The case of the probability of dysfunction

1
The case of the probability of dysfunction

• Françoise Longy (IHPST)

2
Plan

• The origin of the question
• The PHYS and ART proba of dysfunction
• Can ART proba be seen as chances ?
• Intuitively
• Theoretically
• Artefact kinds as non-physical(NP) real kinds
• Probabilities grounded in NP real kinds
• How to present the ART proba of dysfunction?

3
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction

4
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction

Organisms and artefacts
5
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction

6
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction
• Because a causal because

7
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction
• Because a causal because

The probability is grounded in some objective
feature of the world
8
Origin of the question Characterisation of
function

• X has function F only if
• X has an (objective) property O such that
• X is there because of O
• Because of O, X has proba q of doing F
• X has proba p (p1-q) to malfunction
• O ???

The probability is grounded in some objective
feature of the world
9
Which probability is pertinent relative to the
function/dysfunction distinction?

• The DYSF-PHYS-proba of X ?
• The probability that the physical structure
  instantiated by X might not do F
• No

X a coffee machine, a light-bulb, coming out
of the factory
10
Which probability is pertinent relative to the
function/dysfunction distinction?

• The DYSF-ART-proba of X ?
• The probability that X, which has been made in a
  bulb factory, might burn out after 5 minutes of
  normal use
• YES
• ? ? DYSF-ART and DYSF-PHYS probabilities
• DYSF-ART the probability that X might possess
  (inherited) a physical make-up incapable of doing
  F

DYSF-PHYS The probability that the physical
structure instantiated by X might not do F
11

• The PHYS-proba as chances ?
• No particular problem
• The ART-proba as chances ?
• ???

12
ART-Proba as chancesintuitively
X a bulb coming out from a production plant The
probability p that X will be defective

• Great importance of the quality of the production
• Two opposite factors cost/quality of production
• Market ? definite value of p
• The assembly-lines (machines, work procedures,
  control mechanisms, ) are continuously planned,
  verified and adjusted to maintain a stable p

13
ART-Proba as chancestheoretical arguments

• Two ways of considering X
• X as a physical object
• X as an item of artefact kind
• Physical properties/functional properties
• a?b multirealizability
• b is not definable in physical terms

14
ART-Proba as chancestheoretical arguments

• Two ways of considering X
• X as a physical object
• X as an item of artefact kind
• Physical properties/functional properties
• a?b multirealizability
• b is not definable in physical terms

b irreducible to a
15

• to have a function ? to have a capacity
• a historical property
• Organisms, kinds defined by
  historical relations functional kinds
• Artefact kinds and Biological species are
• historical real kinds (Millikan)
• NP (non physical) real kind

16
A NP real kind

• Physical real kind (paradigm water, gold ..)
• A common physical micro-structure explains
• High number of universally shared properties
• NP real kind (paradigm lion, Renault 4 ..)
• External mechanisms explain
• Historical continuity between members
• High number of partially shared properties

17
A NP real kind

• Physical real kind (paradigm water, gold ..)
• A common physical micro-structure explains
• High number of universally shared properties
• NP real kind (paradigm lion, Renault 4 ..)
• External mechanisms explain
• Historical continuity between members
• High number of partially shared properties

• Largely shared (Boyd, Millikan)
• Determinate proportion

18
A NP real kind

• Physical real kind (paradigm water, gold ..)
• A common physical micro-structure explains
• High number of universally shared properties
• NP real kind (paradigm lion, Renault 4 ..)
• External mechanisms explain
• Historical continuity between members
• High number of partially shared properties

• Largely shared (Boyd, Millikan)
• Determinate proportion

19
Types of real kinds and types of causes

• A physical real Kind gold (yellow, )
• A NP real Kind corks (red crests, )

20
Proximate and Ultimate causes
Why has this cork a red crest ?

• The physiological proximate cause
• Because it has such genetic code
• The evolutionary ultimate cause
• Because he has inherited a cork genetic code

Why his genetic code gave rise to a red crest Why
he inherited the capacity to grow a red crest
21
NP Real Kinds and probabilities resulting from
ultimate causes

• Not all probabilistic properties are concerned
• Only those resulting from mechanisms elected for
  regulating a particular ratio
• The probability of being a female(organism)
• Selected mechanisms regulate the male/female
  ratio
• The probability of dysfunction (artefact)
• The proportion of defective items in regulated by
  
• Factory mechanisms
• Firm mechanisms
• The Market mechanisms

22
What nature hasthe DYSF-ART probability ?
The probability that an artefact item is defective

• It is not a real frequency
• it does not result from statistical data
  concerning the actual population of Xs
• It corresponds to the ratio of defective items a
  series of mechanisms (at factory, firm or market
  level) tends to produce under normal conditions
• Frequency in the actual population may differ
  from the one grounded in the NP-real kind
  mechanisms

23
How to qualifythe DYSF-ART probability ?
The probability that an artefact item is defective

• It measures something like a propensity
• It measures an objective feature of the world
  which is not a frequency
• However, speaking of propensity is odd
• An item X could then have 2 ? propensities of
  doing F
• X could have the physical propensity p to do F,
  but the real kind propensity q (q?p) to do F !
• it is not a tendency rooted in what the item is
  all by itself

24
How to qualifythe DYSF-ART probability ?
The probability that an artefact item is defective

• It is rather a property at kind (population)
  level
• If a tendency is involved, it is a tendency
  rooted in the factory (firm or market) mechanisms
• However, calling it simply a frequency is
  misleading
• The objective fact it refers to is not an
  effective frequency in a real population
• It is the frequency that an actual series of
  mechanism, involved in maintaining a NP real
  kind's unity, tend to produce
• Should we then say something like
• "ideal frequency NP real kind mechanisms tend to
  produce"?