On Language Clustering: A Non-parametric Statistical Approach