Holdaway and Wandsnider -- Time in Archaeology

 Created: 27 May 2013  Modified: 23 Jun 2017   BibTeX Entry   RIS Citation  Print

Notes on (Holdaway and Wandsnider 2008)

Murray’s conclusion is very thought provoking (as is Bailey’s chapter).

Murray lumps evolutionary archaeology in with the prevailing paradigm of “archaeology as anthropology”, and that’s probably fair because although we know we need a new metaphysic, and have described the meta-theoretical requirements for one, we haven’t really produced one yet. We keep slipping back into evolutionary ecological theory. Although the cladistics work on branching of the cultural information flow might warrant a serious re-look as one foundation of a timelike paradigm.

With respect to CT itself, the major challenge is to:

  1. Stop exclusively using models which require parameters we cannot estimate or measure, especially independently of each other (e.g., N, or mu).

  2. Develop a time-like set of expectations for things we can actually measure.

  3. Build bridges between ecological time theory and those time-like expectations.

Current Thinking

A really basic problem is that folks have been trying to use CT models that require ecological parameterization to say things about “how much variation we should expect” in the record. And then making claims about CT processes and thus social aspects of past societies based upon deviations from those expectations.

But we haven’t demonstrated that we can securely tie “how much variation to expect” to both a model and the record.

  • This gets to empirical insufficiency (and thus dynamic insufficiency) of models
  • Impedance mismatch between models and the record – time perspectivism/time-like vs. space-like.
  • Lack of proper focus on unit formation and how units, time scales, and model quantities interact


  1. Models where we don’t need to know N and mu, and instead look at “how much variation” there is depending upon time scale of observation and granularity of classification – at at least three levels of granularity to establish scaling relations.

  2. Take an explicitly statistical physical view of these scaling relationships – we’re looking for “phase transitions” and difference phases of behavior in these scaling relationships; models which have the same scaling behavior ultimately can’t be differentiated in a time-perspective record and are thus equifinal.

The central problem is defining CT models which do not require parameters unrecoverable from archaeological sources, and which have direct empirical implication for palimpsest deposits, given flexibility on things like determining deposit duration, choosing classificatory schemes (and possibly more than one).

The result may not be able to make predictions about any specific assemblage, but should be able to make predictions about sets of assemblages relative to one another – flows within metacommunities over time.

Richness itself is highly affected by time-averaging in palimpsests, evenness affected but less so. What we need to study is the scaling of sets of richness and evenness values across groups of assemblages.

References Cited

Holdaway, Simon, and LuAnn Wandsnider. 2008. Time in Archaeology: Time Perspectivism Revisited. University of Utah Press.