< ^

Hypotheses for Future Work

One can subject the same manuscripts to macro-level and micro-level clustering analysis, anticipating that:

1. Strongly similar clusters will emerge from the analysis of the two independent sets of data, suggesting that the clusters reflect natural properties of the manuscripts as a whole on all levels.
2. Because shared transmission is the most natural explanation for similarity at both the micro and macro levels, correspondence in results at both levels will validate the hypothesis that cluster analysis provides insights into textual transmission. This hypothesis can be falsified by the absence of meaning correlation between clustering at the micro and macro levels.