Peirce’s Syllabus of 1903 represents both the culmination of a forty-year research into the nature and the classification of signs, and the beginning of a new phase in the development of his taxonomic investigations. Before the Syllabus, the most important innovation of speculative grammar had been the decision, taken in the Minute Logic of 1902, to view the two semiotic trichotomies (<icon, index, symbol> and <rheme, dicisign, argument>) not as determining classes of signs but as determining parameters that classify signs. I have referred to this innovation as the “first reform” of speculative grammar. The first reform makes it necessary to determine the compossibility of parameters. Thus, after the first reform, the classification of signs becomes a twofold enterprise: first, trichotomies are to be determined which specify the semiotic parameters for the classification; second, the classes of possible signs are to be determined on the basis of rules of combination of the parameters, that is, rules of semiotic compossibility. The rules of the Minute Logic only concern the two trichotomies treated therein, and state that an icon can only be a rheme, while an argument can only be a symbol. In NDTR, the second and final version of the grammatical chapter of the Syllabus, a second innovation was made. This “second reform” of speculative grammar consisted in the addition of a third trichotomy (<qualisign, sinsign, legisign>) to the previous two (<icon, index, symbol> and <rheme, dicisign, argument>). The Syllabus had therefore to generalize the rules of semiotic compossibility of the Minute Logic to account for the expanded system of parameters presented therein: a qualisign can only be an icon, and an icon a rheme, while an argument can only be a symbol, and a symbol a legisign. By applying these rules, Peirce can obtain the ten classes of signs presented at the end of NDTR. As mentioned, these rules can be generalized to a system of any number of basic trichotomies. Peirce proposes a further generalization of the rules in NDTR, but as we have seen (supra, §7.3.4), the generalized rules are unable to determine the ten possible combinations that Peirce has in mind. Peirce would come to a formal and correct statement of the generalized rules of semiotic compossibility in a 1908 letter to Lady Welby: “It is evident that a possible [first] can determine nothing but a Possible [first], it is equally so that a Necessitant [third] can be determined 286by nothing but a Necessitant [third]” (SS 84). These two rules (which I have referred to as R1 and R2, respectively) can also be summarized by saying that a certain combination of semiotic parameters is possible if its elements satisfy the following partial ordering: first element ≥ second element ≥ third element.