The complete sentence (the opening of a novel of C.S. Lewis) can now be parsed - instantaneously of course - by the computer. When a computer takes longer than an instant to do anything something is very wrong.
Two points of interest requiring particular care are exemplified by:
1) the use of one noun attributively as a pre-modifying classifier of a second noun: "chestnut tree."
2) the differences between the grammar of the various prepositions ("on," "from," "of," "into"). The machine is programmed to expect "into" to be governed by a verb, "of" by either a noun or an adverb, "from" by either a verb or a noun, and "on" by either a verb or a noun or the entire clause. The possible governors should be specified not merely as word types, but as the individual verbs, nouns, and so on. (The Collins "Grammar Patterns" series is most helpful with the detail.) A great many ambiguities can be resolved using this approach, which essentially follows the changing expectations of the mind as it moves forward to the apprehension of each word in its turn.
A little more work is needed here, so that the whole analysis might if required fit on one screen, with the use of right and left arrow keys to extract and contract its sub-sections.
The next sentence to be tested - the opening of Oppenheim's
Malefactor - is:
"Tall and burly, with features and skin hardened by exposure to the sun and winds of many climates, he looked like a man ready to face all hardships, equal to any emergency."This one, which an earlier and now superseded version of the reader-parser already handled sufficiently well, contains three points of interest:
1) the string of "free adjectival modifiers" with which it begins;
2) "look like" (sometimes called a "link verb") with its complement;
3) the adjectives ("hardened . . .", "ready . . .", "equal . . .") with their post-modifiers, whose presence causes them to follow the noun rather than precede it.
But I intend also soon to provide some examples of the computer's answering back - an important goal towards which we must all work must we not.
