Words such as ‘trust’ and ‘reliance’ are used as context requires, now in this way, now in that, and they serve to cover loose, overlapping clusters of attitudes and actions. Here I invoke some theoretical licence, however, and use the terms to tag distinct phenomena: ‘reliance’, a generic phenomenon, and ‘trust’, a species of that genus. I want to argue that, while the Internet may offer novel, rational opportunities for other forms of reliance, it does not generally create such openings for what is here called trust.

The chapter is in three sections. In the first, I set up the distinction between trust and reliance. In the second, I outline some different forms that trust may take. And then, in the final section, I present some reasons for thinking that trust, as distinct from other forms of reliance, is not well-served by interactions on the Internet, at least not if the interactants are otherwise unknown to one another. The chapter follows up on a paper I published in 1995; it draws freely on some arguments in that piece (Pettit 1995).

The Internet is exciting in great part because of the way it equips each of us to assume different personas, unburdened by pregiven marks of identity like gender, age, profession, class, and so on. A very good question, then, is whether people can develop trust in one another's personas under the shared assumption that persona may not correspond to person in such marks of identity.