Here, too, I made several notes on matters where Ayn Rand was incomplete or wrong. For instance, language was not invented. Language evolved. I do not know if she was describing her own actual mental activity as an infant, but learning "table" and "furniture" and "man" and "father" were not like that for me. I was more concerned with the FUNCTIONS of objects than their SHAPES. Also, we had only one desk in the house, so "desk" was the name of a unique object, almost a proper name. (When I started school, I learned that many kinds of desks exist.) That said, Rand's main point in Chapter 3 is that the process of concept formation is never-ending. The highest and broadest abstractions follow the same rules as for the simplest.

A concept is a mental integration of two or more units from which are abstracted out their Conceptual Common Denominator and by which isolated characteristic they are integrated into a new concept.

Rand also demonstrates that wider abstractions require more precision and accuracy than do the concepts closer to perceptual primaries. You can know "man" from "table" easily enough, but legalisms and economic theory are more difficult.

She closes the chapter by stating that integrating facts into concepts is induction; and placing new instances into their proper contexts (concepts) is deduction.