In order to track diachronic changes in archaeological sequences, researchers typically partition time into stratigraphic layers defined during fieldwork, which serve as the framework for ensuing analyses. These analytical units have a significant impact on archaeological inference, defining its resolution, and influencing both the study of cultural assemblages and the reconstruction of past environments. However, field layers are seldom re-evaluated after excavation despite the fact that archaeological deposits are now commonly recognised as often containing material ‘mixed’ together by site formation processes, excavation techniques or analytical practices. Although the analysis of intra-site spatial data clearly offers a means to overcome these issues, our literature review of 192 journal articles revealed the potential of this data (notably vertical projections of piece-plotted artefacts) to be under-exploited in Prehistoric archaeology. Here we advocate for the development of a more spatially-informed framework for interpretation that we refer to as Post-Excavation Stratigraphies or PES. After proposing a definition for “PES”, we attempt to develop a framework for theoretical considerations underlying their implication, importance, and potential. Three main benefits of PES are highlighted: ensuring assemblage reliability, increased chronological and spatial resolution, and more reliable interpretations based on a multi-stratigraphic approach. We contend that the stratigraphy defined during fieldwork is insufficient and potentially misleading. By providing a different “stratigraphic view” of the same sequence, each specialist can contribute data that, when combined, produces a better understanding of interactions between changes in, for example, technological or cultural traditions, subsistence strategies, or paleoenvironments.

No abstract

This paper presents SEAHORS, an R shiny application available as an R package, dedicated to the intra-site spatial analysis of piece-plotted archaeological remains. This opensource script generates 2D and 3D scatter and density plots for archaeological objects located with cartesian coordinates. Many different GIS software already exist for this, but they mostly require specific skills and training to be used and are rarely designed for the particular needs of archaeological applications. The goal of SEAHORS is to make the two and three-dimensional intra-site spatial exploration of archaeological data as user-friendly as possible, in order to give the opportunity to researchers not familiar with GIS and R software to utilise such approaches. SEAHORS has an easily accessible interface and can import data from text and Excel files (.csv and .xls/xlsx respectively) without preformatting. The application includes functions to concatenate columns and to merge databases, for example when spatial data (XYZ coordinates) and analytical data (e.g. taxonomical attribution of faunal remains, typo-technological attributes of artifacts, etc.) are stored in separate files. SEAHORS can generate five types of plots: 3D, 2D and density plots, as well as 2D plots by slices (i.e. subdivisions according to a third dimension) and 2D plots with a modification of the angle of projection (i.e. to explore spatial organization without the constraints of the field grid orientation). SEAHORS has visualization tools with several sorting and formatting options (color, size, etc.) applicable to coordinates and all possible analytical variables (i.e. levels, spits, identified species, taphonomical alterations, etc.). Orthophotos can also be imported and directly used in the program. The application also allows the grouping of objects into new variables by selecting items on the interactive 2D plots. We present an overview of the application's functions by using the case study of the Cassenade Paleolithic site (France).

The section labelled 'Conclusions' would benefit from a re-titling -it is really a very thoughtful discission on machines in archaeology more broadly. It makes a series of excellent points about how approaches to machine learning and the labels we apply to it uphold an unhelpful nature/culture dualism. This is a great point and not one that I have seen made elsewhere. Breaking this part of the paper into its own section would help drive this home and make the paper more readable. I suggest shifting the sub-heading 'conclusion' to sit before the final three paragraphs and re-naming the current 'conclusions' section with a different sub-heading.