I do not consider that (B) is an objection for our analysis because you have introduced additional puzzles about the identity conditions of ‘persons’ and our object-dependent view relates primarily to agents instead of person (cf. the end of the paper and the distinction of the object-dependence view with the deflationist view about agent tracking). We are in agreement to state that identity conditions of ‘persons’ might differ from identity conditions of ‘(bodily and embodied) agents’ – and that the metaphysics of ‘personhood’ is a debatable topic. In addition, although it is outside the direct scope of the paper, I find it plausible to consider that ‘personhood’ may be dependent on cultural and social phenomena in a way that would make it dependent not only on bodily or psychological internal continuity but also on emotional and social relations.

Let me consider your first example, the old-friend recognition case. Even though alternative metaphysical options remain possible, the case can be treated as being a standard case of re-identification. Although the two bodies have changed, each is the SAME body according to the framework of the object-dependence view for (1a) it follows the standard organizations rules of biological organisms and their historical transformations (1b) you assume that the two persons are unique namely because although their organism has changed as a function of time, each continuously changing organism has followed a unique and traceable spatio-temporal path during its own history. Moreover most of the social or technological evidence you can use to judge that this agent is the same agent as the person you knew by acquaintance long before may rely on the anchoring provided by the object-dependence view (cf. for instance my note below on biometrics). Please consider your ID card and the ID cards with fingerprints and the causal/historical theories of proper names (Kripke, Donnellan, Perry etc): proper names may originate in something like the ‘baptism’ of a physical object/agent which is the starting point of a social mechanism of automatic tracking of a single and continuing physical object via the proper name and the ID card with the proper name written on it.

Finally, I like the idea you mentioned with respect to the result of Simon and Levin. One should account for these cases in terms of a tracking dynamics based on alternating between using different object, agent and person files. This would account for the incremental characteristics of perceptual identification. It remains highly speculative.