Initial Sketches

The user first specifies a search term in the search bar. This example uses masz “goat”:

1. Shows a histogram of counts associated with this object. The example shows that masz is predominantly counted in very small groups. Can be used to identify items with multimodal distributions, reflecting items with multiple different uses (e.g. small rations for one demographic, and larger rations for a different demographic). If an item can be counted with different systems, radio buttons allow switching between these systems for comparison.
2. Summary statistics show the number of times this object is mentioned in the corpus, how many of the item are recorded overall, and computes the mean and variance of the observed counts.
3. By measuring divergence between the distribution of two items, we can show a list of items with similar distributions to the chosen object. In the first example, we only display the difference between the chosen item (masz) and the similar item (udu): the +10% over the first column tells us that udu is 10% more likely to occur with a small count (between 0 and 30 animals) than masz. In the second example we show the full distribution of both items as histograms with stacked bars. The information displayed is the same, and the user can switch between the two views according to their preference.
4. A concordance shows examples of this word in context: the Value column shows the value of the associated numeral in modern units; Entry shows the text of an entry containing the chosen term; and Occurences shows the number of times this entry occurs in the corpus.
5. A table shows which other items are counted in tablets containing the chosen term. In this example we see that other animals (udu, sila4) and grain products (sze-bi, ziz2) are common in tablets which record masz. This helps the user identify groups of items which are commonly recorded together, which may represent administrative subgenres.
6. A table shows which adjectives/descriptors are used to modify the chosen search term. In this example we see that masz is usually recorded on its own (252 instances with “none” as a modifier), but sometimes it is modified by bar-dul5, other times by bar-dul5 and babbar2, and so on. This helps the user identify the different uses of the chosen object: in this example, masz appears most often as a source of wool for garments (bar-dul5) and white goats (babbar2) are the most commonly mentioned in this usage.
7. The user can filter by adjectives/modifiers by specifying them in the search bar (e.g. “masz mod:babbar2” to search only for white goats). The user can also filter by number system (“szum2 num:weight” to search for instances of szum2 “garlic” that are recorded by weight.)

Alternative: A force-directed graph offers an alternative view of the information in points 5 and 6. For 5, we construct a graph with a node for every commodity which occurs alongside the chosen term masz. Nodes are attracted to one another with a strength proportional to the number of times they occur in a tablet together. The graph will naturally organize itself into cliques of items which tend to be recorded together. In the example, we see a clique of animal terms which presumably represents a genre of livestock texts; some animals such as ansze (an equid) are more distant from this clique (b) reflecting their use in different contexts (probably agricultural texts, with apin “plow”). This gives an immediate impression of the different genres, without having to read through a table.

For 6 we can likewise group modifiers by the frequency with which they occur together. This quickly reveals different uses of the search term masz: as a source of wool for garment production (represented by clique (a): bar-dul5, babbar2, and ge6), and as a recipient of different types of food (clique (b): gu7, sze, and munu4). The distance between these cliques (c) implies that these modifiers are rarely used together: e.g., if an animal is used for its wool, its diet is not recorded.