BNC/VIEW is a new architecture and interface for the 100 million word
British National Corpus. It is freely availably on the web at
http://view.byu.edu
A number of new features have recently been added to the corpus. They
include the following:
** 1) CHARTS
You can now see (in graphical form) the frequency of a word, phrase, or
grammatical construction in the six major registers (e.g. spoken,
academic) and then each of the sub-registers (sermons, poetry, medical,
etc).
** 2) IMPROVED SEARCHES FOR COLLOCATES
The search frame is now much wider when you search for collocates /
surrounding words -- up to ten words on the left and on the right.
Examples: the most common nouns near [kitchen], comparing the nouns near
[uncover] with those near [reveal], and comparing nouns near [chair] in
fiction and academic registers.
** 3) CUSTOMIZED, USER-DEFINED LISTS
You can create an unlimited number of customized lists, containing words
that are related together in any way that you might imagine. Via the web
interface you store these word lists, and you can then re-use them at
any point in the future.
** 4) SORTING BY RELEVANCE (MODIFIED Z-SCORE)
You can now sort by relevance, which provides a much better
understanding of which words are most tightly related together. This
type of query, which is similar to a z-score calculation, takes into
account the overall frequency of collocates and sorts out high-frequency
"noise" words.
Features that were already available include the following:
** 5) BASIC QUERIES BY SUBSTRING, WORD, PHRASE, AND PART OF SPEECH
For example, the frequency of a given word, set of words, phrase,
substring (e.g. *heart*), part of speech (e.g. [av*] [aj*]: very clear),
or combinations of these (e.g. [vv*] it/them [avp]: took them away, give
it up)
** 6) REGISTER-BASED QUERIES
You can find the frequency of words and phrases in any combination of
registers that you define -- on the fly -- e.g. spoken, academic,
poetry, or medical. In addition, you can compare between registers --
for example, verbs that are more common in legal or medical texts,
phrases like [I * that] that are more common in conversation than in
non-fiction texts, nouns near "break" (v) that are found primarily in
academic writings, etc.
** 7) FREQUENCY IN ALL 70 REGISTERS
You can click on a word or phrase in any of the results sets to see the
frequency in all 70 registers. Sorted initially by normalized frequency,
you can re-sort by register name, number of tokens, etc.
** 8) COMPARING COLLOCATES WITH RELATED WORDS
For example, nouns that occurs after [utter] but not with [sheer] or
[total], adjectives within ten words of [man] that do not occur near
[woman], etc. All of this is done via one simple query from the web
interface. This may be quite useful for language learners, to allow
them to compare the uses of competing synonyms.
** 9) INTEGRATION WITH WORDNET (SEMANTICALLY-BASED QUERIES)
For example:
[=bad] [nn*]: any synonym of [bad] followed by any noun, e.g. wicked
witch, foul play, terrible storm, etc
my/your [@body]: [my] or [your] followed by a part of the body: my
leg, your shoulder, etc
[<eat] the [<food]: a more specific word for [eat] followed by a more
specific word for [food], e.g. devour the hamburger, munch the cookies
Again, all of this is done via one simple query from the search form
** 10) COMPLETE RESULTS FAST AND FAST QUERIES
Unlike some other interfaces for the BNC, this one allows you to find
all of the matching strings -- not just those that occur three times or
more. In addition, queries of the 100 million word corpus are quite
fast -- less than one or two seconds for most searches.
Again, the corpus is freely available at http://view.byu.edu. Please
feel free to email me with any questions or comments.
Mark Davies
Brigham Young University
=================================================
Mark Davies
Assoc. Prof., Linguistics
Brigham Young University
(phone) 801-422-9168 / (fax) 801-422-0906
http://davies-linguistics.byu.edu
** Corpus design and use // Linguistic databases **
** Historical linguistics // Language variation **
** English, Spanish, and Portuguese **
=================================================
This archive was generated by hypermail 2b29 : Wed Oct 12 2005 - 23:10:34 MET DST