THE DEVELOPMENT OF THE MODALIZED EXPRESSION PODE SER. A CASE OF (INTER)SUBJECTIFICATION IN PORTUGUESE

Abstracts

Based on a functionalist approach, this paper analyzes the modalized expression pode ser as a complement-taking predicate which embeds a proposition (pode ser₁) and as an independent structure (pode ser₂), in contemporary written and spoken Brazilian Portuguese texts. We aim to identify degrees of (inter)subjectivity, revealing a process of (inter)subjectification (TRAUGOTT, 2010 among others). The analysis carried out in this paper is supported by parameters of (inter)subjectivity of modal elements (TRAUGOTT; DASHER, 2002) and by the notion of modality as a multifunctional category, serving not only to encodethe speaker’s attitude regarding the modalized content, but also as a pragmatic strategy, as a regulator of communicative situation. The exam reveals pode ser as a strongly demanded structure in interaction, a fairly requested set and also productive and useful for interpersonal relationships. The examination of semantic, discursive and morphosyntactic properties indicates a shift from syntax (pode ser₁) to discourse (pode ser₂), interpreted as a development of (inter)subjectification.

Modalization; Pode ser; (Inter)subjectification; Functionalism

Comprometido com um enfoque funcionalista, o trabalho analisa a expressão modalizadora pode ser como predicado encaixador de proposição (pode ser₁) e como construção independente (pode ser₂), em textos contemporâneos do português brasileiro, de fala e de escrita. Busca-se identificar graus de (inter)subjetividade reveladores de um processo de (inter)subjetivização, conforme proposta de Traugott (2010). Sustentam a análise parâmetros de (inter)subjetividade de elementos modalizadores indicados especialmente em Traugott e Dasher (2002) e a noção de modalidade como categoria multifuncional, que não apenas codifica atitude do falante em relação ao conteúdo modalizado, mas que também atua como estratégia pragmática, como reguladora da situação comunicativa. A pesquisa revela pode ser como uma fórmula de grande aproveitamento no jogo discursivo, um conjunto bastante solicitado, produtivo e útil nas relações interpessoais. O estudo de propriedades semânticas, discursivas e morfossintáticas indica deslizamento na operação da construção, da sintaxe (pode ser₁) para o discurso (pode ser₂), interpretado como próprio de (inter)subjetivização.

Modalização; Pode ser; (Inter)subjetivização; Funcionalismo

Introduction

The modalized expression pode ser (may be) is largely used in interpersonal relationships. Native speakers of contemporary Portuguese commonly use pode ser in sentences like (1) and (2)¹, among others.

(1) Doc – pra fora você diz do Brasil?

F2 – pro exterior épode serque eu não tenha tido ainda condição financeira de ir pra fora (NURC/ RJ D2 158)

‘Doc – you mean outside Brazil?

F2 – abroad it’s possible that I haven’t had enough money to travel abroad yet.’

(2)— Você está querendo provar que Deus é justo.

O estrangeiro pensou um pouco.

—Pode ser. (DSP)

‘— You are trying to prove that God is fair.

The foreigner thought a bit.

— May be. ‘

In sentences like these, pode ser is a modal marker, used by the speaker² to show his epistemic stance. In both cases, it signals the speaker is uncommitted to the truth of the propositional content which is within its scope. This propositional content is, then, evaluated by the speaker as an eventuality, as a possibility. In (1), pode ser is a complement-taking predicate which embeds a proposition (pode ser₁) and specifies the speaker attitude towards the embedded propositional content. In (2), pode ser is a syntactically, illocutionally and pragmatically independent construction (pode ser₂), which is used as a complete attitude speech act with scope over the listener’s speech act. The semantic value of pode ser in sentences like (2) is similar to the one from the modal adverb talvez (maybe).

The development of other modality constructions, such as (eu)acho (que) (I think (that)) (GALVÃO, 1999), parece (que) (it seems (that)) (GONÇALVES, 2003), in which the more independent constructions are newer than the complement-taking ones, suggests pode ser₂ is a stage of development of pode ser₁ in which the matrix sentence is not linked to the embedded sentence anymore, becoming an independent marker of modality which is largely used in discourse. Both types of constructions cited above are known as cases of grammaticalization; in the same way, a grammaticalization process may be identified in pode ser.³

Verifying the grammaticalization hypothesis of pode ser is an interesting research direction.⁴ However, it is not the aim of this pape⁵, which investigates the (inter)subjectification paths (TRAUGOTT, 2010; TRAUGOTT; DASHER, 2002; among others) present in contexts with pode ser which explain the development of this construction. More specifically, in this paper, we analyze the syntactic, semantic and discursive properties of the construction in both syntactic contexts considered (complement-taking predicate which embeds a proposition and independent construction), which indicate an (inter)subjectification process in pode ser.

This research is based on a functionalist approach without following any specific functionalist theory, in the same way as Neves (2006, 2011). According to a functionalist perspective, the forms of a language are means to an end and not an end by themselves and, then, structures should be explained by their uses. The analysis carried out here includes not only the linguistic structure, but also the communicative situation, the aim of the speech event, its participants and the discourse content.

The next section presents the methodology of this research. The second and third sections show the concept of modality considering its multifunctionality and a typology of this qualification. The following sections present the notions of (inter)subjectivity and (inter)subjectification, as proposed by Traugott, the analysis of pode ser and the conclusions.

Methodology

The corpus analyzed in this research constitutes of spoken and written texts in contemporary Brazilian Portuguese. In order to build a sample which represents a wide range of Brazilian Portuguese sentences, first of all we used the database from Lexicography Lab (LabLex), available at UNESP (in the city of Araraquara). We used the so-called main corpus 2 (CP2), which contains written texts produced between 1950 and 2000 belonging to different text types (dramatic, oratory, technical and journalistic texts, adds), comprising 15,378,424 words.

After describing the data of written language, we started analyzing spoken language texts. We selected 60 samples from the NURC (Corpus of the Spoken Portuguese Grammar Project) database from five state capitals from Brazil (Porto Alegre, Rio de Janeiro, São Paulo, Recife and Salvador), belonging to three different types (formal elocution (EF), dialogues between the speaker and the interviewer (DID) and dialogues between two speakers (D2)). Spoken language texts comprise 392,510 words.

Occurrences of pode ser₁ and pode ser₂ were searched in the selected texts. In order to control variables, some programs of GOLDVARB computational suite were used, especially the ones that measure frequency and cross-tabulation of factors of analysis. In the discussion presented in this paper, the following variables were verified by means of cross-tabulation carried out by GOLDVARB: type of pode ser (pode ser₁, pode ser₂), modal values, mood and temporal flexion of the embedded clause (only for pode ser₁) and discourse type (same person speech, interlocution).

Modality

Due to the complexity of the theme, the investigation of modality is always a hard task. Among so many metaphors researchers have used in order to explain how difficult it is to understand this category, the one made by Perkins (1983) is worth mentioning: researching modality is like walking in a crowded room without stepping on anyone else’s feet.

First of all, defining modality is not easy, since there is no consensus on its conceptualization. Besides, in despite of the large amount of researches on the topic, the modalization which operates on discourse level is not argued accordingly and this is the theoretical difficulty for the study of pode ser, as we discuss below.

First considerations about modality derive from Aristotelian studies, which propose some modal notions of possibility and necessity and the idea that these concepts are interdefined using negation. In the tradition of logical analysis, modalities are characterized in terms of truth relations, regardless of speaker and, for that reason, the relations established in logic do not hold in linguistic studies. However, as observed by Neves (2006), linguistic studies are very influenced by the logical perspective and hardly ever the concepts of possibility and necessity are absent in the definition of modality.

Another way to define modality is by means of factuality (LYONS, 1977; NARROG, 2005), which is also referred to as realis/irrealis (PALMER, 1986) or validity (KIEFER, 1987). According to Narrog (2005), factuality is the only criterion which distinguishes structures that belong to the category of modality from those that do not belong. In general, modality is characterized by non-factuality, that is, it does not depend on factual status, which means that a modalized utterance is neither positively nor negatively factual. However, as noted by Palmer (1986), factual statements, like the ones which express speaker’s opinion (non-factual), are subjective, representing the speaker’s point of view. Therefore, restricting a modality study to non-factuality is not appropriate.

We assume, in this study, that subjectivity is a fundamental criterion to define modality. The definition based on this criterion, in general, is highly accepted and is differently expressed in scholars’ definitions: “[…] the speaker’s opinion or attitude towards the proposition that the sentence expresses or the situation that the proposition describes” (LYONS, 1977, p.436); “[…] the manner in which the meaning of a clause is qualified so as to reflect the speaker’s judgment of the likelihood of the proposition it expresses being true” (QUIRK et al., 1985, p.219); among others.

This way to understand modality is related to the traditional idea of subdividing the utterance in two parts: dictum (proposition, descriptive part) and modus (non-proposition, modality, evaluation of the speaker towards his verbalization) (LYONS, 1977). This means that every utterance which has proposition consists of modus and dictum, and those which are not explicitly modalized have an evaluation operator in their semantic structure.

Although the notion of subjectivity is in the essence of modality, which involves the expression of something which is part of speaker’s knowledge, this understanding does not imply that this category is used in certain speech situations; this is another aspect to be considered. Modality expresses speaker’s participation in communication, which is carried out under some interactional pressures and with some communicative aims or some discursive effects. Especially in conversation, modality can regulate verbal interaction, and this aspect is not considered in the definitions of modality as the ones presented above. The definition of modality as speaker’s attitude or opinion towards the propositional content of his utterance has a short range, after all, modality:

1. is not only speaker’s attitude/opinion towards his utterance, it can also be his attitude/opinion towards listener’s utterance;

2. is not necessarily oriented to propositional content;

3. can be highly linked to discursive domain in some utterances, and, therefore, in these cases, its main function is not expressing speaker’s position, but regulating verbal interaction.

Modality expressed by means of pode ser₂, such as the occurrence presented in (2), above, is an example of what is said in (a), since in these cases the speaker uses the construction in order to express an epistemic position in relation to the listener’s utterance (in (2), você está querendo provar que Deus é justo (you are trying to prove that God is fair)).

In (3), below, another occurrence of pode ser₂ is presented:

(3)— Você é virgem, Severina?

— Virgem nasci / Virgem me criei / Se comigo não casares / Virgem morrerei – parodiava Severina, inspirada.

Flodoaldo ria, coçava-lhe a vaidade:

— Você é vida como o diabo, Severina!

E ia trepando como gato no cio aquele muro escuro e sólido de resistência. Adoçava a voz.

—Pode serou está difícil, coração? (CR)

‘— Are you virgin, Severina?

— Virgin I was born / Virgin I grew up / If you don’t marry me / Virgin I’ll die – Severina, inspired, parodied.

Flodoaldo laughed, conceitedly:

— You are life as the devil, Severina!

And he was climbing up the dark and solid wall of the house like a cat in rut.

— It can be or it’s hard, sweetheart?’

Pode ser₂, in (3), is oriented to a content which is not linguistically expressed, but is implied in the communicative situation, and can be recovered by the interactional context (“você se deitar comigo” (“you make love with me”)). Modalization, in these cases, represents a pragmatic strategy of attenuation. The speaker knows that the listener resists his intentions, and modalization is one of the resources the speaker uses in order to weaken, or even to avoid, a reaction opposite to what he wants. Thus, modality in (3) is mainly a persuasion strategy, a way to regulate verbal interaction. Consequently, both statements (b) and (c) about modality are justified.

Hoffnagel’s (1997) and Miranda’s (2000) considerations support the treatment of modality as regulator of the verbal interaction.

After analyzing the epistemic modality in spoken Brazilian Portuguese, Hoffnagel (1997) highlights the intersubjective status of modalization: “[…] besides expressing the speaker’s attitudes towards propositions, epistemic modalizers are used, as well, to communicate speaker’s attitudes towards listeners.” (HOFFNAGEL, 1997, p.150). Hoffnagel (1997) analyzes two pragmatic functions of modality. The first one consists of indicating lack of adhesion in relation to the truth of the propositions as part of a facilitation strategy of a more open discussion. The other function is to weaken the utterance strength in order to protect the speaker’s and listener’s faces, when the subject is sensitive.

Miranda (2000) is completely against the understanding of modality as a property of the utterance. According to the author, whose studies are based on a sociocognitive approach, modality should be considered as the speaker’s intention or attitude towards enunciation (not towards proposition or utterance). Modality is an operator which operates on domains within the communicative scene, it “regulates the interaction” (MIRANDA, 2000, p.144), that is, it signals the face-work, regulates the actions of the participants in interaction and the force dynamics established during communication. Modality is understood as a semiosis of the face, since in interaction there is a process of face-protection and face-threatening, and modality, a linguistic category, marks this dramatic construction. In this communicative conflict, this category is an operator of exertion of force/removal of blockage of the expression of force (TALMY, 1988).

A proper treatment of modality should consider its role in text processing, in interaction, or communicative “tension”, which occurs between interlocutors. In this study, we propose an analysis of modality considering two directions, which are different but not dichotomic:

• speaker-oriented modality: refers to what the speaker has in mind (attitudes, beliefs, judgments, used to encode experience, to specify attitude towards the validity of the modalized content in the content level.

• listener-oriented modality: indicates speaker’s communicative strategy, used to regulate verbal interaction, to influence interlocutors.

The first understanding is related to the more traditional way to consider modality, that is, its definition as the speaker’s attitude or opinion towards his utterance (LYONS, 1977; QUIRK et al., 1985, among others). In this case, modality indicates subjectivity.

The listener-oriented modality has the function highlighted by Hoffnagel (1997) and Miranda (2000), that is, regulating the communicative situation, indicating face-saving and face-protection processes (BROWN, LEVINSON, 1987). In this case, the intersubjective status of modality is emphasized.

We understand that both orientations of modality correspond to Halliday’s (1970) ideational and interpersonal functions, respectively, which do not exclude but complement each other. Sentences are built simultaneously as message as well as interaction event and “[…] there is a semantic region where the two functions, the ideational and the interpersonal, overlap, that of speaker’s commentary on the content.” (HALLIDAY, 1970, p. 349).

Both functions are like extreme points of a continuum, within which there are no strong boundaries between the ideational and the interpersonal domains. In other words, there is no dichotomic relation between both functions, although one of them may prevail in certain context.

In the ideational function, language is used to express content related to the speaker’s experiences, in real or mental world (HALLIDAY, 1970). Neves (1997) says that, in the ideational component, experiences from the mental world may involve reactions, cognitions, perceptions, as well as linguistic acts of speaking and understanding.

In the interpersonal function, on the other hand, language is used as a means of participating in speech situation, that is, language is related to the establishment of social roles in communication, it is a means of establishing relations between speaker and listener (HALLIDAY, 1970). The interpersonal element, “[…] in a larger context, establishes and maintains social roles, that are, after all, inherent to language.” (NEVES, 1997, p.13).

Deontic and epistemic modalities

Although there are many typological proposals on modalities, deontic and epistemic domains, the ones important to this study, represent, in general, the basic distinction.

The deontic modality is related to behaviors and expresses values such as obligation and permission; it is characterized by the presence of a (deontic) source of authority, dynamic predicates (actions) and controller subjects. These particularities are recognized by scholars and appear differently in their studies. According to Lyons (1977), the deontic modality indicates the necessity or possibility of acts performed by morally responsible agents. Quirk et al. (1985) associate this modality to a controller subject, and Palmer (1986) relates it to an element of will.

The epistemic modality is concerned with knowledge, it refers to speaker’s (un)commitment towards a certain content, involving beliefs, judgments, opinion (LYONS, 1977; QUIRK et al., 1985; PALMER, 1986; HENGEVELD, 1988, 2004). A relevant discussion for this study is related to the distinction about the subjective epistemic modality and objective epistemic modality.

Lyons (1977) subdivides the epistemic category into objective and subjective to include the alethic modality. Defined in modal logic, the alethic modality refers to the notion of existence and involves determining the truth of propositions. The objective epistemic modality, which Lyons compares to the alethic modality, expresses known or scientifically proven knowledge. The subjective epistemic modality, related to speaker’s attitude towards the propositional content of his utterance, expresses the speaker’s opinion. A sentence such as Alfred may be unmarried (LYONS, 1977, p.797), interpreted as a subjective epistemic modality, indicates the speaker is uncertain about Alfred’s marital status and, then, “be unmarried” is a hypothetical fact. Considering the objective epistemic modality reading of the sentence above, it would indicate that, according to the speaker, there is a mathematically computable chance that Alfred is unmarried and, in this case, the speaker knows, not simply thinks or believes, that there is the possibility of Alfred being unmarried.

The distinction between objective and subjective modalization, in these terms, is maintained by Hengeveld (1988, 2004). The author explains that, in the objective epistemic modality, speaker evaluates the actuality of the state of affairs designated by a predication (HENGEVELD, 1988). To do so, the speaker compares the state of affairs designated by a predication and his knowledge about possible situations or about the actual world. The subjective epistemic modality is characterized in terms of the speaker’s commitment to the truth of the predication content he presents to be considered; the state of affairs is a speaker’s belief, who is the source of information (HENGEVELD, 1988).

Nuyts (1992, 1993, 2001) recognizes the distinction between an objective modal qualification (the one deriving from mathematically estimated results and from verifiable possibilities of a state of affairs occurrence) and a subjective one (the one deriving from subjective suppositions that a state of affairs may occur), but criticizes the distinction between objective and subjective epistemic modalities as defined by Lyons (1977) and Hengeveld (1988). According to Nuyts, the distinction is related to the different sources of information and not to different types of modality, as Lyons (1977) and Hengeveld (1988) consider. Nuyts says that every modal judgment is based on evidence and what may change is the quality of the evidence, but “[…] without evidence, no evaluation of the probability of the SoA is possible – one could then only say that one does not know.” (NUYTS, 1993, p.946).

Nuyts (1992, 1993, 2001) proposes treating the difference between objective and subjective modality in terms of subjectivity and intersubjectivity⁶ (the author prefers this term rather than objectivity). In the subjective epistemic understanding, the speaker is the source of knowledge and, in the intersubjective understanding, the source is a larger group of people, in which the speaker may be included. This way to understand the distinction between subjective and objective epistemic modality seems to be appropriate and, therefore, supports the analysis of pode ser in terms of the distinctions between the objective and subjective epistemic readings.

Subjectification, intersubjectification and modalization

In previous works, Traugott (1982, 1989, 1995a, 1995b) studied subjectification process as being linked to grammaticalization, differently from the author’s more recent studies (TRAUGOTT, 2010), in which, although a strictly relation between the processes is maintained, the boundaries between them are more well-defined.

Grammaticalization is understood as a process of linguistic change by means of which a word or a lexical expression becomes more grammatical or, if it is already grammatical, becomes a yet more grammatical form (HOPPER; TRAUGOTT, 1993). According to Traugott (1982, 1989, 1995a, 1995b), subjectification is linked to grammaticalization in a way that, when changes take place in grammaticalization, there is an increase in subjectivity, during interaction between speaker and listener.

Considering the three functional domains of language as proposed by Halliday e Hasan (1976) (ideational, which serves to denotation; textual, which refers to discourse organization; interpersonal, which is related to the speaker’s attitude), Traugott (1982) suggests that, at early stages of grammaticalization, the main path of change is the development of items which have a more propositional meaning to items with a textual or an expressive meaning, or both, in the following order: (propositional) > [(textual) > (expressive)]. The hypothesis is that change is unidirectional; for instance, the opposite direction, expressive > textual > propositional, is very unlikely in the history of any grammatical item (TRAUGOTT, 1982, 1989). Traugott (1989) proposes the following three tendencies of semantic-pragmatic changes: a) meanings situated in the external described situation > internal (evaluative/perceptual/cognitive) situation; b) meanings situated in the described external or internal situation > textual and metalinguistic situations; c) meanings tend to become increasingly situated in the speaker’s subjective belief of attitude towards the proposition. In this semantic-pragmatic path of change, meanings tend to refer less to descriptions of concrete situations and more to discursive situations, less to objective situations and more to the subjective ones. In other words, meanings show increasingly the way the speaker builds the utterance towards the listener, this is the reason why it is said there is a pragmaticalization of meaning (TRAUGOTT, 1995b).

In spite of the interrelationship, both processes are understood as independent and the strong relationship between them, as Traugott (2010) says, is suggested since grammaticalization involves recruitment of items to mark the speaker’s perspective on different factors, such as “[…] whether the situation is relativized to the speaker’s beliefs (modality, mood).” (TRAUGOTT, 2010, p.40).

Traugott and Dasher (2002) propose the distinction between “subjectivity” and “intersubjectivity”. According to the authors, semantic change follows predictable paths in languages, revealing meanings increasingly based on the speaker’s subjective and intersubjective attitude.

Traugott (2010) says the intersubjectification process follows subjectification and, by means of both mechanisms, “[…] meanings are recruited by the speaker to encode and regulate attitudes and beliefs (subjectification) […]”, and, “[…] once subjectified, may be recruited to encode meanings centred on the addressee (intersubjectification).” (TRAUGOTT, 2010, p.35). Intersubjectification, therefore, does not occur without subjectification. Thus, Traugott (2010, p.35) suggests that the meanings of the items develop following the direction:

non-/less subjective > subjective > intersubjective

Although (inter)subjectification is a historical process of change, verified over time, it is also possible to organize a synchronic cline of (inter)subjectification (TRAUGOTT, 2010), based on criteria that determine a development from less to more (inter)subjective meanings.

As a hypothesis, the processes occur in the flow of speech, in verbal interaction, by means of what Traugott and Dasher (2002) call invited inferencing, a metonymic process, largely accepted as an important mechanism which underlies processes of grammaticalization and (inter)subjectification. The notion of metonym is extended from concrete to pragmatic contexts of conversational and conventional inference. The contiguity involved in this case is based on the discourse world. An example is the development of since, analyzed by Traugott e König (1991). Originally, the preposition only established a temporal relation. In a certain moment, a causal reading was inferred from the temporal one. In some contexts, both readings are possible. With its frequent use, the causal interpretation of since becomes conventionalized and the connective starts to be used in contexts in which the only interpretation is the causal one.

The development from deontic to epistemic meanings, historically proved in languages such as English (TRAUGOTT, 1989; SWEETSER, 1990; BYBEE; PERKINS; PAGLIUCA, 1994; HEINE, 1995; TRAUGOTT; DASHER, 2002; among others), is understood as resulting from the processes of subjectification (TRAUGOTT, 1989; TRAUGOTT; DASHER, 2002) and grammaticalization.

Regarding subjectification of modal elements, the change deontic > epistemic is, by itself, a process of subjectification, since, in this path, the meanings become more centered on speaker’s attitudes and beliefs. However, the gradual status of subjectivity should be highlighted and, therefore, Traugott (1989) refers to modality in terms of “more” or “less” subjective or, yet, “weakly” or “strongly” subjective. The author says that, when epistemic meanings arise, they are weakly subjective and, as they develop, they become more strongly subjective.

In Traugott e Dasher (2002), there is a detailed analysis of modal meanings revealing subjectification. Two factors are defined for verifying distinct degrees of subjectivity: a) type of modal force (deontic, epistemic); b) source of modal force (external to the speaker – religion, laws, social rules, etc. – or internal to the speaker). Considering deontic meanings, for instance, the modal is more subjective if the deontic force is the speaker and less subjective if there is a social, a religion or any other external force implied or specified. Epistemic meanings also reveal a scale in terms of subjectivity, being more subjective when the modal qualification is based on speaker’s personal expectation or shows a speaker’s confidence and less subjective when based on a wider group expectation or if they involve a more general opinion. The distinction between objective and subjective epistemic, as proposed in the previous section, is related to this scale of subjectivity.

Intersubjective meanings, according to Traugott (2010) and Traugott and Dasher (2002), refer to the speaker’s attention to the reception of his utterance by the listener, more particularly refer to the notion of hedges, to the saving of listener’s face.

Based on the aforementioned points, we aim to identify the degrees of (inter)subjectivity of pode ser, which can be interpreted as signals of a process of (inter)subjectification.

The degree of (inter)subjectivity in pode ser

After presenting the theoretical background and the methodology of this research, this section is dedicated to the analysis of the modalized expression pode ser, emphasizing the shift of the construction towards the discourse component, understood as (inter)subjectification (among others: TRAUGOTT, 2010; TRAUGOTT; DASHER, 2002).

From the total of 357 occurrences of pode ser found in the corpus, 333 sentences were collected in the written sample and 24 in the spoken sample. The distribution of the 357 occurrences between the two syntactic contexts considered in this study is the following: 63% (226/357) of the uses of pode ser are cases of complement-taking predicate which embeds a proposition (pode ser₁), exemplified in (4), and 37% (131/357) of the occurrences are uses of pode ser as independent construction (pode ser₂), exemplified in (5) and (6).

(4) — Somos eternos, Talbo. As Vozes me contaram isto, no tempo em que eu podia ver seus corpos e seus rostos.

Talbo conhecia o Dom de sua mulher. Mas fazia muito tempo que ela não tocava no assunto. Talvez fosse o delírio.

— Mesmo assim, nenhuma vida é igual à outra. Epode serque não nos encontremos nunca mais. Preciso que você saiba que te amei a minha vida inteira. Te amei antes de te conhecer. Você é parte de mim. (BRI)

‘— We are immortal, Talbo. The voices told me that, when I could see their body and their faces.

Talbo knew his wife’s gift. But there was a long time since she didn’t talk about the subject. Maybe it was the delirium.

— Even so, no life is the same as another. And may be that we will never meet again. I need you to know that I loved you during my whole life. I loved you before I knew you. You are a part of me.’

(5) Inf – o::sim o cavalo ele naturalmente ele tem o::...ou o freio ou bridão...são duas coisas diferentes também...agora...o que usa normalmente aqui::no interior...é o freio...e o::o freio é...ahn consiste vamos dizer num metal...que entra na boca do cavalo...esse é o freio propriamente dito...agora...para manter esse metal na boca do cavalo...existe::uma cabeçada...feita de couro...que muitas vezes é chamada em conjunto com o freio de freio...mas quando se quer distinguir...o::...aquilo que...o ferro...que entra na boca do cavalo...do restante do freio...então usa-se a palavra cabeçada...a cabeçada por sua vez tem::também algumas partes...porque tem uma que vem...logo::na frente do::...da cabeça do cavalo...que se chama::...cabeção...depois tem uma que vai mais em cima da própria cabeça do cavalo que é a testada...e além disso tem uma...um outro courinho que sai...de cima...e passa por baixo do...da cabeça...ah...próximo ao pescoço...que também tem um nome ahn::... — puxa eu não sei como é que eu posso ter esquecido esse nome...— é alguma coisa como:: ... gargantilha mas não é gargantilha hoje é... ((risos)) é o que as mulheres usam

Doc – pode ser ...daí o exemplo

Inf – mas poderia ser mesmo...talvez até...a::a palavra fosse essa gargantilha...e que agora esteja lembrando mas estou ligando com a coisa que as mulheres estão usando né? (NURC/ SP DID 18)

‘Inf – the::yes the horse it naturally it has the::...either the bridle or the rein...these are two different things also...now...what it’s commonly used here::in the countryside...is the bridle...and the::the bridle is...ahn consists of let’s say a metal...that is put inside the horse’s mouth...this is the bridle itself...but...in order to maintain this metal inside the horse’s mouth...there is::a headstall...made with leather...which is several times called as being part of the bridle...but when someone wants to distinguish...the::...the thing that...the part...which is inside the horse’s mouth...from the other part of the bridle...then it’s used the word headstall...the headstall on its turn has::also some parts...there is one which is...right::in front of::...the horse’s head...there is another one which is over the horse’s head itself...and besides that there is a...another piece of leather that comes... from above...and goes under the...the head...ah...next to the neck...which also has the name ahn::... – I don’t know how I could forget this name...—it’s something like::...necklace but it isn’t necklace nowadays it is...((laughs)) it is what women use

Doc – it may be...give me the example

Inf – but it could be...maybe even...the::word was this necklace...and now I can remember but I am relating it to something women wear.’

(6) — Tem anil?

— É botequim, limitou-se a responder o proprietário, palitando a boca.

— Então me dê uma cachaça, acrescentei como náufrago.

O mulato apanhou a garrafa, estendeu-se um copo embaçado:

— Dupla?

—Pode ser. (BH)

‘— Have you got indigo?

— It’s a pub, the owner only said that, using a stick to clean the teeth.

— Then give me a rum, I added as a shipwrecked person.

The man caught the bottle and, offering me a dull glass, he asked:

— Double?

— May be.’

Occurrences of pode ser₁ are always interpreted as epistemic, such as (4). On the other hand, pode ser₂ expresses both epistemic (example 5) and deontic (example 6) values. The distribution of the analyzed occurrences between both modal domains is presented in Table 1:

Table 1
– pode ser in the epistemic and deontic domains

	Epistemic domain	Deontic domain	Total
Pode ser 1	226 (100%)	-	226 (100%)
pode ser 2	111 (85%)	20 (15%)	131 (100%)

• Source: Made by the author.

The epistemic domain not only predominates in pode ser₁, a consolidated way of epistemic evaluation, but also in pode ser₂: only 20 (15%) occurrences are deontic, as the utterance (6).

The epistemic evaluation marked by pode ser is always related to lack of certainty. By using pode ser, the speaker evaluates the content within its scope as possible, not committing himself to the truth of what is said and, then, revealing little adhesion to the utterance. In (4), the embedded clause (não nos encontremos nunca mais (we will never meet again)) is presented as a speaker’s judgment, who considers uncertain the propositional content. Likewise, in (5), by using pode ser₂, the speaker accepts as possible what was said by the listener (the name of the object described may be necklace).

The evaluative, hypothetical nature of pode ser₁ requires a specific mood-temporal organization for the construction. In the great majority of the embedded clauses, the verb is in the present subjunctive, as can be seen in (4). Table 2 shows the results of the analysis of the mood-temporal flexion of the embedded clause in pode ser₁:

Table 2
– Tense and mood of the verb of the clause embedded in pode ser1

Present subjunctive	176 (80%)
Compound past perfect subjunctive	18 (8%)
Imperfect subjunctive	11 (5%)
Past perfect indicative	7 (3%)
Present indicative	4 (2%)
Future indicative	1 (less than 0.5%)
Total	216 (100%)1

• Source: Made by the author.7

The high frequency of present subjunctive (80%) shows a mood-temporal correlation between the main and the embedded clause in which the semantic value of the matrix clause pode ser (eventuality, doubt, uncertainty) determines the verbal flexion of the embedded clause. Occurrences of pode ser₁ with embedded clause in the indicative mood are not frequent, whereas the ones in the subjunctive mood sum up to 93% (80% in present subjunctive, 8% in compound past perfect subjective, 5% in imperfect subjunctive). This is an expected behavior of embedded clauses: the semantics of the main predicate implies certain features in the embedded clause (NOONAN, 1985; GONÇALVES; SOUSA; CASSEB-GALVÃO, 2008; among others).

Different degrees of subjectivity may be identified depending on the type of knowledge on which the speaker is based in order to express his evaluation, which determines the nature of the modal force. Consider, initially, the occurrence in (7), with pode ser₁:

(7)Infelizmente a tendência atual não é esta. A constante pressão do marketing da indústria farmacêutica aliada à falta de informações e conscientização sobre o problema da resistência bacteriana por parte da classe médica já estão levando a um uso abusivo da gentamicina com o consequente aparecimento de patógenos a ela resistentes. Isto pode levar a uma modificação do quadro epode serque tenhamos que utilizar, em primeira escolha, os novos aminoglicosídeos. (ANT-T)

‘Unfortunately the current tendency is not this one. The constant pressure of the marketing of pharmaceutical industry associated to the lack of information and consciousness about the problem of bacterial resistance by part of the medical class is already leading to an abusive use of gentamicin with the consequent appearing of pathogens resistant to it. This can lead to a modification of the scenery and may be that we have to use, as a first choice, new aminoglicosides.’

In this occurrence, pode ser₁ marks an evaluation based on data, on objective facts, and expresses possibility contextualized in logical reasoning: with the appearance of pathogens resistant to gentamicin, it is possible that there will be the necessity of using new aminoglycosides. This is the evaluation of a specialist based on scientific knowledge about the bacterial resistance. In (8), pode ser₁ shows a higher degree of subjectivity, since the speaker is based on his cultural knowledge and infers about what would be possible to happen:

(8)Chorar é a sua manifestação de desagrado diante de sensações que para ele são insuportáveis. A maior parte das vezes esta inquietação está relacionada com a hora da mamada e pode ser traduzida como fome. Maspode serque o bebê chore e se desespere depois de mamar. Serão cólicas, alguma sensação desagradável relacionada com seu tubo digestivo? (PFI)

‘Crying is the manifestation of dislike of feelings which are unbearable for him. Most of the time this inquietude is related to the time of nursing and it can be understood as huger. But it may be that the baby cries and becomes impatient after nursing. Are these colics any disgusting sensation related to his digestive tube?’

In (8), pode ser₁ does not signal that it is a particular speaker’s thought, but something he knows is possible based on his knowledge about the world, about a baby behavior, in the situation described in the utterance (the baby cries and becomes impatient after nursing).

In (7) and in (8), pode ser₁ expresses objective epistemic modality, in which the source of information is external to the speaker. In this kind of modalization, the speaker evaluates the actuality status of the state of affairs based on his knowledge about the world, about possible situations. The occurrence in (8) is understood as more subjective than the one in (7), since it refers to a common knowledge, which belongs to a community of which the speaker is part, whereas (7) involves knowledge which belongs to scientific domain, totally independent from the speaker.

Pode ser₁ also expresses subjective epistemic modality, in which the source of information is the speaker himself, who transmits his particular belief, as exemplified in (9):

(9)Uma vez no Rio, eu estava de férias passeando no carro da Nesita, quando parou um ônibus ao meu lado. Olhei e tinha uma menina linda me olhando. Dei uma piscada pra ela e ela retribuiu com um beijinho. Então dei uma lambida nos meus lábios e ela me fez uma careta. Depois rimos, e, quando o ônibus partiu, ela mandou um tchauzinho bem íntimo. Fiquei morrendo de vontade de parar o carro, subir no ônibus pra conhecer a garota. Deve ser uma menina legal, pra corresponder assim a uma brincadeira. Mas deixa ela ir embora.Pode serque uma palavra estrague tudo. Essa cena nunca mais saiu da minha cabeça, nem o rostinho bonito dela. (FAV).

‘There was a time, in Rio, I was on vacation riding Nesita’s car, when a bus stopped next to me. I looked at it and there was a beautiful girl looking at me. I blinked at her and she sent me a kiss. Then I licked my lips and she made me a face. After that, we laughed, and, when the bus left, she said a very intimate goodbye to me. I wanted to stop the car, take the bus in order to meet the girl. She might be a nice girl because of the way she behaved. But, let her go away. It may be that a word messes everything up. This scene has never gone from my mind, neither her beautiful face.’

By using pode ser₁, in (9), the speaker expresses a conjecture, a supposition about certain situation (uma palavra estragar tudo (a word messes everything up)), something that he believes is possible.

Pode ser₁, thus, indicates what the speaker knows that is possible, such as in (7) and in (8) – objective epistemic modality – or what he believes is possible, like in (9) – subjective epistemic modality –, these uses of pode ser₁ reflect different degrees of speaker’s subjectivity (lower degree in the first cases and higher in the last one). Either by expressing a personal opinion, or marking an evaluation based on data external to the speaker or even based on logical reasoning, pode ser₁ marks speaker’s uncommitment in relation to the evaluated content, and this feature is also present in pode ser₂, as it was shown in the epistemic occurrences in (2) and in (5), presented above.

However, in pode ser₂ it is not verified the same different degrees of subjectivity which are seen in the analysis of pode ser₁. The epistemic pode ser₂ is always the expression of the speaker’s evaluation based on his own knowledge, that is, it is always the expression of a particular stance, characteristics of the subjective epistemic modality. In the interlocution, recovering something said by the listener, pode ser₂ represents accordance, agreement with the speech act performed before. That is, pode ser₂ is a highly subjective use, since it always reveals a speaker’s particular belief and, at the same time, always expresses subjectivity, for representing the speaker’s evaluation in relation to the listener. Less frequent cases in which pode ser₁ expresses the speaker’s stance in relation to what is said by the listener (in the same way as pode ser₂), are exemplified below:

(10) — Mas ‘tou achando muito custoso Seu Isé chegar vivo até aqui!

— Sei lá...pode serque não, maspode sertambém que sim. (CHA)

‘— But I think it’s very hard Isé arrives here alive!

— I don’t know... it may be not, but it may be yes as well.’

(11) Luiz Raul – Sempre tive a impressão que você tem um profundo desprezo pela gente... Você sempre age como se fosse melhor que todos nós... Tão superior, tão grave, tão coerente...

Léo – Isso não é verdade.

Luiz Raul –Pode serque não seja, mas é a minha impressão... (RE-D)

‘Luiz Raul – I’ve always had the impression that you have a profound contempt for us… You always act as if you were better than all of us… So superior, so coherent...

Léo – This is not true.

Luiz Raul – May be not, but this is my impression...’

In (10) and (11), the speaker uses pode ser₁ in order to stand in relation to the content presented by the listener, what makes very clear the intersubjective status of these two occurrences.

There is, yet, another intersubjectivity scale these utterances highlight. In (10), pode ser₁ introduces vicarious words (yes, no) which refer back to the listener’s speech act expressing its (non-)validity. Both modal sentences, the negative (pode ser que não (may be not)) and the positive one (pode ser também que sim (may be yes as well)), reinforce the doubt expression (sei lá (I don’t know)) used in the beginning of the sentence. Yet, the sentences are connected by the adversative conjunction mas (but), whose meaning effect involves certain disregard for the previous excerpt (NEVES, 1984). All these elements reveal that the speaker has no precise opinion about what was presented to him (‘tou achando muito custoso Seu Isé chegar vivo até aqui! (I think it’s very hard Isé arrives here alive!)) and, because of that, the utterance is modalized in order to mark the speaker’s uncommitted epistemic stance in relation to the content presented by the listener (speaker-oriented modality).

In (11), on the other hand, modalization has another function, it is listener-oriented. The speaker refers back to the listener’s argument, gives it some validity (pode ser que não seja (verdade) – May be it’s not (the truth)), and in what follows introduces the final argument (mas é a minha impressão (but this is my impression...)), which reinforces the stance previously assumed (sempre tive a impressão de que você sente um profundo desprezo pela gente (I’ve always had the impression that you have a profound contempt for us)). In this case, pode ser₁ represents a false admission, which takes place in the conversational world, in which the speaker pretends to agree with a divergent position in order to reduce the negative impact his assertion may have. This movement of “forward and backward” consists of a strategy of controlling divergent points of view in discourse, and involves face-saving process.

In (11), pode ser₁ is used as an attempt to avoid the listener’s possibly unpleasant reactions, that is, it involves the speaker’s concern about the reception of his utterance by the listener, about listener’s image. For this reason, the sentence in (11) is considered more intersubjective than occurrences like the one presented in (10).

The analysis of the data of this study reveals that pode ser₂ is an important discursive mechanism for interpersonal relations. It can also be used to express the speaker’s accordance with situations which imply control. In these contexts, the speaker chooses a deontic source, which can permit or not permit the accomplishment of certain actions which involve offerings, proposals, invitations etc., such as the one in (12), below. Considering that the deontic source is, in these cases, the speaker himself, pode ser₂ expresses a more subjective deontic meaning.

(12) W: Que tal tomarmos alguma coisa, doutor? Uma cerveja, uma batida? Meu colega oferece!

P: Eu?

O: Pode ser ... Pá animar!

P: Mas Waldemar!

W: Ótimo! Vamos entrar aqui mesmo. Depois o senhor procura o tal hotel. (PED-D)

‘W: What about having something to drink, doctor? A beer, a cocktail? My friend is offering us!

P: Me?

O: May be... To cheer up!

P: But, Waldemar!

W: Great! Let’s come into here. Later on you look for a hotel.’

Although pode ser₂, as illustrated in (12), marks the speaker’s agreement in a high level, being compared to an adverb, such as “sim” (yes), the value of doubt, of uncertainty, is, somehow, present. By using pode ser in order to allow an action accomplishment, an invitation acceptance, the speaker reveals certain indifference in relation to the process; there is a lack of conviction in relation to what he wants or, yet, he accepts something unwillingly.

Uses of deontic pode ser₂ with a high degree of intersubjectivity mark speaker’s concern about the reception of his utterance by the listener. These are utterances in which pode ser₂ is placed in the end of the sentence, as a question. This is a way to mitigate the previous speech act which is threatening to listener’s negative face, since it is always something the speaker wants from the listener (a request, and offer, an imposition etc.). The use of pode ser with the attenuation and/or politeness function, in these cases, reflects that the speaker predicts the listener may have an objection or a rejection in relation to what the speaker wants:

(13) No copo dela, a cereja nadava no Martini. Parecia ótima, também

— Escute, estou duro. – esclareci.

— Ah, não se preocupe, eu pago a minha conta – disse ela.

— Não é isso – disse eu – Queria que você me pagasse um Martini, pode ser ?

Ela chamou o garçom, passou o copo dela para mim, e encomendou outro. Depois me olhando e sorrindo do lado de lá da mesa. (LC)

‘In her glass, the cherry was swimming in the Martini. She looked great as well.

— Listen to me, I’m short of money – I said.

— Ah, don’t worry, I’ll pay my bill – she said.

— This is not the problem – I said – I would like you to pay me a Martini, may it be?

She called the waiter, gave me her glass and asked another one. After that, she looked at me and smiled from the other side of the table.’

In (13), pode ser₂ is an attenuation marker used to reduce the negative force the speaker knows may arise from his request (Queria que você me pagasse um Martini (I would like you to pay me a Martini)). By using the interrogative sentence pode ser? (may it be?), the speaker changes the illocutionary force of his previous speech act, making his request more polite. These uses of pode ser₂ correspond to tag questions, as considered by Fraser (1980). According to this author, tag questions are among the attenuation devices, since they attenuate the illocutionary force of the utterances which precede them, mitigating their imposing status. Dik (1989) describes this grammatical operation as illocutionary conversion, which consists of changing the basic illocution of the previous sentence into a derived illocutionary value. Modalization, in cases like the one in (13), clearly serves to protect listener’s face and is, likewise, listener-oriented.

As demonstrated, both pode ser₁ and pode ser₂ express subjectivity and intersubjectivity, but the analysis suggests that pode ser₁ tends to mark subjectivity and pode ser₂ is recognized as intersubjectivity marker. In other words, pode ser₁, in general, expresses speaker’s stance towards his own propositional content (subjectivity) and pode ser₂ always expresses speaker’s stance towards the listener or towards a speech act (intersubjectivity). Although both uses may assume highly pragmatic functions, such as avoiding rejection and seeking approval, pode ser₂ seems to be the most employed form in these cases.

The prototypical context in which pode ser is used is the interlocution situation, what is reflected in the data presented in Table 3:

Table 3
– The distribution of pode ser in continuous discourse and in interlocution

	Pode ser1	Epistemic pode ser2	Deontic pode ser2	Total
Continuous discourse	89 (39%)	24 (22%)	-	113 (32%)
Interlocution	137 (61%)	87 (78%)	20 (100%)	244 (68%)
Total	226 (100%)	111 (100%)	20 (100%)	357 (100%)

• Source: Made by the author.

The functionality of pode ser is highly related to the conversational flow (the real one, in the case of spoken texts, or the simulated one, in the case of written texts), especially when it is not linked to an embedded clause (pode ser₁) and is used as an independent construction (pode ser₂). The interlocution situations predominate in contexts in which pode ser₁ and pode ser₂ are used with epistemic value (61% and 78% respectively) and are the only context in which the deontic pode ser₂ is used (100%). On the other hand, the frequency of occurrence in continuous discourse is expressive in pode ser₁ (39%), low with epistemic pode ser₂ (22%) and zero with deontic pode ser₂.

The frequency of use of pode ser₂ in continuous discourse should be carefully considered, since, even in such a context, the construction represents a speech act in response to another one, most of the time, in a simulation of what would be a dialogue. Continuous discourse is, in principle, a discourse of one only voice, however, in occurrences of pode ser₂, the speaker talks to himself or even brings different voices to his utterance and interact with them. This is the case of occurrences in (14) and in (15):

(14)Não será isto, pergunto-me, um sinal do fim dos tempos, uma evidência de que o Messias está para chegar – tal como mencionado no Livro das Origens (“Quando o Mal atingir o clímax, o Messias chegará, montado em seu cavalo branco, separando com sua espada de luz os justos dos pecadores”)?Pode ser. De qualquer forma, porém, e considerando o que já aconteceu, o Messias está atrasado. O que não seria de admirar, com este infernal trânsito paulista. Como conseguirá ele subir até aqui? Entrará, a cavalo, no elevador? (CEN)

‘Isn’t it? I ask myself, an evidence of the end of times, a signal that Messiah is about to come – as mentioned in Genesis (“when the evil reaches a critical point, Messiah will come, riding his white horse, separating with his light sword the honest people from the sinners”). It may be. Anyway, though, and considering what already happened, Messiah is late. What is already expected, considering the bad traffic in São Paulo. How will he go up here? Will he enter the elevator on horseback?’

(15)Fecho a cortina do box, e o vapor vai me comendo. Vou perdendo de vista o meu corpo e o resto. Um dia, na sauna, meu amigo disse que os antigos chamavam esses banhos de lacônicos.Pode ser. Não sei o que uma coisa tem a ver com a outra. Só sei que vou levar uns bons anos até acertar com outro chuveiro igual a este. (EST)

‘I close the shower curtain, and the steam spreads out all over the place. I stopped seeing my body and everything. There was a day, I was in a sauna, and a friend of mine said the ancient people used to call this kind of bath laconic. It may be. I don’t know what one has got to do with the other. I only know that it’ll take some years until I find a shower like this one.’

In (14), the speaker considers different arguments by using a dialog with himself, that is, pode ser₂ is a speech act appreciative of another speech act, both performed by the same speaker. It is a reasoning in which the speaker makes a questioning (Não será isto, pergunto-me, um sinal do fim dos tempos, uma evidência de que o Messias está para chegar – tal como mencionado no Livro das Origens [...]? (Isn’t it, I ask myself, an evidence of the end of times, a signal that Messiah is about to come – as mentioned in Genesis [...])) which is answered by himself with pode ser (it may be).

The most common situation in which pode ser₂ is used in continuous discourse involves reference of a speech act which is not from the speaker, but is brought to the utterance by means of indirect discourse, as it occurs in (15) (meu amigo disse que os antigos chamavam esses banhos de lacônicos (a friend of mine said the ancient people used to call this kind of bath laconic)). These are utterances very similar to dialogic situations, because there is the indirect insertion of what someone else said.

This way, pode ser₂ is considered as a construction highly related to the communicative interaction, this necessarily implies interlocution, exchange, and it is an expression which encodes intersubjectivity. Since it operates in the discourse level, its scope may be unclear, only recognizable in the interactional flow, as, for example, in occurrence in (3), previously analyzed and repeated below just for convenience, in which pode ser₂ modalizes a content that is identified in the interactional context, but is not linguistically expressed.

(3) — Você é virgem, Severina?

— Virgem nasci / Virgem me criei / Se comigo não casares / Virgem morrerei – parodiava Severina, inspirada.

Flodoaldo ria, coçava-lhe a vaidade:

— Você é vida como o diabo, Severina!

E ia trepando como gato no cio aquele muro escuro e sólido de resistência. Adoçava a voz.

—Pode serou está difícil, coração? (CR)

‘— Are you virgin, Severina?

— Virgin I was born / Virgin I grew up / If you don’t marry me / Virgin I’ll die – Severina, inspired, parodied.

Flodoaldo laughed, conceitedly:

— You are life as the devil, Severina!

And he was climbing up the dark and solid wall of the house like a cat in rut.

— It can be or it’s hard, sweetheart?’

The analyses show, thus, different degrees in terms of (inter)subjectivity, which are interpreted as evidence that an (inter)subjective process, as defined by Traugott (2010; TRAUGOTT; DASHER, 2002; among others), takes place in contexts with pode ser after all, we verified an increase of (inter)subjectivity (pode ser₁ tends to the expression of subjectivity and pode ser₂ is intersubjective), which reflects in different directions of modalization (pode ser₁ expresses, more strongly, speaker-oriented modality, whereas pode ser₂ marks, more strongly, listener-oriented modality).

There is yet a change in the functionality of the expression. Pode ser₁ has its scope over the propositional content of the embedded clause, operating, thus, in the propositional level. Pode ser₂, on the other hand, modalizes the previous utterance or a content implied in the verbal interaction, operating, then, in the discourse level, the level of speech acts. This change is considered by Traugott and Dasher (2002), according to whom the changes in the (inter)subjectivization process typically involve the development of meanings which operate in the propositional level to meanings with scope over the proposition, and from them to meanings with scope over discourse units (TRAUGOTT; DASHER, 2002, p.40). These paths are schematized in the following Figure:

Figure 1
– Correlated paths of directionality in semantic change

truth-conditional	>			non-truth-conditional
content	>	content/procedural	>	procedural
s-w-proposition	>	s-o-proposition	>	s-o-discourse
nonsubjective	>	subjective	>	intersubjective

• Source:Traugott and Dasher (2002, p, p, p, p.40).
• In this figure, s-w: scope within; s-o: scope over.

(Inter)subjective values observed in pode ser correspond, in general, to the second and third columns of this schema. From a subjective value (pode ser₁), with scope over the proposition, the item would have developed into an intersubjective value (pode ser₂), absolutely linked to communication, with scope over discourse units.

According to Traugott and Dasher (2002), subjectification and intersubjectification processes are, hypothetically, related to the speaker-listener dyad, and derive from a mechanism of metonymic inference, a type of pragmatic inference, a conceptual association which is established in discursive context. According to what the authors call Invited Inferencing Theory of Semantic Change, the meanings of the items become increasingly pragmatic and procedural, that is, the items are increasingly used for the organization of communication.

We understand that the conventionalization of conversational implicature (TRAUGOTT; KÖNIG, 1991; TRAUGOTT; DASHER, 2002) may be in the base of the changes in pode ser. Consider the following occurrences, repeated here for convenience:

(10) — Mas ‘tou achando muito custoso Seu Isé chegar vivo até aqui!

— Sei lá...pode serque não, maspode sertambém que sim. (CHA)

‘— But I think it’s very hard Isé arrives here alive!

— I don’t know... it may be not, but it may be yes as well.’

(5)Inf – [...] além disso tem uma...um outro courinho que sai...de cima...e passa por baixo do...da cabeça...ah...próximo ao pescoço...que também tem um nome ahn:: ... — puxa eu não sei como é que eu posso ter esquecido esse nome ... — é alguma coisa como::...gargantilha mas não é gargantilha hoje é...((risos)) é o que as mulheres usam

Doc –pode ser...daí o exemplo [...] (NURC/ SP DID 18)

‘Inf – [...] besides that there is a...another piece of leather that comes... from above...and goes under the...the head...ah...next to the neck...which also has the name ahn::... – I don’t know how I could forget this name...—it’s something like::...necklace but it isn’t necklace nowadays it is...((laughs)) it is what women use

Doc – it may be...give me the example [...]’

(12) W: Que tal tomarmos alguma coisa, doutor? Uma cerveja, uma batida? Meu colega oferece!

P: Eu?

O: Pode ser ... Pá animar!

P: Mas Waldemar!

W: Ótimo! Vamos entrar aqui mesmo. Depois o senhor procura o tal hotel. (PED-D)

‘W: What about having something to drink, doctor? A beer, a cocktail? My friend is offering us!

P: Me?

O: May be... To cheer up!

P: But, Waldemar!

W: Great! Let’s come into here. Later on you look for a hotel.’

(13) — Escute, estou duro. – esclareci.

— Ah, não se preocupe, eu pago a minha conta – disse ela.

— Não é isso – disse eu – Queria que você me pagasse um Martini, pode ser ?

Ela chamou o garçom, passou o copo dela para mim, e encomendou outro. Depois me olhando e sorrindo do lado de lá da mesa. (LC)

‘In her glass, the cherry was swimming in the Martini. She looked great as well.

— Listen to me, I’m short of money – I said.

— Ah, don’t worry, I’ll pay my bill – she said.

— This is not the problem – I said – I wanted to you to pay me a Martini, may it be?

She called the waiter, gave me her glass and asked another one. After that, she looked at me and smiled from the other side of the table.’

The hypothesis is that the change is triggered by the contexts in which pode ser₁, an expression of epistemic modality, is used to express a speaker’s admission towards the listener, such as in (10). In this case, pode ser₁, does not introduce a propositional content, but refers back to the listener’s utterance, admitting its validity and situating it in the uncertainty domain.

The value of speaker’s admission towards the listener is conventionalized in epistemic pode ser₂, which is the prototypical use in interlocution situations, such as in (5). In certain uses of pode ser₂, it no longer implies acceptance in epistemic domain, but implies it in the deontic domain, such as the occurrence in (12).

Conventionalized as an agreement towards an action, the deontic pode ser₂ starts to be used in interrogative utterances, representing a pragmatic resource, an attenuation strategy of a speech act which may be threatening to the listener, such as in (13).

From epistemic to deontic?

This research shows that deontic uses of pode ser are more recent than the epistemic ones, a finding that reveals a very interesting and unexpected path considering historic studies on the development of modal meanings.

Linguists who have dedicated their studies to look for the origins of modalized elements have proved that the development in this area occurs from the deontic domain to the epistemic one, and not the other way around (TRAUGOTT, 1989; SWEETSER, 1990; BYBEE; PERKINS; PAGLIUCA, 1994; HEINE, 1995; TRAUGOTT; DASHER, 2002; among others). About it, Traugott says:

[…]

the history of the modal auxiliaries in English (or indeed any language) is enormously complex, and any brief discussion of it will surely involve gross oversimplification. However, several changes seem fairly solid. It is for example well known in the history of English the auxiliaries in question were once main verbs, and that the deontic meanings of the modals are older than the epistemic ones.

(TRAUGOTT, 1989, p.36).

However, it seems that the communicative tension which is established between speaker and listener, in which modalized elements are used, is not highlighted in these studies. Guo (1995, p.229) observes that “[…] given their methodological limitations, historical studies have seldom investigated the role played by interpersonal discourse functions in semantic change.” and suggests that communicative functions of modals help to structure the semantic content of the item and have influence on semantic change. By analyzing the modal néng, from Mandarin, which corresponds to can, the author demonstrates that the discursive function of the modal is an essential part, inalienable from the meaning of the item and forms an important origin for the semantic change, after all, “[…] what began as a contextual meaning frequently associated with the form has become conventionalized and incorporated into the semantic content of that form […]” (GUO, 1995, p.230), by means of a metonymic process (TRAUGOTT; KÖNIG, 1991).

Sweetser (1990) analyzes epistemic meanings as metaphorical extensions from the root meanings (which includes ability/capacity and deontic values). Change is understood as involving a concept transfer from the external sociophysical domain to the epistemic domain of reasoning and judgment. Motivations for a metaphoric process are cognitive and, in this domain, it seems logic that epistemic meanings are posterior from the root meanings. There is a reasoning underlying this transfer: for instance, when someone says it is probable (epistemic) that someone rides a bicycle, this means that there is no authority (deontic) who prevents this action and that this person is able (capacity/ability) to do so. In these cases, however, modality takes place in the level of propositions.

Pode ser₂ does not establish canonical modal relation in the level of propositions, but it operates at the level of speech acts, the one of the interpersonal relations and, in this case, there is no study attesting the deontic > epistemic path.

It is possible that the pragmatic function of pode ser₂ (as a mark of speaker’s agreement with the listener) has “saturated” its epistemic semantic content resulting in a new meaning (deontic). The idea is that, when the expression starts being used only in verbal interaction (pode ser₂), modal value of epistemic evaluation (pode ser₁), which implies agreement of something in terms of knowledge, may have become usual allowing the use of pode ser₂ also for pragmatic agreement of an action or of an act (deontic value). It is beyond the scope of this study to seek historical confirmation for the epistemic > deontic path in pode ser, but we suggest that this is a strong hypothesis.

Conclusions

The aim of this study was to examine, through a functionalist approach, the expression pode ser as a complement-taking predicate which embeds a proposition (pode ser₁) and as an independent construction (pode ser₂), suggesting that, in these contexts, there is a process of (inter)subjectification (TRAUGOTT, 2010). According to Traugott (2010), meanings are considered increasingly subjective, as they become more based on the speakers’ beliefs and attitudes, and increasingly intersubjective the more they involve the speaker’s attention towards his listener.

Parameters used for analyzing the degrees of subjectivity of pode ser were the type of modal force (deontic, epistemic) and the nature of modal force (legal, social, religious, shared knowledge, belief, opinion etc.). Modal meanings are considered intersubjective when they are used as pragmatic strategies of politeness and attenuation, which aim to save the listener’s face.

The notion of modality as a multifunctional category also underlays this research. The speaker-oriented modality essentially serves to mark his attitudinal stance. The listener-oriented modality serves as a communicative strategy, as a resource of face-saving. Both directions of modality combine in an utterance and one may prevail over the other, what reveals degrees of (inter)subjectivity.

The analyses indicate different degrees in terms of (inter)subjectivity, in which the following aspects are observed:

• pode ser₁ is always an expression of the speaker’s subjectivity, indicating the speaker’s attitude towards the propositional content uttered. Subjectivity may be higher or lower, depending on the evidence the speaker has for his modal qualification;

• pode ser₁ also marks the speaker’s intersubjectivity, when he expresses an admission towards his listener’s speech act;

• pode ser₂ is an independent structure in discourse, typical of interactional situations; epistemic stance does not hold towards the proposition performed by the speaker (as in pode ser₁), but holds towards the listener’s speech act; considering the high frequency of pode ser₂ in discourse, in some cases, the construction does not imply knowledge anymore (epistemic domain), but implies control over actions (deontic domain);

• pode ser₂ starts representing a pragmatic strategy of attenuation, which reveals the speaker’s concern about the reception of his utterance by the listener.

These different degrees indicate a (inter)subjectification process, after all, in relation to pode ser₁, pode ser₂ is more subjective (since the modal force is the speaker himself) and also more intersubjective (since it occurs in the conversational flow, as a reaction to a previous speech act and as a pragmatic strategy).

Some changes in the level in which pode ser operates were also verified. The expression no longer operates in the propositional level (pode ser₁) to operate in the speech act level (pode ser₂). In the first case, modality tends to be speaker-oriented, whereas in the second case it tends to be listener-oriented. These changes, hypothetically, occur in the interaction between the participants of communication and involve a conventionalization of conversational inferences (TRAUGOTT; KÖNIG, 1991; TRAUGOTT; DASHER, 2002).

The analysis carried out here indicated that deontic pode ser is an extension of the epistemic use. We suppose that the development occurs in the speaker-listener dyad, triggered by the mechanism of invited inferencing of semantic change (TRAUGOTT; DASHER, 2002). Although this hypothesis is in the opposite direction from what the historical studies have demonstrated, it should be verified in a diachronic research, which remains to be done.

ABBREVIATION OF THE BOOKS CITED AFTER THE OCCURRENCES

ANT-T –Antibióticos na clínica diária. Fonseca, A.L. 2. Ed. Epume, 1984.

BH –Balbino, O homem do mar... LESSA, O. Rio de Janeiro: José Olympio, 1970.

BRI – Brida. Coelho, P. Rio de Janeiro. Rocco Ltda., 2002.

CEN –Cenas da vida minúscula. Scliar, M. Porto Alegre, L&PM, 1991.

CHA –Chapadão do Bugre. PALMÉRIO, M. Rio de Janeiro, José Olympio, 1965.

CR –Cabra das Rocas. HOMEM, H. São Paulo: Ática, 1973;

DSP –O demônio e a Srta. Prym, Coelho, P. Rio de Janeiro. Rocco Ltda., 2000

EST –Estorvo. Holanda, F.B. São Paulo, Cia. das Letras, 1991.

FAV–Feliz Ano Velho. Paiva, M.R. São Paulo, Brasiliense, 1982.

LC –Lobos e cordeiros. Lopes, E. São Paulo, Moderna, 1983.

PED-D –Pedro pedreiro. Pallottini, R. Revista de Teatro. Rio de Janeiro, n. 458, 1986.

PFI–Pais e Filhos. Várias edições. Rio de Janeiro, Block, 1972. // Propagandas, 1989.

RE-D –A resistência. Amaral, M.A.S. Rio de Janeiro, MEC/DAC/Funarte, 1978.

REFERÊNCIAS

• BROWN, P.; LEVINSON, S. C. Politeness: some universals in language usage. Cambridge: Cambridge University Press, 1987.
• BYBEE, J. L.; PERKINS, R. D.; PAGLIUCA, W. The evolution of grammar: tense, aspect and modality in the language of the world. Chicago: University of Chicago Press, 1994.
• CARRASCOSSI, C. N. S. Usos discursivos e os postulados da gramaticalização: um estudo da expressão modalizadora pode ser. Revista de estudos da linguagem, Belo Horizonte, v.22, n.1, p.207-236, jan./jun. 2014.
• CARRASCOSSI, C. N. S. Gramaticalização e (inter)subjetivização na modalização em português: um estudo de pode ser. 2011. 170f. Tese (Doutorado em Linguística e Língua Portuguesa) – Faculdade de Ciências e Letras, Universidade Estadual Paulista, Araraquara, 2011.
• DIK, S. C. The theory of functional grammar. Part 1: the structure of the clause. Dordrecht: Foris Publications, 1989.
• FRASER, B. Conversational Mitigation. Journal of Pragmatics, Amsterdam, v.4, n.4, p.341-350, 1980.
• GALVÃO, V. C. C. O achar no português do Brasil: um caso de gramaticalização. 1999. 156f. Dissertação (Mestrado em Letras) – Instituto de Estudos da Linguagem, Universidade Estadual de Campinas, Campinas, 1999.
• GONÇALVES, S. C. L. Gramaticalização, modalidade epistêmica e evidencialidade: um estudo de caso no português do Brasil. 2003. 250f. Tese (Doutorado em Linguística) – Universidade Estadual de Campinas, Instituto de Estudos da Linguagem, Campinas, 2003.
• GONÇALVES, S. C. L.; SOUSA, G. C.; CASSEB-GALVÃO, V. C. As subordinadas substantivas. In: ILARI, R.; NEVES, M. H. M. (Org.). Gramática do português culto falado no Brasil: classes de palavras e processos de construção. v.2. Campinas: Ed. da UNICAMP, 2008. p.1021-1084.
• GUO, J. The interactional basis of the mandarin néng ‘can’. In: BYBEE, J.; FLEISCHMANN, S. (Ed.). Modality in grammar and discourse. Amsterdam: John Benjamins Company, 1995. p.205-238.
• HALLIDAY, M. A. K. An introduction to functional grammar. Baltimore: Edward Arnold, 1985.
• HALLIDAY, M. A. K. Functional diversity in language as seen from a consideration of modality and mood in English. Foundations of Language, Dordrecht, v.6, p.322-361. 1970.
• HALLIDAY, M. A. K.; HASAN, H. Cohesion in English. Londres: Longman, 1976.
• HEINE, B. Agent oriented vs. epistemic modality: some observations on German modals. In: BYBEE, J.; FLEISCHMAN, S. (Ed.). Modality in grammar and discourse. Amsterdam: John Benjamins, 1995. p.17-54.
• HENGEVELD, K. Illocution, mood and modality. In: BOOIJ, G.; LEHMANN, C.; MUGDAN, J. (Ed.). Morphology: a handbook on inflection and word formation. v.2. Berlin: Mouton de Gruyter, 2004. p.1190-1201.
• HENGEVELD, K. Illocution, mood and modality in a functional grammar of Spanish. Semantics, [s.l.], v.6, n.1, p.227-269, 1988.
• HOFFNAGEL, J. C. A modalização epistêmica no processamento textual da fala. In: KOCH, I. G. V.; BARROS, K. S. M. (Org.). Tópicos em linguística do texto e análise da conversação. Natal: EDUFRN, 1997.
• HOPPER, P. J.; TRAUGOTT, E. C. Grammaticalization. Cambridge: Cambridge University Press, 1993.
• KIEFER, F. On defining modality. Folia linguística, The Hague, v.21, n.1, p.67-93, 1987.
• LYONS, J. Semantics 1. Cambridge: Cambridge University Press, 1977.
• MIRANDA, N. S. A configuração das arenas comunicativas no discurso institucional: professores versus professores. 2000. 196f. Tese (Doutorado em Educação) – Faculdade de Educação, Universidade Federal de Minas Gerais, Belo Horizonte, 2000.
• NARROG, H. On defining modality again. Language Sciences, Tokyo, v.27, n.2, p.165-192, 2005.
• NEVES, M. H. M. Linguística funcional: princípios, temas, objetos e conexões. Guavira Letras, Três Lagoas, v.13, n.1, p.23-38, 2011.
• NEVES, M. H. M. Texto e gramática. São Paulo: Contexto, 2006.
• NEVES, M. H. M. A gramática funcional. São Paulo: Martins Fontes, 1997.
• NEVES, M. H. M. O coordenador interfrasal mas – invariância e variantes. Alfa, São Paulo, v.28, p.21-42, 1984.
• NOONAN, M. Complementation. In: SHOOPEN, T. (Ed.). Language typology and syntactic description. Cambridge: Cambridge University Press, 1985. p.43-140.
• NUYTS, J. Subjectivity as an evidential dimension in epistemic modal expressions. Journal of Pragmatics, Amsterdam, v.33, n.3, p.383-400, 2001.
• NUYTS, J. Epistemic modal adverbs and adjectives and the layered representation of conceptual and linguistic structure. Linguistics, Hawthorne, v.31, n.5, p.933-969, 1993.
• NUYTS, J. Subjectivity vs objectivity modality: what is the difference? In: FORTESCUE, M. et al. (Ed.). Layered structure and reference in a functional perspective. Amsterdam: John Benjamins, 1992. p.73-97.
• PALMER, F. R. Mood and modality. Cambridge: Cambridge University Press, 1986.
• PERKINS, M. R. Modal expressions in English. London: Frances Pinter, 1983.
• QUIRK, R. et al. A grammar of contemporary English. London: Longman, 1985.
• SWEETSER, E. E. Modality. In: From etymology to pragmatics: metaphorical and cultural aspects of semantic structure. Cambridge: University Press, 1990. p.49-75.
• TALMY, L. Force-dynamics in language and cognition. Cognitive Science, Norwood, v.12, n.1, p.49-100, 1988.
• TRAUGOTT, E. C. (Inter)subjectivity and (Inter)subjectification: a reassessment. In: DAVIDSE, K.; VANDELANOTTE, L.; CUYCKENS, H. (Ed.). Subjectification, intersubjectification and grammaticalization. Berlin/New York: Walter de Gruyter, 2010. (Topics in English Linguistics, 66).
• TRAUGOTT, E. C. Subjectification in grammaticalization. In: STEIN, D.; WRIGHT, S. (Ed.). Subjetivity and subjectivization. Cambridge: Cambridge University Press, 1995a. p.31-54.
• TRAUGOTT, E. C. The role of the development of discourse markers in a theory of grammaticalization. Paper presented at ICHL XII, Manchester, 1995b. Disponível em: http://www.stanford.edu/~traugott/traugott.html>. Acesso em: 12 abr. 2010.
  » http://www.stanford.edu/~traugott/traugott.html
• TRAUGOTT, E. C. On rise epistemic meanings in English: an example of subjectification in semantic change. Language, Baltimore, v.65, n.1, p.31-55, 1989.
• TRAUGOTT, E. C. From propositional to textual and expressive meanings: some semantic-pragmatic aspects of grammaticalization. In: LEHMANN, W. P.; MALKIEL, Y. (Ed.). Perspectives on historical linguistics. Amsterdam: John Benjamins, 1982. p.245-271.
• TRAUGOTT, E. C.; DASHER, R. Regularity in semantic change. Cambridge: Cambridge University Press, 2002.
• TRAUGOTT, E. C., KÖNIG, E. The semantics-pragmatics of grammaticalization revisited. In: TRAUGOTT, E.C.; HEINE, B. (Ed.). Approaches to grammaticalization: Focus on theoretical and Methodological issues. v.I. Amsterdam/Philadelphia: John Benjamins Publishing Company, 1991.

• 1
  The occurrences are from the corpus of this research, presented further in this paper. The abbreviations shown in parentheses identify the text where the occurrences come from. Sentence (1) is from NURC, a Corpus of the Spoken Portuguese Grammar Project, “RJ” means the city of Rio de Janeiro, “D2” is the type of sample, dialogues between two speakers, and “158” is the number of the sample. Sentence (2) is from a corpus of written language and the letters in parentheses indicate the abbreviation of the title of the book, as it was determined by the Lexicography Lab (LabLex) from UNESP (FCL – Araraquara). At the end of this paper there is a list with the titles of the cited books.
• 2
  Throughout this paper, we use the term “speaker” and “listener” to refer to language users, to the verbal interaction participants, irrespective of the linguistic modality, spoken or written.
• 3
  Pode ser refers to both uses under investigation. In order to clarify the identification, we use pode ser₁ and pode ser₂
• 4
  Pode ser in the grammaticalization framework is considered in Carrascossi (2014).
• 5
  This paper is based on the PhD dissertation (CARRASCOSSI, 2011), supervised by Professor Maria Helena de Moura Neves at UNESP – São Paulo State University. Faculty of Sciences and Modern Languages, Araraquara, funded by CNPq (grant 141343/2008-7).
• 6
  Nuyts (1992, 1993, 2001) uses the terms subjectivity and intersubjectivity in a totally different sense compared to the one used by Traugott (2010; among others), whose postulates support the analysis of pode ser. Nuyts defines (inter)subjective in terms of the type of knowledge the speaker has for his modal qualification. According to Traugott (2010), subjectivity refers to the meaning encoding based on the speaker’s attitude and intersubjectivity is related to the meanings which capture the speaker’s attention in relation to the listener.
• 7
  A numberof 226 occurrences of pode ser¹ were analyzed and there were cases in which the embedded clause was suspended (for example: Estas figuras giram; se tiverem algum parafuso no centro, pode ser que... [These pictures turn around; if there is a screw in the center, it may be that…]) or there was no overt verb (for instance: pode ser que sim [may be yes]), for this reason, the total considered in Table 2 is 216.

Publication Dates

• Publication in this collection
  Jan-Apr 2015

History

• Received
  Aug 2013
• Accepted
  Jan 2014