(7)   Under the modified notion of c - command the CP hari ko would c-command uska in (7), but then the pronoun will be bound in its governing category. The theory would predict that the sentence would be ungrammatical, but it is not. In fact, the only possible interpretation is the one in which uska refers to hari. (6) has the following S-structure representation:   (8)   Again, under the modified notion of c-command ram ke bhai ko would c-command uska; however, in that case, uska will be bound within its governing category. The sentence would be wrongly evaluacted as ungrammatical. Under the modified notion of c-command, ram can be the antecedent of the pronoun since it is not a c-commanding antecedent. But the ambiguity of the sentence cannot be accounted for. Suppose we do not modify the notion of c-command and study the relevant phenomena in terms of the existing notion of c - command. That is, ram ne will have the following structure: CP—›NP (ram) C (ne), rather than NP—›NP (ram) C (ne) Then, whereas it will correctly account for the ambiguity of (6) it will incorrectly predict that (5) would be ambiguous because then, like hari, ram too would not c -command the pronoun and would thus be a possible antecedent of the pronoun. Further, the c- command approach cannot account for ram being the antecedent of the anaphor apna in (1) and (2) since ram does not c - command the anaphor in these sentences. Thus, c - comand can account for the relevant facts in neither its present form nor in its suggested modified form. It can, therefore, justifiably be claimed that c - command does not play a crucial role in the theory of binding in the grammar of Hindi.