MLSN logoLast Update: 2009-03-05

Project History

MLSN was started in the Autumn of 2005, and was started to fill a gap: I needed a Japanese thesaurus, and I needed one that was in electronic format and one that was free. Free as in freedom and open source. I was surprised and frustrated to discover there seemed to be nothing even close.

The first public release of the dictionary was mid-February 2006. It was Japanese and English only and there were just over 11,000 entries in the Japanese table, enough for it to be "almost useful". Some of these entries have been added automatically from dictionaries where I have been able to get high confidence, some I have added myself, and some from word lists I have paid other people to create.

February 2007 saw the addition of German and Chinese, and May 2008 say the addition of Arabic. The entries we did not already have from Japanese WordNet were added in February 2009.

I have taken great care to only use sources that are not just free but compatible with the open source MIT license. If you think I have abused your copyright, legally or ethically, please let me know.

Inspirations And Acknowledgments

Wikipedia ( was one major inspiration. It showed how people (the general public, not just us idealistic open source computer programmers...) will come together and work for free on a mutually beneficial database project.

WordNet ( was another major inspiration. When I discovered this I realized what I wanted was not a Japanese thesaurus but a Japanese semantic network. In fact MLSN's skeleton is the English WordNet. See the WordNet Comparison page.

Jim Breen's JMDict ( has been the primary Japanese-English dictionary, and has been an essential resource: without it the current word list would be a small fraction of its current size. You will see JMDict's translation suggestions on the search and add forms. For reference the JMDict license is here

The interwiki links in Wikipedia has acted as a secondary dictionary for all languages. These were automatically extracted from the xml dumps (Learn more here). Just as with JMDict you will see the Wikipedia-sourced suggestions on the search and add forms.

The ALC online dictionary ( has been a third dictionary. It is not free (in the open source sense) so has not been used directly, but I have personally referred to it frequently to double-check the entries I have added. I recommend it: it contains many words not in JMDict, and has many example sentences.

Other sources have then been Google (, Yahoo Japan ( and Wikipedia itself. I have also used Jkensaku ( and suggest you take a look if you have not seen that site before.

For German, BeoLingus ( has been used, and has also been referred to extensively. HanDeDict ( and Wadoku ( have been used for German-Chinese and German-Japan respectively.

For Chinese, Cedict ( and ) has been used. and have also been referred to a lot.

For Arabic the main resource has been AWN, the Arabic WordNet. Due to lack of an open source Arabic dictionary, the Arabic content of MLSN is not much more than a different interface on AWN.

NICT (The National Institute of Information and Communications Technology, Kyoto, Japan) have been building a Japanese WordNet, and in February 2009 finally released it. Thankfully it is under basically the same license as MLSN, so the Japanese noun synsets that MLSN did not already have an entry for have been imported from it.

Site Menu

  • Introduction; News
  • History And Acknowledgments
  • Current Limitations
  • Wordnet Comparison
  • Adding More Languages
  • Making Dictionary From Wikipedia
  • Privacy
  • License
  • Resources, Links
  • Download Database Dump

    Try It!

  • Back to main page

    © Copyright 2006,2007,2009 Darren Cook (