1. The phenomenon of linguistic switching in Latin writers has been recently analyzed by Adams (2003), who discusses Cicero’s practice in his letters on pages 308–47, commenting, “Since this type of code-switching was a game, and since it took place in private between intimates, it is often light-hearted” (p. 345). Compare Hutchinson (1998, 13–16).
2. Compare Kerbrat-Orecchioni (1998, 27) apropos of both letters and conversation: “Dans toute interaction, les phases d’ouverture et de clôture sont des moments particulièrement délicats pour les interactants, d’où d’ailleurs leur caractère fortement ‘ritualisé.’ ”
3. For statistics about the letter corpus, see appendix 1.
4. Demmel (1962, 325 n. 2) noted the rarity of comment regarding family: “allein dem Briefwechsel mit Atticus der Austausch von Familienangelegenheiten vorbehalten war; wenigstens findet sich sonst nichts dergleichen.”
5. Cic. Phil. 11.5. Appian has a more circumstantial and grisly account at BC 3.26.100–101.
6. The social background of Cicero’s correspondents is conveniently set out in the table in Déniaux (1993, 96–108).
7. The exchange of letters between persons who are both in Rome is treated as unusual at Plut. Caes. 17.8 and Nep. Att. 20.2. At most a handful of letters in the Ciceronian corpus are intra-urban, including possibly Fam. 6.15= 322 SB, 7.22= 331 SB, and 11.1= 325 SB. Déniaux (1993, 114) notes the correlation between absence from Rome and risk: “le départ d’un magistrat pour sa province lui impose un éloignement dommageable au soin de sa dignitas.”
8. The Romans were familiar with carrier pigeons (Steier 1932, 2493–94), but then as now, pigeons could not haul much, they traveled in only one direction, and they are scarcely relevant to the topic of letter writing. After a long survey of alternative