Some perplexities about Peirce’s “skeleton ideas”

In seven paragraphs written in 1893, Peirce puts forward a puzzling and thought-provoking claim about the role of rather mysterious “skeleton-sets” in processes of association of ideas: all association of ideas, either by resemblance or by contiguity, requires and involves “skeleton sets”, whose iconic dimension is necessary for these processes to take place. Because it relates to the question of the nature and mode of the existence of ideas, to that of the role of icons in thought, and to that of the content of concepts, this thesis is clearly of great importance for Peirce’s philosophical system. In this paper, I would like to examine two questions: (1) what is the justification of Peirce’s claim that “skeleton-sets”, or “skeleton ideas”, are necessary for an association of ideas – by resemblance or by contiguity – to take place in one’s mind? (2) Is this claim compatible with Peirce’s philosophical system