Introduction to the Special Issue: Investigating Language Variation and Change in Portuguese

1. This Special Issue

This Issue focuses on morphosyntactic variation and change in Portuguese. It features twelve papers based on data from multiple corpora of Portuguese, representing different regional and historical varieties: (contemporary) Angolan, Brazilian, European, Mozambican, Santomean, and Uruguayan Portuguese, as well as Medieval and Classical Portuguese. Several of these varieties have been underrepresented in the literature. This broad coverage makes it possible to showcase both the range of variation and the shared features of Portuguese. The papers adopt various theoretical approaches and use a range of current methods. However, there are also some common trends and connections among the presented studies, which this collection aims to bring to the fore. In this introduction, I highlight those connections and provide some context on the contribution of the papers to the study of Portuguese and more broadly to linguistic theories.

In the preceding decades, several reference works on Portuguese have been published, providing overviews of the language’s subsystems and including descriptions of grammatical structures that display variation (Wetzels et al. 2016; Carrilho and Martins 2016). Some studies focus on specific varieties (e.g., Gonçalves 2010; Bouchard 2017), examine certain phenomena from a variationist perspective (e.g., Malvar and Poplack 2008; Scherre et al. 2018), or highlight the role of pragmatic factors constraining the use of variable structures (e.g., Schwenter and Silva 2002; Posio 2021). However, there are few entire works devoted to variation in Portuguese—a notable exception being the work of (Barbosa et al. 2017). More broadly, there is a need to consider Portuguese data to examine the connection between language variation and change, and to reflect upon the implications of data from Portuguese for theories of syntactic and semantic change. This Special Issue aims to fill this gap.

An additional goal of this volume is to showcase research on topics that have been underexplored in the literature on Portuguese. A case in point is the study of semantic change, which has received significantly less attention than phonetic or morphosyntactic change in the Portuguese context. This volume includes several studies focusing on changes in meaning: these studies examine the change undergone by intensifiers, adversative connectives, free-choice items, and items conveying expressive meaning. All these categories have been understudied in Portuguese, both in terms of synchrony and diachrony. From the perspective of variation, topics that have not received much attention in the literature on Portuguese and are examined in this volume include prepositional accusatives (more broadly known within the phenomenon of differential object marking), variation involving morphological paradigms (such as the case of the second-person singular form of the Preterit—Pretérito Perfeito Simples—and the two forms of past participles), variation in the domain of tense and aspect (including temporal adjuncts of duration), and variants of locative relative clauses.

It is also worth mentioning that this volume showcases innovative methods in linguistics. The studies presented examine data from new corpora of private correspondence, corpora of various genres, and data from social media. As for the new methodological approaches represented in this volume,1 there are several corpus-based studies using distributional approaches and a range of statistical analyses.

From a theoretical perspective, there are some common threads across the papers in this volume. Some articles (e.g., Schwenter et al., Dickinson) highlight the importance of considering lexical factors in phenomena of variation and change, a trend found in other research on Romance, and more developed within studies using a constructional or usage-based approach. Other articles show the different ways in which analogy plays a role in variation and change, an active area of research in studies of language change. The research in the following papers reveals how analogical processes are constrained by various types of language-internal factors: Dickinson, Guilherme, and Delicado Cantero and Amaral. Another set of articles explores the mechanisms of and motivations for semantic change: these are authored by Livio and Howe, Müller, and Delicado Cantero and Amaral. In addition, some articles identify factors at the semantics–pragmatics interface that affect morphosyntactic phenomena, e.g., the research of Hagemeijer et al., and Pinto.

Several papers in this Special Issue highlight similarities and differences between varieties of Portuguese that had not been previously noticed, e.g., in the domain of intensifiers, relative constructions, and the variation between the first-person plural pronouns a gente and nós. Other papers contribute to improving the characterization of previous stages in the history of Portuguese or the identification of new stages. All in all, the studies cover multiple grammatical interfaces, e.g., between syntax and information structure, and between morphosyntax and pragmatics.

In the following, I provide a summary of the papers, starting with those that focus on variation.

2. Summary of the Contributions to This Special Issue

Tjerk Hagemeijer, Rita Gonçalves, and Nélia Alexandre examine variation in relative constructions denoting spatial location, i.e., the use of onde ‘where’ vs. que ‘that’, in spoken corpora of urban varieties of Angolan, Mozambican, and Santomean Portuguese. Their paper shows that these varieties are rather homogeneous in this domain, mainly displaying pied-piping as the main mechanism, and secondarily showing P(reposition)-chopping. As for the examined variants, there are some differences among the three African varieties. Overall, onde tends to be used with locatives and areas, but Angolan Portuguese also uses onde for goals/sources, and Santomean Portuguese displays more variation between both relative markers in terms of areas and containers. Among the semantic factors analyzed, it was found that antecedents that are [-definite] and display less specificity tend to favor the relativizer que in the three varieties, and especially so in Santomean Portuguese. The authors note that, despite the variation described in this study for these African varieties, the main patterns identified are also found in European Portuguese. This study paves the way for more research on the semantic properties of head nouns as well as research on relative constructions using experimental methods.

In a study comparing European and Brazilian Portuguese, Diana Santos examines the variation between temporal adverbials of durationcontaining por, ‘for’, and durante, ‘during’. Using corpora data from the Linguateca resource encompassing different genres (newspaper texts, literary fiction, and oral interviews), the author first shows the overall differences in frequency across varieties, with durante being more frequent than por in European Portuguese, and por being more frequently used than durante in Brazilian Portuguese. Importantly, this paper is the first to empirically test previous claims—based on introspection—about the distribution of these temporal adverbials and to provide a quantitative analysis of corpus data. This study delves into their distributional properties and the sensitivity of duration adverbials to semantic and pragmatic factors. Por and durante show distinct preferences for aspectual classes of verbs: por may co-occur with achievement predicates, and in such cases measures the duration of the result of the event, while durante occurs with states and activities. In addition, por and durante are sensitive to semantic notions like duration length, with por being used more often for short durations. This is reflected both in the co-occurrence of por with nouns denoting short time periods (as determined by their semantic value or by a combination of their encoded meaning and world knowledge) and in its compatibility with the copula verbs ser and estar. The author also discusses the patterns reported in the literature with respect to a preference of por for the expression of planned duration2, but finds no evidence for a categorical choice in such contexts. An interesting connection identified by this article is the relation between these adverbials and the reinforcement of negation through what could be called temporal minimizers, like um momento, ‘one moment’, or um instante, ‘an instant’, under the scope of negation (e.g., nem por um instante, ‘not even for an instant’).

Another paper that uses corpus data to analyze variation is Scott Schwenter, Lauren Miranda, Ileana Pérez, and Victoria Cataloni’s study of agreement with the pronoun tu in Brazilian Portuguese. Unlike previous studies, which used sociolinguistic interviews to study this phenomenon, the authors examine almost 5000 tokens from an online megacorpus of BP that contains data from multiple varieties in terms of region and register. Their study finds that the most common pattern is the co-occurrence of the pronoun tu with a verb form in the third-person singular (e.g., tu é), with agreement in person between pronoun and verb form (e.g., tu és) being infrequent. The statistical analyses performed show that this type of morphosyntactic variation is mostly constrained by lexical factors, with certain verbs showing a strong tendency to lack second-person singular agreement, while other verbs tend to display agreement. The authors also identify “significant constructional effects”, i.e., patterns of variation that arise in combination with sequences involving tense/aspect and auxiliary verbs. Not only do these results reveal an overwhelming tendency to mark person and number features through personal pronouns rather than verbal inflection in BP, but they also confirm a diachronic tendency in this variety to erode verbal morphology.

The paper by Cíntia Pacheco, Ana Carvalho, and Marta Pereira Scherre also focuses on morphosyntactic variation and the pronominal system of Portuguese. This study adopts the approach of comparative sociolinguistics to investigate the variation between the first-person plural pronouns a gente and nós in Uruguayan Portuguese (a variety of Portuguese spoken in northern Uruguay by Spanish–Portuguese bilinguals) and the neighboring monolingual BP variety. The data come from sociolinguistic interviews conducted in two communities at the southern border of Brazil: Aceguá, Brazil, and Aceguá, Uruguay. This study finds that while both pronominal forms are used by bilingual speakers, the form a gente is used less frequently in Uruguayan Portuguese than in Brazilian Portuguese. Nevertheless, the variationist analysis shows that the grammaticalized form a gente is favored in both linguistic systems by the same linguistic constraints. While the focus of this Special Issue is not on language contact, this study shows how certain linguistic factors (e.g., verb tense, the phonic salience of a verb form, discourse persistence of a variant) regulate the use of this pronominal variable, both for monolingual and bilingual speakers.

The papers by Kendra Dickinson and Ana Rita Guilherme look at morphosyntactic variation and specifically discuss analogy and the role of language-internal factors in language variation and change. Kendra Dickinson investigates the variation in past participle forms (i.e., the co-existence of irregular participles and regularized ones, or particípios duplos, e.g., aberto and abrido for the verb abrir ‘open’) in contemporary Brazilian Portuguese. Adopting a usage-based approach, the author examines the role of analogy in the regularization of participial forms, which has had a lasting effect on the history of the language. Using Sketch Engine’s 2011 Portuguese Web Corpus (ptTenTen11), a corpus of 4.6 billion words, the author finds a set of language-internal constraints on the analogical processes affecting participles, whereby certain constructions provide conditioning environments. She finds that Perfect tenses (e.g., the Pretérito Perfeito Composto) promote the regularization of Latinate irregular past participles and are a conserving environment for regular past participles that compete with short-term forms. On the other hand, short-form participles (i.e., innovative forms) tend to occur in adjectival contexts with ser and estar. In addition, some of the variation seems to be lexically constrained, which aligns with other findings in this Issue (cf. Schwenter et al.). More broadly, this paper sheds light on how usage patterns and system-internal factors affect analogy, an important mechanism of language change.

Another paper that highlights the role of analogy in language change is the study by Ana Rita Guilherme. The author focuses on the variation in the form of the second-person singular simple past (Pretérito Perfeito Simples) indicative, e.g., estiveste vs. estivestes ‘you were’, which has been attested since the 18th century in the history of European Portuguese. Using two corpora of private correspondence from the 16th to the 20th centuries (Post Scriptum—PS–and Forgotten Letters—Fly), the author traces the process of the analogical extension of the -s form. Guilherme argues that both morphological and pragmatic factors should be taken into account to explain this analogical extension throughout the history of the language. She presents the following factors: (i) regularization based on other second-person singular forms; (ii) ambiguity and variation within verbal paradigms in Medieval and Classical Portuguese as a consequence of the deletion of the suffix -d (-de/-des) of the second-person plural form in intervocalic contexts; and (iii) changes in the Portuguese system of address forms, involving competition between the pronominal forms tu and vós and the verbal inflection associated with them. The research by Guilherme straddles the study of variation and the study of change. In the following, I summarize the articles that focus more on change.

Lukas Müller investigates non-adverbial uses of lá in European Portuguese (e.g., Acaba lá de comer a sopa! ‘(Just) eat the soup!’) and their diachrony. Adopting the framework of multidimensional semantics (Gutzmann 2013, 2015), the author proposes that these uses of lá do not contribute to the truth conditions of a sentence, and hence should be analyzed as use-conditional. Specifically, in contemporary EP, use-conditional lá contributes illocutionary modification. Using data from the Corpus do Português, the author provides a diachronic study of semantic change, showing that, over time, the adverb conveying spatial deictic meaning underwent a process of pragmaticalization (Traugott 1989). This process yields a type shift, changing from contributing to truth conditions to contributing to use conditions: lá_TC > lá_UC. The author proposes a three-stage diachronic path whereby an implicature of illocutionary modification becomes conventionalized over time, creating a hybrid, two-dimensional item, which is then split into two homophonous items, yielding the contemporary situation of polysemy found in EP. Importantly, this is one of the few studies that analyzes semantic change in Portuguese adopting a multidimensional semantics approach, and more broadly focusing on the emergence of expressive meaning in Portuguese.

Another study of semantic change in European Portuguese is the paper by Manuel Delicado Cantero and Patrícia Amaral. This paper traces the syntactic and semantic shifts undergone by the PP ao passo que, ‘at the step/pace that’. This originally appeared with a relative clause modifying the noun passo and denoted a comparison in pace. It then became a temporal expression denoting a simultaneous gradual change in two dynamic eventualities, and eventually became a contrastive connective. Using data from the Corpus do Português, the authors identify contexts in which the expression conveys both temporal–proportional meaning and a contrast between two events. Over time, this led to a broadening of the aspectual classes that were compatible with the expression as it came to encode just contrast (hence losing aspectual restrictions connected to its source meaning). In addition, the authors examine the internal structure of the expression and the meaning of the noun passo, ‘step’, and discuss how these data can help adjudicate between competing theories regarding the nature of complex categories. This paper fills a gap in our understanding of the diachrony of the expression ao passo que. Overall, there is a lack of studies on the diachrony of contrastive connectives in Portuguese (an exception being the diachronic study of adversatives in the work of Amaral et al. 2023).

Camila Livio and Chad Howe use a Digital Humanities approach to study the interplay between variation and change in the system of degree modifiers in Portuguese. Specifically, they mine Twitter (now X) data to investigate the distribution and meaning of the intensifier bué, possibly a borrowing from Kimbundu, in Angolan Portuguese. Their analysis shows the widespread use of bué, with functions of intensification and quantification. In a broader comparison among varieties of Portuguese with respect to the expression of intensification, the authors argue that the usage patterns of bué correspond to the use of bem (vs. muito) in Brazilian Portuguese. Their study also examines the use of bué in European Portuguese, a borrowing resulting from a well-attested motivation for change, which is innovation and expressivity in the domain of intensification. In this variety, bué has undergone grammaticalization and nowadays co-occurs with both gradable and non-gradable predicates. This paper adopts a distributional approach to advance our understanding of intensification and, more broadly, degree modification in Portuguese, an area that has only recently begun to be explored (see e.g., Gomes 2011; Pinto 2020; Amaral 2020).

Clara Pinto investigates the diachrony of the free-choice item (FCI) qualquer, ‘any’, from qual, ‘which’, and a form of the verb, querer, ‘to want’, in Portuguese, an understudied topic despite the widespread interest in FCIs in the Romance languages. The author focuses on data from Old Portuguese to examine the syntactic and semantic properties of the original forms, adopting a formal syntax approach. She analyzes the different structures that appear in the medieval data and proposes that there are two different diachronic sources for qualquer, underlying the different contemporary interpretations of the word in European Portuguese. Specifically, Pinto argues for the existence of two items in the historical data: (i) prenominal qualquer displays the properties of a quantifier in the 13th century, while (ii) postnominal qualquer should rather be considered an adjectival form resulting from the reanalysis of an appositive relative clause. This paper not only advances our knowledge of FCIs in Portuguese, but it also furthers our understanding of understudied domains of the grammar of Portuguese, like the syntax and semantics of polarity items as well as the properties of determiners that defy classification (e.g., outro, ‘other’, and its cross-linguistic counterparts).

Ana Regina Calindro’s study examines the variation and change in a-marked arguments in the history of Portuguese. Overall, it is shown that a-marked accusatives (“prepositional accusatives”) are much less frequent than bare accusatives, and this is true both for European and Brazilian Portuguese. Using the Tycho Brahe Corpus of Historical Portuguese, the author analyzes texts from the 16th to the 19th century and confirms previous claims that a-marked accusatives increased in Portuguese in the 17th century, and then decreased in the 18th century. Calindro examines specific verb classes and the prepositions associated with their arguments in European and Brazilian Portuguese, noting the differences between these varieties. Finally, the author proposes an analysis of psychological verbs, where contemporary European Portuguese displays a structural marked dative corresponding to the semantic role of Experiencer (O filme agradou ao João/agradou-lhe), while Brazilian Portuguese does not display overt marking for this role (O filme agradou o João/agradou ele). First, this study provides further empirical evidence that differential object marking is a marginal phenomenon in Portuguese, unlike what happens in other Romance languages. Second, it brings to the fore the behavior of specific verb classes with respect to argument structure, a topic deserving further study using both historical and lexical semantic approaches.

In their research, Aroldo Leal d’Andrade and Lara da Silva Cardoso explore a recent hypothesis that Brazilian Portuguese may have evolved from a grammatical system, Colonial Brazilian Portuguese, which should be distinguished from Classical Portuguese. The authors focus on word order as they compare two texts from the 17th century, one written by a Brazilian author (Eusébio de Matos), and another written by a Portuguese author (Padre António Vieira). The texts are similar from both a content and a literary perspective, but they display differences that allow Andrade and Cardoso to propose an emerging distinctiveness in the grammar of their authors. Specifically, they show that in main clauses Colonial Portuguese has more V2 sentences, while Colonial Brazilian Portuguese displays a higher percentage of clauses with V1 word order. The authors also present other differences in information structure configurations with various types of constituents, arguing that there is less variation in the types of marked constructions in Colonial Brazilian Portuguese than in Colonial Portuguese. Adopting a generative diachronic syntax approach, the authors conclude that the frequency differences observed in these two texts are compatible with the potential microparametric changes that have been proposed for BP and EP.

3. Looking Ahead

First and foremost, the contributions of this Special Issue highlight the need for further research on underrepresented varieties of Portuguese. Beyond those studied in this volume, I would mention, e.g., the importance of learning more about Judeo-Portuguese. Second, the articles in this Issue open the way to research on domains like the syntax–semantics of degree expressions, polarity, indefinites, and tense and aspect, which have been generally understudied in Portuguese (and usually limited to descriptions of European or Brazilian Portuguese). In this respect, the range of methodologies used to investigate these phenomena in this Special Issue are a good indication of the potential and vitality of the field. Third, much work remains to be conducted on the history of Portuguese. From the point of view of data, it would be important to expand diachronic corpus data to include genres and texts that go well beyond canonical and literary sources. Having a broader range of texts would allow us not only to better date changes in the language but also to gain a more fine-grained understanding of factors affecting different grammatical subsystems over time. From the point of view of subjects, I would underscore the critical need for more historical and diachronic studies on the pronominal system(s) of Portuguese, the evolution of temporal and aspectual categories, and the modal system (including both the verbal and the adverbial domain).