In British English, a phonologically deficient us allows for first person singular reference. While descriptions of singular us in a variety of regional dialects abound (Wright 1905, Miller 1993, Edwards 1993, Beal 2004, Anderwald 2004, Snell 2007 a.o), detailed investigation into its syntactic and semantic behavior is non-existent. This paper presents two empirical observations about singular us: (i) it cannot be locally c-commanded by I in dream reports and (ii) it cannot be a bound variable. It is argued that singular us is phi-clitic (Dechaine & Wiltschko 2002)—that is, specified for person, but not number, features (it is underlyingly number neutral). It is shown that this featurally deficient clitic analysis is able to account for the behavior of singular us in (i) and (ii). The paper also discusses Snell's (2007) claim that singular us functions to establish solidarity between speaker and addressee (i.e., it is a polite pronoun, Brown & Levinson 1987). It is shown that the proposed analysis supports this claim.