Dispositions and probabilities
Chris Mungall
I've always been a bit troubled by dispositions. What are the identity
criteria? Are the following equivalent or distinct?
1. The disposition of this glass to break in two when dropped 5ft onto
this concrete floor
2. The disposition of this glass to break in two when dropped 5.2ft
onto this concrete floor
3. The disposition of this glass to break in two when dropped 5ft
onto that other concrete floor
4. The disposition of this glass to break into three pieces when
dropped 5ft onto this concrete floor
etc
Also, when does a disposition come into being? Is every object in
front of my eyes perpetually surrounded by an infinite cloud of
dispositions?
I have found that I have an easier time of this if I cease thinking of
dispositions as realist instantiated entities and switch to thinking
of dispositions as abstract reifications of conditional statements.
There's possibly some connection to modal logic here, but what is more
interesting is the connection to probability theory, especially well-
axiomatized probability theory e.g. bayes theory, graphical models,
etc. These are tremendously powerful and useful in bioinformatics (and
I imagine in the sciences in general).
Here is a rough sketch. With an expressive enough logic such as IKL we
can refer to propositions using the 'that' operator
event1 = that( (dropped(g1, 5ft, t1) & hits(g1, floor1, t2) &
precedes(t1,t2)& precedes(t2,t3) -> broken(g1, t3) )
Even with less expressive languages there are reification techniques.
We can now do really useful things, such as talk about p(event1)
We can also break it down:
pre1 = that( dropped(g1, 5ft, t1) & hits(g1, floor1, t2) &
precedes(t1,t2)& precedes(t2,t3) )
post1 = that( broken(g1, t3) )
so we can talk about p( post1 | pre1) - giving us our link to bayes
theorem. So by equating bfo dispositions with either propositions or
causal pairs of propositions we gain some powerful axiomatisations, we
solve the identity criteria question and so on.
This seems fairly obvious, so I'm guessing someone has thought about
this more deeply and either documented it, or can explain why it won't
work.
One problem is that OGMS states that diseases are dispositions. It
would seem odd to say that 'diseases' are conditional propositions,
and there are probably some negative implications of pursuing this
path. This could be fixed by having OGMS equate the 'disease' with the
(well thought out) OGMS:disorder. This would reduce the tri-fold
inflation of entities without any loss of expressive power, and I
think this would be to OGMS' benefit.
Cheers
Chris
Barry Smith
>I've always been a bit troubled by dispositions. What are the identity
>criteria?
What, one may similarly ask, are the identity criteria for organisms,
specifically for people? Does the fact that we cannot state such
criteria imply that people do not exist?
>Are the following equivalent or distinct?
>
>1. The disposition of this glass to break in two when dropped 5ft onto
>this concrete floor
>2. The disposition of this glass to break in two when dropped 5.2ft
>onto this concrete floor
>3. The disposition of this glass to break in two when dropped 5ft
>onto that other concrete floor
>4. The disposition of this glass to break into three pieces when
>dropped 5ft onto this concrete floor
>
>etc
>
>Also, when does a disposition come into being? Is every object in
>front of my eyes perpetually surrounded by an infinite cloud of
>dispositions?
Some of the most promising, and influential, philosophical treatments
of causality these days appeal to dispositions (aka 'powers'); some
of these do indeed accept an infinite cloud of dispositions
associated with every object; some of them even identify objects with
infinite clouds of dispositions. Fortunately we can avoid both of
these outcomes, and remain neutral as to the numbers and types of the
dispositions associated with objects of given types, by following our
usual policy that we will accept names for those types of
dispositions into our ontology only if such types are recognized by
scientists and documented in their data. This also resolves the 5 ft
glass dropping disposition problem raised by Chris above. Influenza,
I believe (with OGMS), as a disposition (type). 'Influenza' is also a
term used by scientists.
I agree with Chris that we need to explore how to understand
dispositions in relation to probability assertions.
However, he himself points out that there is the following objection
to his proposal to regard dispositions as abstract reifications of
conditional statements:
>One problem is that OGMS states that diseases are dispositions. It
>would seem odd to say that 'diseases' are conditional propositions,
>and there are probably some negative implications of pursuing this
>path. This could be fixed by having OGMS equate the 'disease' with the
>(well thought out) OGMS:disorder. This would reduce the tri-fold
>inflation of entities without any loss of expressive power, and I
>think this would be to OGMS' benefit.
I agree that it is tempting to identify disease with disorder.
However, there would still be other dispositions biology would need
(of bananas, to ripen, for instance). Moreover, for this particular
identification to work we would need to have a good way of securing a
one-one correspondence between disease and disorder. Consider, now,
John, who has two diseases (say cancer, pulmonary fibrosis -- the
details do not matter), affecting one and the same portion P of one
and the same lung. The disorders differ at, say, the molecular and
cellular levels. But I do not see a good way, in keeping with our
current view of disorder, of doing justice to this difference, since
in both cases the relevant parts would be absorbed within the larger
whole P, and the differences would disappear. We could of course say
that the disorders in question here, too, are abstract reifications
of statements and thereby secure the difference. But now it seems
that we are progressively abandoning the whole ontological approach,
and substituting instead some sort of fancy model-building. Moreover,
to identify diseases (or disorders) with abstract reifications of
statements would, on the face of it, seem to imply that there were no
diseases before there were statements. That reifications of
statements can cause vomiting, can be cured, etc.
BS
Ludger Jansen
"I've always been a bit troubled by dispositions. What are the identity
criteria? Are the following equivalent or distinct?
1. The disposition of this glass to break in two when dropped 5ft onto
this concrete floor
2. The disposition of this glass to break in two when dropped 5.2ft
onto this concrete floor
3. The disposition of this glass to break in two when dropped 5ft
onto that other concrete floor
4. The disposition of this glass to break into three pieces when
dropped 5ft onto this concrete floor"
If we talk about types of dispositions, the answer seems to be clear: A
thing can have a type-2-disposition without having a type-1-dispositon. But
such types are rather defined classes but not scientifically approved
universals. What, e.g., should "this" and "that" (in, e.g., [1] and [3])
imply causally?
Analogous problems arise with sizes:
i. Being 5 inch long
ii. Being more than 4 inch long
iii. Being between 4 and 6 inch long
iv. Being less than 10 inch long
etc.
The same size or different sizes? One and the same size token can make any
predation of such size predicates true, but the predicates have different
extensions. Nevertheless, the things we see with our eyes do have sizes.
"Also, when does a disposition come into being?"
This is an empirical question. No general answer applying to all
dispositions is available.
"Is every object in front of my eyes perpetually surrounded by an infinite
cloud of
dispositions?"
No, not "surrounded": Dispositions inhere in their bearers like other
properties. Despite of i.-iv. above, you wouldn't say that things are
"surrounded by an infinite cloud" of properties like sizes, would you?
"I have found that I have an easier time of this if I cease thinking of
dispositions as realist instantiated entities and switch to thinking
of dispositions as abstract reifications of conditional statements."
Sure, their is an intimate connection between contrafactual (!) conditional
sentences and dispositions. But mind the direction of fit: The conditionals
are true because of the dispositions, not the other way round.
"There's possibly some connection to modal logic here, but what is more
interesting is the connection to probability theory, especially well-
axiomatized probability theory e.g. bayes theory, graphical models,
etc. These are tremendously powerful and useful in bioinformatics (and
I imagine in the sciences in general)."
This is all true. But probability theory cannot do the whole job. We have,
e.g., to distinguish
a. the probability that an instance of a certain type has a certain
disposition
b. the probability with which a disposition is realized
If you can only speak about the probability of the realization period
("p(event1)"), you cannot draw this distinction.
The following papers might be of interest in this context:
"Tendencies and other Realizables in Medical Information Sciences", in: The
Monist 90/4 (2007) 534-555.
http://home.arcor.de/metaphysicus/Texte/Tendencies-preprint.pdf
"Molecular Interactions. On the Ambiguity of Ordinary Statements in
Biomedical Literature" (togehter with Stefan Schulz), in: Applied Ontology 4
(2009) 21-34.
http://dx.doi.org/10.3233/AO-2009-0061
http://home.arcor.de/metaphysicus/Texte/Schulz-Jansen_Interactions.pdf
Best,
Ludger
Dr. Ludger Jansen
Institut f�r Philosophie
Universit�t Rostock
18051 Rostock
http://www.iph.uni-rostock.de/Vita-Jansen.22.0.html
Zentrum f�r Logik, Wissenschaftstheorie und Wissenschaftsgeschichte
(ZLWWG)
http://www.zlwwg-rostock.de
Cristian Cocos
> current view of disorder, of doing justice to this difference, since
> in both cases the relevant parts would be absorbed within the larger
> whole P, and the differences would disappear.
Should we understand, then, that dispositions owe their existence to a
simple granularity restriction, and that we could dispense with them as long
as we are allowed to dig deeper? Are they, in other words, just an artifact
of our choosing to stick with a certain level of granularity?
C
Albert Goldfain
I think there is good reason to believe that we will always find
dispositions even (and especially) if we are allowed to dig deeper.
Alexander Bird gives a very convincing argument of this in Nature's
Metaphysics:
http://books.google.com/books?id=_y6ru_-g5TQC
and I have seen nothing from modern physics indicating that
dispositions ever go away at any scale. Moreover, since they are so
tightly linked with theories of causality, I'm not sure why we would
ever want to dispense with them.
-AG
Wacek Kusnierczyk
> [...] such types are rather defined classes but not scientifically
> approved universals.
what are scientifically approved universals?
vQ
Ludger Jansen
>> approved universals.
>
>what are scientifically approved universals?
>
>vQ
universals refered to in scientific theories
LJ
Matthew Pocock
Matthew
--
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.