METHOD
Our aim was to reduce the diversity of expression both on lexical and syntactical level. On lexical level we have reduced the diversity of expression by choosing one word or phrase to substitute all equivalent meanings. For example
dispossess = expropriate = nationalize = TAKE
confiscate = forfeit = impound = lay hands on = seize = sequestrate = TAKE
etc.
On syntactical level we have reduced the the diversity of expression by
a) substituting phrases with single words, preserving the meaning, for example
give a promise = promise
get the knack of = master
go in for sport = train
etc.
b) standardizing the syntactical structure of the sentence. This means that we have tried to use always one and the same sequence of prepositional and adverbial phrases, depending on meaning.
For example
Richard went to (the) school by car = Richard went to (the) school with his Fiat =
Richard went by car to the college
will be translated always "went to school by car". If we want to be more abstract, we can make the
program translate
a) "went to school using private transport" or
"went to school using public transport", if we want to include such phrases as "by bus", "by train", etc.
Instead of using codes for the meanings, as we have done in SOFTHESAURUS (Georgiev, 1993), we have used natural language, English words. We have chosen one word to represent a whole group of synonyms. We have chosen such a word that can assume different grammatical endings (Noun, Verb, Adjective, Adverb) without changing its initial part (the root). For example "wave" can be a Noun, Verb or Adjective and can replace such words as billow, ripple, undulate, undulation, in context. The word "wave" can substitute also "sea wave", "electromagnetic wave", "radio wave(s)", "motion with the hand", etc. The word try in combination with other words changes its meaning. For example:
I tried a piece of the apple pie my mother made.
will mean
I ate a piece of the apple pie my mother made.
In another context
John tried a cigarette.
would mean
John smoked a cigarette.
Therefore, it would be better to use the actual contextual meaning of try.
Wherever possible, we have preferred an unambiguous word to act as a substitute.
In the examples below "fragile" and "fragility" can replace all other words from the same group. The same applies to "swamp".
FRAGILE apt to break breakable breakableness brittle
brittleness crisp crumbliness crunchy easy to break
fragility
SWAMP bog boggy marsh marshy
marshy ground mire morass moss bog peatbog
peatmoss quagmire slough sloughy spongy ground
swampiness swampland swampy
So, the initial part of the word (the part before the grammatical ending) acts as a code for the meaning. We can use a rule in the program "compare string", when we want to know what meaning
the word has. This saves us a lot of programming effort and finds a way to code the meaning successfully, using a natural language word, which everybody can understand, without using a code, which only the program and the programmer can understand.
From a set of synonyms we have chosen a single word for the translation, not a phrase. For example:
bring to safety = rescue
brought to safety = rescued
As a result, many phrases were replaced with a single word. For example
union jack = flag
impose a fine = fine
the white scourge = consumption
By substituting entire phrases with single words, equivalent in meaning, we have often changed the syntactic structure of the sentence. For example
with little margin (Prepositional Phrase) = narrow (Adjective)
a Prepositional Phrase becomes an Adjective used as an Adverb, or
pay a visit (Verb plus Nominal Phrase) = visit (Verb)
fallen woman = prostitute
fast girl = prostitute
In the latter case we have resolved also an ambiguity (pay and fallen, fast).
You can download an evaluation copy, which will work ten times.
Download link: