‘occupies spatial region’ definition and mereological change

[gregfowlerphd]

Describe the (possible) bug

It’s possible this issue is an artifact of the difficulty of dealing with time in OWL, but I think the definition for occupies spatial region implies that I needn’t occupy the spatial region (SR) jointly “taken up” by my head, torso, legs, arms, etc. at the present time. Allow me to explain.

The definition reads:

b occupies spatial region r =Def b is an independent continuant that is not a spatial region & r is a spatial region & there is some time t such that every continuant part of b occupies some continuant part of r at t and no continuant part of b occupies any spatial region that is not a continuant part of r at t

Let b be me and r be SR. Is there any time t such that every continuant part of me occupies some continuant part of SR at t? Well, by the definition of ‘continuant part of’--

b continuant part of c =Def b and c are continuants & there is some time t such that b and c exist at t & b continuant part of c at t

--a skin cell (SC) that sloughed off my arm yesterday continuant part of me, since there’s a time t (e.g., some time yesterday before the sloughing occurred) such that SC continuant part of me at t. However, SC does not occupy any continuant part of SR at the present moment, nor need there be any time at which SC occupies some continuant part of SR. Hence, there need be no time t such that every continuant part of me occupies some continuant part of SR at t, in which case the occupies spatial region definition has the consequence that I needn’t occupy SR, the spatial region jointly “taken up” by my head, torso, legs, arms, etc. at the present time.

Is this an intended implication of the definition? It’s certainly odd, but given that the continuant part of definition itself has the odd implication that SC continuant part of me (despite having sloughed off yesterday), I’m not sure that shows anything.

[alanruttenberg]

Al relations that have a temporal index in BFO FOL, have OWL counterparts in BFO-core that have an "at some time" interpretation. continuant-part-of and occupies-spatial-region are both such relations..

Does it help to remember the appropriate reading of the OWL relation is "at some time?

-a skin cell (SC) that sloughed off my arm yesterday continuant part of me at some time, since there’s a time t (e.g., some time yesterday before the sloughing occurred) such that SC continuant part of me at t.

This seems correct. Am I missing something?

However, SC does not occupy any continuant part of SR at the present moment, nor need there be any time at which SC occupies some continuant part of SR

"the present moment" is outside the expressivity of OWL. There needs to be some time at which SC occupies the SR, which there is. What OWL can't say, but the FOL can, is that the two existential times(when part, and when occupies) are in fact the same time.

It is confusing, IMO, that the OWL relation in BFO-core has the same name as the relation in BFO FOL, but mean different things, but that's the choice that was made. I recommend using the fully temporalized version in which there are both a some time and an all time relation for each time-indexed relation in BFO FOL.

[gregfowlerphd]

@alanruttenberg: Rather than respond to all the points you made--which I mostly agree with--I'll focus on the point of contention, and we can loop back to the others as necessary. You say:

There needs to be some time at which SC occupies the SR, which there is.

The first part is correct: Given the definition of 'occupies spatial region', for me to occupy spatial region SR, there needs to be some time at which SC occupies some continuant part of SR. However, the second part needn't be correct.

Remember that SR is the spatial region jointly “taken up” by my head, torso, legs, arms, etc. at the present time. SC needn't have occupied any continuant part of SR at any of the times that SC was continuant part of me, since I might not have had SR as my spatial location at any of those times. And SC clearly needn't have occupied any continuant part of SR at some time that SC wasn't continuant part of me. Hence, there need be no time at which SC occupies some continuant part of SR.

Therefore, given the first part of what you said--the part I agreed with--the definition of 'occupies spatial region' has the consequence that I needn't occupy SR. (These considerations also seem to show, by the way, that as currently defined, occupies spatial region is not equivalent to occupies spatial region at some time.)

[alanruttenberg]

I'm trying to grok this. I believe that the definition might have problems. I can't remember whether I or Barry wrote those, but I suspect Barry. They haven't got a ton of review, so I'll look more closely at that.

But before that, I am having trouble with

Remember that SR is the spatial region jointly “taken up” by my head, torso, legs, arms, etc. at the present time

What is this "present time" you are talking about. There's no such thing in either the OWL or the FOL.

[gregfowlerphd]

I'm trying to grok this.

Not sure if this’ll help, but here’s a formalized statement of the argument:

1. Given the definition (of ‘occupies spatial region’), for all x and y, x occupies spatial region y only if there is some time t such that every continuant part of x occupies some continuant part of y at t.
2. Therefore, given the definition, I occupy spatial region SR only if there is some time t such that every continuant part of me occupies some continuant part of SR at t. [From (1)]
3. SC continuant part of me. [From the definition of ‘continuant part of’ and the fact that SC continuant part of me at some time]
4. Therefore, given the definition, I occupy spatial region SR only if there is some time t such that SC occupies some continuant part of SR at t. [From (2) and (3)]
5. There needn’t be any time t such that SC occupies some continuant part of SR at t (since I needn’t have had SR as my spatial location when SC was continuant part of me and SC clearly needn’t have occupied some continuant part of SR at any other time).
6. Therefore, given the definition, I needn’t occupy spatial region SR (despite that SR is the spatial region I most paradigmatically occupy--the one currently jointly “taken up” by my head, torso, legs, arms, etc.).

What is this "present time" you are talking about. There's no such thing in either the OWL or the FOL.

This isn’t a surprise though, right? The present time is an individual (an instance of a class), and BFO asserts no individuals, so one wouldn’t expect it to be in the OWL or the FOL. I take it, however, that definitions are still supposed to apply to individuals.

Along similar lines, my understanding is that we’re modeling entities out in the world; indeed, I believe that’s something you’ve noted on more than one occasion in working group meetings. Surely we can agree that there is, out in the world, a spatial region jointly “taken up” by my head, torso, legs, arms, etc. at the present time (or currently/right now/however you want to put it)? What I want to know is what the definition of ‘occupies spatial region’ implies about whether I occupy spatial region that spatial region (and I’ve argued it implies that I needn’t occupy it).

I suppose one might have concerns about talk of the present time involving the lack of absolute simultaneity in relativity theory. Two thoughts here.

First, even someone who rejects absolute simultaneity for such reasons needs an account of how our talk of what’s going on at the present time (or now/currently) connects up to the world, since it appears there’s some connection, unlike in the case of, say, phlogiston talk. For instance, such an account might hold such talk is indeterminate in its reference between the members of a class of “slices” of spacetime, which each slice determined with respect to a different frame of reference. Given this account, we can simply arbitrarily select one of the slices indeterminately referred to by my talk of the present time and replace SR in the argument with the spatial region jointly “taken up” by my head, torso, legs, arms, etc. with respect to that slice.

Second, talk of the present time is inessential to the argument. Simply take the spatial region jointly “taken up” by my head, torso, legs, arms, etc. at some time, consider some skin cell that sloughed off my arm 24 hours before that time, and run the argument with that spatial region and that skin cell in place of SR and SC.

I believe that the definition might have problems.

I think I concur. Looking at the definition, I believe the problem stems from its apparent use of the defined “at some time” object property continuant part of rather than the triadic parthood relation in terms of which that object property is defined. Consider the following alternative definition, which uses the latter:

b occupies spatial region r =Def b is an independent continuant that is not a spatial region & r is a spatial region & there is some time t such that every continuant part of b at t occupies some continuant part of r at t and no continuant part of b at t occupies any spatial region that is not a continuant part of r at t

Unlike the original definition, this one doesn’t imply that in order for me to occupy spatial region SR, there must be some time t such that SC occupies some continuant part of SR at t, since although SC continuant part of me, SC needn’t be continuant part of me at the relevant time. In other words, if we replace the original definition with the alternative and modify premises (1) and (2) of the argument accordingly, the inference from (2) and (3) to (4) fails.

[alanruttenberg]

I believe that the definition might have problems.

I think I concur. Looking at the definition, I believe the problem stems from its apparent use of the defined “at some time” object property continuant part of rather than the triadic parthood relation in terms of which that object property is defined. Consider the following alternative definition, which uses the latter:

b occupies spatial region r =Def b is an independent continuant that is not a spatial region & r is a spatial region & there is some time t such that every continuant part of b at t occupies some continuant part of r at t and no continuant part of b at t occupies any spatial region that is not a continuant part of r at t

Unlike the original definition, this one doesn’t imply that in order for me to occupy spatial region SR, there must be some time t such that SC occupies some continuant part of SR at t, since although SC continuant part of me, SC needn’t be continuant part of me at the relevant time. In other words, if we replace the original definition with the alternative and modify premises (1) and (2) of the argument accordingly, the inference from (2) and (3) to (4) fails.

And indeed it does fail. I've already written to Barry that the definitions that mention instances of object, fiat object part, and object aggregate must have a time qualification. In fact all mentions mentioning an instance of a class should qualify temporally qualify the instantiation. The reason the definitions get away with not is that all the classes in BFO other than object, fiat object part, and object aggregate are rigid.