Innateness is Canalization: In Defense of a Developmental Account of 
Innateness
Andre Ariew
University of Rhode Island



0. Introduction 
Lorenz proposed in his (1935) articulation of a theory of behavioral instincts 
that the objective of ethology is to distinguish behaviors that are "innate" 
from behaviors that are "learned"  (or "acquired").  Lorenz's motive was to 
open the investigation of certain "adaptive"  behaviors to evolutionary 
theorizing.  Accordingly, since innate behaviors are "genetic", they are open 
to such investigation.  By Lorenz's light an innate/acquired or learned 
dichotomy rested on a familiar Darwinian distinction between genes and 
environments.  Ever since Lorenz, ascriptions of innateness have become 
widespread in the cognitive, behavioral, and biological sciences.  The trend 
continues despite decades of strong arguments  that show, in particular, the 
dichotomy that Lorenz invoked in his theory of behavioral instincts is 
literally false: no biological trait is the product of genes alone.  Some 
critics  suggest that the failure of Lorenz's account shows that innateness is 
not well-defined in biology and the practice of ascribing innateness to 
various biological traits should be dropped from respectable science.  
	Elsewhere (Ariew 1996) I argued that despite the arguments of critics,  
there really is a biological phenomenon underlying the concept of innateness.  
On my view, innateness is best understood in terms of C.H. Waddington's 
concept of "canalization", i.e. the degree to which a trait is innate is the 
degree to which its developmental outcome is canalized.  The degree to which a 
developmental outcome is canalized is the degree to which the developmental 
process is bound to produce a particular endstate despite environmental 
fluctuations both in the development's initial state and during the course of 
development. The canalization account differs in many ways to the traditional 
ways that ethologists such as Konrad Lorenz originally understood the concept 
of innateness.  Most importantly, on the canalization account the distinction 
between innate and acquired is not a dichotomy, as Konrad Lorenz had it, but 
rather a matter of degree difference that lies along a spectrum with highly 
canalized development outcomes on the one end and highly environmentally 
sensitive development outcomes on the other end.  Nevertheless, I justified 
the canalization account on the basis of a set of desiderata or criteria that 
I suggested falls-out of what seemed uncontroversial about Lorenz's account of 
innateness (briefly): innateness is a property of a developing individual, 
innateness denotes environmental stability, and innate-ascriptions are useful 
in certain natural selection explanations (more below).  From that same set of 
desiderata I argued (in my 1996) that neither the concept of heritability nor 
of norms of reactionsÑtwo concepts from population geneticsÑsuffice to ground 
innateness.  
	In this essay, I wish to provide further support of the canalization 
account in two ways.  First, I wish to better motivate the desiderata by 
revisiting a debate between Konrad Lorenz and Daniel Lehrman over the meaning 
and explanatory usefulness of innate ascriptions in ethology.  Second, I wish 
to compare my canalization account of innateness with accounts proposed by 
contemporary philosophers, one by Stephen Stich (1975), another by Elliott 
Sober (forthcoming), and a third by William Wimsatt (1986).
	
1. Lorenz's theory of instincts.
Within Lorenz's theory of instinct the theoretical significance of identifying 
innate traits is three-fold: it is useful in taxonomy, individual explanation, 
and in evolutionary explanation (Richards 1974).  (1) Taxonomy: to 
characterize species-specific traits as opposed to traits that pick out 
differences between individuals in a species or a population.  (2) Individual 
explanation: citing innate behavior may explain why individual organisms act 
as they do in certain circumstances.  For example, why does a stickleback 
attack a wax model of a fish that lacks all structural resemblance to a rival 
stickleback?  The answer is that the model has a red-underbelly which serve in 
sticklebacks to release an innate attacking behavior (Richards 1974).  (3) 
Innate traits are genetically determined and hence are subjected to 
evolutionary investigation and explanation.  Lorenz sought to provide natural 
selection explanations for certain "adaptive" species-specific behaviors.
	According to Lorenz field observations and "isolation experiments" 
constitute the empirical investigation of behaviors.  In an isolation 
experiment the organism under study is deprived of the opportunity to "learn" 
or acquire the candidate behavior from environmental cues.  If the organisms 
undergoing isolation develops the trait "normally" then the trait or behavior 
is said to be innate.  For example, Lorenz observed that female mallards 
raised to reproductive age in exclusive company of pintail ducks show no 
sexual affinity for the pintail drakes.  But upon seeing a male mallard for 
the first time, the females immediately engage in the sexual courtship 
behavior particular to its species.  Remarkably, a mallard expresses courtship 
behavior even if it had no opportunity to learn it.  That is, mallards which 
are naturally or experimentally deprived of the opportunity to observe 
courtship behavior through experience, tend to develop it nonetheless.  
Determining that the mallard's courtship behavior is innate allows the 
investigator to: (1) include the behavior as part of the taxonomy of mallards, 
(2) explain why mallards exhibit courtship behavior, and (3) open the study of 
courtship behavior to evolutionary investigation.
	As Lorenz's critics have demonstrated, however, Lorenz's account is 
unsatisfactory. Attributing behavior to the genes is to misunderstand the 
nature of how biological traits are manifested.  No biological trait is the 
instantaneous product of genes alone or of the environment alone.  Biological 
traits are the product of a developmental process which depends on the actions 
and interactions of both genes and environments.  Lorenz thought that the 
results of deprivation experiments showed that some traits appear "out of a 
vacuum" (Lehrman 1953).  Yet, even behaviors manifested in deprivation depend 
on some environmental interaction, namely the environment of the deprivation 
device (assuming that the creature is kept alive).  It follows from the truism 
that no phenotype is the product of genes alone. 
	
2. The Lorenz/Lehrman Debate.
Lorenz devoted a small book to respond to his critics (Lorenz 1965).  In his 
response Lorenz makes some concessions though he maintains that the category 
of innateness is useful to ethology.  Lorenz's main concession was to 
acknowledge that attributing behavior to the genes is misleading in the 
context of a developmental explanation of behavior.  Nevertheless, Lorenz 
asserts that the categories of innate and acquired (or learned) are 
explanatorily useful (and hence ought not be dropped) as they serve to 
articulate the "adaptiveness" of certain behaviors.  Lorenz writes, 

As students of behaviour, we are not interested in ascertaining at 
random the innumerable factors that might lead to minute, just bearable 
differences of behaviour...What we want to elucidate are the amazing 
facts of adaptedness.  Life itself is a steady state of enormous general 
improbability and that which does need an explanation is the fact that 
organisms and species miraculously manage to stay alive (ÔInnate Bases 
of Learning' in On the Biology of Learning, K. Pribram, 1969).
 
I think that Lorenz's response is significant.  It reveals to me that the 
exchange between Lorenz and his critics is in part a clash between divergent 
explanatory strategies or approaches to explain the nature of organic life.
	What's clear from the passage just quoted is that Lorenz's interest in 
the category of innateness is to explain an evolutionary phenomenon, 
adaptiveness.  What does it mean to be interested in explaining adaptiveness?  
Adaptiveness is a relation between an organic form and its habitat.  Within a 
particular habitat, an adaptive trait is one that confers relative 
survivability and/or reproductive success to individuals that possess them.  
Darwinists explain adaptiveness by appeal to the theory of natural selection.  
More precisely, since evolution is a population-level phenomenon, what 
Darwinists explain is the origins and prevalence of adaptive traits in a 
population of organisms.  	
	In contrast, Lorenz's critics are in large part developmental biologists 
interested, not in the relation between organic form and its habitat, but in 
the organic form itself.  Further, developmentalists are not necessarily 
interested in a prevalence of organic form, but its origins in an individual 
embryo.  
	Given this distinction between what evolutionists and developmentalists 
seek to explain, I think we can clarify Lorenz's response to his critics.  He 
admits that attributing behavior to the genes is an error if the point is to 
explain how a trait comes to develop in an individual (the developmentalist 
project).  However, if the point is to explain the characteristics of 
adaptiveness (the evolutionist project) then attributing behavior to the genes 
is part of a standard Darwinian evolutionary explanation.  Let me briefly 
explain why Lorenz might have thought this is so.
   	While there are many points of intersection between evolution and 
development, evolutionary explanations often "black-box" development.  That 
is, evolutionists often talk about the genetic transmission of phenotypes 
without regard to what might be a complex correspondence between genotypes and 
phenotypes.  From the evolutionist's point of view what is important is not 
the knowledge of the causal factors that lead individuals with a given 
genotype to develop particular phenotypes, but that genotype and phenotype are 
correlated in a way that allows investigators to examine an evolutionary 
phenomenon (Sober 1984, Amundson 1994).  In other words, the developmental 
story that links genes to phenotypes is not the central issue for 
evolutionists though it is important that that there is some developmental 
story to tell.   
	Although both explanatory strategies, the developmentalist's and the 
evolutionist's, appeal to the concept of "gene" and "environment", in the case 
of genes what counts as the significant causal role diverges between 
developmentalists and environmentalists, and in the case of the "environment", 
meanings (reference) differ.  For evolutionists (especially adaptationists) 
genes serve as the unit of inheritance necessarily, and sometimes as the unit 
of selection.  The laws that dictate how genes play that role (e.g. Mendelian 
and non-Mendelian laws of segregation and assortment) are insensitive to the 
molecular composition of genes (i.e. the fact that genes are bits of 
chromosomes composed of DNA in the form of a double-helix, etc.).  But for the 
developmentalist, genes play a crucial bio-chemical role in the determination 
of organic form.  Genes code for amino acids that determine the 3-D folding 
structure of protein molecules, etc.  Further, in regard to what counts as the 
"environment", what distinguishes modern-day epigeneticism from pre-19th 
century preformationism is the tenet (truism) that genes alone do not produce 
biological traits.  Development involves complex interactions between genes, 
among genes, between genes and other components of a cell, etc.  Everything 
outside of a particular gene is called "the environment" for 
developmentalists.
	The evolutionist (again, especially the adaptationist) regards the 
environment as part of the "selective regime" including all conditions 
external to the organism itself (or outside the unit of selection): resources 
(food, shelter), existence of predators, climatic conditions, topography, 
location of mates, etc.  The evolutionist typically gloss over the role of the 
environment in the development of individuals.  Are they justified in doing 
so?  
	To answer, consider how natural selection explains prevalence and 
maintenance of biological traits.  Selection operates over variants in a 
population.  Certain organisms, because of particular traits they possess, are 
better suited to survive and reproduce than their conspecifics in the 
environment they all inhabit.  The key is that an adaptationist is interested 
in trait differences between organisms exhibited in an environment where 
selection is going on.  In a particular selective regime the factors of the 
environment that instigate selection (climate, resources, etc.) are fixed as 
far as the evolutionist is concerned.  The phenotypic differences are all due 
to genetic differences (hopefully much of that genetic differences is 
responsive to selection).  That is not to say that gene and environment 
interactions do not occur, in other words it is not to say that the 
development of the organism is solely due to the genes.  That would violate 
the central tenet of epigenetics and make evolutionary theory incompatible 
with the best theory of development.  Rather, it is assumed that these various 
ways in which environment effects development does not account for the 
phenotypic variation.  It is this sense in which, vis-ˆ-vis the evolutionary 
project that the environmental role in development is said not to matter and 
is often taken for granted.
	If the evolutionary project were the extent of Lorenz's methodology, I 
think that Lorenz's response to his critics is relatively uncontroversial.   
For example, it might interest an evolutionist to consider the selection 
pressures that might have been present to favor individuals possessing innate 
behavior capacities rather than individuals who must acquire or learn their 
behaviors.  In the spirit of adaptationism a model may be constructed to 
speculate on possible answers (see Sober 1994a).  
	But Lorenz fails to distinguish the question of explanatory significance 
from the question of what makes a trait innate.  Lorenz, you may recall, 
originally identified innateness with the ability of an individual organism to 
develop a trait in isolation of environmental cues.  In response to his 
critics, Lorenz asserts the value of innateness from an evolutionist's point 
of view.  But his response fails to answer the original question, what sort of 
developmental phenomenon is innateness?
	Lorenz recognizes this problem and provides an answer consistent with 
his evolutionary approach.  His answer is roughly as follows: for adaptive 
behavior to develop in an individual the organism requires information about 
its environment.  There are two possible sources for this information.  Either 
it is provided by the individual's interaction with its environment, or it is 
provided by the evolutionary process of natural selection in which case it is 
encoded in the organism's genes.  When the source of the information is 
provided by natural selection, Lorenz argued, the trait is said to be 
"innate".  So, the category of innateness is not only important to the study 
of evolutionary biology, but it is also important to the study of development. 
It allows one to distinguish traits whose developmental information is due to 
natural selection from those traits whose developmental information is in the 
individual organism's environment.    
	Setting aside Lorenz's associations between innateness and "information" 
let us focus on his views on the explanatory role of natural selection.  I 
take it that the key point is: knowing that a trait is a product of natural 
selection helps explain its development in an individual.  But as the 
psychologist Daniel Lehrman (1970) pointed out, Lorenz's theory about the role 
of natural selection in development is mistaken.  According to Lehrman, 
knowing that a trait is a product of natural selection does not settle any 
questions about the developmental process by which the phenotypic 
characteristic is produced in an individual organism.  
	Lehrman writes, "the clearest possible genetic evidence that a 
characteristic of an animal is genetically determined in the sense that it has 
been arrived at through the operation of natural selection does not settle any 
questions at all about the developmental processes by which the phenotypic 
characteristic is achieved during ontogeny" (1970, 25-28).  Lehrman's argument 
rests on an example of a trait, species-specific mating preferences that is 
widely under selective control.  Compare the development of the trait between 
doves and cowbirds.  For both the trait is a product of natural selection, yet 
individual doves develop their mating preferences through "learning" while the 
cowbird does not.
	I think the most perspicuous way of illustrating Lehrman's point is to 
consider an example used by Elliott Sober (1984), which I alter to fit our 
purposes.  Imagine that entrance to a school requires that individuals read at 
the third grade level.  Suppose that Sam, Aaron, Marisa, and Alexander pass 
the test and so are admitted, whereas other students do not pass, and so are 
excluded.  The selection process explains why the room contains only 
individuals that read at the third-grade level.  But the selection process 
does not explain why the individuals in the room (Sam, Aaron, and so on) read 
at the third-grade level.  As Lehrman put it, the selection process favors 
certain outcomes not the process by which those outcomes are achieved.  Notice 
that it is quite possible that the story about how each individual came to 
read at the third grade level varies.  For example, suppose Sam and Aaron have 
parents who taught them how to read at a young age, Marisa took a pill which 
enhances her reading ability, and Alexander reads well despite living with 
parents that did nothing to aid his progress.  These individual differences 
are not part of the selection-explanation for the prevalence of children who 
read at the third-grade level. But they are part of the developmental 
explanation for how each child came to read at the third-grade level.   
	Lorenz's mistake was to miss-apply natural selection explanation.  
Selection explanations are taylor-made for the investigation of evolutionary 
phenomenon.  Yet they are inappropriate for the investigation of what is 
essentially a developmental phenomenon, the "innateness" of biological traits.  
Further, by defining innateness as a product of natural selection (via genetic 
transmission), we are left without an adequate definition of innateness as a 
developmental phenomenon.  
	Let us put the point in a different way.  Attributing biological traits 
to genes as Lorenz would have it does not answer the central question that 
innateness ascriptions were meant (by Lorenz) to address.   The issue is, 
broadly, why does an individual have the traits that it does?  (Or, how does 
an individual come to manifest the traits that it has?)  To say that a trait 
is innate, that is, to say that it is genetically determined is only to deny 
that it is, in some unclear sense, not environmentally acquired.  Identifying 
traits as innate as opposed to acquired says nothing of the developmental 
process involved in the production of the trait (Lehrman, p. 344).  If we want 
to know why an individual has the traits it has we need to ask how environment 
and genes interaction to produce the trait in question.  Attributing a trait 
"mostly" to the genes says nothing about the developmental process that gets 
us from genes to the trait in question.

3. Apportioning casual responsibility between genes and environments.
Is there another way to preserve Lorenz's genetic account and still have it 
that innateness is a feature of development?  Some authors, most notably 
Robert Richards (1974) and Michael Levin (1992), have suggested that the 
distinction between innate and acquired traits is in fact useful to the 
developmental project of decomposing phenotypes in terms of genetic and 
environmental components.  The thought is that the fact of development does 
not preclude the possibility that the causal role of each may be 
distinguished.  As both Richards and Levin point out the decomposition of 
phenotypes is consist with standard practice in physics and chemistry.  Levin 
(1992) writes "the causal role of genes can be isolated, just as the causal 
role of any single chemical in a reaction can be" (p. 50).  Richards (1974) 
concurs, "The discernment of the sources of behaviour, a little like the 
discernment of spirits, is rarely easy...The situation is no different, of 
course, in any of the other sciences...the problem of holding constant or 
eliminating factors other than those of interest is one which...all 
experiments in natural science share" (p. 127).  
	Unfortunately, Levin's and Richards' approach won't work.  The method of 
decomposition to which Levin and Richards refer is not applicable to the 
investigation of phenotypes.  Let us see why. 
	In chemistry and in physics one can meaningfully ask "How much did two 
causal factors contribute to an event?"  We can ask this about chemical or 
physical systems because physics and chemistry both obey J. S. Mill's 
principle of the composition of causes.  Consider a Newtonian example where 
the event is a particle's acceleration, one causal factor is an effect of 
gravity, and the other is an effect of electricity.  By Mill's principle the 
result of the two forces is just the sum of what each would have achieved had 
each acted alone (Sober 1994b, 184).  To answer the question, "how much did 
gravity and electricity contribute to the particle's acceleration" we ask two 
further questions: (1) How much acceleration would there have been, if the 
gravitational force had been present, but the electrical force been absent?  
(2) How much acceleration would there have been if the electrical force had 
been present, but the gravitational force had been absent?  
	Such questions have no analogue in developmental biology. The 
investigation of phenotypes does not obey Mill's principle.  It follows from 
the fact of development that one cannot ask what effect genes or the 
environment would have had on the production of a biological trait if genes or 
environment had acted in isolation.  For example, one cannot ask of Sally how 
tall would she have been if genes acted in isolation or the environment acted 
in isolation.  
	Interestingly, Sober (1994b) argues that Mill's principle is inessential 
to the issue of whether it is possible to ask how much genes and environment 
contributed to a biological trait.  If that's right, then the truism of the 
phenotype does not automatically yield the negative conclusion.  What is 
essential to the conclusion is that the contributions of genes and 
environments to the production of a biological trait are incommensurable; as 
Sober puts it, "their contributions are not made in a common currency" (Sober, 
193). 
	Consider the following example to illustrate the notion of 
commensurability between two causal factors and to demonstrate why the truism 
of genes and environments is inessential to the conclusion.   Suppose two 
brick-layers are responsible for building a wall.  Each begins work on one end 
of the wall and work his/her way to the middle, both involved in the task of 
laying down the mortar and the bricks.  Since both perform the same task, it 
is possible to ask which brick-layer contributed more to the job.  To answer 
we simply count the number of bricks each worker contributed to the job.  
There is a way to determine the relative causal contribution of each worker 
for each makes her contribution in a common currency.  Notice that to ask how 
much each brick-layer contributed to the wall, it is not necessary to ask how 
much would each contribute had she acted alone.  That is, to answer the 
question about causal contribution we do not need to appeal to Mill's 
principle of contribution of causes.  We just need to compare the numbers of 
bricks each brick-layer laid.
	One might object that counting each brick-layer's contribution just 
amounts to asking how much each would have contributed had she acted alone.   
A slightly different example puts some doubt to that intuition.  Suppose the 
masons coordinate their work load such that they each take turns laying down 
one brick before the other lays down theirs.  So, one mason sets down brick 
number 1, 3, 5, etc. while the other sets down brick number 2, 4, 5, etc.  
After the wall is built, we cannot answer the question how much did each 
contribute had each acted alone because each worker's contribution does not 
constitute a wall at all.  A wall with successive bricks missing is not a wall 
but a heap of bricks!  If we want to know the relative contribution of each 
worker we simply count the bricks each mason laid down.
	Now suppose that the division of labor is such that each brick-layer has 
a specific task, one lays down the mortar and the other sets the bricks.  Is 
it meaningful to ask which brick-layer contributed more to the task of 
building the wall?  Since, in this scenario, each brick-layer does not make 
her contribution in a common currency, this question has no answer.  The 
interaction between genes and environments is akin to the second brick-laying 
example whereby the contribution the factors make are incommensurable.  
	What would it be like for biological traits to be decomposable and their 
relative contributions have a common currency?  Suppose height were the result 
of an accumulation of height particles whereby both genes and environments 
contributed a certain number of particles to make up a person's height.  If 
so, then we could compare the number of particles genes and environments 
contributed to, say, Jane's six-foot stature (Sober, 193).  Getting back to 
innateness, if the term is meant to denote a trait that owes a great deal of 
its structure to genes, and if the contributions genes and environments made 
were commensurable in this way (e.g. there are height particles) then 
innateness ascriptions would be helpful in apportioning causal responsibility 
between genes and environments.  We could determine the relative number of 
particles genes and environments contributed to the trait by elaborating 
various deprivation experiments.   Alas, there are no height particles; the 
contributions that genes and environments make to biological traits are not 
commensurable; we cannot answer the question, how much each factor, genes and 
environments, contributed to a trait.  Hence, innateness cannot be ascribed to 
answering such questions. 
	The nature/nurture case is not unique.  The distance a cannonball 
travels depends on both the angle of the cannon's barrel and the muzzle 
velocity.  But the two factors do not make their contributions in a common 
currency; there is no telling how much muzzle velocity and barrel angle 
contribute to the distance the cannonball travels (Sober 197).
	So, I conclude that Levin's and Richards' suggestion intended to save 
Lorenz's genetic account of innateness fails.  Does it follow, as Lehrman and 
his modern-day counterparts (Oyama, Johnston, Griffiths and Gray, etc.) would 
have us believe, that the gene/environment dichotomy upon which the 
innate/acquired dichotomy rests is unexplanatory?  Is there no refinement of 
the notion of innateness that is useful for any field of inquiry?
	No.  Lorenz's naive account is not the only one available.  There are 
other biological definitions of innateness, the best of which are not 
falsified by the fact that genes alone do not produce biological traits.  I 
will offer one and criticize three others in this essay.  But first I need 
some criteria of evaluation.  I take it that the best source of these criteria 
comes from what remains fairly uncontroversial about Lorenz's theory.
(a) Development: An account of innateness should make it a feature of 
development.  In the birdsong literature, for example, you often find a 
contrast between birds that develop their songs "innately" vs. those that 
don't.  Recall that part of Lorenz's project was to pick-out certain adaptive 
behaviors that particular animals exhibit even if they had no opportunity to 
"learn" or "experience" the behavior in the wild.  Ethologists are still 
interested in such behaviors.  For example, ethologists of bird song have 
discovered, via the use of deprivation experiments, various ways in which 
species-specific birdsong develops in different species of birds.  Certain 
birds develop their species-specific song regardless of whether or not they 
hear another bird sing that song.  Other birds seem to require experiencing 
the song before themselves singing it.  Here we have a contrast in 
developmental requirements some birds require a specific environmental cue for 
the development of their own song while others do not require that specific 
environmental cue.

(b) Environmental Stability: Innateness should denote an environmentally 
stable trait.  In the ethology and biology literature innateness seems to have 
something to do with what environment does not do to influence development in 
an individual.  Evidence from deprivation experiments and observations in the 
wild suggests that some traits develop normally in a range of environments, 
including impoverished and non-normal ones.  In such cases, the environment 
does not prevent the trait from being manifested.  Ethologists since Lorenz, 
sometimes associate innateness with environmentally stable, as opposed to 
labile, traits (Hinde, 1982, 86).    

(c) Natural Selection explanandum.  An account of innateness should make clear 
how natural selection can effect prevalence of some adaptive traits.  Recall 
that Lorenz was interested in applying natural selection to explain the 
prevalence of certain highly adaptive species-specific traits.  For example, 
he was interested in how selection might explain the prevalence of ducklings 
that display courtship behavior without any need of actually experiencing 
other duckling displaying that behavior.  I object only to defining innateness 
solely as a product of natural selection.  I have no in principle objection to 
natural selection explanations of the prevalence of innate features.  The 
natural selection project seems promising although it is an open question 
whether all such traits are explainable via natural selection.

	First, I'll propose my own account of innateness stemming from C.H. 
Waddington's concept of "canalization".  After demonstrating how the 
canalization account of innateness captures the desiderata I'll compare it 
with three other accounts, one put forward by Stephen Stich, another by 
Elliott Sober, and a third offered by William Wimsatt.
	
4. The canalization account.
In 1936 the developmental biologist C.H. Waddington along with his associates 
(Joseph Needham and Jean Branchet) made a remarkable discovery (see Gilbert 
1991).  They had set-out to determine the specific substance that induces 
competent ectoderm tissue into developing neural plates.  Waddington had 
thought that the substance was a particular steroid and only that steroid 
could induce neural plate development from competent ectodermal tissue.  
However Waddington's crew discovered, rather than a specific steroid, a large 
variety of natural and artificial compounds induced neural plate development.  
Since lots of different substances served as an inducer and not all cells 
could form neural plates, Waddington proposed that there must be something 
special about the competent tissues of the ectoderm that allowed it to respond 
to the range of the inducing chemicals.  What is special, Waddington 
conjectured, is that competency to respond to a range of inducing agents is a 
genetically controlled feature of certain tissues.  The wide-ranging 
competency of the ectoderm allows neural plates to develop in a wide range of 
environments.  
	If competency is genetically inherited it might be susceptible to 
natural selection forces.  For instance, neural plates develop out of ectoderm 
in a wide range of developing environments, even environments lacking 
steroids.  For certain selective regimes the capacity to develop in a range of 
environmental conditions might confer reproductive success to individuals 
possessing them, say, in unstable environments.  In such a case, Waddington 
said that the pathway became canalized by natural selection.  Canalization 
denotes a process whereby the endstate (the product of development) is 
manifested despite environmental perturbations (Waddington 1940).
	According to Waddington, canalization explains why developing organisms 
tend to produce a number of distinct and well-defined body types despite 
environmental variation between the individuals.  Waddington envisioned 
development as "an epigenetic landscape," a branching out of various 
developmental pathways each leading to the production of a distinct end-state.  
Once development starts (e.g. in the egg) any number of a range of inducing 
agents force the developing organism down one or another canalized pathway 
(Waddington 1957).  
	Further, the canalization model could account for why, for some 
quantitative characters, there appears to be a "normal" range, such that 
deviant morphologies are difficult to produce.  Waddington writes: "for most 
animals there seems to be a Ônormal' size, to which the adult individual often 
closely approximates even though it may have suffered various enhancements or 
retardations of growth during its lifetime as a consequence of factors such as 
the number of its litter-mates, the level of its feeding and so on" 
(Waddington 1975, 99).
	In regard to our project of providing an adequate biological account of 
innateness, canalization fits the bill.  Recall the three desiderata for an 
adequate account of innateness.  An adequate account ought to: (a) make 
innateness a feature of development, (b) contrast traits that are 
environmentally stable rather than labile, and (c) be part of natural 
selection explanations for the prevalence of certain traits.  Waddington 
provided an account of the development of environmentally stable traits that 
occur in individuals and whose prevalence can be effected by natural 
selection, satisfying the three desiderata.  I propose the following as an 
adequate account of innateness:

The degree to which a biological trait is innate for individuals 
possessing an instance of a genotype (or set of genotypes) is the degree 
to which the developmental pathway for individuals possessing an 
instance of that genotype (or set of genotypes) is environmentally 
canalized.  

The degree to which a developmental pathway is canalized is the degree 
to which development of a particular endstate (phenotype) is insensitive 
to a range of environmental conditions under which the endstate emerges.
 
The concept of canalization is useful because there often exists a high degree 
of constancy ("robustness") of phenotypes over fairly well-defined "normal" 
range of environmental conditions. 
	Several features of the canalization account of innate traits are worth 
noting.  First, the canalization account preserves the idea that traits are 
innate with respect to certain genotypes.  It may turn out that, for example, 
some of the genotypes that typically express blue-eyes are canalized while 
others are more sensitive to environmental fluctuations. 
	Further, although Waddington's account of canalization was intended to 
provide an illustration of the genetic control of phenotypic characters, it is 
important to note that Waddington recognized that development is an 
interactive process involving both genes and environmental conditions.  
Certain environmental conditions are essential for the development of certain 
traits.  No developmental pathway is resistant to all possible environmental 
perturbations.  As Waddington states, "even the most well-canalized character 
is, of course, never entirely invariant" (Waddington 1975, 100).  So, that 
there is a limit in the environmental ranges to which a developmental pathway 
produces a single phenotypic endstate should come as no surprise given the 
fact of development.  
	To reflect the fact of development on the canalization account, 
innateness is a matter of degree: the greater the environmental range against 
which a developmental pathway is buffered, the more canalized is the 
developmental pathway.  The steepness of an entrenched pathway (or "chreode") 
in Waddington's representation of development represents the degree to which a 
pathway is canalized.  The steeper the sloping sides of the pathway, the more 
resistant the pathway is to disturbances.    
	On the canalization account, innateness too is a matter of degree.  
Loosely speaking, limb development in many organisms is highly canalized; 
limbs develop in all but only the harshest environments.  However, a trait 
requiring several very specific environmental cues to develop in an 
individual, e.g. learning French (as opposed to learning languages), is not 
highly canalized.
	Finally, consider that Waddington's comparisons are made within a 
specific environmental range.  This leads to the question, what counts as an 
appropriate environmental range to determine innateness?  I doubt that there 
is a uniquely correct answer to the question.  We probably cannot avoid having 
to determine the appropriate range on pragmatic considerations, say, depending 
on the interests of the biologist.  As Waddington himself recognized, as a 
consequence of the fact that no biological trait develops independently of 
biological factors no trait is canalized simpliciter.  Mutatus mutandis for 
innateness.  Recall that Lorenz's mistake was to think innateness denoted 
traits that appear in vacuo.  An adequate account of innateness should not 
make the same error.  
	However, biologists and ethologists tend to ascribe innateness to traits 
whose development are to some surprising degree insensitive to certain 
environmental conditions (we made this environmental stability condition as 
one of our desiderata of an adequate account of innateness).  Let us look at 
three examples illustrating that what constitutes the appropriate range 
depends on the trait in question.
	First, Lorenz was interested in understanding the development of 
goslings who exhibit a "follow-mother" behavior even in the absence of mother 
geese.  For Lorenz, the exhibition of a gosling's follow-mother behavior is 
insensitive to whether mothers are present or not.  Lorenz was not interested 
in how fluctuating embryonic temperatures effect the development of the 
behavior although it may be an open question as to whether temperature does 
effect the outcome.  Lorenz was impressed that a behavior that is specifically 
designed to follow mother (that is, a well-adapted activity) is exhibited 
without the need of the actual mother or any other gosling present.  According 
to Lorenz, there is something interesting about the behavioral development of 
goslings that allow such a trait to emerge.  In the terms of the canalization 
account, what's interesting is that the end-state is canalized within the 
range of environments whereby mothers are present or not. (Note: it may be 
that the development of the follow behavior is canalized or innate within 
environments where mothers are either present or not, but not canalized within 
environments in which, say, embryonic temperature is fluctuating.) 
	Second, birdsong ethologists are interested in the contrasting 
developmental pathways between individuals of different sparrow species.  Some 
sparrows produce their characteristic song even if they are reared in silence 
while among other species, a sparrow produces its song only if it first hears 
that song performed (Gould and Marler 1991).  Birdsong ethologists are 
interested in contrasting developmental pathways of individuals representing 
different songbird species, i.e. ones that develop when reared in silence with 
ones that require their specific song as a precondition for development.
	Third, Waddington expressed interest in traits whose development 
appeared to be insensitive to "normal" environmental conditions.  Waddington 
sought to provide selection explanations for why certain traits persist in 
"normal" environments and express variants only in "stressful" ones.  What 
counts as a "normal" and "stressful" environmental range is vague, but it 
probably denotes the environments a population of individuals typically shared 
throughout their selective history.  
	To reflect the varied interests of biologists, what counts as an 
appropriate range might change depending on the trait in question.  But at 
least one caveat is in order: at minimum biologists should restrict themselves 
to environments in which the organism can develop.  This is an important 
condition vis-ˆ-vis an account of innateness.  For example, one does not prove 
that hair color is not innate (for individuals possessing a genotoken) by 
showing that there are environments in which the individual does not develop 
hair at all (Sober, pers. com.).

5. Innateness and natural selection.	
Finally, in accepting that Waddington's notion of canalization applies well to 
the idea of innateness, we open the doors to a range of parallels between the 
work of Waddington and the work of ethologists like Lorenz.  A striking 
example will further strengthen the case for thinking that innateness is 
canalization.  In this instance, we find Waddington solving a problem that 
Lorenz failed to solve: how can we invoke natural selection to explain the 
prevalence of highly adaptive traits, such as sexual affinity in mallards, 
that are seemingly environmentally acquired yet turn out upon further 
investigation (e.g. by performing an isolation-rearing experiment) to manifest 
in isolation?  For Waddington, the key is the concept of canalization.
	Recall Waddington's notion of competence.  Competent tissues are 
inducible by a range of compounds.  Most importantly, the range may extend to 
compounds found either to be internal or external to the developing organism.  
Waddington hypothesized that once a developmental pathway is canalized, the 
competent tissue may transfer its responsiveness from one inducing agent to 
another (Gilbert 1994, p. 851).  The possibility of transfer of competency is 
significant for evolutionary biology.  For example, calluses are typically 
environmentally induced by friction between skin cells and some external 
surface.  Here, the genes play a role in proliferating cells to form the 
callus.  Suppose the competence of the skin cells to form a callus structure 
when induced by friction is both a product of natural selection and the 
pathway initiated by the friction that leads to the callus structure is 
canalized.  The canalization of callus formation depends on an external agent 
in this instance.  But if the skin cells are competent for a range of agents 
it is possible for an internal agent to substitute for the external agent.  
For instance, suppose a mutation appears that enables the skin cells to 
respond to a stimulus within the developing embryo.  Then, the ability to form 
a callus due to friction may become part of the "genetic heritage" of the 
organism.  That is, what starts as an externally induced trait undergoes a 
mutation that in effect transfers its competence to internal inducers.  The 
end result is an organism that develops a callus in the absence of a friction.  
In some selective regimes the transfer of competence from external to internal 
inducer may confer fitness to organisms possessing the capacity.  Waddington 
called the transfer of competence "genetic assimilation".  He hypothesized 
that selection aiding in the genetic assimilation of the capacity of ostriches 
to develop calluses explains why ostrich embryos possess calluses (Waddington 
1975).  If the transfer of competence is due to a mutation, genetic 
assimilation is said to be an instance of what is known as the "Baldwin 
effect".
	Unfortunately, Waddington did not back his hypothesis concerning ostrich 
calluses with experimental evidence.  However, Waddington did find evidence of 
genetic assimilation in experiments on thorax development in Drosophila.  To 
illustrate the phenomenon, Waddington (1975) managed to induce an extreme 
environmental reaction in the developing embryos of Drosophila.  In response 
to ether vapor a proportion of embryos developed a radical phenotypic change, 
a second thorax.  At this point in the experiment we would say that bithorax 
isn't innate; it is a kind of chimera induced by an unusual environment.  But 
then Waddington continually selected for Drosophila with the developmental 
capacity to respond to the environmental stress.  After about 20 generations 
of selection, some Drosophila were obtained that developed bithorax without 
being exposed to ether treatment.  What happened, according to Waddington, is 
that selection favored a particular pathway that led to the production of the 
optimal (in this case desired) effect.  Eventually the pathway became 
canalized, hence the end-state, bithorax, appeared regardless of environmental 
conditions.
6. Critique of Stich's, Sober's, and Wimsatt's theories of innateness.
Now I wish to compare the canalization account of innateness with three other 
developmental accounts, one by Stich (1975), one by Sober (forthcoming), and a 
third by Wimsatt (1986).  Using certain counter-examples, I hope to 
demonstrate the superiority of the canalization account over the others.  
Further, in demonstrating why the canalization account is superior, I 
highlight its key features.  For example, I have emphasized that canalized 
development is (relatively) insensitive to developmental environments.  The 
resiliency, not the interaction with environmental conditions, accounts for 
the developmental invariance of the phenotype in question.  This feature is 
not clearly a part of either Stich's or Sober's developmental account of 
innateness. 
	Stich's aim is to provide an account of innate ideas.  Following 
Descartes, his strategy is to provide a more general account of innateness 
(actually, one of innate diseases) and then apply it to the case of ideas: "in 
calling ideas innate, Descartes tells us, he is using the same sense of the 
word we use when we say certain diseases are innate.  So let us launch our 
analysis of innateness by pursuing Descartes' hint and asking what it is to be 
afflicted with an innate disease" (p. 3).  
	Stich's account resembles Descartes' own; both are "disposition 
accounts".  On Descartes' view, individuals that suffer an innate disease "are 
born with a certain disposition or liability to acquire them" (Stich 1975, 
p.6).  Stich's account unpacks the disposition to provide the following 
account:

A person has a disease innately at time t, if and only if, from the 
beginning of his life to t it has been true of him that if he is or were 
of the appropriate age (or at the appropriate stage of life) then he has 
or in the normal course of events would have the disease's symptoms 
(1975, p. 6).

The Cartesian approach (of which Stich's is a piece) is sensitive to the 
intuition that "inborn" or "present at birth" is not a necessary condition for 
"innateness".  Muscular dystrophy is, intuitively, an innate disease, yet the 
symptoms do not appear until later in a child's life.  Stich's Cartesian 
account preserves this intuition: a child born with muscular dystrophy has the 
disposition to experience the associated symptoms hence muscular dystrophy is 
an innate disease.  Or, more precisely, it is true of the diseased child that 
at the appropriate age she will begin to experience the symptoms (Wendler, p. 
91).
	Stich's account does not fare well with diseases that are associated 
with parasites that are acquired in the normal course of one's development as 
Wendler (1996) demonstrates.  Humans typically possess an abundant supply of a 
particular species of bacteria clostrium difficile (c. diff.) in our 
intestines.  Humans are not born with c. diff., we typically acquire it from 
ingesting food and water.  C. diff. is not harmful to healthy humans, but it 
can make sick people on antibiotic treatment sicker by colonizing in areas 
which the antibiotic treatment (to which c. diff. may be resistant) is killing 
off harmful bacteria.  Unabated colonization of c. diff. produce toxins that 
lead to diarrhea and associated symptoms.  Is the possession of c. diff. in 
one's intestines innate?  
	Innateness is generally and intuitively contrasted  with environmentally 
acquired (albeit as a matter of degree).  C. diff. is acquired from ingesting 
food and water, so, intuitively the possession of c. diff. is not innate.  
However, in the normal course of events humans eat and drink water and hence 
acquire c. diff.  As Descartes would say, humans have a disposition to suffer 
the symptoms of c. diff. in the intestines.  It follows, on Stich's Cartesian 
account, c. diff. is an innate disease of the intestines.  This is an 
unfortunate consequence of Stich's account.  
	Stich recognizes the loophole in his account: "there are commonly a host 
of necessary environmental conditions for the appearance of the symptoms of a 
disease.  If these conditions all occur naturally or in the normal course of 
events, the symptoms will be counted as those of an innate disease" (p. 7).  
Stich's response amounts to hoping for vagueness in one's intuitive judgments 
per case: "it is often unclear whether the occurrence of a certain necessary 
[environmental] condition is in the normal course of events.  So it will often 
be unclear whether a person is afflicted with an innate disease or is, rather, 
susceptible to a (noninnate) disease" (p. 7).  But Wendler's counter-example 
is a paradigm.  Clearly c. diff. is clearly in the normal course of event, at 
least part of the normal task of eating and drinking.
	It must be admitted that the problem that Stich alludes to is difficult 
for any account of innateness.  Stich's problem can be generalized as follows: 
for the development of any biological trait, disease, or otherwise, there will 
be a host of necessary environmental conditions for the appearance of that 
trait.  This just follows from the fact of development, that genes do not 
operate in vacuo.  
	The canalization account handles this fact in two ways.  First by making 
innateness a matter of degree, and second by noticing that for the development 
of some (special) traits, once a certain environmental (or genetic) condition 
is met, development becomes resilient to further environmental perturbations; 
in other words, development becomes canalized.  For instance, suppose a 
developmental system is found to corresponds to the developmental Ôrule of 
thumb': "Develop X no matter what the state of the world happens to be".  This 
exemplifies an extreme form of canalization.  Less extreme are systems that 
conform to the conditional rule: "Develop X if experience is C" or "Develop X 
if experience is C, develop Y if experience is D".  The more sensitive to the 
environmental condition the developmental system is the less inclined we are 
to say that the developmental endstate is canalized.  Let's see why this is an 
improvement over Stich's account.
	Stich's account fails to take into account the causal dependency of 
development on the environmental conditions that constitute the normal course 
of events.  Sober's account suffers from a similar problem.  For Sober, "a 
phenotypic trait is innate for a given genotype if and only if that phenotype 
will emerge in all of a range of developmental environments" (forthcoming). In 
other words, innateness amounts to phenotypic invariance across a range of 
environmental conditions.  Both Stich and Sober fail to recognize that there 
are two ways in which a trait emerges (invariantly) in a (e.g. normal) course 
of development:

1) By means of strict genetic control over development so that the outcome of 
development is insensitive to the conditions under which it occurs.  Such 
outcomes are said to be strongly canalized against environmental perturbation.

2) By means of a developmental is sensitivity only to environmental factors 
that are themselves invariant within the organism's (e.g. normal) 
developmental environment (both conditions come from Johnston 1982, 420).  

The second outcome is not canalized in Waddington's sense but it invariant 
under the normal conditions of development.  The c. diff. case is a paradigm 
example of the second case.  Contracting the symptoms associated with c. diff. 
(or even contracting c. diff. itself) is part of one's normal course of 
environment.  Further, acquiring the disposition to suffer the symptoms 
associated with the possession of c. diff. emerges across quite a range of 
(normal) human environments.  So, we would say the outcome is invariant and 
hence innate on Sober and Stich's account.  But acquiring c. diff. and having 
the disposition to suffer the consequences are both normal and developmentally 
invariant because humans are susceptible to an environmental condition that is 
itself invariantly part of human development.  C. diff. is in the food we eat 
and the water we drink, and we obviously need to eat and drink to develop at 
all.  Although the outcome is invariant it is acquired, not innate and not 
canalized.  
	Next, I turn to Wimsatt's "generative entrenchment" account of 
innateness. I reject it because generic entrenchment is not at all an 
essential condition for innateness.
	The "crucial feature" of Wimsatt's account is the idea that, "features 
that arise early in development have a higher probability of being required 
for features that appear later...and tend to have a larger number of 
downstream traits depending on them" (1986, p. 198).  Accordingly, Wimsatt 
defines "generative entrenchment" of traits "to the degree that they have a 
number of later developing traits depending on them" (p. 198).  There appears 
to be two central claims here.  One is that innate traits tend to appear early 
in development.  The second is that innate traits have a number of later 
developing traits depending on them.  I'm not altogether sure how Wimsatt 
conceives of the relationship between these two concepts, though it does seem 
that the concept of generative entrenchment is most central to Wimsatt's 
notion of innateness judging from his quip, "on this analysis, if it is 
generatively entrenched, it is 'innate'" (p. 200).  Nevertheless I think 
neither are necessary features of innate traits.  To see what's wrong with 
Wimsatt's theory, consider the innate development of pubic hair in 
adolescents.  Because: (i) pubic hair appears late in development, and (ii) 
pubic hair has no further trait (at least prima facie) depending upon it 
Wimsatt must say that pubic hair is not innate.  Whether or not a trait is 
innate has nothing to do with how late in development it appears or what 
further trait depends on it.   
	Wimsatt might retort that pubic hair is in fact not innate.  But, that 
would contradict his motivation to unify many philosophical and ethological 
accounts of innateness under the concept of generative entrenchment.  The idea 
that pubic hair is innate is supported by several of these accounts, including 
this one: "Innate behavior for a given species is universal among normal 
members of that species in their normal environment either: (a) because the 
behavior has a genetic base [or] (b) because the behavior is Ôcanalized' or 
homeostatically regulated in development" (p. 187). 
	According to Wimsatt, an interesting consequence of his developmental 
model is that some environmental experiences count as being innate.  For 
example Wimsatt claims that "not only is the imprinting mechanism of the 
greylag goose at birth 'innate'...but the object of imprinting is also 
'innate'" (p. 200).  Take one of the minimal requirements for an environmental 
condition to be called innate: "the acquisition of that kind of information at 
that stage of development is deeply generatively entrenched with respect to 
subsequent behavior" (p. 200).  Accordingly, imprinting is innate because a 
lot of later developing traits depend on it: "there is a very high probability 
that the young goose will properly imprint on its mother and will, in short 
order, learn to distinguish her cries and her appearance from that or other 
female greylag geese nearby" (p. 201).  
	To see what's wrong with this account, try applying it to the case of 
the fetal deformities resulting from mother's taking thalidomide during 
pregnancy.  For those who wish to preserve the contrast between innate and 
acquired, the effects of thalidomide are not at all innate when we consider 
that the presence of thalidomide, an environmental condition, makes all the 
difference between a child being born with a limb or without.  Whether or not 
a lot of later traits depend on the taking of thalidomide does not sway our 
intuition that the child's deformities are not innate.  I conclude, whether or 
not a trait is innate has little to do with whether it is generatively 
entrenched.

7. Conclusion
Taking my cue from critics of Lorenz, I have presented an account of 
innateness that avoids the fallacy of claiming that traits can develop by 
genetic causes alone.  Innateness, on the canalization account, is a property 
of a developing individual.  At the same time, the proposal captures what is 
thought to be distinguishing features of innateness: satisfying the ontogeny 
condition, referring to the capacity to produce environmentally stable traits, 
and making sense of the idea that natural selection can install innate traits 
in a natural population.

WORKS CITED

Amundson, R. (1994), "Two Concepts of Constraint: 
	Adaptationism and The Challenge from Developmental 
	Biology", Philosophy of Science, 61: 556-578.

Ariew, A. (1996), "Innateness and Canalization", Philosophy 
	of Science, 63 (Proceedings) pp. S19-S27.

Gilbert, S. (1991), "Epigenetic Landscaping: Waddington's Use 
	of Cell Fate Bifurcation Diagrams". Biology and 
	Philosophy 6: 135-154.

Gilbert, S. (1994), Developmental Biology 4th edition. 
	Sunderland, MA: Sinauer Associates, Inc.

Griffiths, P. and Gray, R. (1994), "Developmental Systems and 
	Evolutionary Explanation", Journal of Philosophy 91: No. 
	6. 277-304.

Gould, J. and Marler, P. (1991), "Learning by Instinct", In 
	D. Mock (ed.) Behavior and Evolution of Birds, San 
	Francisco: Freeman, 4-19.

Gottlieb, G. (1991), "Experiential Canalization of Behavioral 
	Development: Theory", Developmental Psychology, Vol. 27, 
	No. 1, 4-13.

Hinde, R A. (1982), Ethology.  New York: Oxford University 
	Press.


Johnston, T. D. (1988), "Developmental Explanation and the 
	Ontogeny of Birdsong: Nature/Nuture Redux", Behavioral 
	and Brain Sciences 11:  617-629.

Johnston, T. D. (1987), "The Persistence of Dichotomies in 
	the Study of Behavioral Development", Developmental 
	Review 7, 149-182.

Johnston, T. D. (1982), "Learning and the Evolution of 
	Developmental Systems", In H.C. Plotkin (Ed.), Learning, 
	Development, and Culture. Chichester: Wiley, 411-442.


Kuo, Z. (1921), "Giving Up Instincts in Psychology", Journal 
	of Philosophy, 18, 645-664.

Lehrman, D. S. (1953), "A Critique of Konrad Lorenz's Theory 
	of Instinctive Behaviour".  The Quarterly Review of 
	Biology 28: 337-363.

Lehrman, D. S. (1970), "Semantic and Conceptual Issues in the 
	Nature-Nuture Problem". In L.R Aronson, E. Tobach, D. S. 
	Lehrman, and J. S. Rosenblatt (Eds.), Development and 
	Evolution of Behavior. San Francisco: Freeman. 17-50.

Levin, M. (1994), "A Formal Treatment of Gene Identity, 
	Genetic Causation, and Related Notions", Behavior and 
	Philosophy. Vol. 22, No. 2. 49-58.

Lewontin, R. (1974), "The Analysis of Variance and the 
	Analysis of Causes", American Journal of Human Genetics, 
	26, 400-411.

Lorenz, K. (1957), "The Nature of Instincts".  In C. H. 
	Schiller (Ed.), Instinctive Behavior.  New
York: International University Press, 129-175.

Lorenz, K. (1965), Evolution and Modification of Behavior. 
	Chicago: University of Chicago Press.

Lorenz, K. (1935), "Der Kumpan in der Umwelt des Vogels", 
	Journal fŸr Ornithologie, 83, 137-213.

Oyama, S. (1985), The Ontogeny of Information: Developmental 
	Systems and Evolution. Cambridge: Cambridge University 
	Press.

Oyama, S. (1988), "How Do You Transmit A Template?", 
	Behavioral and Brain Sciences 11: 664-645.

Richards, R. J. (1974), "The Innate and the Learned: The 
	Evolution of Konrad Lorenz's Theory of Instinct".  
	Philosophy of Social Sciences 4: 111-133.

Sober, E. (forthcominga), "Innate Knowledge" in 
	Routledge Encyclopedia of Philosophy.

Sober, E. (forthcomingb), "Separating Nature and Nuture" in 
	D. Wasserman and R.Wachbroit (Eds.).

Sober, E. (1994a), "The Adaptive Advantage of Learning and A 
	Priori Prejudice", in E. Sober From a Biological 
	Point of View. New York: Cambridge University Press.

Sober, E. (1994b), "Apportioning Causal Responsibility", in E. 
	Sober From a Biological Point of View. New York: 
	Cambridge U. Press.

Sober, E. (1984), The Nature of Selection. Cambridge, MA: MIT 
	Press.

Stich, S. (1975), "The Idea of Innateness". In S. Stich (Ed.) 
	Innate Ideas. Los Angeles: U. of California Press. 

Vance, R. (1996), "Heroic Antireductionism and Genetics: A 
	Tale of One Science", Philosophy of Science, 63 
	(Proceedings) S36-S45.

Waters, C. (1994), "Genes Made Molecular", Philosophy of 
	Science 61, 163-185.

Waddington, C. H. (1975), The Evolution of an Evolutionist.  
	Ithaca: Cornell U. Press.

Waddington, C. H. (1940), Organizers and Genes. Cambridge: 
	Cambridge U. Press.

Waddington, C. H. (1957), The Strategy of the Genes. London: 
	Allen and Unwin.

Wendler, D. (1996), "Locke's Acceptance of Innate Concepts", 
	Australasian Journal of Philosophy Vol. 74, No. 3. 467-
	483.

Wimsatt, W. (1986), "Developmental Constraints, Generative 
	Entrenchment, and the Innate-Acquired Distinction", in 
	W. Bechtel (ed.), Integrting Scientific Disciplines. 
	Dordrecht: Martinus Nijhoff Publ.


  Parts of this paper were read at ISSHP in Seattle, the University of 
Arizona, the University of Rhode Island, Arkansas State University, Virginia 
Tech University.  I wish to thank audiences there for helpful comments.  In 
addition I wish to thank Denis Walsh, Elliott Sober, Chris Stephens, Robert 
Cummins, Tom Bontly, Joel Pust, Paul Bloom, and Henry Byerly for extensive 
comments on earlier drafts.
  E.g. Kuo (1921), Lehrman (1953), Oyama (1985), Johnston (1988),  Griffiths 
and Gray (1994).
  Especially Oyama (1985), Johnston (1988), and Griffiths and Gray (1994).
  Johnston writes on behalf of Lehrman: Òthe main point of LehrmanÕs argument 
was not to take a stand on the issue of whether, or how much, behavior is due 
to the environment as opposed to the genes, but rather that any such stand 
simply reflects a misunderstanding about the nature of developmentÓ (1987, 
153).
  My demonstration is borrowed heavily from Sober (1994b).
  The example is taken from Lewontin (1974).
  I thank Joel Pust and Tom Bontly for pointing this out to me and helping me 
with a solution.
  Note that the distinction between the two means of invariant outcomes is not 
a dichotomy but a matter of degree.
  I should point out that both intuitively and on my canalization account, the 
universality condition is not central or necessary to identify innateness.  
Consider those unfortunate persons who possess the/a gene that increases oneÕs 
chances of developing Tay-SachÕs syndrome.  It appears that Tay-Sachs syndrome 
is canalized to a large degree for those members possessing the gene.  Neither 
possessing the gene or developing the syndrome is universal among humans.
	HardcastleÕs Volume