Analysis of Prehistoric Iconography with the R package iconr

By definition, prehistorical societies are characterized by the absence of a writing system. During, the largest part of human history, and everywhere in the world, symbolic expressions belong mostly to illiterate societies which express themselves with rock-art paintings, pottery decorations, figurines, statuary, etc., and a lot of now disappeared carved woods, textile design, etc. These graphical expressions are the most significant remaining part of humankind’s symbolism. At the composition level, the presence of recurrent patterns of signs (i.e., graphical syntax) in meaningful associations indicates the existence of social conventions in the way to display and to read these expressions. Well-established and shared methods to record and study these graphical contents would open the possibility of cross-cultural comparisons at a large scale and over the long-term.


Background
By definition, prehistorical societies are characterized by the absence of a writing system. During, the largest part of human history, and everywhere in the world, symbolic expressions belong mostly to illiterate societies which express themselves with rock-art paintings, pottery decorations, figurines, statuary, etc., and a lot of now disappeared carved woods, textile design, etc. These graphical expressions are the most significant remaining part of humankind's symbolism. At the composition level, the presence of recurrent patterns of signs (i.e., graphical syntax) in meaningful associations indicates the existence of social conventions in the way to display and to read these expressions. Well-established and shared methods to record and study these graphical contents would open the possibility of cross-cultural comparisons at a large scale and over the long-term.

Statement of need
Ancient iconography is often perceived as different from other 'current' archaeological remains (lithics, potteries, settlements, etc., Chenorkian, 1995). Indeed, the inherent variability of ancient iconography has led to considerable problems in its study, drastically limiting the possibility to draw a synthesis of graphic expressions at a large scale and over the long-term: • Spatial proximities between the graphic units are not precisely quantified. Graphical units are attached to sub-areas of the support (e.g. upper part of a rock, neck of a pottery, centre of a stele). • Groupings -like graphical units grouped into figures, figures grouped into patterns, patterns grouped into motives, etc. -are not self-explanatory and introduce a tedious number of groups and hinder their systematic analysis. • Relationships and similarities between these groups are often not self-explanatory and unquantified. • Descriptive vocabularies and methods of analysis are site-dependent or perioddependent.
Even the reevaluation of semiotics paradigms following the scientific trends -structuralist turn during the Processual archaeology period, ca 1960-1980 (Binford, 1962;De Saussure, 1989), iconic turn during the Post-processual archaeology period, ca 1980-2010 (Gell, 1998;Hodder & others, 1982), did not led to the development of efficient tools for ancient iconography studies, such as common descriptive variables, or common interpretation grids.

Core functionality
The R package iconr is designed to offer a greater normalization of quantitative indexes for iconography studies (Alexander, 2008;Huet, 2018). It is grounded in graph theory and spatial analysis to offer concepts and functions for modeling prehistoric iconographic compositions and preparing them for further analysis: clustering, typology tree, Harris diagram (i.e. temporal succession of archaeological contexts, Harris, 2014), etc. The main principle of the iconr package is to consider any iconographic composition (here, 'decoration') as a geometric graph of graphical units. Geometric graphs, also known as planar graphs or spatialized graphs, allow to model the neighborhood of these graphical unit which are the fundamental relationships of visual semiotics (  library(iconr) dataDir <-system.file("extdata", package = "iconr") site <-"Cerro Muriano" decor <-"Cerro Muriano

Plot
Plot the Cerro Muriano 1 stele decoration graph with the function plot_dec_grph().

Compare
Compare and classify the iconr decoration training dataset according to pairwise comparisons between decorations based on their common nodes and common edges; functions list_dec() and same_elements().