ELAN-Corpa is an extended version of ELAN version 4.7.3 from the Max Planck Institute for Psycholinguistics, Nijmigen, Neitherlands. It has been developed by Coralie Villes and Christian Chanard from the CNRS research laboratory LLACAN (umr 8135 of CNRS), for the research program ANR ‘CorpAfroAs’ (A Spoken Corpus for AfroAsiatic Languages), coordinated by Amina Mettouchi.

This version contains an additional tab, ‘Interlinearize,’ which allows the management of a XML lexicon (of extension .eafl) and the interactive segmentation into morphems of words from a tier. Those morphemes are then annotated on 2 additional tiers (gloss and category), with the contents of the lexicon.

A second lexicon (Parse Lexicon, with an .eafp extension) containing the words segmented and annotated from previous texts can be created or extended to speed up the annotation process.

Cleaning a local Parse Lexicon
Extracting an ELAN-CorpA lexicon from a local Parse Lexicon

In addition, the table showing the hits found from a search into the annotated texts can display 2 more columns, ‘parent’ and ‘children’ (in addition to ‘before’ and ‘after’) which contains respectively the parent and the children of a search element (for example the word from which a morpheme has been found, the gloss and category of this morpheme). Note: Actually, this has been incorporated into ELAN since 4.6.1 release

In the grid display, there is a new option between 'multiple tiers with symbolic association' or 'multiple tiers with symbolic subdivision' allowing the display of the children of a tier depending on its stereotype. Note: this has been incorporated into ELAN since 4.6.1 release

In the version 4.7.4, a new item can be found in the Search menu : CorpA Multiple files Search where searching can be done using CorpA Query Language. It also adds a zoom functionnality to the interface (cf. Edit, Preferences... Platform/OS). When you change the factor of zooming, close and re-open ELAN-CorpA to have it taken into account.

The last version 4.8.0 adds a new tab, Groups & Links allowing to

  • create a Group Table in which annotations taken from the annotation area can be bundled in new lines of the table. A group (a line in the table) can be given a Type (eventually taken from a controlled vocabulary) and a Name (i.e an identifier).
  • create a Link Table, allowing the creation of links between groups taken from the Group Table. A link has a Source, a Target and can have a Type and a Name.
  • The Tables can then be sorted by Type for counting the different structures, for example, or by Name for counting the number of similar identifiers for example.
  • The content of these tables can be exported (with the time boundaries of the annotations involved in their lines) in CSV format to be exploited into a spreadsheet.
  • A multiple files search result can be filtered by constraining a hit to Exist or Not Exist into a Group or a Link Source or Target as a Type or Name.

  ELAN-CorpA   (click to download ELAN-CorpA)

Windows :
  • After downloading, double-click install.exe
  • You do not need to install any other software. A Java virtual machine is included with this download.
Mac OS X :
  • After downloading, double-click install
  • Be sure you have Java installed. You can download Java from Apple's site
  • The compressed installer should be recognized by StuffIt Expander and should automatically be expanded after downloading. If it is not expanded, you can expand it manually using StuffIt Expander 6.0 or later
  • If you have any problems launching the installer once it has been expanded, make sure that the compressed installer was expanded using StuffIt Expander. If you continue to have problems, please contact technical support
  • Once ELAN-Corpa is installed in a Mac, the program may hang up during the loading of the media file you chose to annotate ('Initializing...' problem) or at the opening of the media file of an ELAN file you chose to open ('No player...' problem). Here are 2 things to do to fix this
    • First go to the Java preferences, and drag the 32 bits java machine in the first position (if there is also a 64 bits). If Java preferences doesn't exist: Go to Applications Folder, Right-click on ELAN-CorpA, Select Get Info, Check the box run on 32 bits
    • Make a shortcup of the ELAN-corpa.jar and drag it to the desktop.

    Now use this shortcut to launch ELAN-Corpa.

You will find here a PDF document on how to use ELAN-CorpA

You will find here (PC / Mac) an ELAN template file Corpo1.etf that can be imported into ELAN to set up the types and tiers according to the CorpAfroAs model for one speaker, and there (PC / Mac) Corpo2.etf for two speakers. If the contents of those files open in your browser when downloading, just save them (file save as) on your computer.
To create a new ELAN file according to the CorpAfroAs template :

  • File, New, Browse to your Wav file, >> , ok
  • Delete the default tier (right-click on its label, delete default tier)
  • Tier, import Tiers
  • Browse to the downloaded file (Corpo1.etf for one speaker, Corpo2.etf for 2 speakers)
  • Import, Close