![]() | Adding to the CYC® Lexicon using WALES |
![]() | Adding Irregular Nouns |
For each morphological variant string associated with an instance of #$EnglishWord, there will be at least one assertion specifying a denotation in the KB, i.e. its semantic interpretation(s).
Let's begin by adding a fictitious new word to the lexicon, "bliik", a synonym for "parakeet". To create a new word in the lexicon, enter its root form in the box labeled: "Enter word". For simple nouns, the root form is singular. In our example, "bliik" is the root form.
Next, select the "Add regular strings to lexicon" button and press "Add Info". Recall that a noun is regular if its plural is formed by adding an "s" or an "es" to its stem.
Since no constant for the word "bliik" previously existed in CYC®, WALES displays the "Create New Word Unit" page. If "bliik" was already in the lexicon, then WALES would ask whether you wanted to add information to the existing word unit, or to create a new one .
To create the new constant, press the "Create" button on the "Create New Word Unit" page. You can modify the name of the word unit, if necessary, by first editing the word unit that is in the "Unit name:" box and then by pressing the "Create" button. If you make a mistake, the "Reset" button will reset the word unit in the box to its original form. After pressing the "Create" button, WALES will display the "Add Part of Speech Forms" page.
At the "Noun Form sing" box, enter the singular form. You need only to fill out this box, as the parser recognizes plural forms of regular strings. If there are other part-of-speech forms that you want to link to this constant, you may enter those at this time. Click on the "Update Lexicon" button and then "Do It".
At this point, WALES returns you to the original "Add to CYC® Lexicon" page. The root string of the word you just created will remain in the box so that you can easily move to a page to add its denotation with minimal typing. You will also notice that throughout this process, the word unit that you created will be present on the pages as a clickable link to its representation in the KB. As soon as your operations are processed, you should see all of your word unit assertions represented in the KB.
After your string assertions are made, you will need to add denotations for your strings. First, click on the "Add Other Lexical Information" button and then "Add Info". Using WALES, denotation assertions may only be added to one part of speech at a time. In this example, since there is only one part of speech already specified for the word constant, WALES will display the "Denotation Entry" page. Had more than one part of speech been specified for "Bliik-TheWord", then the user would be prompted to which part of speech s/he intends to use.
On the "Denotation Entry" page you will see boxes which contain the denotations for various word senses. Since our example is a newly created word, all boxes should be blank. For general guidelines on selecting an appropriate denotation, see the "WALES User's Guide". In this case, we want "bliik" to denote the CYC® constant Parakeet, so enter "Parakeet" into a denotation box.
After you have typed your denotations in the appropriate boxes, click on the "Update Lexicon" button. If you make a mistake, you can always click on the "Restore Values" button, which will restore the information in the boxes to their original values.
The CYC® NL system requires that all word senses have some denotation which reflects their meaning, even in cases where a more complicated semantic translation rule will be needed to fully express the meaning. In such cases, the denotation of a word sense will be a CYC® constant which is close in meaning to the word, or which captures some aspect of that word's meaning. The actual word meaning is specified with a semantic translation rule. For example, the noun form of "anchorwoman" denotes #$Anchorperson. The fact that an anchorwoman is an anchorperson who is also female is captured with a #$nounSemTrans rule.