In the topic ' ' clulas' ' the author tries to establish it researcher criteria to be observed, given that they show the character of the research as: sex, age, interviewed social classroom and the amount of. David Zaslav addresses the importance of the matter here. It appears, then, one asks, what to make with the not-natural data? The author locates itself in the following way: the natural data could be used to establish estilstica hierarchy of performance of the informer and beyond these data that will be recorded, language tests could be elaborated on the lingustica variant. The diagramming of the interview in the following way also sends regards: question-reply, provoked, spontaneous and subsequentes numbers of interventions, narratives. In chapter 3, the author comes to show the detailed description of the variants that it calls of envelope of variation. According to proposal of Labov the conception of the sociolingustico model can be synchronous and diacrnico and that to compare the variants of same nature in different languages it can have double objective: to describe, to analyze and systemize the envelope of variation and to compare the results of the analyses. The variants correspond the certain contexts that are favored by conditioning factors.

The hypotheses are given from the survey of all the contexts that influence in the accomplishment of an 0 variable. To facilitate to the quantification of data the author it considers some steps for cluster analysis of facts: fonolgico context, position of the 0 variable, morphologic classroom of the word and the morphologic statute of the word that contains the 0 variable. The lingustico encaixamento of the 0 variable is the motivation of the hypotheses of the groups of factors. Already the extralingustico factor is everything what he will not be strict lingustico. The formality or informality, socioeconmico level, the escolaridade, the etria band and the sex of the falante can be considered conditioning factors. In chapter 4, the author argues the evaluation of lingusticas variants of the point of view of the informer.