Estudo quantitativo da complexidade gramatical absoluta em diferentes línguas e suas implicações tipológicas

Abstract

Measuring the complexity of a language is a strenuous task; one runs into several problems such as the issue of comparability (CROFT, 2003), commonplace in typological studies, as well as the central issue of defining complexity, obviously necessary before we measure it (MIESTAMO, 2008). Besides that, comparative studies in linguistics often use terminology from 19th-century holistic classification of languages, classifying them wholly as isolating, agglutinative, fusional and so on. Also, most analyses are based on a cross-linguistically inconsistent unit, the word (MATTOS; LAZZARINI-CYRINO, 2020). With all this in mind, the present study, using morphemes as analysis units, lists the most frequent grammatical morphemes of 19 languages sampled worldwide, characterizing them by semantic criteria that don’t use the notion of word (CROFT, 2000), and quantifying them according to absolute complexity measures, that is, not referring to the language’s perceived difficulty, but to the number of parts of the system. In order to do so, we used McWhorter’s taxonomy of complexity (2007), Dahl’s notion of verbosity (2004) and Trudgill’s complexifications (2011), classifying each morpheme based on five parameters – the first one, if a given morpheme is glossed as free or bound, and the four subsequent ones properly being about complexity: if a given morpheme realizes contextual inflection, its degree of verbosity, if it presents irregularities and/or structural elaboration, and if it is polyexponent. Relative frequencies for each parameter were taken, and the results were processed using the software R and compared. We found a strong correlation between a higher proportion of unbound grammatical morphology and smaller scores on the four used complexity parameters, as well as the existence of dramatically simplified languages compared to the rest; it’s hypothesized, following McWhorter (2016), that this is due to intense contact events in their past histories. Moreover, we noticed behavior similarities between American languages, and we propose relinquishing holistic classification terminology, as well as not portraying less complex languages negatively.