Fuzzy Chronologies in Longitudinal Archaeological Networks

Abstract

"Fuzzy chronologies are a challenge for computational methods. While vague time concepts translate the imprecision of current archaeological dating as faithfully as possible, they conflict with the crisp timestamps required for longitudinal network studies. The chronological uncertainties of the archaeological record are difficult to express in precise numbers, yet their suppression would considerably change the informational value of both the data set and the reconstructed networks. Datasets with well-delimited dates have been fitted to time-windows through cumulative probabilities in the past, but this process relies on a fixed interval and cannot be readily applied to less certain timestamps. Fuzzy boundaries need to be preprocessed by a complementary method to be solidified, while still preserving their ambiguities. This article will explore the use of fuzzy logic to model the uncertainties of fuzzy timestamps. Attributing different membership values to subintervals of the timestamp allows us to transcribe varying levels of certainty mathematically. These weighted subintervals can be used to calculate adjusted cumulative probabilities for a more faithful representation of data records within longitudinal network studies. The application will be tested in a case study based on imported fine pottery from Amathus in the 8th and 7th century BC."