Most of the forms and protocols of information visualization come from disciplines outside the humanities. The basic principles of empirical and quantitative methods on which visualizations and the construction of data depend are antithetical to the interpretative methods of the humanities. In visualization techniques, much of the process of construction becomes obfuscated by the display, which is in turn, read as an artifact with self-evident properties. This is just the first in a number of problems with using the bar charts, statistical graphs, and network diagrams that have become ubiquitous. What epistemological, critical, and aesthetic considerations need to be brought to bear on their use for projects in the humanities? Should we accept that the adopted methods can reveal something about humanistic corpora, or should we instead be pushing humanistic principles into the shaping of new visual tools and approaches? This talk summarizes some of the basic issues in using borrowed methods, outlines some of the fundamental epistemological fallacies, and proposes a few speculative alternatives to current practices.