Re: Buffalo
Barry Smith
>At 02:59 PM 3/1/2010, you wrote:
>>On Monday, March 1, 2010, Phillip Lord <philli...@newcastle.ac.uk> wrote:
>> > Alan Ruttenberg <alanrut...@gmail.com> writes:
>> >>>> If a bug hides in hole, then a portion of the bug's life (a process)
>> >>>> occurs in hole.
>> >>>> Lots of digestion related processes happen in FMA:14585
>> Cavity of stomach.
>> >>>
>> >>> If there is a bug in it, then it's not a hole right?
>> >>
>> >> I'm still trying to understand how all this fits together, Phil, but
>> >> according to my best undertanding, it's still a hole. The bug crawls
>> >> out, it's there again. You might want to check with an anatomist about
>> >> their intuitions about cavities, but from my experience, they are
>> >> going to say that the cavity of the stomach is still there when you
>> >> have a meal in it. .
>> >
>> >
>> > Sure, just not one that is made of an immaterial entity.
>>
>>I'm not sure that "made of" is the way of speaking here.
>>
>>However note that the fma says the cavity of stomach is an immaterial entity.
A hole can be empty or filled
It is still the same hole
>>I don't think that the equation hole = empty should be made. If it
>>were we wouldn't be able to have holes anywhere but in a vacuum.
>>
>>Instead I think the analogy is to boundaries, which are dependent on
>>some material entity. Move the thing that the hole is in, and the hole
>>moves. Destroy the thing that the hole is in and you destroy the hole.
>>
>>This dependency seems to differentiate immaterial continuants from space.
Exactly
BS
>>-Alan
>>
>>
>> >
>> > The point is to avoid getting a situation where you were out that
>> > processes can happen in nothing, because, in general, they can't.
>>
>>
>>
>> >
>> > Phil
>> >
>> > --
>> > You received this message because you are subscribed to the
>> Google Groups "BFO Discuss" group.
>> > To post to this group, send email to bfo-d...@googlegroups.com.
>> > To unsubscribe from this group, send email to
>> bfo-discuss...@googlegroups.com.
>> > For more options, visit this group at
>> http://groups.google.com/group/bfo-discuss?hl=en.
>> >
>> >
>>
>>--
>>You received this message because you are subscribed to the Google
>>Groups "BFO Discuss" group.
>>To post to this group, send email to bfo-d...@googlegroups.com.
>>To unsubscribe from this group, send email to
>>bfo-discuss...@googlegroups.com.
>>For more options, visit this group at
>>http://groups.google.com/group/bfo-discuss?hl=en.
Bjoern Peters
I thought 'site' was essentially defined as a spatial region whose boundaries are determined relative to some material entities. That makes sense to me, and would make the 'hole' in X a site, independent of bug occupancy.
I thought 'spatial region' was a Newtonian view of absolutist space that exists without any reference to material entities (and so would not be applicable for any of the examples discussed in this thread).
Now I understand from what BS wrote that 'site' is supposed to consist of material entities (molecules of oxygen in the mole hole) and would be modified if those were removed. That seems completely unnecessary. We can perfectly well define the material entities located in the hole relative to the spatial region hole being defined relative to what it it a hole in.
If I understand this right, there is a proposal to have four things:
A material entity: scattered aggregate of oxygen molecules:
A site: scattered aggregate of molecules contained in a hole (I see no difference to the above)
An immaterial entity: 'hole location' the spatial region occupied by the site(?)
A spatial region: At any given instance, a different part of space (assuming that the hole is on earth, which moves in relation to absolutist space)
I hope I misunderstood what I read and believe that all that is needed is 'site' defined as a spatial region whose boundaries are determined relative to some material entities, and a relation 'located in' of material entities to sites.
- Bjoern
--
Bjoern Peters
Assistant Member
La Jolla Institute for Allergy and Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters
Barry Smith
>I was hoping that this discussion would clarify some of the issues I
>have had with bfo:spatial region in the past. No such luck.
>
>I thought 'site' was essentially defined as a spatial region whose
>boundaries are determined relative to some material entities. That
>makes sense to me, and would make the 'hole' in X a site,
>independent of bug occupancy.
The problem is that sites being determined relative to material
entities as boundaries, and sites being spatial regions, yield
conflicts when the material entities which serve as boundaries move
in space, e.g. in the case of ships' hulls.
I prefer to go with the material entities as boundaries approach --
which also corresponds to the standard view as formulated e.g. in
Casati and Variz's book, Holes.
A site is an immaterial entity that is a hole, cavity, determined by
some material entity or entities which serve as its boundaries. Sites
may also have fiat boundaries, e.g. in the case of an open mouth, a
room with an open door, a open hole in the ground.
Each site, so defined, is projected on some spatial region.
BS
Phillip Lord
> At 11:04 PM 3/1/2010, Bjoern Peters wrote:
>> I was hoping that this discussion would clarify some of the issues I have
>> had with bfo:spatial region in the past. No such luck.
>>
>> I thought 'site' was essentially defined as a spatial region whose
>> boundaries are determined relative to some material entities. That makes
>> sense to me, and would make the 'hole' in X a site, independent of bug
>> occupancy.
>
> The problem is that sites being determined relative to material entities as
> boundaries, and sites being spatial regions, yield conflicts when the material
> entities which serve as boundaries move in space, e.g. in the case of ships'
> hulls.
I am a little bit confused about this. Things don't move in space, they
move relative to other things. Space can only be specified relative to
things.
So, what is the difference between the spatial region defined by the
inside of a hull of a ship and the site defined by the same hull? What
is the use case that requires you to make this distinction?
Phil
Alan Ruttenberg
<philli...@newcastle.ac.uk> wrote:
> Barry Smith <phis...@buffalo.edu> writes:
>> At 11:04 PM 3/1/2010, Bjoern Peters wrote:
>>> I was hoping that this discussion would clarify some of the issues I have
>>> had with bfo:spatial region in the past. No such luck.
>>>
>>> I thought 'site' was essentially defined as a spatial region whose
>>> boundaries are determined relative to some material entities. That makes
>>> sense to me, and would make the 'hole' in X a site, independent of bug
>>> occupancy.
>>
>> The problem is that sites being determined relative to material entities as
>> boundaries, and sites being spatial regions, yield conflicts when the material
>> entities which serve as boundaries move in space, e.g. in the case of ships'
>> hulls.
>
>
> I am a little bit confused about this. Things don't move in space, they
> move relative to other things. Space can only be specified relative to
> things.
I'm confused. Are you saying there is no such thing as "space".
Physics recognizes acceleration independent of any other object. What
is that acceleration relative to?
-Alan
>
> So, what is the difference between the spatial region defined by the
> inside of a hull of a ship and the site defined by the same hull? What
> is the use case that requires you to make this distinction?
>
> Phil
>
Barry Smith
Bjoern Peters
----- "Alan Ruttenberg" <alanrut...@gmail.com> wrote:
> On Wed, Mar 3, 2010 at 6:35 AM, Phillip Lord
> I'm confused. Are you saying there is no such thing as "space".
> Physics recognizes acceleration independent of any other object. What
> is that acceleration relative to?
>
> -Alan
Acceleration: Yes. Movement: No.
And yes, it is meaningless (or at least not necessary) in physics to talk about space independent of any material entity. That is why I believe all we need are 'sites' (Some being e.g. earth) which form reference frames relative to which motion should be defined.
Will read Barry's link next.
- Bjoern
>
> >
> > So, what is the difference between the spatial region defined by
> the
> > inside of a hull of a ship and the site defined by the same hull?
> What
> > is the use case that requires you to make this distinction?
> >
> > Phil
> >
> > --
> > You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> > To post to this group, send email to bfo-d...@googlegroups.com.
> > To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com.
> > For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
> >
> >
>
> --
> You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
--
Bjoern Peters
----- "Barry Smith" <phis...@buffalo.edu> wrote:
OK, did not have to read long: If we are doing a realist ontology, why is Newton's magical view of an absolutist reference frame of any relevance?
>
> >Phil
> >
> >--
> >You received this message because you are subscribed to the Google
> >Groups "BFO Discuss" group.
> >To post to this group, send email to bfo-d...@googlegroups.com.
> >To unsubscribe from this group, send email to
> >bfo-discuss...@googlegroups.com.
> >For more options, visit this group at
> >http://groups.google.com/group/bfo-discuss?hl=en.
>
> --
> You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
--
Phillip Lord
>> I am a little bit confused about this. Things don't move in space, they
>> move relative to other things. Space can only be specified relative to
>> things.
>
> I'm confused. Are you saying there is no such thing as "space".
> Physics recognizes acceleration independent of any other object. What
> is that acceleration relative to?
There is no absolute space. If you want to describe a particular piece
of space, then it must be specified relative to some object. So you can
say "the piece of space which the earth is in" or "the piece of space
halfway between the earth and the sun", or "the piece of space inside
this ship hull". The only exception I can think of this is if you want
to say "all of space", but then I think that this is specified relative
to all objects.
From what Barry is saying these are sites, because they can move. But,
given his example of a ship hull moving through, for example, water, it
seems equally plausible to state that the hull is stationary and the
water is moving.
Phil
Phillip Lord
>> I'm confused. Are you saying there is no such thing as "space".
>> Physics recognizes acceleration independent of any other object. What
>> is that acceleration relative to?
>>
>> -Alan
>
> Acceleration: Yes. Movement: No.
And acceleration doesn't give us an absolute notion of space either.
Even here, all we do is that all the entities in this accelerative frame
have a force upon them. But this could be a graviational force.
> And yes, it is meaningless (or at least not necessary) in physics to
> talk about space independent of any material entity. That is why I
> believe all we need are 'sites' (Some being e.g. earth) which form
> reference frames relative to which motion should be defined.
>
> Will read Barry's link next.
Well, if the distinction between site and spatial region requires us to
read a 22 page document, I suspect that in the context of a simple upper
ontology like BFO, almost no one is going to be able to apply the
distinction cleanly.
Personally, I'd go for saying we don't need "sites", rather than spatial
regions. My own feeling is that the solution that works for spatial
regions also needs to work for spatiotemporal regions. Does, for
example, "the site inside a hull of a ship" relate to a spatial region
or a spatiotemporal region?
Phil
Barry Smith
Since sites are independent continuants, then they are projected on
spatial regions -- just on different spatial regions at different times.
We need sites to deal with FMA immaterial entities, such as for
example the lumen of the gut.
As you move through different spatial regions so also does the lumen
of your gut.
There the lumen of your gut cannot BE a spatial region.
BS
Bjoern Peters
----- "Barry Smith" <phis...@buffalo.edu> wrote:
Phil: We are on the same page. When I say 'all we need is sites', I meant in the current BFO sense of sites being defined in relation to material entities vs. spatial regions defined against a magical Newtonian absolut reference frame.
Barry: The projection of sites to spatial regions can just as well be dealt with as the location of sites in other sites.
- Bjoern
Phillip Lord
>>Personally, I'd go for saying we don't need "sites", rather than spatial
>>regions. My own feeling is that the solution that works for spatial
>>regions also needs to work for spatiotemporal regions. Does, for
>>example, "the site inside a hull of a ship" relate to a spatial region
>>or a spatiotemporal region?
>
> Since sites are independent continuants, then they are projected on
> spatial regions -- just on different spatial regions at different
> times. We need sites to deal with FMA immaterial entities, such as for
> example the lumen of the gut. As you move through different spatial
> regions so also does the lumen of your gut. There the lumen of your
> gut cannot BE a spatial region.
That doesn't follow. Why can two spatial regions not move through each
other? The spatial region defined by the lumen of your gut moves through
the spatial region defined by the earth.
Alternatively, we could say that the site defined by the lumen of the
gut is moving wrt to the site defined by which ever part of the earth we
are in.
There appears to be no clear difference between the two here. In your
example, tell me how you have determined that the lumen of the gut is
moving through a spatial region, as opposed to the lumen staying still
and the surrounding moving?
Phil
Phillip Lord
>> Since sites are independent continuants, then they are projected on
>> spatial regions -- just on different spatial regions at different
>> times. We need sites to deal with FMA immaterial entities, such as
>> for example the lumen of the gut. As you move through different
>> spatial regions so also does the lumen of your gut. There the lumen
>> of your gut cannot BE a spatial region. BS
>>
>
> Phil: We are on the same page. When I say 'all we need is sites', I
> meant in the current BFO sense of sites being defined in relation to
> material entities vs. spatial regions defined against a magical
> Newtonian absolut reference frame.
Yes, I agree with you.
Phil
Barry Smith
>Barry: The projection of sites to spatial regions can just as well
>be dealt with as the location of sites in other sites.
>
>- Bjoern
I am tempted by this idea. One objection:
some spatial regions (e.g. all the 0-, 1- and 2-dimensional ones) are not sites
BS
Michel Dumontier
Phillip Lord
Why?
Phil
Phillip Lord
Michel
They could be, but then do we really need "container" as a term in BFO.
And how would this be different from any other spatial region which can
contain entities of the same or lower dimensionality; or a boundary
which is something that does contain entities of the same or lower
dimensionality.
Phil
Michel Dumontier <michel.d...@gmail.com> writes:
> Sites could be those that contain entities of the same or lower
> dimensionality.
>
> m.
>
> On Wed, Mar 3, 2010 at 1:20 PM, Barry Smith <phis...@buffalo.edu> wrote:
>
>> At 12:33 PM 3/3/2010, Bjoern Peters wrote:
>>
>>> Barry: The projection of sites to spatial regions can just as well be
>>> dealt with as the location of sites in other sites.
>>>
>>> - Bjoern
>>>
>>
>> I am tempted by this idea. One objection:
>>
>> some spatial regions (e.g. all the 0-, 1- and 2-dimensional ones) are not
>> sites
>>
>> BS
>> --
>> You received this message because you are subscribed to the Google Groups
>> "BFO Discuss" group.
>> To post to this group, send email to bfo-d...@googlegroups.com.
>> To unsubscribe from this group, send email to
>> bfo-discuss...@googlegroups.com<bfo-discuss%2Bunsu...@googlegroups.com>
>> .
>> For more options, visit this group at
>> http://groups.google.com/group/bfo-discuss?hl=en.
>>
>>
--
Phillip Lord, Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics, Email: philli...@newcastle.ac.uk
School of Computing Science, http://homepages.cs.ncl.ac.uk/phillip.lord
Room 914 Claremont Tower, skype: russet_apples
Newcastle University, msn: m...@russet.org.uk
NE1 7RU
Barry Smith
contain = part_of here?
At 01:30 PM 3/3/2010, you wrote:
>Sites could be those that contain entities of the same or lower
>dimensionality.
>
>m.
>
>On Wed, Mar 3, 2010 at 1:20 PM, Barry Smith
><<mailto:phis...@buffalo.edu>phis...@buffalo.edu> wrote:
>At 12:33 PM 3/3/2010, Bjoern Peters wrote:
>Barry: The projection of sites to spatial regions can just as well
>be dealt with as the location of sites in other sites.
>
>- Bjoern
>
>
>I am tempted by this idea. One objection:
>
>some spatial regions (e.g. all the 0-, 1- and 2-dimensional ones)
>are not sites
>
>BS
>--
>You received this message because you are subscribed to the Google
>Groups "BFO Discuss" group.
>To post to this group, send email to
><mailto:bfo-d...@googlegroups.com>bfo-d...@googlegroups.com.
>To unsubscribe from this group, send email to
><mailto:bfo-discuss%2Bunsu...@googlegroups.com>bfo-discuss...@googlegroups.com.
>For more options, visit this group at
><http://groups.google.com/group/bfo-discuss?hl=en>http://groups.google.com/group/bfo-discuss?hl=en.
Barry Smith
Our purpose is to support scientific researchers. If you can find a
community of researchers who take the lumen of your gut as their
frame of reference I would be happy to provide you an argument, here.
Otherwise, I prefer to go with: this is evidently the practical way.
BS
Phillip Lord
>>That doesn't follow. Why can two spatial regions not move through each
>>other? The spatial region defined by the lumen of your gut moves through
>>the spatial region defined by the earth.
>>
>>Alternatively, we could say that the site defined by the lumen of the
>>gut is moving wrt to the site defined by which ever part of the earth we
>>are in.
>>
>>There appears to be no clear difference between the two here. In your
>>example, tell me how you have determined that the lumen of the gut is
>>moving through a spatial region, as opposed to the lumen staying still
>>and the surrounding moving?
>
> Our purpose is to support scientific researchers. If you can find a community
> of researchers who take the lumen of your gut as their frame of reference I
> would be happy to provide you an argument, here. Otherwise, I prefer to go
> with: this is evidently the practical way.
From Wikipedia...
"In mammals, food enters the mouth, being chewed by teeth, with chemical
processing beginning with chemicals in the saliva from the salivary
glands. Then it travels down the esophagus into the stomach, where acid
both kills most contaminating microorganisms and begins mechanical break
down of some food (eg denaturation of protein)"
Food travels through the gut. So, there we have it.
You will not support scientific researchers by ignoring science. If you
can find me a community of researchers who can agree on the frame of
reference that everybody should use to define spatial region, then I
will be happy to listen.
Your solution is not practical, it's wrong. You are splitting hairs
between site and spatial region that neither need to be split and,
critically, ignores experimental data that has been available for over
100 years. As it's wrong, it's also going to be hard to apply in
practice, and people will use one randomly; so, therefore, it's bad
modelling also.
Site or spatial region, pick one.
Phil
Barry Smith
>Barry Smith <phis...@buffalo.edu> writes:
> >>That doesn't follow. Why can two spatial regions not move through each
> >>other? The spatial region defined by the lumen of your gut moves through
> >>the spatial region defined by the earth.
> >>
> >>Alternatively, we could say that the site defined by the lumen of the
> >>gut is moving wrt to the site defined by which ever part of the earth we
> >>are in.
> >>
> >>There appears to be no clear difference between the two here. In your
> >>example, tell me how you have determined that the lumen of the gut is
> >>moving through a spatial region, as opposed to the lumen staying still
> >>and the surrounding moving?
> >
> > Our purpose is to support scientific researchers. If you can find
> a community
> > of researchers who take the lumen of your gut as their frame of reference I
> > would be happy to provide you an argument, here. Otherwise, I prefer to go
> > with: this is evidently the practical way.
>
>
> From Wikipedia...
>
>"In mammals, food enters the mouth, being chewed by teeth, with chemical
>processing beginning with chemicals in the saliva from the salivary
>glands. Then it travels down the esophagus into the stomach, where acid
>both kills most contaminating microorganisms and begins mechanical break
>down of some food (eg denaturation of protein)"
>
>Food travels through the gut. So, there we have it.
This is indeed a good example, but unfortunately it does not satisfy
the challenge, which -- in your own terms -- is to find an example of
science which takes the lumen as frame of reference and sees its
surroundings as moving ("... the lumen staying still and the
surrounding moving ...")
>You will not support scientific researchers by ignoring science. If you
>can find me a community of researchers who can agree on the frame of
>reference that everybody should use to define spatial region, then I
>will be happy to listen.
BFO is, for obvious reasons, not offering or promoting a single
spatial or temporal or spatiotemporal frame of reference that
everyone should use. Astronomers, evolutionary biologists,
geographers, clinical trial designers, people performing
observational studies of animal behavior, will each take somewhat
different frames of reference. The point is that in every case the
regions in question will be distinct from sites. Thus they will not
move, they will not change their shapes ...
>Your solution is not practical, it's wrong. You are splitting hairs
>between site and spatial region that neither need to be split and,
>critically, ignores experimental data that has been available for over
>100 years. As it's wrong, it's also going to be hard to apply in
>practice, and people will use one randomly; so, therefore, it's bad
>modelling also.
I agree that all of these concerns are important, but I do not think
either side has the right to be confident that it has, yet, a working
solution.
>Site or spatial region, pick one.
Picking would be too easy.
Can you, if you truly are a defender of the 'sites can do the job of
spatial regions', tell me how we would treat 0-, 1- and 2-dimensional
spatial regions?
And how we would deal with the fact that the lumen of your gut can
change its shape through time?
BS
Bjoern Peters
'site' a primitive, but roughly: A continuant consisting of a spatial shape in relation to some arrangement of other continuants.
Two separate relations:
site has_part only site
material entity has_location only site
zero-dimensional site: a site that has no extension quality. Example: The tip of the needle. An intersection of two one-dimensional sites. The center of mass of a material entity.
one-dimensional site: a site that has a length extension quality.
two-dimensional site: a site that has an area extension quality.
three-dimensional site: a site that has a volume extension quality.
Sites can only be part_of sites of equal or higher dimension (Michel's point).
This is a first attempt, I am sure there are problems, but don't see that we are missing a separate 'spatial region'.
- Bjoern
children:
three-dimensional site,
# a geometric element that has position but no extension; "a point is defined by its coordinates"
----- "Barry Smith" <phis...@buffalo.edu> wrote:
--
Barry Smith
immaterial continuant in the FMA, and elsewhere), e.g. the lumen of
Phil's gut? This can move, can change its shape, etc.
BS
Bjoern Peters
- Bjoern
Barry Smith
>I thought the same way we deal with material entities whose mass and
>shape changes over time. The lumen of Phil's gut is an instance of a
>three-dimensional site which is part of site-Phil (the site in which
>the material entity Phil is located) and bounded by the interior of
>Phil's gut wall + some fiat boundaries at the ends.
let lumen of Phil's gut = L
Dr McX is studying the strange contracting and expanding behavior of
Phil's gut over time using a coordinate spatial frame of reference
defined by his imaging device, telling him that
spatial region occupied by L at time t1 = R1
spatial region occupied by L at time t2 = R2
R1 =/= R2
what is the relation between L and R1 and R2?
BS
Bjoern Peters
So far this is trivial. The interesting bit is defining movement of L within the ImageSite from t1 to t2. This would have to be done carefully, but roughly:
Within ImageSite, a continuous grid of one-dimensional sites indexed by x,y,and z can be constructed, with x increasing from left to right, y from bottom to top, and z from front to back. This construction can be done reproducibly at t1 and t2. The part of ImageSite occupied by L can be determined in terms of x,y,z. Therefore, a site exists at t2 which consists of the volume occupied by L at t1.
This says that movement can only be reasonably defined in reference to a site in which it is possible to compare past and current locations. To me this seems a reasonable requirement.
We can also single out sites in which movements of material entities follow Newton's Laws as Inertialsystems.
Barry Smith
>The area in which the device can record images (e.g. the inside of a
>CT tube) defines the ImagingSite. The lumen of Phil's gut L is part
>of the ImagingSite at t1, t2. The volume of L has two different
>quantitative values at t1, t2, similar to a mass quality.
Yes. And the relevant part of the imaging site at t1 is R1. The
relevant part of the imaging site at t2 is R2. The question, we will
recall, is: what is the relation between L, R1 and R2? See below
>So far this is trivial. The interesting bit is defining movement of
>L within the ImageSite from t1 to t2. This would have to be done
>carefully, but roughly:
>
>Within ImageSite, a continuous grid of one-dimensional sites indexed
>by x,y,and z can be constructed, with x increasing from left to
>right, y from bottom to top, and z from front to back. This
>construction can be done reproducibly at t1 and t2. The part of
>ImageSite occupied by L can be determined in terms of x,y,z.
>Therefore, a site exists at t2 which consists of the volume occupied
>by L at t1.
L, as originally defined, is this site. The question still remains.
What is the relation between the lumen of Phil's gut and the two
regions it occupies at t1 and t2?
BS
Matthew Pocock
Or perhaps I am missing something obvious here.
Matthew
Barry Smith
I think you are missing the FMA, which has all of these things
because this is they way scientists talk about anatomical structures.
The very fact that referring to these 0-, 1- and 2-dimensional parts
is useful means that we need to include terms for them in our ontology.
BS
Bjoern Peters
----- "Barry Smith" <phis...@buffalo.edu> wrote:
> At 03:31 PM 3/4/2010, Bjoern Peters wrote:
> >The area in which the device can record images (e.g. the inside of a
>
> >CT tube) defines the ImagingSite. The lumen of Phil's gut L is part
> >of the ImagingSite at t1, t2. The volume of L has two different
> >quantitative values at t1, t2, similar to a mass quality.
>
> Yes. And the relevant part of the imaging site at t1 is R1. The
> relevant part of the imaging site at t2 is R2. The question, we will
> recall, is: what is the relation between L, R1 and R2? See below
>
I don't believe there is a separate R1, R2. See below.
>
> >So far this is trivial. The interesting bit is defining movement of
> >L within the ImageSite from t1 to t2. This would have to be done
> >carefully, but roughly:
> >
> >Within ImageSite, a continuous grid of one-dimensional sites indexed
>
> >by x,y,and z can be constructed, with x increasing from left to
> >right, y from bottom to top, and z from front to back. This
> >construction can be done reproducibly at t1 and t2. The part of
> >ImageSite occupied by L can be determined in terms of x,y,z.
> >Therefore, a site exists at t2 which consists of the volume occupied
>
> >by L at t1.
>
> L, as originally defined, is this site. The question still remains.
> What is the relation between the lumen of Phil's gut and the two
> regions it occupies at t1 and t2?
> BS
>
Let me try again, and expand the example. I can ask my device to display what parts of the ImageSite were occupied by L at t1, t2. For lack of a better word, I will call them projection sites IS(L,t1), IS(L,t2).
Now assume the imaging was done in a moving ambulance. For an observer on the street there are projection sites O(L, t1), O(L, t2). That observer would describe the ImageSite itself to be moving as well O(IS,t1), O(IS,t2). There are trivial transformations that would allow the observer to construct IS(L,t1), IS(L,t2) are.
There is just one Lumen of Phil's gut L. But asking what 'spatial region was occupied at time t1, t2 like you are doing is only meaningful when also specifying the reference site (Bezugssystem) used.
- Bjoern
Bjoern Peters
Barry Smith
Let us do as you say
In frame-of-reference A, we have R1 at t1 and R2 at t2.
R1 and R2 both exist at other times; that is the nature of regions.
We can assert that for all times, R1 and R2 are not the same region.
In frame-of-reference Phil, we have L at t1 and L at t2.
These are both L, they are identical.
The question then arises: what is the relation between L, R1 and R2.
BS
The question again
Bjoern Peters
In my view, what you call R1 and R2 are just sites, which can be described by a constant coordinates within the ImageSite. No need to refer to 'regions' as something different from sites.
> We can assert that for all times, R1 and R2 are not the same region.
> In frame-of-reference Phil, we have L at t1 and L at t2.
> These are both L, they are identical.
> The question then arises: what is the relation between L, R1 and R2.
> BS
I hope I am getting your question, as this seems trivial: at t1, R1 and L exactly overlap. At t2, R2 and L exactly overlap.
Barry Smith
> > >
> > >There is just one Lumen of Phil's gut L. But asking what 'spatial
> > >region was occupied at time t1, t2 like you are doing is only
> > >meaningful when also specifying the reference site (Bezugssystem)
> > used.
> >
> > Let us do as you say
> > In frame-of-reference A, we have R1 at t1 and R2 at t2.
> > R1 and R2 both exist at other times; that is the nature of regions.
>
>In my view, what you call R1 and R2 are just sites, which can be
>described by a constant coordinates within the ImageSite. No need to
>refer to 'regions' as something different from sites.
>
> > We can assert that for all times, R1 and R2 are not the same region.
> > In frame-of-reference Phil, we have L at t1 and L at t2.
> > These are both L, they are identical.
> > The question then arises: what is the relation between L, R1 and R2.
> > BS
>
>I hope I am getting your question, as this seems trivial: at t1, R1
>and L exactly overlap. At t2, R2 and L exactly overlap.
Exact overlap is, surely, identity.
But then we have a contradiction.
BS
Bjoern Peters
----- "Barry Smith" <phis...@buffalo.edu> wrote:
Surely not. The identity criteria for two sites should be to exactly overlap at all times. At t2, R1 and L don't exactly overlap. At t1, R2 and L don't exactly overlap.
Barry Smith
Do you agree that, in (most of) biology, we are dealing with frames
of reference which are trivially translatable into each other, just
as degrees Fahrenheit and degrees Celsius, and the Christian calendar
and the Jewish calendar, are trivially translatable into each other?
BS
Bjoern Peters
For most of biology, dealing with non-relativistic transformations between inertial frames of reference will be sufficient. The Galilean transformations between them are straightforward, but I would not consider them trivial or compare them to the scale transformations you give as examples. The answers for seemingly trivial questions change based on them, e.g. "did x move?".
Also, let me pose this related challenge: Taking the example you gave, and describing it using BFOs view of absolute space. Then R1, R2 as 'spatial regions' which the site L occupies at t1, t2. Can you tell me what the distance between R1, R2 is? (hint: you can't)
- Bjoern
Matthew Pocock
The very fact that referring to these 0-, 1- and 2-dimensional parts is useful means that we need to include terms for them in our ontology.
BS
I entirely agree with this. Ontologies should ideally cover exactly those things that the users wish to discuss and any ancilliary things that enables this discussion (perhaps numbers, units, opinions, ...). However, this doesn't strengthen the view that these 2-d and 1-d things exist in any sense other than as a convenience for clear discussion, and it certainly does not mean that we should confuse coordinates with the places they index or a coordinate system with actual volumes of space. Being explicit about the indexing of space (coordinate systems and the degrees of freedom they provide) as distinct from the space itself has to be better than muddling them up. It seems like a classic case of use-mention to me.
Matthew
Barry Smith
Can we agree on the following:
that even if we accept inertial frames of reference as indispensable,
there would still be a difference between the objects located in
regions defined by such frames of reference and the regions themselves?
objects move relative to regions
that objects change shape (absolutely?)
that regions do not move
that regions do not change shape
the immediate issue is then: is the lumen of Phil's gut more like a
region or a like an object
BS
Jessica Turner
--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To post to this group, send email to bfo-d...@googlegroups.com.
To unsubscribe from this group, send email to bfo-discuss...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/bfo-discuss?hl=en.
--
Jessica Turner, Ph.D.
Research Scientist
MIND Research Network
1101 Yale Blvd. NE
Albuquerque, New Mexico 87106
Ph: 505-272-1570
email: jtu...@mrn.org
Alan Ruttenberg
> Objects can be located within other objects, right? So the original point,
> that stuff can move through the lumen, doesn't determine whether it's more
> like a region than an object?
In the current relation ontology the object/object relation located_in
is defined in terms of the region/region located_in, by first
projecting to the region (using the object/region located_in
relation), and then relying on the spatial sense of part_of. Doing so
gets us the definition of located_in (object/object) in terms of
something we seemed to (before this conversation) better understand -
how space is part of space.
At question is how the relations part_of, location_of, and contains_in
are defined - how many must be primitively asserted in order to
determine the others - how they are inter-related.
-Alan
Bjoern Peters
Unfortunately we can't agree on this. My main argument is that I don't recognize why we need two types of 'spatial entities' (site and spatial region). When you say 'region' in the above, I assume you want to maintain that distinction. The point 'that regions do not move' sounds very much like magical 'absolute space' again, which simply does not exist.
Here is a rewrite of the points above substituting in my thinking about sites instead.
- inertial frames of reference are sites with certain properties
- there is a difference between an object (which I take to be short for material entities) and the site it is located in
- objects move relative to other objects, which corresponds to movement within larger sites.
- objects can change shape
- sites move relative to sites
- sites can change shape
- The lumen of Phil's gut is an instance of a site which is part of site-Phil (the site in which the material entity Phil is located) and bounded by the interior of Phil's gut wall + some fiat boundaries at the ends. At different time points, different material entities have locations that are contained in the lumen of Phil's gut.
I appreciate this discussion. At this point, I would like to know if you still think there is strong argument for needing two separate 'spatial entities' (site and spatial region), or if we can start together trying to formalize in earnest the use of just one.
- Bjoern
>
> --
> You received this message because you are subscribed to the Google
> Groups "BFO Discuss" group.
> To post to this group, send email to bfo-d...@googlegroups.com.
> To unsubscribe from this group, send email to
> bfo-discuss...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/bfo-discuss?hl=en.
--
Bjoern Peters
----- "Alan Ruttenberg" <alanrut...@gmail.com> wrote:
> On Tue, Mar 9, 2010 at 6:10 PM, Jessica Turner <jtu...@mrn.org>
> wrote:
> > Objects can be located within other objects, right? So the original
> point,
> > that stuff can move through the lumen, doesn't determine whether
> it's more
> > like a region than an object?
>
> In the current relation ontology the object/object relation
> located_in
> is defined in terms of the region/region located_in, by first
> projecting to the region (using the object/region located_in
> relation), and then relying on the spatial sense of part_of. Doing so
> gets us the definition of located_in (object/object) in terms of
> something we seemed to (before this conversation) better understand -
> how space is part of space.
>
> At question is how the relations part_of, location_of, and
> contains_in
> are defined - how many must be primitively asserted in order to
> determine the others - how they are inter-related.
>
> -Alan
I believe you will get the same relational framework when substituting in 'site' for 'region'. There is a 'projection relation' which links a material entity to the site it occupies. If this site is contained in a larger site occupied by a second material entity, then the first material entity is contained in the second etc.
I would love to work jointly on formalizing this in detail.
- Bjoern
>
> >
> > --Jess
> > On Tue, Mar 9, 2010 at 4:02 PM, Barry Smith <phis...@buffalo.edu>
> >> --
> >> You received this message because you are subscribed to the Google
> Groups
> >> "BFO Discuss" group.
> >> To post to this group, send email to bfo-d...@googlegroups.com.
> >> To unsubscribe from this group, send email to
> >> bfo-discuss...@googlegroups.com.
> >> For more options, visit this group at
> >> http://groups.google.com/group/bfo-discuss?hl=en.
> >>
> >
> >
> >
> > --
> > Jessica Turner, Ph.D.
> > Research Scientist
> > MIND Research Network
> > 1101 Yale Blvd. NE
> > Albuquerque, New Mexico 87106
> > Ph: 505-272-1570
> > email: jtu...@mrn.org
> >
Pierre Grenon
below will make sense but here's a try
It's pretty hard to argue with doctrine. I'm not sure what prevents
you from saying "I don't find regions useful, but I'm happy to take
them as abstractions." From a practical point of view, at the
knowledge representation level, they might serve the purpose of
defining what you seem to be exclusively interested in, i.e. Sites,
which are not, by the way, just susbtantial versions of regions
(regions in the category of substance or 'object'). In passing, places
a la Aristotle are just a degenerate or limit case of sites. Also,
from a knowledge acquisition point of view, they allow for securing
the mapping of certain kind of quantitative locative data (in relation
to 'abstract coordinate systems' if you want).
BFO is in fact rather generic and flexible on purpose here and to some
extend the site/region distinction secures the possibility of adopting
either an absolutist or relational view. Also, BFO has tended to be
careful not to impose too much structure on space, although by
default, something along the lines mentioned by Barry has been
intended (regions don't change, don't move around and so on). More
importantly, however, BFO is, I'd say 'only slightly', biased in the
sense that regions are comparatively easier and sites comparatively
hard. The assumption regarding knowledge representation is that we can
offload a lot of spatial (and spatiotemporal) KR on regions and
contribute to a characterisation of sites, in so far as sites are
located in space (and in a mediated way in spacetime, on the side of
occurrents BFO has 4D environments). Other characteristics of sites,
such as their material bulk and the fact that they change parts, are
much more difficult to deal with and for them a reduction to region is
not enough.
> Here is a rewrite of the points above substituting in my thinking about sites instead.
>
> - inertial frames of reference are sites with certain properties
I'm not sure where this is going but it's a bit of a monodimensional
view on sites. Sites are susceptible of domain categorisation. Also,
sites are the buidling blocks for dealing with things such as
environments and surroundings. Sites are overlapping (materially, not
just spatially, even if possibly only at lower dimensions) with their
'host' and with their hosts' surroundings.
> - there is a difference between an object (which I take to be short for material entities) and the site it is located in
yes, that's fine, but it doesn't say much
> - objects move relative to other objects, which corresponds to movement within larger sites.
That, if it is true or insightful in any way, could be a theorem.
> - objects can change shape
> - sites move relative to sites
if sites are like objects you get that for free
> - sites can change shape
"and parts" is actually the most important bit overlooked in this discussion
> - The lumen of Phil's gut is an instance of a site which is part of site-Phil (the site in which the material entity Phil is located)
To my mind the issue with the catgeory of Site is that of its somewhat
relational character. But what this means, perhaps, is only that a
host can participate in defining a site for something at times and for
another at other times. There is however indeed a mereology of sites.
> and bounded by the interior of Phil's gut wall + some fiat boundaries at the ends. At different time points, different material entities have locations that are contained in the lumen of Phil's gut.
There's nothing that contradicts BFO's approach to sites and regions
here. And why do you use 'location' now?
> I appreciate this discussion. At this point, I would like to know if you still think there is strong argument for needing two separate 'spatial entities' (site and spatial region), or if we can start together trying to formalize in earnest the use of just one.
So, again, BFO separates the two. Mostly because sites play a role in
the realm of causation and so on which regions do not. Sites are bulky
and of course, as substantial, they have spatial properties. This is
no reason for dropping regions.
As to formalisation, BFO uses regions to express some of the theory of
sites. But other parts of the theory of sites require more than
regions (causal relations, energy exchange and so on).
Your argument against regions does not need to focus on sites. It is
not clear that some spatial relations can not be dealt with already at
the level of objects (distance, orientation and so on). So if you want
to get rid of region, your argument should focus on doing spatial
representation with objects, including sites as a special case. In
fact, not only objects are spatial entities, 'dependent continuants'
too are spatial entities in the broadest sense, which in BFO more or
less means 'being located in space'.
Leaving aside deep philosophical issues, there are, to my mind,
practical knowledge representation and acquisition motivations for
regions. Other motivations include accounting for empty space, which
arguably may not be immediately relevant to biology, and the
distinction between being a part of and being located in. For the
later, sites are useful and, in relation to this, regions allow for
teasing out the way sites support the distinction.
I'm probably not cautious enough here, but the practical bottom line
is that it's hard to see a motivation for eliminating a category on
the basis that, in the worst case, one only finds it useful for the
axiomatisation of another, rejecting any other possible use. If
anything this seems to be the road for an increase in complexity of
the representation.
cheers
pierre
Pierre Grenon
>
> ----- "Alan Ruttenberg" <alanrut...@gmail.com> wrote:
>
>> On Tue, Mar 9, 2010 at 6:10 PM, Jessica Turner <jtu...@mrn.org>
>> wrote:
>> > Objects can be located within other objects, right? So the original
>> point,
>> > that stuff can move through the lumen, doesn't determine whether
>> it's more
>> > like a region than an object?
>>
>> In the current relation ontology the object/object relation
>> located_in
>> is defined in terms of the region/region located_in, by first
>> projecting to the region (using the object/region located_in
>> relation), and then relying on the spatial sense of part_of. Doing so
>> gets us the definition of located_in (object/object) in terms of
>> something we seemed to (before this conversation) better understand -
>> how space is part of space.
>>
>> At question is how the relations part_of, location_of, and
>> contains_in
>> are defined - how many must be primitively asserted in order to
>> determine the others - how they are inter-related.
>>
>> -Alan
>
> I believe you will get the same relational framework when substituting in 'site' for 'region'. There is a 'projection relation' which links a material entity to the site it occupies. If this site is contained in a larger site occupied by a second material entity, then the first material entity is contained in the second etc.
>
> I would love to work jointly on formalizing this in detail.
BFO's axiomatisation here has been minimal and has consisted in laying
primitive or working primitives (that is vocabulary which is primitive
only as far as the axiomatisation goes) and some necessary conditions,
such as in relational to the locational properties of sites on the
basis of spatial regions. The occupies relation between susbstantial
and sites is thus hardly more than constrained on its domain and
range.
The theory of sites is in the main that of Barry's and Achile's niches
and holes.
p
Bjoern Peters
----- "Pierre Grenon" <pierre...@gmail.com> wrote:
> sorry to pitch in so late in the discussion, not sure how much of the
> below will make sense but here's a try
>
I was using a modified definition of 'site' developed in earlier emails in this rather long thread. Obviously that does not help to catch up on what my argument is, as your email indicates (which makes assumptions of what I mean that I did not intend). I will try to re-formulate and avoid the use of existing labels (site/region).
- Bjoern
> On Wed, Mar 10, 2010 at 1:32 AM, Bjoern Peters <bpe...@liai.org>
> wrote:
> > Unfortunately we can't agree on this. My main argument is that I
> don't recognize why we need two types of 'spatial entities' (site and
> spatial region). When you say 'region' in the above, I assume you want
> to maintain that distinction. The point 'that regions do not move'
> entities have locations that are contained in the lumen of Phil's