The Paratext Glossing Tool is a program which works closely with Paratext and provides a facility to gloss a word from one text to its closest equivalent in another text. It is intended as an aid for those tasked with helping to check new translations of Scripture for consistency and accuracy. It allows the user to explore the new translation semantically, checking for possible inconsistencies in usage. The Glossing Tool can analyse a single semantic or it can help the user review a complete passage of Scripture via the automatic interlinear back-translation which it generates.

Information
The Glossing Tool works only with texts held within Paratext projects, and will work optimally with more complete texts.

How does it work?
The Glossing Tool compares each verse of the model and project text, looking for words that tend to appear together. For example, the English word "food" and the Spanish word "comida" tend to appear in the same verses. Using some statistical techniques, it decides which word is most likely. It also does a morphological analysis of the project text, looking for stems and affixes, depending on the language type selected.

History
The Glossing Tool is a Paratext deployment of the statistical glossing engine technologies created by Jon Riding & David Robinson of the British & Foreign Bible Society (BFBS) Machine Assisted Translation Team (now BFBS Linguistic Computing). The tool in its present form was created by Clayton Grassick, working with BFBS LC on behalf of the UBS TPCRG and the Institute for Computer Assisted Publishing.

Here's a short timeline of Glossing Tool Development:

• Late 1980's – BFBS creates the Concordance to the Good News Bible. Statistical Glossing techniques were developed by Jon Riding and David Robinson of the BFBS Machine Assisted Translation team (MAT). These enabled automatic lemmatisation of texts as a pre-requisite for concordance creation. The best way to achieve this was found to be glossing between two (or more) parallel corpora.
  
• Early 1990's – Early systems ran on mainframe computers at BFBS.
  
• 1997 – The first PC based system (geared towards glossing highly inflected project texts) was tested in the field.
  
• 2005 – The Paratext Glossing Tool project further develops these techniques and is closely coupled with Paratext 6.

Future
We want to find out what the limits and capabilities of the Glossing Tool are, and with your feedback, to improve the user interface.

As far as future versions and features, those are just hopes and dreams. We'd like to see the Glossing Tool become more flexible, more efficient, and more accurate. We'd like to see it work together with other tools that could assist its task or gain from its results. We'd like to see fewer limitations, better support for Right-To-Left languages and other language types.

This is an initial release of the Glossing Tool. Its ultimate usefulness is largely dependent on those who begin to use it. As you use the tool you may well find yourself thinking, "it would be good if it could..." or perhaps, "I don't like the way it does..." or even "I wonder if it could...". Please don't keep these thoughts to yourself. The Glossing Tool developers are keen to hear your comments. Please send them to the development team by email at emailglossing@ubs-icap.org.

If you are an associate of ICAP you can find out more about the Glossing Tool here http://confluence.ubs-icap.org/display/GT/Home .
And the full user manual is available here: http://confluence.ubs-icap.org/download/
attachments/9523/GlossingTool1_0.pdf?version=1 .

A paper ( Statistical Glossing & Bible Translation v0.1) on the development and benefits of the Glossing Tool can also be found in the proceedings for the Conference Translating & the Computer 30 (2008).
cite this work:


And: Glossing Technology in Paratext 7 v0.3 Glossing technologies as implemented in Paratext 7.