‘Wiktionary’ and the Limitations of Collaborative Sites

On: September 20, 2008
Print Friendly, PDF & Email


Wiktionary is probably one of the oddest projects currently run by the Wikimedia Foundation. Started in 2002, it is a free and open dictionary working on the collaborative content creation model of Wikipedia. It currently boasts more than 900,000 lemmata for both English and French, as well as several hundreds of thousands of entries for Russian, Turkish, Vietnamese, Tamil, Ido (a language created by disgruntled Esperantists) and several others. But in contrast to Wikipedia, most of these articles were not written by enthusiastic individuals, but compiled automatically using bots. These articles were apparently harvested from public-domain dictionaries and converted into the Wiktionary architecture (Wikipedia). Other articles have been manually edited by the Wikipedia brand of enthusiastic collaborators.

Although the idea of creating an open dictionary on the basis of Wikipedia seems admirable, even a casual user of Wiktionary will quickly discover its faults. A good dictionary, after all, requires a lot of arduous, precise and methodical manual editorial work. The reason why many languages have “old”, tried-and-true dictionaries like the Oxford English Dictionary or Le Petit Robert as their standard reference is that it takes an incredible amount of work to compile a lexical inventory of even the most basic words that is sufficiently clear, while at the same time being both brief and inclusive of most common meanings. The Oxford English Dictionary, the standard and most exhaustive record of the English language took forty years to compile for its first edition.

It is difficult for a free collaboratively-edited dictionary to attain a similar level of intellectual rigour and inclusiveness. Compiling a dictionary is pretty dull work, as a young J.R.R. Tolkien found when working on one of the later volumes of the first edition of the OED. Are there really enough people who would do that kind of work for free in their spare time? Well, the Wikimedia empire seems to be a veritable treasure-trove of dedicated crackpots like the person who created the Wikipedia entry on Einstein’s theory of General Relativity in Esperanto. But Wiktionary’s problem is that even if you do have dedicated people who collaborate freely and you do have lots of automatically generated lemmata, that doth not a dictionary make: a dictionary, like Wiktionary, that is not complete, exhaustive and intelligently manually edited is actually worthless. What use is a dictionary if you cannot rely on it? Why would you use a dictionary if you know beforehand that there is a chance that the lemma you are looking for will be incomplete and that you will not find the real meaning of a word that you are looking for?

Take for instance the Wiktionary lemma of English “man“. The basic description of the meaning of the word seems allright, and there is some additional (though arbitrary) extra information like antonymia. But a standard commercial one-volume dictionary contains many more, much more nuanced meanings of this common word. A dictionary that is not methodically compiled may lead you to interpret a word incorrectly. For instance, “horse” contains several correct definitions, but also, as a sixth meaning, “(slang, pejorative) a Tongan.” I personally could not imagine anyone except maybe a professional rugby player from the southern hemisphere using this word. Imagine you have children of school age; would you not buy them a real printed dictionary rather than risk that they interpret words incorrectly by using a tool like Wiktionary?

Wikipedia works as an encyclopedia because everybody has an interest in something, and if plenty of people with a common interest collaborate to write an encyclopedia entry, odds are you will end up with a reasonable piece of content. But compiling a dictionary requires a completely different kind of effort. Every intelligent native speaker of a language possesses knowledge about the meaning of words, but to compile a dictionary that has any value at all, you also need to be inclusive, consistent and methodical. Wiktionary’s reliability for any sort of serious study seems questionable: some entries, for instance from French, seem pretty extensive, but the Latin Wiktionary looks pretty useless: even a common verb like “to run” misses all of the secondary, metaphorical meanings that are also contained in this semantically rich language.

Compiling a usable dictionary is something that requires effort antithetical to the free, open, “ad-lib” nature of collaborative content creation. That does not mean that the people who created Wiktionary deserve to be hanged: I would still use Wiktionary in a pinch if I had no other reference available. But anyone who is seriously trying to learn their own or any other language better, would be well advised to buy an old-fashioned paper dictionary. After all, many lexicologists suffered for years to bring you that reliable piece of reference.

Comments are closed.