ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Current Semantic Web Layer Cake

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "Adrian Walker" <adriandwalker@xxxxxxxxx>
Date: Wed, 8 Aug 2007 16:22:50 -0400
Message-id: <1e89d6a40708081322i1e33b250l95dd8da85ded5177@xxxxxxxxxxxxxx>
Hi John --

Good, we are agreed.  Vocabulary should be open.

But IMHO so should the syntax, to the largest extent possible, as in the IBL system online at the site below.

So, this raises questions:

Q1.  I wish to use the new verb "to change up".  (It's established government jargon, and it's not about shifting gears in a car).  In a "controlled" approach, would I have to update a dictionary, and possibly also a grammar? 

Q2.  Then I'd like to write executable English sentences like "we will change up the capacity building withholding under programmatic disinterested restabilization".  (Invented, but the kind of thing you find in government  documents.)  Would I need to find a computational linguist to update the "controlling" grammar for me?  For each new sentence of this kind of complexity?  It starts to look impractical.

These kinds of "controlled" systems are often brittle, and do not seem to be widely used for practical real world tasks.  For example, I once tested one of the most sophisticated known dictonary-grammar based "controlled English" systems with the sentence "carpooling is a good thing".  Sad to say, it died.

At one level, this is a discussion about  two different pragmatic approaches (controlling dictionary-grammar versus neither) to making English computationally useful.  However, underlying that, there may be some rather fundamental questions about the nature of natural language in context.  What do you think?

                                    Cheers,  -- Adrian

Internet Business Logic (R)
A Wiki for Executable Open Vocabulary English
Online at www.reengineeringllc.com    Shared use is free

Adrian Walker
Reengineering

On 8/8/07, John F. Sowa <sowa@xxxxxxxxxxx> wrote:
Adrian,

I was most definitely arguing for an "open" vocabulary:

AW> John's point that you can describe anything in English is
> well taken.  However, although the tradition is to do this
> computationally by means of a "controlled vocabulary", there
> is another way.  It is in fact possible to get strict meanings
> using an *open* vocabulary.

The controlled natural languages, such as ACE, CLCE, etc.,
definitely have an open vocabulary.  The list of "reserved words"
is very small:

    and, or, not, if, then, some, every, is, has, a, an, the,
    of, for, such that.

The above list is a minimum, and there are some additions that
one could add to make the style more readable.  The list of
content words is totally open-ended.

As I said in my previous note, please look at the Rolf Schwitter's
web page for controlled NLs, which has pointers to a variety of
different versions:

    http://www.ics.mq.edu.au/%7Erolfs/controlled-natural-languages/
    Controlled Natural Languages - Homepage

John




_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto: ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx



_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (01)

<Prev in Thread] Current Thread [Next in Thread>