ABSTRACT Five structural linguistic features give strong promise of being useful as genetic markers, i.e. of having sufficient diachronic stability, resistance to borrowing, and independence from other structural properties that they can point to deeper genetic relatedness between families of the greatest age now reachable by the standard comparative method. All five are abstract properties of the verbal lexicon, realized in verbal derivations, verbal inflection, and text frequencies of valence patterns and derivational types. This project will trace the five features through 120 or more languages (using a pre-existent worldwide sample, genetically based at a controlled time depth) and survey a dozen language families and three high-diversity language areas in much more depth. This data base will suffice to establish the degree of variation of these features within language families, the extent of borrowing between families in contact situations, and any statistically significantly correlations between these features and standard typologically relevant features (already surveyed in previous work). The anticipated result will be a demonstration that these features are useful genetic markers, demonstration of their structural independence from other features and of their usefulness in typology, and concrete hypotheses of deeper genetic relatedness between certain language families, chiefly of northern Eurasia.
