Some unreliable interpretations of statistical chemical correlations in applied earth science and in archaeometry