Saving the Intuitions: Polylithic Reference Ioannis Votsis, 100408

1
Saving the IntuitionsPolylithic Reference
Ioannis Votsis, 10/04/08
2
Plan
(1) Introduction
(2) Descriptive, Causal and Hybrid Theories
(3) Conflicting Intuitions
(4) Polylithic Reference
(5) Reference and Truth
3
Preamble

• A realist view of science seems to require that
• - Explanatory and empirical success (of a given
  grade) is a good indicator of the approximate
  truth of a theory or of certain theory-parts.

- Approximately true theories/theory-parts
describe the unobservable world, its entities and
processes.
- Successor theories (with success of a given
grade) are closer to the truth than their
predecessors and preserve (at least in some limit
form) their predecessors approximately true
parts (if they have any) and hence any related
ontology.
- Referential stability or continuity through
theory change ensues after the rise of a theory
with approximately true parts.
- A theory of scientific term reference provides
not only the mechanics of reference but
explicates this referential stability.
4
Theories of Reference

• Aim
• I will explore a largely ignored corner of the
  space of possible relations between scientific
  realism and theories of reference.

• Three central questions
• (A) How is reference fixed?
• (B) How does reference and truth relate?
• (C) How does reference and meaning relate?

• I will focus on questions A and B. Although
  question C is related and important, I will
  largely ignore it for the purposes of this paper.

• I will also mainly focus on scientific term
  reference (e.g. observational, theoretical and
  natural kind terms). What I will say here need
  not apply to an analysis of everyday language
  term reference.

5
Descriptive Theories
Main idea Reference is fixed by virtue of a
terms associated descriptions.
Sample A term t occurring in scientific theory T
refers to an entity a if and only if a satisfies
the descriptive claims T associates with t.
Some intuitions satisfied - I can successfully
talk about a thing only when all my ideas about
it hold. Otherwise, Im not really talking about
that thing.
- if you and I don't share a consistent set of
ideas about some thing we could not be talking
about exactly the same thing!!!
- to attempt to really talk about some thing we
at least need to attempt to somehow describe it.
Notable Proponents Frege, Russell and Searle.
6
Problems
(a) Too strict None of our theories are exactly
right, so no scientific term refers. Moreover,
two theories that assign conflicting claims to a
term t cannot be referring to the same object.
Reply Modified descriptivism, e.g. a term t
refers to an object a iff a satisfies a certain
special subset of the descriptions associated
with t.
(b) Unnecessary We seem perfectly capable of
referring to things about which we have little to
no knowledge.
Reply Some descriptive knowledge is required,
otherwise reference becomes too easy.
(c) Ntity-lity How many and/or which
descriptions are enough? Otherwise put, how do we
define the special subset?
Replies majority, mature science, central
concepts, essential properties, etc.
7
Causal Theories
Main idea Reference is fixed by virtue of a
terms causal history.
Sample A theoretical term t refers to an entity
a if and only if either t is used to originally
dub a (typically via ostension) xor t was used to
originally dub a and any speakers subsequent use
of t is causally linked back to the original
dubbing.
Qualification The dubbing event may include
descriptions and may even be based solely on
descriptions (e.g. Neptune).
Some intuitions satisfied - I can successfully
talk about a thing even when all my ideas about
it are incorrect.
- you and I may talk about the same thing even if
we have conflicting ideas about it.
- to talk about some thing typically involves
causal contact with the world.
Notable Proponents Kripke, Putnam and Devitt.
8
Problems
(a) Inescapability In scientific contexts, at
the very least the following description is
always assumed to be satisfied The entity that
caused our current use of the term _____ .
Reply Not the same idea of descriptive content
envisioned by descriptivists.
(b) Too liberal Any old term ends up referring
to modern-day entities.
Reply Further conditions apply, i.e. the
principle of the Benefit of Doubt.
(c) Ntity-lity How many and/or which
descriptions are enough? Otherwise put, when are
we justifiably charitable?
Replies majority, mature science, central
concepts, essential properties, etc.
9
Causal-Descriptivist Theories
Main idea Reference is fixed by a combined
causal-descriptive strategy.
Notable Proponents Evans, Lewis, Kroon and
Psillos.

• Psillos view A term t refers to an entity x if
  and only if x satisfies the core causal
  description associated with t (1999, p. 296).

Q What is the core causal description?
A It is the description of the
kind-constitutive properties by virtue of which
it i.e. a posited entity plays its causal role
vis-à-vis the relevant set of phenomena (p.
295).
Q What are kind-constitutive properties?
A They are those whose presence in an item
makes that item belong to a kind (p. 288).
10
Psillos View (1)
Referential Continuity Two terms t and t
denote the same entity if and only if (a) their
putative referents play the same causal role with
respect to a network of phenomena and (b) the
core causal description of t takes up the
kind-constitutive properties of the core causal
description association with t (p. 296).
Some intuitions satisfied - some sort of causal
contact with the world is maintained.
- I can successfully talk about a thing only when
my central ideas about it hold.
- we can talk about the same thing over time so
long as our past central ideas about it still
hold.
Example Ether refers to the electromagnetic
field.
11
Psillos View (2)

• Problem 1 Psillos two definitions of reference
  dont always match.

Example Suppose that James Ladyman has, among
other things, two kind-constitutive properties,
creature with a heart of gold and creature
with a fist of steel. The core causal
description of term t contains only creature
with heart of gold' and the core causal
description of term t contains only creature
with a fist of steel.

• Problem 2 Psillos correctly notes But, an
  objector may ask, how (and when) is the core
  description to be singled out? (p. 297). The
  ntity-lity problem strikes again!!!

Psillos reply These issues are serious enough.
However, they are not intractable For instance,
some descriptions associated with a term are less
fundamental in view of the fact that the posited
entity would play its intended causal role even
if they were not true (ibid.).
12
Psillos View (3)

• Problem 3 Though a reasonable answer, it still
  doesnt give us a non-ad hoc and unambiguous way
  to determine the fundamental from the
  non-fundamental descriptions.

Example We might contest Psillos account of the
ethers core causal description, arguing that the
following fundamental properties were not taken
up by the mature electromagnetic fields core
causal description, e.g. that the ether is a
material and mechanical medium capable of
containing energy as it is transmitted from one
body to another, that ether molecules oscillate,
that transverse waves require a solid medium,
etc. (Stanford 2003, Holton and Brush 2001).

• Desideratum An adequate account of reference
  should be able to provide a non-ad hoc and
  unambiguous answers to questions of referential
  failure or success (including referential
  continuity) .

Note I am not arguing here that Psillos theory
is a lost cause. Im simply pointing out some of
its difficulties.
13
Saving the Intuitions

• Some of the aforementioned intuitions are
  conflicting, e.g. whether we can successfully
  talk about a thing when all of our non-trivial
  ideas about it are false, whether a single
  contradictory idea entails that we cannot be
  talking about the same thing, whether causal
  contact is necessary, sufficient, both or
  neither, etc.

Dilemma EITHER reject the conflicting
intuitions as somehow faulty and on that basis
choose a theory of reference that satisfies the
remaining intuitions this is the option we were
just looking at.
OR save all the intuitions and find some way to
reconcile them.

• I will explore the latter option by supposing
  that the concept of reference is not monolithic.

NB That reference is not monolithic can still be
explored under the auspices of the former option.
14
Polylithic Reference

• Motivation Naturally to save our varied and
  contrary intuitions.

• Qualifications
• I will not here argue that this is the only or
  even the best way to understand the notion of
  reference.

(2) I will not here present an exhaustive list of
notions of reference or even tell you how many
are required.
(3) I will present an outline of three candidate
notions each satisfying different standards and
different sets of intuitions. The three notions
may require modification, replacement or even
deletion.
(4) Intricacies involving a speakers intentions
have been avoided, though they may be integrated
to further enrich the resulting formulations.
Note Kitchers (1993) token-type account of
reference comes closest to conveying the same
idea.
15
ReferenceICS (1)

• Ref(ics) A term t employed during period x
  refers(ics) to an entity a iff t was used during
  period x to identify the causal source of a
  certain set of phenomena associated with a,
  EITHER via some description OR via actual contact
  with the source.

Motivation Saves our intuitions about not
neglecting causal contact and about our ability
to talking about things despite false (or wildly
diverging) associated ideas.
Qualification A term that refers(ics) might also
refer in some other way. Also, even though it is
a very easy notion of reference to satisfy, it is
not trivially satisfied.
Note This kind of reference establishes a very
thin referential continuity thats not sufficient
to tackle historical objections.
16
ReferenceICS (2)

• Referential success example Aristotles tends
  towards its natural place refers(ics) to gravity
  as spacetime curvature.

- Aristotles notion was postulated to explain
among other things phenomena relating to falling
objects. Since such phenomena are associated with
our modern understanding of gravity, the term
tending towards its natural place refers(ics)
to gravity as spacetime curvature.
Note Friedman (2001) offers an even more radical
account of the continuity between Aristotelian
and classical physics.
in the sublunary region close to the surface
of the earth, where the earth is by far the
principal gravitating body, heavy bodies do
follow straight paths directed towards the centre
of the earth, again to an extremely good
approximation (p.379).
17
ReferenceICS (3)

• Referential failure example The term celestial
  sphere does not refer(ics) to gluons.

• The celestial spheres were postulated to explain
  celestial phenomena like the daily motion of the
  fixed stars but not the kind of phenomena we
  associate with gluons, namely that it mediates
  strong interactions between quarks, binds protons
  and neutrons in atomic nuclei, etc.

18
ReferenceMND (1)

• Ref(mnd) A term t employed during period x
  refers(mnd) to an entity a iff t was used during
  period x to identify the causal source of a
  certain set of phenomena associated with a AND a
  satisfies some of the non-trivial theoretical
  claims the relevant scientific theory associates
  with t.

Motivation Saves our intuitions about the
importance of identifying a causal source and
about successfully talking about a thing when at
least one of our ideas about it holds.

• RefCont(mnd) Term t employed during period x
  and term t employed during period y are
  referentially continuous(mnd) with respect to an
  entity a iff t' and t refer(mnd) to a AND t'
  inherits (some/all) of the satisfied descriptive
  claims of t".

Note This last definition does not aim to
re-describe the conditions of reference and hence
avoids the mismatch error.
19
ReferenceMND (2)

• Note1 This notion is a bit harder to satisfy. It
  establishes a thicker kind of referential
  continuity capable of tackling some historical
  objections, e.g. part of the old theory about the
  object must be (i) correct and (ii) preserved in
  the new theory.

• Note2 No ntity-lity problem!!! One description
  is enough to establish ref(mnd) and referential
  continuity(mnd) between two theories.

Example of referentialMND and referential
continuityMND success Early 19th century's
'caloric' refersMND and is referentially
continuousMND to modern thermodynamical accounts
of heat because all the correct descriptive
content associated with the caloric is preserved
in the modern account.
Example of referential continuityMND failure
The term humour is not referentially
continuousMND with the term leukocytes.
20
ReferenceMXD

• Ref(mxd) A term t employed during period x
  refers(mxd) to an entity a iff t was used during
  period x to identify the causal source of a
  certain set of phenomena associated with a AND a
  satisfies all (and only) the descriptive claims
  the scientific theory associates with t.

Motivation Saves our intuitions about the
importance of identifying a causal source and of
reference that gets all and only the scientific
theorys descriptions right.
Note This notion is very hard to satisfy. A
referential continuity notion based on
reference(mxd) would not be of help to problems
of conceptual change.
21
Ad hoc / Post hoc Problems

• A problem facing any theory of reference that
  incorporates clauses about correct descriptions
  is the following

No prospectively applicable criteria One can
only identify the correct descriptions by virtue
of the fact that these have been preserved in
subsequent accounts of the object. Hence the
condition that new theories preserve the correct
descriptions of the old ones is trivially
satisfied. (Stanford, 2006, pp.168-183).

• Replies
• (1) We can run partly-independent checks on the
  descriptions.

(2) We can also run a historical check A set of
descriptions preserved in the transition between
theories T1 and T2 typically must also be
preserved in the transition between T2 and T3.
22
Does Realism require Reference?

• Laudan A necessary condition for a theory
  being close to the truth is that its central
  explanatory terms must genuinely refer (1981,
  p.33).

• Hardin and Rosenberg (1982) deny the view that
  reference is a necessary condition for
  approximate truth.

Example there is nothing like a gene in the
sense envisioned by Mendel (p. 607).

• Cruse and Papineau (2002) The cognitive content
  of a scientific theory is captured by its Ramsey
  sentence and the latter makes reference
  irrelevant.

Problem To the extent that the Ramsey sentence
of a theory or an unramsified theory are approx.
true, they are surely telling us something about
unobservable entities and hence some sort of
reference to these entities is unavoidable.
23
In Memoriam
Delft, June 2007