The development of word frequency distribution in first language acquisition. An analysis on a spoken language corpus of french children

French speaking children, which aims to evaluate how their lexical output is related to a standard word-frequency distribution: Zipf’s law. We adopted a set of spoken language transcripts of French children named CoLaJE: by using Python tools we turned original transcripts into strings that allowed us to estimate the exponential parameter of the wordfrequency distribution (alpha) for each child, as well as for parental input. We show how alpha values tend to converge to 1 during later development which is coherent with current literature. We also estimate the exponential parameter for parental input and we found that Spearman’s rho shows a fairly positive correlation between child’s alpha and parents’ alpha in later ages. Finally, we discuss our results in the light of previous studies on the CoLaJE corpus and we compare the obtained values to similar works on children’s spoken language transcripts that were sampled in an analogous way, before outlining possible future directions.