karlsruhe court, but they also decree: There cannot be any voluntary subject area LER/Religion.]S
Justify[I will never eat this cake]N [because I have diabetes.]S (CE)
Purpose[In order to protect their troops,]S [the U.S. do not comment on news reports on the start of a ground operation in Afghanistan.]N
Reason[With each new day of air raids, the military operations of the U.S. lose more and more credibility.]N [By means of comprehensive area-wide destruction you can’t hit the Taliban, nor can you eliminate bin Laden.]S
Result[It was raining heavily;]N [therefore, we stayed indoors.]S (CE)
Solutionhood[With the anti-smoker regulations being passed, many pubs will be caught in a trap.]S [They should start looking into possibilities for having separate rooms.]N
ConductionPreparation[One thing is evident:]S [With each day of air raids, the mili- tary operations of the U.S. lose more and more credibility.]N
Summary[The project faced many delays, budget issues, and staffing problems.]N [In short, it did not succeed.]S (CE)
Restatement[The mayor gave all the information to the councillors,]N [kind of filling them in completely.]S
ComparisonAntithesis[At one point they seemed to have succeeded.]S [But the buyer didn’t pay.]N
Contrast[My first car was small.]N [The second was already a sizable
limousine.]N
Concession[Sanitary facilities nowadays are a standard on big campgrounds.]S [But Radewege has a hard time with upgrading to such standards.]N
ElaborationBackground[Until 1984, Burkina Faso was called Obervolta.]S [According to an EMNID poll, many Europeans today believe that they are two different countries.]N
Circumstance[When Veag came under pressure because of the deregulation of the electricity market,]S [they compensated for this by squeezing their suppliers.]N
Elaboration[Diepensee will relocate.]N [No question about that.]S
InterpretationEvaluation[Its past seemed to shadow the big hotel like a curse.]S [For many years the Potsdam local court had tried to find a buyer, without success.]N
Interpretation[Now for the first time we bought a Cadillac.]N [That probably
makes us part of the bourgeoisie.]S
Motivation[Running the Olympic Games are a win-win situation for any city.]S [Berlin must now apply for the 2026 games!]N
SequenceSequence[At nine o’clock the teacher entered the classroom.]N [Five minutes later she announced that a test will be written.]N
Temporal
(succession)
[She had left her house in a hurry,]S [and then she arrived at the station.]N (CE)
Subjective/
Dialogical
Hesitation[We could approve the proposal today.]N [Well… I’m not sure if that’s the best decision.]S
Inference[All the shops were closed, and the streets were empty.]S [It must have been a public holiday.]N
Topic-shift[We’ve finished discussing the budget.]S [Now, let’s talk about next quarter’s marketing strategy.]N
Source of the examples: Stede et al. (2017), except those marked CE (=constructed example).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Andrade, A.L.d.; Oliveira, G.M. The Back-and-Forth of assim que in the History of Portuguese. Languages 2026, 11, 57. https://doi.org/10.3390/languages11030057

AMA Style

Andrade ALd, Oliveira GM. The Back-and-Forth of assim que in the History of Portuguese. Languages. 2026; 11(3):57. https://doi.org/10.3390/languages11030057

Chicago/Turabian Style

Andrade, Aroldo Leal de, and Glayson Martins Oliveira. 2026. "The Back-and-Forth of assim que in the History of Portuguese" Languages 11, no. 3: 57. https://doi.org/10.3390/languages11030057

APA Style

Andrade, A. L. d., & Oliveira, G. M. (2026). The Back-and-Forth of assim que in the History of Portuguese. Languages, 11(3), 57. https://doi.org/10.3390/languages11030057

Article Metrics

Back to TopTop