| {vo'a, -e,-i,-o,-u} refer to the sumti occuppying the corresponding {fV}
places in the uppermost of nested bridi, i.e., the bridi in which the others are nested. This clarifies an apparent conflict between the Book and the cmavo list. {no'a [xiPA]} This repeats the bridi PA levels up from the place where it occurs. The default {no'a} = {no'axipa} the bridi in which the occurrence is immediately nested. The topmost bridi in the nesting chain (the one to whose sumti {vo'V} refer) is always reachable as {no'axiro}. For counting purposes, a new level begins as soon as a subordinate bridi is guaranteed: at NU or NOI [are there others? - LE had best not count or this whole thing gets to be too complicated]. The ordering of the levels (from the bottom up rather than top down) and of the default case (lowest rather than highest) were based on practical considerations: what would most likely be used and which could be calculated most easily. The rule about when a new level starts is controversial, since it allows for paradoxes: reference to incomplete bridi, to sumti that have not yet appeared, and even self-referencing. However, given that this system defines reference by level, any other version is totally arbitrary, and every version allows these same problems at some point (indeed, in intrasentence anaphora of this sort, every reference to the present or higher bridi must be to an incomplete object, since the bridi of which the present reference is a part, cannot be complete until after this reference is done). {nei} repeats the bridi in which it occurs. This leads to more immediate paradox, since {nei} standing alone is presumably a complete bridi, namely itself -- desperately hard to interpret. However, things like {le nei} are needed to repeat sumti in that bridi for reflexives and the like when the bridi involved is not the topmost one, for which {vo'V} are used. [I can't help wondering if, were we not now frozen in, this whole system could have been rendered somewhat less messy. For example, to get sumti from various levels, perhaps {vo'V[xiPA]} could have been used, avoiding the messy bridi anaphora altogether ({vo'Vxiro} = {vo'V}). But if bridi anaphora is needed, perhaps it would be better to recognize that LE too starts a subordinate bridi and then do without {nei}, thus avoiding one round of paradoxes and yet covering all the practical cases (I think, but have not pushed the process too far). ] |