Re: SUO: Try coding protocols
Josiane,
At 12:25 AM 8/26/2001 +0200, josiane caron wrote:
>Adam,
> First I want to answer to this:
>
>At 13:22 24/08/01 -0700, Adam wrote :
>>As a separate issue, I should point out that these statements are not
>>English comments, or an informal notation. Spaces, capitalization,
>>argument order are all significant, so
>>
>>(in scope of interest
>>Person1
>>(penetrates Peg-B Disk1))
>>
>>is not a meaningful expression because this uses a predicate 'in' that
>>takes 5 arguments.
>>
>>(inScopeOfInterest
>> Person1
>> (penetrates Peg-B Disk1))
>>
>>is a meaningful statement in SUO-KIF and SUMO because the relation
>>inScopeOfInterest is defined on two arguments, the first of which is an
>>'Agent' and the second of which is an 'Entity' (and a forumla is a
>>subclass of Entity).
>
>O.K. I thought is was. But after I forgot and it is sure that I am not yet
>familiar with. I will be more careful
>So all the suggestions I did to John Bateman are wrong.
Don't worry, it can take a while to understand this stuff. I appreciate
your hard work in learning it.
>>Josiane,
>>
>>At 12:23 AM 8/24/2001 +0200, josiane caron wrote:
>>>Adam,
>>>
>>> Moreover I would have like to code the formulation of the disk
>>> and of the peg:
>>>In example 1: 'the pink disk'
>>>you have introduced the notion of attribute, but if I use this form I
>>>find it is not right. It 'll give us 'the pink disk1' instead of ' the
>>>pink disk'
>>>In fact I need to know both concept and the occurrence of the concept
>>>(type and occurrence). I proposed a re-writing for sentence 1, below (I
>>>left sentences 2 and 4 wirth your writing). But the problem is I need
>>>some different formulation for 'instance'; I wrote something by
>>>introducing the word 'naming'; what can I do ?
>>
>>If you want to avoid naming the instance, you could state
>>
>>(exists (?X)
>> (instance ?X Disk)
>> (attribute ?X Pink))
>>
>>; "There exists something that is a disk and is pink"
>
>I don't realy understand what you mean by 'avoid naming the instance' ?
>Nevertheless I try to do something.
>The example can be wrote in this way, could you tell me if it is right ?
>
>1 - I take the pink disc I put it on the disc b #2
>(4 3 2 ) (1) ( )
>4321 0 0 ----> 432 1 0
>
>(instance Person1 Human)
>(instance Take1 Removing)
>(instance Put1 Putting)
>(destination Put1 Peg)
>(exists (?X)
> (instance ?X Peg-B
the syntax of the expression is wrong. It needs two closing parentheses -
(exists (?X)
(instance ?X Peg-B))
However, although with that fix the expression is syntactically ok, it
loses the information in the sentence above. There is nothing in the
logical statements that you list which relate the action to the particular
peg. If you wish not to refer to the peg by name, but rather to
existentially quantify it as well, you need a compound expression as
follows, and you might as well also not reify the person or events:
(exists (?PER ?PEG1 ?PEG2 ?EVT ?EVP ?D)
(and
(instance ?PER Human)
(instance ?EVT Removing)
(instance ?EVP Putting)
(instance ?D Disk)
(destination ?EVP ?PEG2)
(origin ?EVT ?PEG1)
(patient ?EVT ?D)
(patient ?EVP ?D)
(agent ?EVT ?PER)
(agent ?EVP ?PER)
(after ?EVP ?EVT)
(attribute ?D PinkColor)))
>(exists (?Y)
> (instance ?Y disk-B)
>
>Here the problem is that the destination peg is peg-B at physical level.
>but at metal level (is it that you call abstract ?) it is 'disk-B', i e
>the disk which is moved to the destination peg. It is what I called a
>disk-piton. But that it is the story.
>In order to kept the information in the codage, can I write (exists (?X)
>and (exists (?Y) ?
>if no I will simplifly; if yes I would like to write that disk-B = peg-B
>at physical level
>
>It seem important because here the concept-disk and the concept-peg are
>linked. It is that that I want to code in this complicated case.
>
>Now I continue the example:
>
>(origin Take1 Peg-A)
>(patient Put1 Disk)
>(exists (?Z )
> (instance ?Z it)
This statement needs a closing parenthesis. Also, you must define what the
term "it" is. In a logic statement, any term is just a token or string of
characters. Each term needs definitions to give it meaning. It is also
essential to understand the relations from SUMO that you are
using. 'instance' is a relation defined in SUMO. It takes two arguments,
the first of which is an individual, the second of which is a class. So,
for 'it' to appear in the second position and be correct, it must be a class.
>(agent Put1 Person1)
>(patient Take1 Disk)
>(instance Disk1 Disk)
>(exists (?X)
> (instance ,X Disk)
> (attribute ?X Pink)
The statement should be
(exists (?X)
(instance ?X Disk)
(attribute ?X Pink))
>(agent Take1 Person1)
>(after Put1 Take1)
>
>(inScopeOfInterest
> Person1
> (penetrates Peg-B Disk1))
>
>(inScopeOfInterest
> Take1
> (on Disk4 Peg-A)
> (on Disk3 Disk4)
> (on Disk2 Disk3)
> (on Disk1 Disk2))
inScopeOfInterest is a relation that takes two arguments. In the
expression above you've given it 5 arguments. I suspect you meant
(inScopeOfInterest
Take1
(and
(on Disk4 Peg-A)
(on Disk3 Disk4)
(on Disk2 Disk3)
(on Disk1 Disk2)))
>(inScopeOfInterest
> Put1
> (on Disk4 Peg-A)
> (on Disk3 Disk4)
> (on Disk2 Disk3)
> (on Disk1 Peg-B))
same as above
>>You could use the 'documentation' relation instead of having to create a
>>new relation called naming.
>
>I don't understand this: I found the documention which gives an overview
>of the SUMO. It's fine. But I don't understand how to use it 'instead of
>having to create a new relation called naming. Sorry, could you explain me ?
In your previous message you introduced a relation that you called
'naming'. I think that you could use the relation 'documentation'
instead. For example, instead of stating
(naming Peg1 "Peg-A")
you could say
(documentation Peg1 "Peg-A")
>>However, I'm not sure why you would want to avoid creating a token in a
>>logic expression to signify a concept. In fact, in most deduction
>>systems even the existentially quantified term would get automatically
>>turned into a skolem function which names the concept. So, I don't see
>>why your proposed relations of 'specification-naming', 'instance-naming',
>>and 'naming' are needed.
>
>I don't want to avoid creating a token in a logic expression to signify a
>concept. I already know that it works like that.
>I need to specify the linguistic expression of the concepts because it
>introduces a relation with another concept and I guess I need it later at
>mental level.
Ok, maybe I misunderstood you before, but can you explain why (if)
'specification-naming', 'instance-naming', and 'naming' are needed?
>>> Furthermore there are the problems of connectives in 3 and 4.
>>
>>By the 'connectives' do you mean the word 'again'? Like 'so', 'then' or
>>many other English utterances, this word (used in this way) may not have
>>an actual meaning in terms of what is being communicated, but rather
>>functions to regulate the act of communication itself. 'Again' just
>>emphasizes that this is an action that has been performed before. So,
>>
>>"I hit the drum."
>>"I hit the drum again."
>>
>>The 'again' just lets the listener know that there is no mistake in the
>>utterance. The sentence really means exactly what it says. So, the
>>appearance of 'again' would not change the logical expression that can be
>>translated from the surface natural language form.
>>
>>Adam
>
>Yes I mean these word, but perhaps it is not useful to consider them now.
>i am O.K.
>Thanks for your help.
glad to be of assistance.
Adam
>Josiane
>_________________________________________________
>
>Josiane Caron-Pargue
>Laboratoire Langage et Cognition LaCo, UMR 6096
>Maison des Sciences de l'Homme et de la Société, MSHS
>99 Avenue du Recteur Pineau, F-86022 Poitiers cedex
>tel bur 05 49 45 46 23, fax 05 49 45 46 16
>tel secr. 05 49 45 46 10, 05 49 45 46 13
>http://www.mshs.univ-poitiers.fr/laco/
>josiane.caron@mshs.univ-poitiers.fr
>_________________________________________________
Adam Pease
Teknowledge
(650) 424-0500 x571