In criticizing J. L. Austin's theory of speech acts, Jacques Derrida and Jonathan Culler have argued, in part, that Austin misunderstands how signatures function. They claim that he ignores the essentially formal and "iterable" structure of a signature -- a structure that betrays the absence of any subjective consciousness on the part of the signer. I argue that their concept of iterability does not fully apply to this case. Rather, legal practice in England and the United States countenances a wide range of variation for signatures, variation that is consistent with and legitimized by Austinian assumptions about personal agency. The fact that the legal situation is somewhat different in France may explain the structure of Derrida's argument.