from Gerdes p. 336 in Powell and Frankenstein:

Early stages of mathematics education offer a very efficient way of instilling in the children a sense of failure and dependency. In other words the mathematical competencies, which are lost in the first years of schooling, are essential at this stage for everyday life and labor opportunities. But they have indeed been lost. The former, let us say spontaneous, abilities have been downgraded, repressed, and forgotten while the learned ones have not been assimilated either as a consequence of a learning blockage, or of an early drop-out, or even as a consequence of failure or many other reasons. The question which arises then is what to do: "should we . . . give up school mathematics and remain with ethnomathematics? Clearly not . . .one should, however compatibilize cultural forms, that is the mathematics in schools shall be such that it facilitates knowledge, understanding, incorporation and compatibilization if known and current popular practices into the curriculum. In other words, recognition and incorporation of ethnomathematics into the curriculum, it is necessary to identify within ethnomathematics a structured body of knowledge. 

D’Ambrosio, U. (1985). Socio-cultural bases for mathematics education.  Campinas: UNICAMP. p.45 - 47)