1 / 28

Ptolemy's Britain and Ireland: A New Digital Reconstruction

Ptolemy's Britain and Ireland: A New Digital Reconstruction. Presenter: Dmitri Gusev Associate Professor Computer and Information Technology Purdue University. Co-authors: Corey Abshire , Data Science, Indiana University Anthony Durham , Greenwich, United Kingdom

deloresp
Télécharger la présentation

Ptolemy's Britain and Ireland: A New Digital Reconstruction

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Ptolemy's Britain and Ireland: A New Digital Reconstruction Presenter:Dmitri Gusev Associate Professor Computer and Information Technology Purdue University Co-authors:Corey Abshire, Data Science, Indiana University Anthony Durham, Greenwich, United Kingdom Sergey Stafeyev, VNIIGeoSystem, Moscow, Russia

  2. Ptolemy’s “Geography” • 6,300+ objects (rivers, mountains, cities, etc.) with ancient coordinates • The coordinates need georeferencing to become usable in the modern GIS • 170+ objects with coordinates in Britain and Ireland: • Albion – 133 points • Hibernia – 40 points

  3. Our Prior Work • Abshire, C., Gusev, D. A., Stafeyev, S.K. (2017). The Fertile Crescent in Ptolemy’s “Geography”: A New Digital Reconstruction for Modern GIS Tools. In Proceedings of the 12th ICA Conference on Digital Approaches to Cartographic Heritage (pp. 152-167). Venice, Italy: International Cartographic Association. • Abshire, C., Gusev, D.A., and Stafeyev, S.K. (2016). Enhanced Mathematical Method for Visualizing Ptolemy’s Arabia in Modern GIS Tools. In Proceedings of the 11th ICA Conference on Digital Approaches to Cartographic Heritage (pp. 133–154). Riga, Latvia: International Cartographic Association. • Abshire, C., Gusev, D. A., Papapanagiotou I., Stafeyev, S. K. (2016). A Mathematical Method for Visualizing Ptolemy’s India in Modern GIS Tools. e-Perimetron, 11(1), 13-34. • Gusev, D. A., Stafeyev, S. K., Filatova, L. M. (2005). Iterative Reconstruction of Ptolemy’s West Africa. The 10th International Conference on the Problems of Civilization, Moscow, Russia. This work was also presented (2006) at the GIS Seminar, Rochester Institute of Technology, Rochester, New York.

  4. The Geography of Roman Britain(55 BC to 400 AD) Ancient names have survived for about 570 distinct locations of • forts & camps • towns & settlements • rivers & estuaries • promontories & islands • “tribes” Where do these names belong on the map? What did these names originally mean? Art: Mateo Maté. Prophetia. La aparición de lo real (2015)

  5. The Ancient Sources Ptolemy’s Geography Ravenna Cosmography– 307 names listed in geographically logical sequences AntonineItinerary– 110 places with soldiers’ marching distances between NotitiaDignitatum– 49 forts where army units were posted Tabula Peutingeriana– the fragment shown above various inscriptions, coins, and textual mentions with lots of spelling variation, due to copying errors, etc.

  6. Ptolemy’s Criticism of a Round Mappa Mundi • 8.1.2. Now that we have seen what rendering of the whole ecumene in a single map would be suitable, the next thing is to set out the summary outlines to be if we divide it into several maps in order to put in the actual data in full and in scale for clarity. For in a single drawing where we must keep the proportion of the parts of the ecumene to each other it is necessary for some of the parts to be crowded because of the wealth of the data being shown and for others to be wasted for lack of data to be shown. • 8.1.3. To evade this most were forced by the maps themselves, but not by the matter, to distort the sizes and shapes of the countries extensively. Thus those who allotted the greatest part of the map to Europe in both longitude and latitude for the wealth of data being shown, and the least part in longitude to Asia and in latitude to Libya for the contrary. For this reason they turned the Indian ocean beyond Taprobane northward as the map prevented their extending it eastward while they had nothing to put in against Scythia lying to the north, and they turned the western ocean eastward as the map prevented their extending it southward while here too the depth of interior Libya and of India did not have anything to be put in to continue the western coast. In this way the notion of the whole earth surrounded by ocean began from errors in drawingand ended in unproved doctrine. • Source: Aubrey Diller’s 1984 translation of Book 8, http://www.dioi.org/vols/w50.pdf

  7. Objectives • Estimate modern coordinates for every location described in Ptolemy’s Albion and Hibernia • Develop modern GIS assets for that set of locations and other related ancient sources (the Antonine Itinerary and Ravenna Cosmography) • Study the linguistic aspects of toponym identification specific to ancient Britain and Ireland • Continue to validate our two main mathematical methods: triangulation and flocking

  8. Method 1: Triangulation We compute a Delaunay triangulation over all the known points, and then choose the 3 points of the triangle in which each unknown point falls. We then estimate the modern coordinates of an unknown place by combining known coordinates of the three surrounding places. The known locations are combined by a weighted average using the areas of the three spherical triangles formed. Problem: Each of the unknown places to be predicted must be enclosed by a spherical triangle formed by known points

  9. Method 2: Modified Flocking k’1 k’2 u’ k1 k2 u k’3 k3 Center all ancient points over known points. Take the k-nearest neighbors for each unknown(we used k=3) For each of those three, find the vector that describes the movement of the point in ancient coordinates to its position in modern coordinates. Take a weighted average of those vectors to derive a new vector to describe the movement of the unknown point. Apply this new vector to estimate the modern coordinates of the unknown point. Codenamed “flocking” approach since it reminds us of Reynold’s Boids flocking model.

  10. Post-Processing: Bayesian Adjustment x = Our prior belief is that the places we’re trying to locate are somewhere in Britain or Ireland. Based on known locations of surrounding places, our “evidence” is that the point is somewhere in a given location. We treat this as a bivariate normal distribution over the sphere. Using Bayes rule, we combine the evidence with our prior to come up with our posterior distribution. We take the MAP as our adjusted point.

  11. Point Classification • Known points • Unknown points • Tentatively identified points • Duplicates

  12. Prior Art: The Place-Names of Roman Britainby Leo Rivet and Colin Smith (1979) is a superb book, by brilliant authors, but: it cannot explain and/or locate 20% of names, it goes actively wrong with many more names. Where are the problems? Can we do any better?

  13. How to Make Progress • Exploit all Internet resources: maps, archaeology, texts, dictionaries, etc. • Dump the dogma: distrust theories and double-check all data. • Treat names collectively, as sequences or groups, and with parallels elsewhere. • Ask what were the salient features of each place to people there at the time.

  14. A Gigantic Game of Battleships!Each square has a certain probability of being correct plus a set of logical interactions with other squares

  15. Name Meanings Are names built from recognisable words? How were they built grammatically? Can meanings help to find locations? Modern investigators often assume that: 1. Indigenous people created early names. 2. They used their native languages. 3. Iron-Age Britons all spoke Celtic languages. Any or all of these assumptions may be untrue!

  16. Languages Present in Early Britain Latin – top dog, used for most writing & admin Greek – surprisingly prevalent Germanic – mother tongues of many soldiers Celtic – precursors of Welsh and Cornish But no one truly knows where on the track from proto-Indo-European (origin before 3000 BC) to Old Welsh and Old English (written down after 600 AD) native British dialects were in Roman times.

  17. Easy to translate, but not always easy to locate Anicetis = ‘undefeated’ ?= Cadbury Castle Antrum = ‘cave’ = Whitby (narrow harbour entrance) Βουλλαιον ≈ ‘council chamber’ ?= in south Wales Burrium = ‘orangey red’ = Usk (stone fort) ΚαινοςΛιμην = ‘new harbour’ ?= Newhaven? Portslade-by-Sea? (see the next slide!) Cunis = ‘wedge’ = Ailsa Craig (volcanic plug) Epocessa = ‘settlement of a colony’ = ? Iuliocenon ≈ ‘treeless’ = Hardknott (bare hilltop fort) Λουεντινον ≈ ‘washing’ = Pumsaint gold mines Mancunio ≈ ‘maimed, on the left’ = Manchester fort Οχθηυψηληυ = ‘high banks’ ?= Clyth Ness Praetorio = ‘commander’s residence’ ?= on Yorkshire coast Punctuobice ≈ ‘point throw’ ?= Avebury stone circle Ρηγνοι ≈ ‘broken asunder’ = tribe round Chichester Stodoion = ‘storehouse’ = Birrens (granary fort) Salinae = ‘salt works’ – three of them; tentatively the one near Wisbech Tripontium = ‘three bridges’ – on Watling Street Vernalis = ‘of spring’ ?= Cornwall Ουικτωρια = Victoria = site of monsGraupiusbattle(All Oversimplified!)

  18. Finding Καινος Λιμην (‘new harbour’) Ptolemaic map Our map Our methods point toward the modern Lympne, known in antiquity as Portus Lemanis, or simply Lemanae.

  19. Some Elements Used in Early Names Alauna, Isca, Navis = ‘river’ – but which type?? -ar ≈ ‘joint’ – at river confluences etc Bri-/Briga ≈ ‘brow/brink/bridge/brim’ Duro- ≈ ‘door’ – at river crossings -dunum ≈ ‘hill fort’ – usually a native power centre Ebur = ‘ivory’ – boars’ tusks (military emblem) Leuk-/Lug- ≈ ‘light’ – at places with a good view Mag-/-magus ≈ ‘power’ Ocelus = ‘lookout’ – resembles Latin for ‘little eye’ Tin-/Ten-/Tan- = ‘extending, sticking out’ –uba/–oba = ‘water, river’ ?= ‘bank’ Uxela = ‘high point’ -vicus = ‘outlying’ – often a Romanised admin unit Vindo- = ‘fair, pleasant’

  20. Ireland The Roman army never reached Ireland, so all we have is Ptolemy, who supplied 53 names in Hibernia. They are mostly pure Greek, or general Indo-European (“pre-Celtic”). It is at least as easy to find later parallels in English as in Gaelic. Much the same holds for the western isles of Scotland, but there the Ravenna Cosmography supplies names too. Some real puzzles remain. Castlereban = Ptol. Raeba?? The River Fane = Ptol. Vinderis R.??

  21. Hibernia Before and After Triangulation

  22. Unsolved Problems with the Algorithms • Sometimes the unknown points that Ptolemy places along the shoreline are pulled inland by our current methods, following a false concavity • Some islands joined the mainland (e.g. Ptolemy’s Tanatis I. = Isle of Thanet) • In other regions, such as India and Mesopotamia, the need to place unknown points close by rivers continues to pose a significant challenge • Our current methods do not explicitly track rivers • Rivers are well-known to change their course • Canals and reservoirs were built within the historical time period

  23. Model Validation / Precision Analysis Error histogram: India before the Ganges, triangulation w/ Bayesian adjustment Error histogram: Arabia Felix, triangulation w/ Bayesian adjustment

  24. Model Validation / Precision Analysis Error histogram: Judaea Palestina, flocking w/ Bayesian adjustment Error histogram: Syria, flocking w/ Bayesian adjustment

  25. Model Validation / Precision Analysis Error histogram: Hibernia, flocking w/ Bayesian adjustment Error histogram: Albion, flocking w/ Bayesian adjustment

  26. Some General Lessons on Names • Old thinking that name elements have parallels only in the modern Celtic languages is generally unhelpful. • It is best to stay agnostic about languages and to think mainly in terms of human geography. • Religion has been overstated as a motivator of names. People have always been pragmatic. • Ptolemy's names of tribes tend to be outsiders' descriptions of a lifestyle (e.g., Ichthyophagi = ‘fish eaters’) more than insiders' proclamations of identity and/or political unity. • Anyone with a critical mind and an Internet connection can still make new discoveries.

  27. Future Work • Improve models: • Use more data to estimate unknown coordinates • Incorporate river, mountain, and coastal data • Improve tentative identification verification • Improve known coordinates • More linguistic analysis • Apply the models to more regions in Ptolemy’s Geography • Place the tribes listed by Ptolemy • Improve deliverables: • Better GIS assets (e.g., more feature-rich KML files, ArcGIS filesets, etc...) • Translations and internationalization work • Website (http://www.claudiusptolemy.org/) and feedback mechanisms • Dr. Anthony Durham’s website http://www.romaneranames.uk/ − exceptionally good!

More Related