zhengj2007
commented
9 years ago From albertgo...@gmail.com on February 05, 2013 13:15:03
Regarding the proposal to expand the range of 'has material basis' to include qualities...
One of the original motivations for choosing 'has material basis' as the label for this relation was to highlight the fact that the range was bfo:material entity. This was an improvement on talk of a 'physical basis' for a disposition (since almost all entities can be described as 'physical').
I have mentioned several times that there should be a corresponding 'has qualitative basis' relation with range bfo:quality that would relate dispositions to the essential qualities of their bearers. Most recently, the need for 'has qualitative basis' came up on the ogms-discuss thread: https://groups.google.com/d/topic/ogms-discuss/7SYug0AWAcE/discussion I think for clarity and intuitive ease-of-use, it would be best to have two relations 'has material basis' (domain: disposition, range: material entity) and 'has qualitative basis (domain: disposition, range: quality).
cmungall
From cmung...@gmail.com on July 17, 2012 01:04:20
ref states:
ELUCIDATION: a has_material_basis b at t means: a is a disposition & b is a material entity & there is some c bearer_of a at t & b continuant_part_of c at t & c has_disposition d at t because b continuant_part_of c at t. [071-001]
EXAMPLES: the material basis of John’s disposition to cough is the viral infection in John’s upper respiratory tract; the material basis of the disposition to wear unevenly of John’s tires is the worn suspension of his car.
I think there might be a disconnect between what this elucidation is actually saying and what people want it to say. Consider the 2nd example:
"the material basis of the disposition to wear unevenly of John’s tires is the worn suspension of his car."
Let us call the suspension of John's car "s", and let us say it has the following 3 qualities or characteristics:
q1: being worn q2: being purchased from Barry's autoshop q3: being colored bright pink
Note that according the elucidation, the basis is "s", not any of the qualities of "s".
This means the example could equally well be written as:
"the material basis of the disposition to wear unevenly of John’s tires is the suspension that was purchased from Barry's autoshop."
"the material basis of the disposition to wear unevenly of John’s tires is the bright pink suspension."
These are logically the same but carry different connotations of where the blame in the worn tires lies. If the example is not to mislead, it should be re-written as:
"the material basis of the disposition to wear unevenly of John’s tires is the suspension."
This would then be a non-misleading example.
But note that this exposes the fundamental weakness of this relation. In fact, s has many qualities, and only a subset of these are in any way causally related to the disposition. Presumably the pinkness of the suspension (q3) does not have any causal role in the wear of the tires. And knowing Barry is not a dodgy mechanic we presume the causal basis is not in s's characteristic of being bought from Barry (q2). We have no way to "pin the blame" of the worn tires on q1. And AFAICT there is no way to compose this from existing relations - it has to be a primitive.
I propose that has_material_basis is generalized to include qualities of material emtities as the range of the relation, OR a new fundamental relation has_qualitative_basis is introduced.
Original issue: http://code.google.com/p/bfo/issues/detail?id=113