If your measures are all interval or ratio scale, an easy way to make your own mds in spss is to perform dimensional reduction on your measures using the principle components method and keep the measurement item loadings on the first two or three components. Select the first pair of components' item loadings and plot them on a scatter graph - (maybe easier to do in a spreadsheet / office productivity application) where each point is plotted as the x and y item loading for  each variable. If the measurement items are categorical you can do it but with more steps.

The correlations are distances (more technically, dissimilarities), so there should be no problem using the coefficients as pairwise resemblances for the construction of a multidimensional scaling ordination (metric or non-metric).  Personally I'd do it in Primer, but if SPSS does MDS then it should be straightforward.

Thank you, I did it in SPSS (defining them as similarities instead of dissimilarities). Then I was able to choose Pearson's correlation index as "distances".