EXCERPT Realization of Language In the previous chapter, we witnessed the development of K into what Rosenberg terms a CHILD, whose interactions with its environment are bound up with and inferred by its interactions with its linguistic guardians. Next we saw that the transition from CHILD status to fully fledged general intelligence requires certain cognitive regimens or educational methods through which K becomes increasingly competent in expanding its outlook by imputing structure to the world (universe) and to its own thoughts and actions. However, this development looked suspiciously straightforward, for there was in fact a sleight of hand in our thought experiment. With the introduction of the multi-agent system, we assumed that K’s adult guardians were full concept-having language-using AGIs—that is, we assumed we had already constructed general intelligence. In other words, we made too great a leap from the goal of our thought experiment—the realization of general intelligence—to the presupposition that it had already been attained. Nevertheless, there is nothing inherently erroneous in this assumption, since we could easily swap the role of the linguistically proficient automata S and M with their linguistically proficient human counterparts S´ and M´. However, while this rectification is easy and sound, it misses a point: the condition of possibility of discursive apperceptive intelligence rests on the condition of possibility of language—or, in other words, the realization of general intelligence is constituted by the realization of language. Even though the introduction of linguistic agents to the thought experiment is not an illegal move, then, it does occlude the key role played by the realization of language—both in terms of its evolution and its autonomous and sui generis rule-governed functions—in the realization of the conditions of possibility of general intelligence in the first place. …