Also Fijian keda ‘we (plural incl.)’. Forms that appear to reflect *i-kita have simply incorporated the closely bound personal case marker for first and second person reference (less commonly, some languages have extended a reflex of *si, the bound personal case marker for third person reference to the base *ita or *kita. Since this was not an affix in PAn it is not treated as part of morphology here. It has long been known that first person plural pronouns have two distinct shapes (*ami/kami, *ita/kita), but both the morphology and the semantics that distinguish these pairs have remained elusive. Some languages have k-initial forms for all non-third person forms, as with Amis kako ‘I’, kiso ‘you’, ciŋra ~ ciira ‘s/he’, kita ‘we (incl.)’, kami ‘we (excl.)’, kamo ‘you (pl.)’, caŋra ~ caira ‘they’. Others, as Karo Batak or Toba Batak, contrast k-initial forms which function as more formal expressions of plural reference, or as polite 2nd sg. reference, with vowel-initial forms which function as familiar or intimate expressions of plural reference, and are not used at all in singular reference. There is little evidence that the forms *kita and *kami are bimorphemic, and the best option based on present evidence appears to be that vowel-initial and *k-initial forms were variants that conveyed distinctions of formality in PAn. Reflexes that are ambiguous for *ita or *kita have been assigned to the latter, since it is somewhat more common than the former.