The Semantic Approach and Its Application to Evolutionary Theory, Filozofia, Filozofia - Artykuły
[ Pobierz całość w formacie PDF ]
Philosophy of Science Association
Philosophy of Science Association
The Semantic Approach and Its Application to Evolutionary Theory
Author(s): Elisabeth A. Lloyd
Source: PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association,
Vol. 1988, Volume Two: Symposia and Invited Papers (1988), pp. 278-285
Published by: The University of Chicago Press on behalf of the Philosophy of Science
Association
Stable URL:
Accessed: 26/10/2008 14:43
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press
,
Philosophy of Science Association
,
The University of Chicago Press
,
Philosophy of Science Association
are collaborating with JSTOR to digitize, preserve and extend access to
PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association.
The Semantic
Approach
and its
application
to
Evolutionary Theory1
ElisabethA.
Lloyd
University
of
California-Berkeley
1. Elements of the SemanticView
is
simply
a methodof
formalizing
thecontentof scientifictheories.
2.
Application
of the SemanticView to
EvolutionaryTheory
In a seriesof articlesanda
book,
I have
analyzed
the structure
of moder
evolutionary
theoryusing
the semanticview as a framework
(Lloyd
1984, 1986a, 1986b, 1987a, 1987b,
1988,
forthcoming;
cf.
Thompson
1983, 1985,1988).
I shall
brieflyrecap
the
analyses
I
have
done,
in orderto demonstrate
the
range
of
problems
accessible to the semanticview.
First,
I have used the semanticview to
analyze
the structureof
populationgenetics
models,
including
kin and
group
selection models. Thereis a heateddebate
in
the
genet-
ics literatureaboutwhetherkin selection models shouldbe
interpreted
as
group
or as
organismic
selection
models. I have offered an
analysis
of this
problemutilizing
the
semanticview
(1988,
Ch.
5).
The semanticview is not
applicableexclusively
to mathematical
models,
however.
I
have also
analyzed
the basic structureof naturalselection models,
andthe interrelations
among
the
components
of these
models;
these models arecharacterizednon-mathemati-
cally
(see esp. 1988,
Ch.
6).
I
use this characterizationof the basic structureof selection
models to offer a
new definitionof a unitof selection in termsof its actualrole in models.
This
definition,
in
turn,
allows a
precise
formulationof severalcontroversial
problems
involving
units of selection.
For
instance,
I use
my
structuraldefinitionof a unitof selection to
comparespecies
selection models with otherhierarchical
selection models
(Eldredge
andGould
1972;
Gould and
Eldredge
1977;
Eldredge
andCracraft
1980;
Vrbaand
Eldredge
1984;
Vrba
1984;
VrbaandGould
1986).
I find thattherearecertain
discrepancies
between the struc-
tureof
species
selection models andotherselection
models,
and
I
suggest
a new formula-
tion of
species
selection models which is consistent
with the
general
structureof selection
models
(1988,
Ch. 6;
Lloyd
andGould
ms.).
PSA
1988,
Volume
2,
pp.
278-285
Copyright
? 1989
by
the
Philosophy
of Science Association
The semanticview
of
theory
structure,
as
developedby Suppes
(1957,1967),
Suppe
(1974,
1976, 1977, 1988),
andvanFraassen
(1970, 1972, 1980),
represents
theoriesas class-
es of modelsor structures.Thesemodels
are,
on the versionof the semantic
approach
used
here,
defined
by specifying
their
laws, parameters,
andvariables.
The semantic
approach
to
theory
structure
279
My analysis
of the structureof
populationgenetics
models is also useful in under-
standing
what is
wrong
with
genic
selectionism
(1988,
Ch.
7).
The semanticview allows
a
precise
formulationof
exactly
how and
why genic
selection models are boundto be
empiricallyinadequate,
underone
interpretation,
or
trivially
differentfrom other
models,
underan alternate
interpretation.
In
particular,
I demonstrate
why
the
attempted
resurrec-
tion of
genic
selectionism
by Sterelny
andKitcher
(1988)
misses the
point
of the debate.
Use of the semantic
approach
to
theory
structureallows
precise
formulationof various
different
questions
aboutunitsof selection.
Questions
aboutwhich entitiesare
functioning
as
replicators,
and
which as
interactors,
need to be
keptcompletely
distinct,
andthe seman-
tic view allows the
precise
translationof thisdistinctionin termsof laws andstate
spaces.
Finally,
I also use the semanticview of
theory
structureto
develop
a schemaof
theory
confirmationthatis more subtle and sensitive to
genuine
scientific concernsthantradi-
tional
approaches
in
philosophy
of science
(1988,
Ch.
8).
3. Is the SemanticView
doing any
Real Work?
Anyone
who wantsto
argue
thatthe semanticview has done
nothing
for
analyzing
bio-
logical
theoriesneeds to show eitherthatthe use of the semanticview is
doing
no realwork
in the
analyses
reviewed
above,
or thatthe
analyses
themselvesareineffective.I wouldcer-
tainly
not claim that
using
the semanticview
formally
is the
only way
to make
progress
on
issues in
philosophy
of
biology.
I do find it
suggestive,
however,
thatthe fine
analytic
work
of
Brandon,
for
instance,
is
completelycompatible
with,
andin
manyways, suggestive
of,
an informal
use
of the semanticview of
theory
structure
(e.g.,
Brandon
1981, 1982).
At
any
rate,
I have no attachmentto the
"manifest
destiny"
of
the semanticview. Workers
using
the
semanticview havecreatedcarefulreconstructionsof some
importantparts
of
evolutionary
theoryusing
this
approach,
andtheuse of this
analytic
framework
has,
I would
argue,
con-
tributedto the clarificationof a numberof issues in the
philosophy
of
biology.
I know of a numberof
people
who have dismissed the
utility
and value of the seman-
tic
view on accountof some basic
misconceptions.
I would like to take some time
today
to review what I see as the threemost
common
misconceptions
aboutthe semanticview.
4. Three Common
Misconceptions
aboutthe SemanticView
By
far the most common
complaint
is the
following:
"The semanticview
necessarily
involves anti-realism.I am a
realist,
thereforeI must
reject
the semanticview".
As a matterof
fact,
the semantic
view,
as a view
about
theorystructure,
is neutralon
the issue of scientific realism.Thereis
nothing
in the claim that
scientific theoriesare
usefully
and
clearly
reconstructedas classes of model
types
which
entails eitherrealism
or anti-realism.The
epistemic
attitudetowardsthese models and the entities
within them
is distinctfrom the
description
of the structureof the models themselves.
It is
especially surprising
that
people cling
to the belief thatthe semanticview of theo-
ries is
necessarily
anti-realist,
in
the face of RonaldGiere's visible and
long standing
advocacy
of both the semanticview of
theory
structureandrealism
(see
Giere
1988).
Fred
Suppe's quasi-realist
approach
is
yet
another
possible epistemic approach
associated
with the semanticview
(Suppe 1988).
I do take it thatBarbaraHoranbelieves thatthe semanticview
andanti-realismmust
go together.She,
at
least,
has
arguments
for this
view,
ratherthan
just assumingguilt by
association. I shall address
her
argument
later.
280
The second most common
myth
aboutthe semanticview is that"thesemanticview
can
only
work with mathematicaltheories".This was never
true,
as evidenced
by Suppe's
analysis (1974)
of
biological taxonomyearly on,
andalso
by my
work
on Darwin
(1983)
andon
species
selection, and
most
recently,by
JamesGriesemer'swork on
laboratory
andmuseummodels
(Griesemer
andWade
1988,
Griesemer
ms.).
The
supposed
restrictionto mathematicaltheoriesis no more truefor the semantic
view of theoriesthanit is for model
theory
in
general.
It is easier to see how the semantic
view would
represent
mathematical
models,
but this is in no
way
exclusive.
The thirdmost
popular
misunderstanding
of the semantic
view, according
to
my
infor-
mal
survey
is: "thereis no substantivedifferencebetween
the semanticview and the stan-
dard,
received view of
theory
structure."
First,
one
logical point.
If the received view is takento
require
the use of first order
logic only-which
is the
way
it is often conceived-then
any theoryinvolving
the real
numbersis not
representable
within
it,
butwould be within the semanticview. This for-
mal
point
does
not, however,
strikeme as
getting
to the heartof the matter.The sensible
response
is
simply
to lift the restrictionto firstorder
languages.
Then the two
approaches
would seem to be
equivalent.
But thisis not
quiteright,
becauseone of the
advantages
of the semanticview is thatit,
unlike
syntacticapproaches,
is notrestrictedor committedto a
particular
linguistic
formula-
linguistic
formulationsof a
theory.
But let us
step
aside from these formalissues for a moment.
Suppose
someone were to
say
thatthe
problems
with
committing
to a
particularlanguage
did not botherthem.Let
us
imagine
thatthe two
systems
of
description
of scientific theoriesare
formallyequiva-
lent,
that
they
contain
precisely
the same
information,
and are
completely
intertranslat-
able. Are therestill reasonsfor
preferring
the semanticview? That
depends
on what
you
want to do. If
you
want a view of
theory
structurethatcan be used to
analyze
the
empiri-
cal contentof theories,discuss theirinterrelations
with other
theories,
andexamine how
they
areused
by
scientists,
the semanticview has a clear
advantage.
First,
the semanticview is closer
to
the
practices
of
scientists,
and it
provides
a natural
andconvenient
way
of
reconstructing
theoriesandclaims aboutthose theories.
Second,
the semanticview does not
require
laws of nature-a
problematic
concept,
especially
in
evolutionarybiology-though
it also does not
preclude
the formulationand
use of laws of nature.
Third,
the semanticview has a betterchanceof
representing
scientists'
problems
in
termsaccessible to
them,
becauseit is closer to the formof scientists' own
reasoning.
Fourth,the semantic
view allows
either
a realistor anti-realist
(or quasi-realist)
inter-
pretation
of theories.
issues
in
theory
identification
and
theorychange.
On
the
syntactic
view,
a
change
in
linguistic
formulationmeansthatthere
is now a new
theory.
Onecan
imaginemany
cases in whichsucha
change
wouldbe
trivial,
yet
the
syntacticapproach
wouldstillcall the
entity
a new
theory.
On the semantic
view,
in
contrast,
two differentaxiomatic
systems
thatare
semanticallyequivalent-that
is, they
share
the sameset of models-constitute one
theory,
nottwo.
In
other
words,
the factis that
you
can
change
thenameswithout
changing
the
meaning
relations,
andunderthe semanticview
it is therelationswhich aretakento be essentialin the
description
of thetheoretical
systems.
Hence,
thereis an
important
differencebetweenthesemanticview andvarious
syntactic
approaches,
namely
thatthe semanticview avoidstheworriesabout
theorychange
anddefi-
nitionwhicharisefrom
havingparticular
tion.This
point
has
consequences
for somerather
important
281
Fifth,
the semanticview
provides
the frameworkfor a much more
subtle,faithful,
and
powerful analysis
of how datacan
support
or confirman
empirical
claim.
In
summary
so
far,
I claim that
rejections
of the semanticview of
theory
structure
often
rely
on mischaracterizationsof this view.
Furthermore,
I thinkthatBarbara
Horan's
objections
to the semanticview lie in this
camp,
andI am
especially sorry
about
that,
because I also thinkHoranraises some
important
issues abouttheoretical
explanations,
testing,
and the constructionof scientific theories.
5. Horan'sCriticismsof the SemanticView
to
generalize..."(1989,
p.
269.)
Horanclaims thatthese models are
unrealistic,
since
they
are so abstractandremoved
from the
biological
details and
complexity
of each case.
Therefore,according
to
Horan,
the use of models is
supposed
to resultin anti-realism.This is the basis of her
argument
thatthe semanticview is
necessarily
anti-realist.In her
words,
if we
accept
the semantic
view, "which makes theoreticalmodels an
integralpart
of our
conception
of
theories,
we
must
accept
anti-realismas
well".
(1989, p. 274.)
But the claim thatthe semanticview of
theoriesdemands
high-level
theoreticalmod-
els is
simply
false. Whata semanticview theorist
actuallysays
is this:
give
me a scientif-
ic
description
of a
system-any system-and
I
prefer
to
represent
that
system
in
termsof
its
meaning
structure.The semantic
approach
can be used at
any
level of scientific
theory,
includingextremely detailed,
low-level theories.
In
fact,
I take this
flexibility
of the semantic
view to be one of its virtues.
Furthermore,
in
my
discussion of
genic
selection,
I criticize the
reductionist
approachby
showing
thatthe
genic
level models do not
ordinarily
contain
enough
informationto
describe
accurately
the
systems they
areintendedto
explain
(1988,
Ch.
7).
In other
words,
the fact thatmore
complexity
anddetail are
required
in the model is
clearly
delin-
eated and
defended
using
the semanticview of
theories.
I
conclude thatthe
complexity
of
biological phenomenaemphasizedby
Horan
is no
problem
at all for the
semanticview. In
fact,
it is rather
good
for business. With
many,
many
mid- to low-level
descriptions
of
natural
systems
running
around,
we are
unlikely
to run out of work
anytime
soon.
the details of othercases aboutwhich one wants
I
see two main
problems
with Horan's
argument.
The first is thatshe does not distin-
guish
between
philosophers'
reconstructionsof scientific theoriesandthe scientific
theo-
ries themselves. A
philosophical preference
for
reconstructing
and
representing
theoriesa
particularway,
as classes of models
(semanticentities)
is identified
by
Horanwith
the
value of the
scientific
preference
for certain
abstract,
general
theories.But normative
claims made
by
scientists about
theory
constructionarenot the same as normativeclaims
made
by philosophers
about
theory
reconstruction.Withoutthis
identification,
further-
more,
Horanhas no case
against
the semanticview.
The basic
problem
is thatHoranidentifies
high-level
theoreticalmodels
in
population
biology
with the semanticmodels of the semanticview.
Hence,
she seems to think
that
(1)
the semanticview demandsthese
very high
level
models,
and
(2)
thatthe semantic
view can
fairly
be saddledwith
any problems
associatedwith these
high
level models.
Let us
considerthe first
claim,
in
which Horanidentifies the semanticview with the
demandfor
very high-level,
abstractmodels. Horan
writes,
"whentheoretical
models are
constructedone eliminates as much of the detail aboutthe case one wants to
explain
as
possible...instead,
the model builderselects a few variablesand uses them to make the
model a successful
predictor...ignoring
[ Pobierz całość w formacie PDF ]