Considering that there is no break separating minha mãe and ela morou em vários lugares and that both blocks of text are uttered in a single intonational contour, we argue that minha mãe is a Subact of Reference within the Nuclear Act, and not itself a separate Act. This kind of construction received special attention in de Groot and Limburg (1986) as one of the pieces of evidence for the diachronic development of pronominal clitics through the cliticization of free pronouns. According to the authors, there are three different types of language: the free pronoun type, the clitic type and the appositional type. In their analysis, BP should be a language which is in a transitional stage between the clitic type and the appositional type, since the Subsidiary Act is drawn into the Nuclear Act. However, the fact that the pronominal element occupying the Referential Subact slot within the Nuclear Act (ela, in (01)) is not developing into a clitic at all suggests that BP does not fit in the diachronic cycle proposed by de Groot and Limburg (1986).

The process of integration of Subsidiary Acts seems to be related to another phenomenon: the tendency, in contemporary BP, to overtly express the the subject constituent with verbs that occur in the third person form. The reason for this tendency might be changes in the verbal inflectional paradigm in BP: the second person pronouns – tu (singular) and vós (plural) – are not used anymore, and were replaced by the pronouns você and vocês; and the first person plural – nós – is gradually losing space for the expression a_gente. Both você(s) and a_gente require a verb in the third person. Consequently, the verbal paradigm in BP has become restricted to three distinct forms: first person singular, third person singular and third person plural. Taking this new paradigm into account, it is not possible to distinguish the second and third person singular and the first person plural on the one hand, and the second and third person plural on the other by means of verb inflection. On the basis of this restructuring of the verbal inflectional paradigm, we may predict that the subject tends to become obligatory when the verb appears in these forms.

Since constructions with an integrated Subsidiary Act occur with third person referents only, we propose the following analysis for a sentence like the one presented in (01): since verb inflection alone is not sufficient to indicate subject agreement, the pronominal element becomes obligatory as an agreement marker. At the same time, the intonational break which separates the Subsidiary Act from the Nuclear Act is suppressed and the NP is integrated into the Nuclear Act as a Referential Subact. Since it refers to the same entity as the pronominal element, an appositional relationship between the NP and the pronoun is established.

We conclude that there is a grammaticalization process under development in BP, according to which the free pronouns in constructions with integrated Subsidiary Acts are becoming agreement markers. However, since no signal of cliticization is observed in BP, the grammaticalization process in this language does not conform to the diachronic cycle proposed by de Groot and Limburg, which should be refomulated to cover the BP facts.