Alexander Martin edited r-score-res.tex  about 9 years ago

Commit id: d02e4b4d6bf1199e943c9af83a4028cf11726287

deletions | additions      

       

\subsection{R-score Results}  We applied our new measure of functional load to the French lexicon, looking both at the structure of the entire lexicon, but and  also considering the syntactic category of nouns separately. The results of our calculations can be seen in Figure~\ref{fig:rscores}. A one-way ANOVA revealed no significant differences across features in the whole lexicon, but a statistically significant effect was found when considering only the noun class ($F = 9.107$, $p < 0.001$). A Tukey Honest Significant Difference post-hoc test revealed that the functional load of the \emph{place} feature was significantly higher than that of the \emph{manner} feature ($p < 0.001$), and marginally higher than that of the \emph{voicing} feature ($p = 0.684$).