In this work, the European, Brazilian, Cape Verdean varieties of Portuguese were studied in comparison between them and with Italian, with a focus on the null subject parameter (NSP).

Beside the big number of studies on the latter, its nature is not completely clear, as so for concurrent properties and activation factors.

The work is therefore a paper to study the micro variation over proximate languages.

 

The generative theory base itself in the idea of a Universal Grammar (UG) thanks to children are able to learn languages. Without this notion, many concepts would not be explicatable. The key idea is that every human being is born with a set of unconscious knowledge. In correlation with UG, the Principles & Parameters theory (P&P) was formulated and associated. Its base is the existence of universal principles that all the languages have in common (that build the UG) and parameters language-specific (that build the specific grammar of that language). These parameters should be defined in a Boolean form (yes or no). These values are defined in the first years of life thanks to experience, and, once they are set, they are correct and unchangeable.

Knowing that the subject may not always be pronounced, there should be a way in which the child understands if it exists or not, given that he is actually acquiring information on something that is missing. For this reason, one of the principles of the P&P theory states that every sentence has a subject, that can be phonetically expressed or not. This possibility would derive from the null subject parameter (NSP), according by which there are two kinds of languages in the world: the one with a phonetic expressed subject, and the one without, exactly based on the Boolean nature of the parameter.

The positive value in the parameter should lead to other concurrent proprieties identified by Perlmutter and Rizzi: null subject, Verb-Subject order inversion, That-Trace effect violation and rich agreement inflection. On the other hand, the negative value should reject them.

Since the subject may not be pronounced, there should exist other factors to allow the setting of the NSP, called the activation factors. For Rizzi and Taraldsen(1980), a verbal system with sufficient morphological richness can let the subject unpronounced. For example, in Italian, for every person the verb has a different morphological form, and therefore every verbal form allow the subject to be identified without its explicit presence but using only the termination of the verb itself. When the morphology is neutralized, the subject needs to be present to avoid ambiguity or ungrammaticality. The problem now becomes, what is enough? Besides, the reasoning does not hold for languages like Chinese, where the subject can not be expressed even if the verb doesn’t have verbal flexion, Hebraic in which the subject can be null only in some person and tenses, and French, in which the verbal morphology is rich but it does not allow null subject. Consequently, the richness of the morphology is not a sufficient characteristic to allow the subject to be null.

Jaeggli and Safir (1989) tried then to save the hypothesis of the morphology being in count for the NSP, considering the morphological uniformity: “An inflectional paradigm P in a language L is morphologically uniform if P has either only underived inflectional forms or only derived inflectional forms”, where a flectional form is given by root + affix. In other words, for them, the mixed verbal systems do not allow null subject. If we try to attack their theory, as regard French language, Jaeggli and Safir (1989) refer to phonetics: in this case French is a mixed language, and therefore require to have an expressed subject. But the problem persists if we consider Brazilian Portuguese.

Looking deeper into the problem, we cannot consider all the null subjects identical, since sometimes it is optional to leave it blank and sometimes it is obligatory. Indeed, the languages that allow null subject, admit it only when the verb is conjugated, since when the verb is in infinitive all the languages require it to be not pronounced.

Holmberg & Roberts (2009) defined thus different kinds of NSLs (null subject languages): canonical NSL, partial NSL, radical NSL, expletive NSL, non-NSL. If the value for the parameter can only be binary, how can exist so many types of languages? In addition, there could not even be a difference in the kind of subject. The problem is that, being the theory universal, there could not be an exception for the binarity only for NSP, and therefore the whole P&P theory should be completely re formulated or even deleted.

In this sense, the schema proposed by Veríssimo (2017) is very interesting, because we can notice that there still exist a binary choice for every level.

Verissimo

 

Let’s study more in detail the languages under consideration.

  • European Portuguese (EP): canonical NSL where NSP is fixed as positive. However, in contexts of contrast or indeterminism it is obligatory to have the subject expressed, while in expletive senteces it is forbidden. These rules are of semantic and pragmatic nature.

  • Brazilian Portuguese (BP): BP derives from EP, and should therefore be a canonical NSL, however over the past years there is the tendency of pronounce the subject and in the meantime, the verbal inflexion is weakening. BP is so changing the value of the NSP in the direction of a non-NSL. Null subject resists more in the speech of the eldest or of the higher educated. There are two way to see this situation: an ongoing changing (Duarte, 1995) or the idea that the partiality is a consolidated state (Holmberg, Nayudu & Sheehan, 2009; Veríssimo, 2017). The evidence shows that the choice of whether or not pronounce the subject depends on semantics and pragmatics.

  • Cape Verdean Portuguese (CVP): The Cape Verde territory was occupied by the Portuguese and since the beginning, there emerged the creole language, that coexists until now with Portuguese. CVP is the official language but the creole is the mother tongue for the population. The creole lost the possibility for the referential subject to be null, having removed as well the verbal flexion of both person and number. Nevertheless, it is interesting to notice that it did not lost the possibility to have a post-verbal subject, That-Trace effect, and it does not use an expletive pronoun (Lucchesi, 2009). It does not exist a Portuguese spoken in the same way by all the population, so there is not a real standard over the territory. Notwithstanding, there is a trend to realize the subject in the PCV, mainly for the influence of the creole. Also in this case, the more the population is educated, the more the subject is null; but again, the percentages of null subjects depends on semantic and pragmatic context.

  • Italian (IT): prototypical canonical NSL. In the majority of cases, the subject is null and the expressed form is limited due to a situation of necessity: “Given the existence of a zero pronominal option in languages like Italian, the overt form will be limited to the cases in which it is necessary, i. e. when the pronominal subject, being focal or contrastive, must bear stress (evidently, the zero element cannot bear stress)” (Rizzi, 1988; Chomsky, 1981). We can say that the phonological realization of the subject is a marked, functional option, used in situations of emphasis, contrast or to avoid ambiguity. For example, in the subjunctive mood, there are situations in which the subject needs to be pronounced to avoid confusion. A first reason could be because in this mood the first, second and third person singular have the same inflexion, but the phonetic realization is not that straightforward. Changing the person to which the pronoun refers, we can have different behaviors. This makes think that the null subject situation is more of semantic nature, instead of morphological. Or again, the speaker, when include himself in the speaking, uses for the majority of cases an expressed subject (Duranti & Ochs, 1979). In this case, the pronoun appear because we have new information. It looks like semantics and pragmatics are taken into account in the subject realization.

In all these languages, expletive subjects have to be null and semantics and pragmatics are involved in the realization of referential subjects.

 

The main problem is than the NSP in the way it is composed. It is impossible to explain all the possible language typologies with the old idea of NSP and also, we cannot consider the NSP to have a binary value. The hypothesis, according to Soares da Silva (2006), is the existence of a level scale, where on one side we have the non-NSL and on the other the radical NSL.

scale

Trying to formulate the new idea of NSP, we need to remember that there is the need of a Universal Grammar model that is general to include the language universals but also as flexible to include the variations between languages (Jaeggli & Safir, 1989). The solution would be a base rule (the NSP) on which apply other rules. The aforementioned reflections have shed a light over semantics and pragmatics. Until now, the most reliable factor to fix the NSP would be morphological uniformity: once the NSP is fixed, we can use semantics and pragmatics to define the situation more in detail. We can consider the languages in question as null subject languages, where its phonological realization depend on later context of semantics and pragmatics. We can therefore delineate this schema, where by “exception” we mean a particular situation semantically or pragmatically speaking.

Final

We can conclude that semantics and pragmatics would allow differentiating between the different kinds of language and maybe are also the key in understanding in a complete and stable way the micro variation over languages.

 

Bibliography

Chomsky, N. (1981). Lectures on Government and Bilding. Dordrecht: Foris.

Duarte, M. E. (1995). A perda do princípio “Evite Pronome” no português brasileiro. Tese de doutorado. Campinas: Unicamp.

Duranti, A., & Ochs, E. (1979). Left-dislocation in Italian conversation. Em T. Givón, Syntax and Semantics, vol. 12: Discourse and Syntax (pp. 377 - 415). New York: Academic Press.

Jaeggli, O., & Safir, K. J. (1989). The null subject parameter and parametric theory. Em O. Jaeggli, & K. J. Safir, The null subject parameter (pp. 1 - 44). Dordrecht: Kluwer Academic Publishers.

Holmeberg, A., Nayudu, A., & Sheehan, M. (2009). Three partial null-subject languages: a comparison of Brazilian Portuguese, Finnish and Marathi. Studia Linguistica, 63(1), 59 - 97.

Holmberg, A., & Roberts, I. (2009). Introduction: parameters in minimalist theory. Em T. e. Biberauer, Parametric variation: null subject in minimalist theory (pp. 1 - 57). New York: Cambridge University Press.

Lucchesi, D. (2009). A realização do sujeito pronominal. Em D. Lucchesi, A. Baxter, & I. Ribeiro, O português afro-brasileiro. Salvador: EDUFBA.

Rizzi, L. (1988). The new comparative syntax: principles and parameter of universal grammar.

Soares da Silva, H. (2006). O parâmetro do sujeito nulo: confronto entre o português e o espanhol. Rio de Janeiro.

Taraldsen, K. T. (1980). On the Nominative Island Constraint, Vacuous Application and That-trace filter. Indiana: Bloomington.

Veríssimo, V. (2017, jan./jun.). A evolução do conceito de parâmetro do sujeito nulo. Entrepalavras, 76 - 90.

 

This has been my Bachelor thesis for my degree in Language, Civilisation and the Science of Language at Ca’ Foscari University of Venice in 2018 written under the supervision of the professor João Miguel Marques da Costa.

View paper PORTUGUESE: pdf