1 Spell Checking Oriented Word Lists (SCOWL)
2 @`if [ "$SCOWL_VERSION" ]; then echo -n "Version $SCOWL_VERSION"; fi`
3 @`git log --pretty=format:'%cd [%h]' -n 1 --`
4 by Kevin Atkinson (kevina@gnu.org)
6 The SCOWL is a collection of word lists split up in various sizes, and
7 other categories, intended to be suitable for use in spell checkers.
8 However, I am sure it will have numerous other uses as well.
10 The latest version can be found at http://wordlist.aspell.net/.
12 The directory final/ contains the actual word lists broken up into
13 various sizes and categories. The r/ directory contains Readmes from
14 the various sources used to create this package.
16 The misc/ contains a small list of taboo words, see the README file
17 for more info. The speller/ directory contains scripts for creating
18 spelling dictionaries for Aspell and Hunspell.
20 The other directories contain the necessary information to recreate the
21 word lists from the raw data. Unless you are interested in improving the
22 words lists you should not need to worry about what's here. See the
23 section on recreating the words lists for more information on what's
26 Except for the special word lists the files follow the following
28 <spelling category>-<sub-category>.<size>
29 Where the spelling category is one of
30 english, american, british, british_z, canadian,
31 variant_1, variant_2, variant_3,
32 british_variant_1, british_variant_2,
33 canadian_variant_1, canadian_variant_2,
34 Sub-category is one of
35 abbreviations, contractions, proper-names, upper, words
37 10, 20, 35 (small), 40, 50 (medium), 55, 60, 70 (large),
38 80 (huge), 95 (insane)
39 The special word lists follow are in the following format:
40 special-<description>.<size>
41 Where description is one of:
42 roman-numerals, hacker
44 The perl script "mk-list" can be used to create a word list of the
45 desired size, its usage is:
46 ./mk-list [-f] [-v#] <spelling categories> <size>
47 where <spelling categories> is one of the above spelling categories
48 (the english and special categories are automatically included as well
49 as all sub-categories) and <size> is the desired size. The
50 "-v" option can be used to also include the appropriate
51 variants file up to level '#'. The normal output will be a sorted
52 word list. If you rather see what files will be included, use the
55 When manually combining the words lists the "english" spelling
56 category should be used as well as one of "american", "british",
57 "british_z" (british with ize spelling), or "canadian". Great care
58 has been taken so that only one spelling for any particular word
59 is included in the main list (with some minor exceptions). When two
60 variants were considered equal I randomly picked one for inclusion in
61 the main word list. Unfortunately this means that my choice in how to
62 spell a word may not match your choice. If this is the case you can
63 try including one of the "variant_1" spelling categories which
64 includes most variants which are considered almost equal. The
65 "variant_1" spelling category corresponds mostly to American variants,
66 while the "british_variant_1" and "canadian_variant_1" are for British
67 and Canadian variants, respectively. The "variant_2" spelling
68 categories include variants which are also generally considered
69 acceptable, and "variant_3" contains variants which are seldom used
70 and may not even be considered correct. There is no
71 "british_variant_3" or "canadian_variant_3" spelling category since
72 the distinction would be almost meaningless.
74 The "abbreviation" category includes abbreviations and acronyms which
75 are not also normal words. The "contractions" category should be self
76 explanatory. The "upper" category includes upper case words and proper
77 names which are common enough to appear in a typical dictionary. The
78 "proper-names" category includes all the additional uppercase words.
79 Finally the "words" category contains all the normal English words.
81 To give you an idea of what the words in the various sizes look like
82 here is a sample of 25 random words found only in that size:
84 @`src/rand-samples | iconv -f iso-8859-1 -t utf-8`
86 And here is a count on the number of words in each spelling category
87 (american + english spelling category):
91 (The "Words" column does not include the name count.)
93 Size 35 is the recommended small size, 50 the medium and 70 the large.
94 For spell checking I recommend using 60. Sizes 70 and below contain
95 words found in most dictionaries while the 80 size contains all the
96 strange and unusual words people like to use in word games such as
97 Scrabble (TM). While a lot of the words in the 80 size are not
98 used very often, they are all generally considered valid words in the
99 English language. The 95 contains just about every English word in
100 existence and then some. Many of the words at the 95 level will
101 probably not be considered valid English words by most people. I use
102 the 60 size for the English dictionary for Aspell, and I don't
103 recommend anyone use levels above 70 for spell checking. Levels above
104 70 contain rarely used words which can hide misspellings of similar
105 more commonly used words. For example the word "ort" can hide a
106 common typo of "or". No one should need to use a size larger than 80,
107 the 95 size is labeled insane for a reason.
109 Accents are present on certain words such as café in iso8859-1 format.
113 From Version 2015.05.18 to 2015.08.24 (Aug 24, 2015)
117 From Version 2015.04.24 to 2015.05.18 (May 18, 2015)
119 Added some new words found to have a high frequency in the COCA
120 corpus. (http://corpus.byu.edu/coca/).
122 Fix en spelling suggestions for 'alot' and 'exersize' in hunspell
123 dictionary (upstreamed from the changes made in Firefox).
125 From Version 2015.02.15 to 2015.04.24 (April 24, 2015)
127 Added some new words.
129 Convert hunspell dictionary to UTF-8 in order to handle smart
132 From Version 2015.01.28 to 2015.02.15 (February 15, 2015)
134 Added a large number of neologisms (newly invented words)
135 such as "selfie" and "smartwatch" thanks to Alan Beale.
137 Various other new words.
139 Clean up the special-hacker category by removing some words that
140 didn't exist in the Google Book's Corpus (1980 - 2008) and
141 originated from the "Unofficial Jargon File Word Lists".
143 From Version 2014.11.17 to 2015.01.28 (January 28, 2015)
145 Various new words, many from analyzing the Google Book's Corpus
146 (1980 - 2008). See http://app.aspell.net/lookup-freq.
148 Moved some uncommon words that can easily hide a misspelling of a
149 more common word to level 70. (calender, adrenalin and Joesph)
151 Removed several -er and -est forms from adjectives that were so
152 uncommon that they were not found anywhere is the Google Book's
153 Corpus (1980 - 2008).
155 From Version 2014.08.11.1 to 2014.11.17 (November 17, 2014)
159 Fix typo in Hunspell readme.
161 From Version 2014.08.11 to 2014.08.11.1 (August 13, 2014)
163 Forgot to mention this important change from 7.1 to 2014.08.11:
165 Shifted the variant levels up by one: variant_0 is now variant_1,
166 variant_1 is now variant_2, and variant_2 is now variant_3.
168 Other minor fixes in this README.
170 No changes to the contents of the lists.
172 From Revision 7.1 to Version 2014.08.11 (August 11, 2014)
174 Added some missing possessive forms.
176 Added some new words and proper names.
178 Clean up the categories (words, upper, proper-names etc) so that they
181 Convert documentation to UTF-8. For now, the wordlist are still in
182 ISO-8859-1 to prevent compatibility problems.
184 Add schema and scripts for creating a SQLite database from SCOWL.
185 Add some utility and library functions using them. This database is
186 used by the new web app's (http://app.aspel.net/lookup & create).
188 Enhance speller/make-hunspell-dict. The biggest improvement is that
189 it that it now generates several more dictionaries in addition to
190 the official ones. These additional dictionaries are ones for
191 British English and larger dictionaries that include up to SCOWL
194 From Revision 7 to 7.1 (January 6, 2011)
196 Updated to revision 5.1 of Varcon which corrected several errors.
198 Fixed various problems with the variant processing which corrected a
201 Added several now common proper names and some other words now
204 Include misc/ and speller/ directory which were in SVN but left
205 out of the release tarball.
207 Other minor fixes, including some fixes to the taboo word lists.
209 From Revision 6 to 7 (December 27, 2010)
211 Updated to revision 5.0 of Varcon which corrected many errors,
212 especially in the British and Canadian spelling categories. Also
213 added new spelling categories for the British and Canadian spelling
214 variants and separated them out from the main variant_* categories.
216 Moved Moby names lists (3897male.nam 4946fema.len 21986na.mes) to 95
217 level since they contain too many errors and rare names.
219 Moved frequently class 0 from Brian Kelk's Wordlist from
220 level 60 to 70, and also filter it with level 80 due to, too many
223 Many other minor fixes.
225 From Revision 5 to 6 (August 10, 2004)
227 Updated to version 4.0 of the 12dicts package.
229 Included the 3esl, 2of4brif, and 5desk list from the new 12dicts
230 package. The 3esl was included in the 40 size, the 2of4brif in the
231 55 size and the 5desk in the 70 size.
233 Removed the Ispell word list as it was a source of too many errors.
234 This eliminated the 65 size.
236 Removed clause 4 from the Ispell copyright with permission of Geoff
239 Updated to version 4.1 of VarCon.
241 Added the "british_z" spelling category which is British using the
244 From Revision 4a to 5 (January 3, 2002)
246 Added variants that were not really spelling variants (such as
247 forwards) back into the main list.
249 Fixed a bug which caused variants of words to incorrectly appear in
250 the non-variant lists.
252 Moved rarely used inflections of a word into higher number lists.
256 Shifted the variant levels so that level 0 is now 1, level 1 now 2,
259 Added other inflections of a words based on the following criteria
260 If the word is in the base form: only include that word.
261 If the word is in a plural form: include the base word and the plural
262 If the word is a verb form (other than plural): include all verb forms
263 If the word is an ad* form: include all ad* forms
264 If the word is in a possessive form: also include the non-possessive
266 Updated to the latest version of many of the source dictionaries.
268 Removed the DEC Word List due to the questionable licence and
269 because removing it will not seriously decrease the quality of SCOWL
270 (there are a few less proper names).
272 From Revision 4 to 4a (April 4, 2001)
274 Reran the scripts on a never version of AGID (3a) which fixes a bug
275 which caused some common words to be improperly marked as variants.
277 From Revision 3 to 4 (January 28, 2001)
279 Split the variant "spelling category" up into 3 different levels.
281 Added words in the Ispell word list at the 65 level.
283 Other changes due to using more recent versions of various sources
284 included a more accurate version of AGID thanks to the work of
287 From Revision 2 to 3 (August 18, 2000)
289 Renamed special-unix-terms to special-hacker and added a large
290 number of commonly used words within the hacker (not cracker)
293 Added a couple more signature words including "newbie".
295 Minor changes due to changes in the inflection database.
297 From Revision 1 to 2 (August 5, 2000)
299 Moved the male and female name lists from the mwords package and the
300 DEC name lists form the 50 level to the 60 level and moved Alan's
301 name list from the 60 level to the 50 level. Also added the top
302 1000 male, female, and last names from the 1990 Census report to the
303 50 level. This reduced the number of names in the 50 level from
306 Added a large number of Uppercase words to the 50 level.
308 Properly accented the possessive form of some words.
310 Minor other changes due to changes in my raw data files which have
311 not been released yet. Email if you are interested in these files.
313 COPYRIGHT, SOURCES, and CREDITS:
317 The variant word lists were created from a list of variants found in
318 the 12dicts supplement package as well as a list of variants I created
321 The Readmes for the various packages used can be found in the
322 appropriate directory under the r/ directory.
326 The process of "sort"s, "comm"s, and Perl scripts to combine the many
327 word lists and separate out the variant information is inexact and
328 error prone. The whole things needs to be rewritten to deal with
329 words in terms of lemmas. When the exact lemma is not known a best
330 guess should be made. I'm not sure what form this should be in. I
331 originally thought this should be some sort of database, but maybe I
332 should just slurp all that data into memory and process it in one
333 giant perl script. With the amount of memory available these days (at
334 least 2 GB, often 4 GB or more) this should not really be a problem.
336 In addition, there is a very nice frequency analyze of the BNC corpus
337 done by Adam Kilgarriff. Unlike Brian's word lists the BNC lists
338 include part of speech information. I plan on somehow using these
339 lists as Adam Kilgarriff has given me the OK to use it in SCOWL.
340 These lists will greatly reduce the problem of inflected forms of a
341 word appearing at different levels due to the part-of-speech
344 There is frequency information for some other corpus such as COCA
345 (Corpus of Contemporary American English) and ANS (American National
346 Corpus) which I might also be able to use. The former will require
347 permission, and the latter is of questionable quality.
349 RECREATING THE WORD LISTS:
351 In order to recreate the word lists you need a modern version of Perl,
352 bash, the traditional set of shell utilities, a system that supports
353 symbolic links, and quite possibly GNU Make. The easiest way to
354 recreate the word lists is to checkout the corresponding Git version
355 (see the version string at the start of the file) and simply type
356 "make" (see http://wordlist.aspell.net). You can try to download all
357 the pieces manually, but this method is not no longer tested nor
360 The src/ directory contains the numerous scripts used in the creation
361 of the final product.
363 The r/ directory contains the raw data used to create the final
364 product. If you checkout from Git this directory should be populated
365 automatically for you. If you insist on doing it the hard way see the
366 README file in the r/ directory for more information.
368 The l/ directory contains symbolic links used by the actual scripts.
370 Finally, the working/ directory is where all the intermittent files go
371 that are not specific to one source.