--- /dev/null
+#
+
+WORDLIST_LEN=3000
+
+all: words.list
+
+o= >$@.tmp && mv -f $@.tmp
+
+words.final: words.interim Makefile
+ grep -Ff forbidden-words $< >$@.1.tmp
+ head - $o
+
+words.interim: massage-lemmas lemma.al
+ ./$^ $o
--- /dev/null
+#!/usr/bin/perl -w
+use strict;
+
+our %allow_class =
+ ((map { $_=>1 } qw(v n a interjection conj pron prep
+ modal infinitive-marker)),
+ (map { $_=>0 } qw(det adv)));
+
+our %words;
+
+while (<>) {
+ # frequency sort rank, frequency, word, word-class
+ m/^\d+ (\d+) (\S+) (\S+)$/ or die "$_ ?";
+ my ($freq,$word,$class) = ($1,$2,$3);
+ my $allow = $allow_class{$class};
+ if (!defined $allow) {
+ warn "$class ?";
+ $allow = $allow_class{$class} = 1;
+ }
+ next unless $allow;
+ $words{$word} += $freq;
+}
+
+our @words = sort { $words{$b} <=> $words{$a} } keys %words;
+print $_,"\n" or die $! foreach @words;