Lewis' early graduate student work was a criticism of Bertrand Russell's definition of "material implication" (a → b as "a implies b") in the Principia Mathematica.

Lewis proposed a definition for "strict implication" which prevented false antecedents from implying true consequents. In this work he defined an intensional modal operator that marked the beginning of modern modal logic and possible worlds analysis of various truths.

Lewis' first office as a Harvard Professor was the room in Widener Library that contained all the manuscripts of Charles Sanders Peirce. Lewis began his career as an Idealist, following his thesis adviser Josiah Royce.

Influenced by Peirce's arguments, Lewis in 1923 proposed the radical idea that a priori systems of thought are invented by humans and validated empirically.

In a short paper, A Pragmatic Conception of the A Priori, Lewis claimed that there is not just one logic, but many logics from which we must choose the one that best explains our experience. This eventually is the basis for Quine's naturalization of epistemology, and perhaps Quine's attack on the analytic/synthetic "dogma," which distinction was defended by P. F. Strawson and H. Paul Grice.

After a decade browsing through Peirce's papers, Lewis declared himself a "conceptual pragmatist" in his 1929 book Mind and the World Order. The Roycean, Kantian, and Hegelian transcendental absolute mind imposing order and regulative principles on the world disappeared from Lewis' thought, replaced by the action of human minds constructing various a priori conceptual schemes to make sense of the raw perceptions of the world that Lewis called the "given."

Following Peirce, Lewis took the "truth" or meaning of a concept to be its consequences in experience, and he thought of concepts as relations. Influenced by Percy Bridgman's operationalism, Lewis could bypass the problem of different subjective experiences for different persons, something his Harvard colleague Ralph Barton Perry and fellow "New Realists" had called "the egocentric predicament."

One such subjective epistemological problem is spectrum inversion - "what I see as red, you see as green."

What is important is how you act on the perception. As long as you always stop at a traffic light that I see as red, we have the same relation of information in our minds to information in objects and processes in the world. Lewis concluded:

"If your hours are felt as twice as long as mine, your pounds twice as heavy, that makes no difference, which can be tested, in our assignment of physical properties to things."

Lewis thought the verification of empirical meaning depended on experience, and more importantly on the common relational experiences of many observers. This came to be called pragmatic meaning. The positivist idea was called semantic meaning.

The Positivists hoped to build up their system of knowledge starting with "logical atoms" that could be the foundation on which all knowledge could be made certain. They believed that logic and mathematics could be shown to have an a priori truth and many basic linguistic concepts could be regarded as analytic - since they were true by definition, such as "a bachelor is an unmarried man."

Lewis pointed out that all such analytic linguistic definitions could have been otherwise (indeed they are in other languages, as well as in other contexts) and the Kantian notion of a priori truths had failed. Kant's idea that Euclidean geometry and deterministic Newtonian physics were fundamental categories of the understanding was simply wrong.

We choose from among the possible geometries based on their usefulness in explaining experience. And so must we choose among possible logics, by empirically testing their applicability in the world. Long before Quine's Two Dogmas of Empiricism, Lewis knew that analytic truths have an empirical or "synthetic" basis. And that knowledge depends on relations between concepts and not atomistic reductions of statements to their supporting observation sentences.