Figures
1.1 Topology of the region’s trade routes in Antiquity. Map: author 25
1.2 Example 1: Iatrippa to Hegra. LCP black, reconstructed routes dashed, Lava Fields hatched. Map: author 32
1.3 Example 2: Hegra to Tayma. LCP black, reconstructed routes dashed, Lava Fields hatched. Map: author 34
1.4 Reconstructed trade routes and Least Cost Paths in comparison. LCP black, reconstructed routes dashed, Lava Fields hatched, dune seas dotted. Map: author 37
2.1 Copper alloy implement found at Pyla-Kokkinokremos. Photograph: Joachim Bretschneider 47
2.2 3D model of the Pyla-Kokkinokremos plateau with indication of the different sectors investigated by the C-PEPL project (Cyprus—Pyla Excavation Project Louvain/Gent). The copper alloy implement was found in Sector 3. 3D model: Nicolas Kress 49
2.3 Alabaster vase and associated objects found in Sector 3 at Pyla-Kokkinokremos a) the object of this study, b) copper alloy object, c) bone or ivory fragments. Photograph: J. Bretschneider 50
2.4 Bronze figurines with stylistic and chronological similarities to the copper alloy implement found in Sector 3 at Kokkinokremos. a) Catling 1971, fig. 3; b) Voyatzis 1985, pl. XX; c) Karageorghis and Demas 1984, pl. XXV 51
2.5 Orthographic snapshots of the 3D model of the implement in different views. Image: authors 54
2.6 Mesh density check of the 3D model. Six samples from different locations on the 3D models were taken to compare their surface and point density. Image: authors 55
2.7 Two different views of the braided hair or headpiece. Gap circled in red. Image: authors 57
2.8 Two different views of the broken braids A and B at the back of the head. Image: authors 57
2.9 Small circular protuberances above the eyes are indicated by the red arrows. Image: authors 58
2.10 Transition line highlighted in red between the upper part of the implement and the blade. Image: authors 58
2.11 Colorization of the surface according to curvature (a) in MeshLab and roughness (b) in CloudCompare. Images: authors 59
2.12 Roughness comparison between the anthropomorphic head and the blade in Cloudcompare. Image: authors 60
2.13 Orthophotos of the blade’s edge: a) backside with smoothed edge; b) colourised Meshlab curvature analysis, highlighting the smoothed edge; c) front side non-smoothed edge. Images: authors 61
2.14 a) left side of the implement’s anthropomorphic head, with rough surface, cheekbone and jawline visible. b) right side exhibiting a smooth surface. Image: authors 61
2.15 Virtual reconstruction of the suggested use of the implement. 3D reconstruction: Charalambos Ioannou in 3DsMax 62
3.1 An exemplar setup of a MLR acquisition dome. Figure created in SketchUp Make v16.0.19912 64bit. © KU Leuven, Hendrik Hameeuw 78
3.2 From top to bottom: a Greek silver coin (o.i. 522, © KU Leuven Art Collection), a cylinder seal impression (O.00861, © Art & History Museum, Brussels), impressions on the bottom of a funerary cone (E.03984, © Art & History Museum, Brussels), and a cuneiform tablet (O.00095, © Art & History Museum, Brussels). For a high-quality colour version of this figure, see <https://doi.org/10.5281/zenodo.2445220>. For each artefact four visualizations were generated with the PLD MLR viewer; from left to right: textured (colour), shaded, automated sketch and normal map. Image: authors 83
3.3 Study of ancient smoothing marks left by fingers on the surface of an Egyptian execration figurine (E.07453, © Art & History Museum, Brussels), PLDviewer shaded exaggerated mode + measure tool (annotation with arrows in Photoshop). Image: authors 85
3.4 Recto of an Egyptian figurine (E.07450, © Art & History Museum, Brussels), 3D models generated in the PLD viewer. Top right: example of the measuring tool in the same viewer. Image: authors 85
3.5 Inspecting the 3D surface of a Neo-Assyrian relief fragment from Nineveh, Iraq (O.01923, © Art & History Museum, Brussels), PLDviewer shaded mode + measuring tool. Image: authors 86
3.6 3D model of a Punic coin with minnaert-shader in MeshLab based on PLD MLR dataset (Naster1869, © Royal Library of Belgium). Image: authors 87
3.7 Reverse of an Egyptian execration figurine (E.09076, © Art & History Museum, Brussels), pre-processing and post-processing MS results. (For a high-quality colour version of this figure, see: <https://doi.org/10.5281/zenodo.2445458>; consult this DOI especially for the ‘MS MLR PLD result, false-colour PCA’) Image: authors 91
3.8 Trichromatic false-colours representations of a detail on an ancient Egyptian jar (E.02458, © Art & History Museum Brussels), top: Red Green Blue rendering, bottom: IR-Red-Green rendering. Image: authors 93
3.9 MS MLR PLD-generated histograms of similar appearing black/brown pigments on the original and restored sections of an ancient Egyptian jar (E.02458, © Art & History Museum, Brussels). Image: authors 94
3.10 Detail on a fragment of an ancient Egyptian coffin (Thebes, end twenty-second or early twenty-fifth Dynasty; © Archaeological Collections Faculty of Arts, KU Leuven). A: photograph with white light; B: green-induced infrared luminescence (523 nm radiation); C: red-induced infrared luminescence (623 nm radiation); D: normal map (surface relief estimations) of isolated Egyptian blue, based on result in C; see also Hameeuw, Van der Perre, and Boschloos forthcoming. Image: authors 96
3.11 Normal maps of the surface of a gold coin (o.i.613, © KU Leuven Art Collection), based on five different MLR datasets (all produced during the same MS PLD recording) + PLD measuring tool. Image: authors 98
3.12 MS PLD recording on a section of an ancient Egyptian coffin (of Butehamun, Egyptian twenty-first Dynasty; E.05288, © Art & History Museum, Brussels) with visualisation of the coffin relief’s uppermost surface (d. with UV) and inside the varnish (c. with IR) + PLD measuring tool. Image: authors 99
4.1 Simplified model of the Dig ANES/ACAWAI-CS database. Image: authors 120
4.2 Modern impression of an Old Babylonian cylinder seal (after Collon 1986: no. 38) illustrating the labeling of picture elements used by the projects Dig ANES/ACAWAI-CS. Image: authors 124
4.3 Word clouds can be used to visualize the main components of Old Babylonian cylinder seals. The examples given derive from the labeling and quantitative analysis of 8646 pictorial elements from 1870 seal images dating to the Old Babylonian period: a) all pictorial elements; b) gender distribution of human figures; c) gender distribution of divine figures; d) hairdo of male figures; e) beard styles of male figures. Image: authors 127
4.4 Variants of the figure conventionally called “priest” in Old Babylonian seal iconography. The diagram above is based on the total number of occurrences in the Dig ANES/ACAWAI-CS database; the diagram below splits the data according to the publications used. The calculations are based on the labeling of 8646 pictorial elements from 1870 seal images dating to the Old Babylonian period in the Dig ANES/ACAWAI-CS database. Image: authors 129
4.5 Contextualization of the visual features typical for the “priest” through quantitative methods. Image details in the centre of this figure taken from scans of modern roll outs of cylinder seals housed at the LMU Institute of Near Eastern Archaeology, Munich. Original seals from the British Museum: BM 129508, 129524, 89432, 89058. Image: authors 130
4.6 Network visualization of pictorial elements co-occurring with the “monkey” on the same seal image. Created with the network visualization tool Gephi. Image: authors 131
5.1 A three-dimensional space with three arrows, the tips of which represent vectors. Image: first author 144
5.2 Querying for fastText results for the fifteen most similar words for the divine name Nergal in our dataset; vowel length is not indicated in proper nouns, and thus the spellings of some divine names differ from the spellings used in the chapter text. Image: first author 150
5.3 An imaginary two-dimensional vector space with vectors for the Akkadian words šarru (“king”), sinništu (“female, woman”), and zikaru (“male, man”). The difference between the coordinates for sinništu and zikaru (in black) is added to the vector for šarru and the vectors of the words that are closest to the position found are amtu (“maid, female slave”) and mārtu (“daughter”). Image: first author 154
5.4 The most similar gods to Nergal visualized as an ego network graph. Image: first author 160
5.5 The thirty semantically most similar words to the average position between the word vectors of Nergal and digi.du visualized as an R scatter plot. Image: first author 165
6.1 Qoh. 1:1,5. Text and critical apparatus from BHS and BHQ editions. Image: authors 196
6.2 An example of DSL apparatus entry from Qoh. 1:1. Image: authors 196
6.3 Parse tree applied to Qoh DSL. Image: authors 207
6.4 Dendrogram to visualize similarities among witnesses. Image: authors 207
6.5 Multidimensional Scaling Map. Image: authors 208
6.6 Phylogram of the textual tradition. Image: authors 208
6.7 An example of critical apparatus created in LaTeX. Image: authors 209
6.8 SQL search through the web interface. Image: authors 215
6.9 Domain ontology: modelling ritual actions. Image: authors 215
6.10 Domain ontology: relevant relations. Image: authors 216
6.11 Example of a query using the ontology. Image: authors 217
7.1 J.M. Hutton imaging a small Roman-era object at the Muzeul Banatului in Timișoara. Photograph: Pruett, April 2018 231
7.2 Image of PAT 959 with conventional photography. (Image available at UWDC, <https://digital.library.wisc.edu/1711.dl/DXP5S3BH4LGQG8C> [accessed June 28, 2022], CC BY-NC 4.0) Photograph: Greene and Hutton, January 2015 233
7.3 Detail of RTI-image of PAT 959. (Image available at UWDC <https://digital.library.wisc.edu/1711.dl/DXP5S3BH4LGQG8C> [accessed June 28, 2022], CC BY-NC 4.0) Photograph: Greene and Hutton, January 2015 233
7.4 Layout for PAT 1773 on UWDC. (Images available at <https://digital.library.wisc.edu/1711.dl/DB2NPHXXDA3TP8M> [accessed June 28, 2022], CC BY-NC 4.0) Photograph: Bonesho and Greene, July 2013 237
7.5 Line drawing of PAT 1773 by Hutton. Hutton 2016, 146 239
8.1 North-West Semitic world and Greek world. Image: ERC MAP project 260
8.2 Embryonic example. Image: ERC MAP project 265
8.3 Conceptual data model of the MAP database. Image: ERC MAP project 268
8.4 Extract from a “Source” entry in the MAP database. Image: ERC MAP project 272
8.5 Extract from a “Testimony” entry (“To the lady Tinnit face of Baʿal and to the lord Baʿal Ḥammon”) in the MAP database. Image: ERC MAP project 272
8.6 Extract from an “Element” entry (Baʿal) in the MAP database. Image: ERC MAP project 273
9.1 A straightforward hierarchy of levels for a Bible. Diagram: author 310
9.2 A slightly more complex hierarchy of levels for a Bible. Diagram: author 310
10.1 Reduced conceptual scheme of the EbDA database. Diagram: authors 326
10.2 Schematic representation of two texts having a word in common. Diagram: authors 328
11.1 Simple diagram of a RDF triple. Image: author 340
11.2 Extended diagram of a RDF triple. Image: author 340
11.3 Example of instance level representation. Image: author 341
11.4 Knowledge graph based on an Assyriological example. Image: author 344
11.5 Schema level representation of the example in 11.3. Image: author 345
11.6 Virtuoso SPARQL endpoint querying DBpedia. Image: author 353
11.7 Pubby interface from a digital musicology project. Image: authors 354
11.8 Protégé UI displaying the CIDOC-CRM as a hierarchical structure. Image: author 356
11.9 Protégé UI first stage. Image: author 357
11.10 Inserting the first
11.11 Naming the first
11.12 The Equivalent Class icon in Protégé. Image: author 358
11.13 Uploading datasets into Web-Karma. Image: author 360
11.14 Web-Karma recognises the uploaded file as an OWL ontology. Image: author 360
11.15 The Web-Karma UI can be used to assert properties. Image: author 361
12.1 Excerpt of Topographic Points from Ugaritica IV. Image: after Schaeffer 1962, Plan I 376
12.2 Data Model in the Ras Shamra Tablet Inventory. Image: author 379
12.3 Akkadian texts in the Royal Palace of Ugarit. Image: author 380
12.4 Texts from the Royal Palace attesting the words “list” and “king”. Image: author 381
13.1 HUMOT: homepage of the project. Image: author 392
13.2 The database interface and its tools. Image: author 396
13.3 Outcomes of the search for “medicus”. Image: author 396
13.4 Six main entities of the databases (“categories”) that can be used as search criteria. Image: author 397
13.5 Example of biographical entry in HUMOT. Image: author 401
13.6 Example of prosopographic tabular comparison of the two main geographic areas. Image: author 405
14.1 Hierarchy for the top-level concept of “Humanity”. Image: author 417
14.2 Hyperlinked prose for the concept “Human Capacities”. Image: author 419
14.3 LSTO data for the concept “God’s Omnipotence”. Image: author 421
14.4 Word counts for the “Humanity” branch of the hierarchy. Image: author 424
14.5 Counts for the 20 Most Common Lemmata. Image: author 426
14.6 Top TF-IDF values for
14.7 The number of documents in each of the eight top-level branches of the hierarchy. Image: author 428
Maps
1 General map XLI
2 Toponyms for Mesopotamia, Syria, and the Levant XLII
3 Toponyms for Cyprus XLIII
4 Toponyms for Arabian Peninsula and Egypt XLIV
5 Toponyms for Italian Peninsula, Mediterranean Basin, Hellas, Balkans, and Asia Minor XLV
Tables
1.1 Values established for humans in literature and the approximations for camels 29
8.1 Relational way of organising data 264
8.2 Table of evaluation coefficients for the accuracy of data 277
9.1 Examples of surface forms, lemmas, and morphology from English, Hebrew, and Greek 294
9.2 Sample queries for searching an English Bible, illustrating Boolean operators 294
9.3 Sample queries for consonant-only searches of the lemmas of a Hebrew Bible, illustrating proximity and ordering options 295
9.4 Sample queries for searches of an English Bible illustrating the use of levels 296
9.5 Partial list of types in an English Bible. This ordered list forms a lexicon 297
9.6 A partial inverted index. For each type in the lexicon (conquer, conquering, conqueror, conscience, etc.), it shows the location of every token of that type in the Bible, here represented by book, chapter, verse, and token within the verse 298
9.7 Partial expanded lexicon for an English Bible 302
9.8 Encodings of some small integers using various variable-bit codes 304
9.9 Encodings of some small integers using various variable-bit codes 305
9.10 When storing the types to disk, a type may be compressed by recording the number of characters that are the same as the previous type (compressed), the number of additional characters (compressed), and then those actual additional characters 307
9.11 Instead of storing the entire expanded lexicon on disk or in memory, one can simply store the lexicon and then store the index of each type and number of character rotations for each entry in the expanded lexicon 308
14.1 Church eras in which aligned works were published 423
14.2 Top ten features for branch classification. Each row indicates how much statistical certainty this term contributed to the decision (with 1.0 indicating perfect certainty) 430
14.3 Cross-tabulated classifier results. Each cell is a count of the documents that have the pair of actual and predicted values 430
14.4 Precision and recall by branch 431