Print

Print


On Fri, May 11, 2012 at 10:32 AM, Patrick Dunn <[log in to unmask]> wrote:
> On Fri, May 11, 2012 at 12:22 PM, Wm Annis <[log in to unmask]> wrote:
>
>> On Fri, May 11, 2012 at 11:34 AM, Patrick Dunn <[log in to unmask]> wrote:
>> > Well, your approach sounds like generative grammar, which is fine.  If I
>> > were to draw a syntax tree, "no hay de que" would stand on its own node.
---
>> I think I'm safe in saying most generativists would actually object very
>> strongly to trying to shove a phrase into a single node.
---
> Yes.  But then I'm saying "no hay de que" and "thank you" are not phrases.
>  They're compound words.

Speaking as an engineer who was once involved in trying to program a
computer to "understand" natlang sentences, I would put "no hay de
que" in the dictionary as a single entry. Just as I actually did with
"New York", "Boy Scout", "natural selection", and "To Kill a
Mockingbird". Granted these CAN be analyzed, but in practical terms,
analyzing them is a waste of time since they occur as complete units.
Since biological systems seem to always favor parsimony, my guess is
that so as to not waste precious processing cycles, the human brain
treats them as compound units as well.

As for how they MIGHT be interpreted within one theoretical framework
or another, that is far less interesting to me than knowing how they
are actually processed in the brain. Because that's what leads to
understanding how best to handle them in a computer.

--gary