References:ND 1633 *ŋ[U] 'thing, what', 2616 *ya 'who' 2655 *[y]iyo 'which'. The existence of interrog. *jV is in fact quite dubious (despite ND, Altaic certainly has no trace of it, Mong. ja- certainly going back to *ŋia-, and TM forms like Evn. ē̂- - to TM *xia-); IE *i̯o- may well be < *ńo- < *ŋi̯V- (since *i̯- normally < *ń-); the Drav. and Kartv. variation *j- / *n- is also quite suspicious ( < *ŋi̯-?); the only case with *j- is thus FV.