linguistic bug in chapter 17

[Jon Kiparsky]

Since we're posting bugs, here's one. This is "just semantics" but of
course clarity is a matter of getting the words right, so "just
semantics" is maybe important. If so, the definition

(defun weld (subject object) ... )

is incorrect, since we're not talking about a subject at all. You
might use "object" and "indirect-object" or "object1" and "object2" or
you could reach into the bag of thematic roles and pull out, maybe
"patient" or "theme" or some such, or you could simply punt and say
"object1" and "object2" - depends on what aspects of the sentence
"weld the chain to the bucket" you want to highlight.
In no case, however, is the subject of the sentence either the chain
or the bucket.

Just sayin'...

[Sequiturian]

It might look a little long to write it out as:

(defun weld (direct-object object-of-the-preposition-to) .... )

Especially, since it's tedious and boring to do such things in programming.

Just sayin', too...

[Jon Kiparsky]

What about avoiding the parts of speech entirely?

(defun weld (this that) ... )

As in "weld this to that". I know, they use it in those nasty imperative languages, but it's a perfectly good construction.

[Jochen Heller]

Am Donnerstag, 10. November 2011 08:00:53 UTC+1 schrieb Jon Kiparsky:

the definition

(defun weld (subject object) ... )

is incorrect, since we're not talking about a subject at all.

Well ...

(game-action weld chain bucket attic
  (if (and (have 'bucket) (not *chain-welded*))
      (progn (setf *chain-welded* t)
         '(the chain is now securely welded to the bucket.))
      '(you do not have a bucket.) ))

That's where it ends:

`(the ,subject is now securely welded to the ,object.)