Recent Advances in Salzburg Dialectometry
Department of Romance Philology, University of Salzburg, Austria
Correspondence: Hans Goebl, Fachbereich für Romanistik, Akademiestrasse 24, A-5020 Salzburg, Austria. E-mail: Hans.Goebl{at}sbg.ac.at
This paper documents the many taxometric and cartographic achievements of the Salzburg school of dialectometry. The paper discusses the following topics: (1) problems of measurement of linguistic atlas data (with particular consideration of Romance linguistic atlases), (2) establishment of the data matrix, (3) choice of the similarity index (Relative and Weighted Identity Value), (4) generation of the respective similarity and distance matrices, (5) their subsequent cartographic exploitation, which encompasses the following cartographic tools: similarity maps, parameter maps, dendrograms (and their spatial projection), and correlation maps. The ultimate purpose of these highly sophisticated cartographic techniques (choropleth and isopleth maps) is to increase our knowledge of the complex mechanisms of the dialectal management of space by man. From a methodological point of view our paper deals with problems related to (Romance) dialectology and linguistic geography, historical linguistics, numerical classification, statistics and statistical cartography. The examples are drawn from the French linguistic atlas ALF (Atlas linguistique de la France) published by Jules Gilliéron and Edmond Edmont (Paris: Champion, 1902–1910, 10 volumes) more than one hundred years ago. The taxometric calculations and their respective visualizations are realized by a powerful computer program called ‘Visual DialectoMetry’ (VDM), created by Edgar Haimerl (Blaustein, Germany) between 1997 and 2000 in Salzburg, which is freely available for research purposes.
1 We refer to the important centres of dialectometry in the Netherlands and the United States: cf. for instance the works of John Nerbonne (2001, with W. Heeringa, and 2003, with W. A. Kretzschmar), Wilbert Heeringa (2002, with J. Nerbonne and P. Kleiweg, and 2004) and the brothers Cor and Geer Hoppenbrouwers (2001) on the one hand, and those of William A. Kretzschmar (1989, with E. W. Schneider and E. Johnson, and 2003) on the other hand, realized with the data of the East American linguistic atlas LAMSAS.
2 For reasons of space the map types honeycomb map and beam map cannot be analysed here: for a detailed discussion, see Goebl, 1984, I:183–196, and 1983b: passim.
3 Note that the different targets or objectives have to be defined previously by geolinguists or linguists.
5 It is our hope that this research interest should create a strong link between areal, historical and general linguistics through its insistence on an empirical basis in linguistic behaviour.
6 Between 1973 and 1982 I worked mainly at Regensburg University (Bavaria, Germany).
7 Cf. Clue Julve, 1999: passim, and Viaplana, 1999.
8 Cf. Aurrekoetxea, 1992: passim.
10 The ALF-map 14 aiguille (<lat. *ACÚCLA) has been analyzed five times and thus resulted in five working maps: 1: referring to the phonetic results of the pretonic A-, 2: referring to the phonetic results of the intervocalic -C-, 3: referring to the phonetic results of the stressed Ú, 4: referring to the phonetic results of the intervocalic nexus CL, 5: referring to the phonetic results of the final -A.
11 See the respective curves in Goebl, 2002 (13–15) and 2003 (68, 70–71).
12 Note that each atlas site is compared to each of the remaining N – 1 sites. The symbol j is used for the reference site, and the symbol k for the atlas site that j is compared to.
13 I.e. the reference site with the symbol j.
14 This value corresponds to the lower threshold of the interval 1: see the numerical legend of Map 2.
15 This value corresponds to the upper threshold of the interval 6: see the numerical legend of Map 2.
16 This value corresponds to the upper threshold of the interval 3: see the numerical legend of Map 2.
17 The acronym MINMWMAX refers to three German statistical concepts: MIN (=Minimum), MW (=Mittelwert, arithmetic mean), MAX (=Maximum). The use of well defined statistical algorithms for cartographic aims has a long tradition which goes back to the 19th century: see Dickinson, 1973: passim, and Palsky, 1996: passim.
18 The use of 6 (coloured) intervals is recommended by cartographers and geographers for psychological and optical reasons. Nevertheless our dialectometric software VDM allows the current (and rapid) generation of ten different types of mappings (based on 2 up to 20 intervals) using three different algorithms of visualization and a completely free colour spectrum.
19 See Goebl, 1983a: passim; cf. also Goebl, 1984, I: 83–86 and 1987: passim.
20 See the respective (coloured) visualizations in Goebl, 2003: 110–111.
21 Cf. the seminal book of W. Besch Sprachlandschaften und Sprachausgleich of 1967.
22 See the respective coloured figures in Goebl, 2002: 54–59 and 2003: 112–155.
23 Implemented hierarchic-agglomerative procedures: besides Ward also Complete Linkage, Single Linkage, Simple Average Linkage, Average Linkage-UPGMA, and the Centroid Method.
24 In this ‘game’ Northern French (Langue or Domaine d'Oïl) took over the active part and expanded, whereas Occitanian (Langue or Domaine d'Oc) had the passive role and retreated.
25 Cf. Goebl, 2005 and Scapoli et al., 2005. See also Manni et al., Literary and Linguistic Computing, 21(4): 507–527.
26 See Haimerl's paper, Literary and Linguistic Computing, 21(4): 437–444.
27 Cf. Goebl/Schiltz, 2001: passim.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Manni, W. Heeringa, and J. Nerbonne To What Extent are Surnames Words? Comparing Geographic Patterns of Surname and Dialect Variation in the Netherlands Lit Linguist Computing, November 1, 2006; 21(4): 507 - 527. [Abstract] [Full Text] [PDF]
|
|