A geometric interpretation of Milnor's triple linking numbers (Q1408555)

In [Princeton Math. Ser. 12, 280--306 (1957; Zbl 0080.16901)], \textit{J. Milnor} gave a definition of higher order linking numbers, or \(\bar{\mu}\)-invariants, which measure the depth of the longitudes of the components of a link in the lower central series of the link group. These invariants are obtained by identifying the longitudes of the link with certain power series and reading off appropriate coefficients from the power series, interpreting the coefficients in the integers modulo the greatest common divisor of the pairwise linking numbers of the relevant components of the link. In [Mem. Am. Math. Soc. 427 (1990; Zbl 0705.57003)], \textit{T. Cochran} gave a more geometric formula for calculating the first non-vanishing invariant. In the paper reviewed here, the authors give a geometric interpretation of the triple linking numbers which generalizes Cochran's work for links with trivial pairwise linking numbers. Given a choice of Seifert surfaces for the link, the triple linking number for each triple of components of the link is interpreted as a difference between the signed count of the triple points in the intersections of the Seifert surfaces and certain numbers obtained from the intersections of one link component with the Seifert surfaces for other two link components. Moreover, it is shown that the Seifert surfaces may be chosen so as to make either term in the difference trivial. The proof also shows how one could construct a 3-component link with any prescribed triple of pairwise linking numbers and any prescribed triple linking number.