In this paper, we analyze the first three sections on mtshan nyid kyi mtshan nyid (the “definition of definition”) by examining five logical works from four scholar-monks who preceded Sakya Pandita. The headings of these sections are:
1. A criticism of the argument that the definition of definition is not necessary.
2. A criticism of rNgog lo tsa ba’s idea of the definition of definition.
3. What is the definition of definition?
In the expression “definition of definition,” the second instance of the word “definition” is not a real definition, but a definiendum (mtshon bya), which should be defined by the real definition (that is, the first instance of the word “definition”). This idea was commonly accepted by later scholar-monks in Tibet, but the Kadam pa scholars may have also acknowledged it.
rNgog lo tsa ba argued that the definition of definition is “the reason for attributing definiendum to the subject (mtshan gzhi).” However, Phya pa and later scholars criticized rNgog’s idea. They stated that the definition of definition was not a genuine reason for attributing the definiendum, but was rather a simple synonym of definiendum.
Phya pa conceived the definition of definition as having the following three characteristics (chos gsum):
(a) It should be a substantial being.
(b) It is not the reason for attributing names other than the definiendum.
(c) It should exist in the subject.
Kadam pa scholar-monks who succeeded Phya pa maintained the idea of the definition containing the three characteristics, but did not mention the details.