What is a “Speaking [Language Atlas] ”?

The aim of the “Speaking” was and remains to provide the users of the four map volumes of the language atlas ALD-I with an audible equivalent to the transcriptions visible on the pages of the ALD-I. The “Speaking” has 21 reference points that cover the five historical valleys of the old Tyrolean-Brixen Ladinia (Val Badia, Gherdëina, Fascia, Fodom, Anpezo). Unfortunately, for reasons of work economy, it was not possible to extend the “Speaking” to the entire set of 217 ALD-I reference points or to implement this project also for the ALD-II.


Data collection:

The audible sound samples here were recorded in Ladinia in the years 1992-1994 by the ALD-I researcher Helga Böhmer with the help of particularly advanced technical equipment (directional microphone, DigitalAudioTape tapes). The recording of particularly high-quality sound was necessary because, when listening contrastively to different individual words, the background noise inevitable in analog recordings proved very annoying to the human ear.


Data processing:

The DAT recordings provided by Helga Böhmer were subsequently segmented electronically (1994-1997) in the ALD archive of the University of Salzburg. The isolated sound segments were then stored in a database, which is accessed by user software specially designed for this purpose by the software company RST in Essen. This software from the last century has obviously been replaced in this “Speaking” with a solution updated to the new conditions.

The segmentation work (very labor-intensive) of the years 1994-1997 involved: Susanne Heißmann, Brigitte Rührlinger, and Slawomir Sobota (all from Salzburg).


Phonetic transcription, transcriptions, drum rolls (or melodic chords):

The transcriptions visible here (always in ALD standard mode) correspond exactly to those found in the four volumes of the published ALD-I. Since these transcriptions are about 6-7 years older than the DAT recordings made by Helga Böhmer, there may occasionally be discrepancies between the responses audible here and the transcriptions shown.


Two more technical notes:

  1. Information on the phonetic transcription used (Böhmer-Ascoli-AIS system) can be found at the beginning of all volumes of both parts of the ALD (ALD-I and ALD-II) in detailed diagrams and explanatory lists;
  2. Drum rolls or melodic chord: these are audible wherever Helga Böhmer was unable to obtain a quality sound for an existing transcription in 1992-1994.

On the origin and history of the “Speaking ”:

The first ideas and preliminary work on the “Speaking” date back to 1989. Since then, four versions of the “Speaking” have been created, of which only the last three are based on the DAT recordings made by Helga Böhmer in 1992-1994:

  1. a pilot version with an audio CD from 1990, published as a supplement in Ladinia 14 (1990)
  2. a version on two CD-ROMs (1998-2002), included with the ALD-I published in 1998
  3. a version on DVD (from 2002), presented (with a DVD included) in Ladinia 29 (2005) and in Mondo ladino 29 (2005)
  4. an online version (from spring 2005)

Particular thanks go to Roland Bauer (Salzburg) for creating and maintaining versions 1) to 3). The user software used for the DVD version is from the company RST in Essen (Director: Reinhard Köhler, University of Trier).

The IT steps necessary for the online placement of the “Speaking” were taken in spring 2005 by Marcel Müller (Freiburg im Breisgau). For this purpose, the idea and layout were partly borrowed from the Berlin (HU) project VIVALDI (Director: Dieter Kattenbusch, Carola Köhler, Marcel Müller). Thomas Frankewitsch (Chair of Medical Informatics, University of Erlangen) and Fabio Tosques (then: HU Berlin) made a significant contribution to the development of the Java applet for the display of map data. This version worked perfectly for more than 10 years. Unfortunately, it fell silent due to the failure of the ALD server set up at the Salzburg computing center in 2019.

For further information, please see the preface and bibliography at the beginning of the first volume of ALD-I (1998, pp. VII-XXI).

Technical notes on the “Speaking ” visible and audible here:

The work screen consists of two parts: on the left, you can see a list of the 806 (numbered) questions from the ALD-I questionnaire. By including all individual elements of these 806 questions, this table contains 1266 rows.

In the upper right half, you can see a light blue map of Brixino-Tyrolean Ladinia with its 21 reference points (PP. 81-101).

Directly below it, you can see a line for adjusting certain modalities of the transcriptions and the presentation of the selected responses: type of transcription (with three options), volume adjustment during sound playback, adjustment of pauses (in milliseconds) between playing the selected responses, choice between playing the responses in a dictionary style (= always the same place, but different responses) or in the style of a language atlas map (= varying responses to the same question, but different places).

At the bottom, there is a two-dimensional matrix (“Listening Matrix”) which is initially always empty. However, it can be filled as follows:

  • horizontally (with questions or their responses): by clicking on the relevant questions (on the left side of the screen);
  • vertically (with locations): by clicking on the interesting locations on the Ladinia map (on the right side of the screen, at the top).

In this way, the hitherto empty cells of this listening matrix are filled as follows:

  • with the relevant transcription and the associated sonagram, as well as three other service buttons:
    • top right: (magnifying glass symbol) to enlarge the relevant cell;
    • bottom right: to download (for further electro-acoustic processing) the relevant mp3 file;
    • in the middle: a triangle to trigger the playback.

Both the questions (or the responses obtained) and the locations can be selected multiple times, in various combinations and arrangements, and entered into the listening matrix.

In this way, based on the list on the left of all 806 questions, fictional sentences can be formed, the responses to which are grammatically correct sentences, but which are completely artificial (fictional) and have never been produced by our informants in this form.

Here are two examples:

  • morphological-lexical: voglio (795/1) mangiare (380/1) l’oca (803/6) grassa (326/1)
  • phonological: il carro (102/1) caro (101/1)