Dear Haines, I think the first thing to do here is to un-buzz-word the whole discussion and focus on specific questions. Epigenetics means too many different things to be useful for real conversation. The same can apply to Lamarckism. (Oddly, in Origin of Species Darwin largely assumes Lamarck is right, although nobody ever mentions this. He wanted to add natural selection as an explanation of speciation, which is the one thing it is not - it only addresses phenotypic drift within species.) 


The general pattern in genetics is that more or less all possible mechanisms can and probably do apply somewhere but in most cases only rarely so drawing broad conclusions from oddities is unwise. In vertebrates by and large inheritance is just based on DNA sequence. Issues of differentiation of stem cells are irrelevant because they do not affect the germline cells that give rise to the next generation. Generally speaking vertebrate behaviour cannot affect DNA structure even in terms of methylation and certainly not in terms of sequence. Now that we understand activation induced deaminase and 'programmed mutation' we have a pretty clear picture why Lamarckian behaviour-to-DNA feedback is off the menu, even if the buzz-worders are muddying the waters. If such Lamarkian feedback was a reality Homo sapiens would have capitalised on it big time at least by ancient Greece and certainly in Soviet Russia. It is a myth.


Very complex patterns of vertebrate behaviour must be encoded in DNA sequence. If I remember rightly it is the Ancient Murrelet chick which hatches in a cave in a warm dung heap and walks hundreds of yards to the sea and behaves just like the parents it never ever met - total behavioural encoding, including knowing how to search for caves and build dung heaps later. Most turtles are the same. Old World cuckoos sing a song they have never heard. This lends no support for Lamarckism - I cannot think of any arguments that would point in that direction. It makes complete sense on the basis of a combination of natural selection and the shifting of macro-formatting of DNA that separates species. 


Darwinian-style natural selection occurs all the time in the development of the individual - the adaptive immune system works entirely on a Darwinian basis. That is standard dogma.

Best wishes, Jo E

Joe E.

Thanks for your comments. I realize that the term "epigenetic inheritance" is a can of worms,
and so tried to avoid ambiguity by defining (neo)-Lamarckianism by saying it means that
individual behavior can alter its DNA.  Agreed about "buzz-words", but I thought my distinction
here was helpful.

You point out that "by and large" inheritance is determined by DNA sequence, and you say that
"generally speaking" behavior can't affect DNA structure. My question was not whether it is
likely, but whether it is possible.

"Now that we understand activation induced deaminase and 'programmed mutation' we have
a pretty clear picture why Lamarckian behaviour-to-DNA feedback is off the menu". This I'd
appreciate your explaining (I'm an ignoramus). It seems that this is simply a case where there
is a chemical modification of DNA, and it is a specific instance (involving production of anti-
bodies), not relevant to the individual organism's behavior. So why does this case argue
against the possibility of behavior affecting DNA?

Then you appear to say the opposite, that complex patterns of veterbrate behavior must be encoded in DNA sequence. Are you contradicting your earlier judgements, or are you saying
that a genetic mutation in favor of this behavior favored reproductive sucess? A wasp, for
example, is hard-wired to build a sophisticated nest, but it is hard to understand that all these
nests are essentially identical when genetic variation would suggest an array of nest forms
that are equally functional.

You offer some examples of remarkable learned behavior (Ancient Murrelet, Old World Cuckoos,
turtles), but then conclude these are not arguments in support of a Lamarckianism. I assume
this behavior is not imitative or learned by trial-and-error. If not, then how did these complex behaviors manage to get written to DNA? 

Haines

  

Haines, you ask: "does not ethology suggest that some rather sophisticated behaviors are hard wired in terms of DNA? For example, wasp or bird nest-building techniques. That such complex behaviors are generally inbred rather than learned by observation, and are passed along for many generations, is remarkable. Do such behaviors imply DNA encoding, or can they explained otherwise? Also, if hard-wired in DNA, does it lend support to  neo-Lamarckianism?"

It's a little bit unclear to me what the argument behind this question is. Certainly, there is complicated behaviour in insects and birds, and some of this behaviour  -- e.g., birdsong -- may be imprinted, which is to say it will follow a pattern that is evinced by other organisms in the vicinity. And there is language-learning in humans -- which is innate and imprinted (though probably this is too weak a characterization).

Presumably any innate behaviour is in some sense or other DNA encoded, or is so to some degree. So if you put it in the terms of the above question, the question might be: "Does human language learning, which is to some degree and in some way hard-wired in DNA, lend support to neo-Lamarckianism?"  As far as I know, nobody has suggested this. But I am open to instruction. Why would imprinting lend support to neo-Lamarckiansism?

Secondly, you say: "the probability of which cells are replicated during growth can arise from structural constraints, and therefore DNA variations that are probable in relation to that structure will be favored."

Again, I am puzzled. Certainly, events taking place during development inside and outside the uterus will influence the outcome of development. For instance, Werker and Tees showed that infants can discriminate phonemes in any language, but by one year some of these discriminatory patterns are extinguished. You could label this developmental Darwinism, because some (useful) discriminatory abilities are retained and other (useless) discriminatory abilities are extinguished by environmental events. It's not clear to me, however, that this is Lamarckianism, which would entail (would it not) that in the next generation, the infants should have a limited range of phonetic discrimination abilities.

Could you explain?

Mohan

Thank you for the reply.

You mention examples of imprinted complex behavior that is learned. "Presumably any innate behaviour is in some sense or other DNA encoded, or is so to some degree". The issue is,
how does this imprinted behavior become encoded in DNA?  

Lamarck is not an easy person to pin down, and so I should put my question independently
of him: is there a permanent (vs. epigenetic) inheritance of behavior? I assume that "permanent" means genetic encoding, and I gather the real issue is not so much the fact that organisms
inherit complex behaviors, but finding a mechanism that can explain how behavior can influence
DNA the first place.

Often arguments are made that a complex organ that is functional as a whole cannot be
readily explained in Darwinian terms (natural selection) because the emergent whole contains
more information than that held by its constituents. In the social sciences this issue
comes up in criticisms of functionalist explanations, which describe a functional system
in terms of unobservable extrinsic functional properties of its parts that depend on their relation
to the whole, but which fails to explain how something can have a functional property before
that whole that emergences from its its functional properties. 

I had speculated about such a mechanism and unfortunately was not clear. You note that
embryonic development is subject to outside influences. True, a newborn comes into the
world with, say, some basic ontology (it seems knowing the difference between categories
of things and individual things, causal relations, seeing self as another other, etc.). But
humans are an animal with a long period of dependency on parents, and so come into
the world with rather generalized abilities, as beings very able to learn quickly. Most
animals (chickens vs. crows, for example), come into the world with rather hard-wired
intelligence suited to a rather specific niche. To keep things clear, it seems best to speak
of chickens rather than human embryos. 

I assume these more specific hard-wired behaviors are encoded in DNA, originally through
natural selection. However, it often seems these specific skills are just one choice among a
range of equally functional adaptive possibilities. Just how a wasp builds its nest must be
explained by genetic rules that go beyond mere adaptation to specify just how that function
will be instantiated. Mere functionality would only need a broad notion of nests. The choice
of mud over some other building material may be adaptive, but not the exact form, which
to a degree is non-adaptive information hard-coded in genes.

You point out that it is unlikely that skills developed by a child will be passed along to its
descendants. True, but if the inherited ability of the child is an open ended flexible
intelligence, your point follows from this. But the behavior of the wasp is hard wired
and is passed along in its genes. It seems that the specificity of a nest design goes
beyond mere being a functional response to the environment.

There are studies (you will have to pardon my ignorance of matters biological - not my
field) of a correlation between environmental complexity or variability and genetic
mutational frequency. That is, in a challenging environment, more genetic mutations
will be generated to provide a wider probability distribution for "expression". Is this in
agreement with your understanding?

If this be true, then the development of a particular organism can select among the
possible cells those which are most adaptive. I'm not sure if totipotent stem cells offers
an example. In any case, this would accommodate Darwinianism and Lamarckianism,
but there are problems with it. For one thing, E. coli replication and that of eukaryotes
is quite different, and it is difficult to understand how in the latter there can be feedback
to influence encapsulated cell replication during ovum formation. The issue here then
becomes, how can acquired behavior influence the development of the ovum?

Let me try to put my question this way. Except for random mutations, a zygote has the
same chromosomes at its parents, but following fertilization, it must replicate. This
replication carries out rules that are carried in genes. But genetic rules don't specify
unequivocally the outcome of replication, but are only rules that are subject to contingency
or circumstance (proteins) that affect the outcome. Assuming that the replication can
and does involve some genetic variation, and that the rate of mutation is influenced by
proteins and ultimately by circumstance, so the question becomes, can the behavior
of the parent during the growth of the embro somehow influence or select which variant
possible replication is favored, so that the resulting cells are a function of the parent's
behavior?  My sense is that while no such mechanism is known, the observable results
suggest there probably is some such mechanism.

Haines    

Dear Haines, I think we have some fairly simple wire crossing here. For the molecular biologist 'encoding of behaviour in DNA' simply means that the DNA sequences provides the code that determines behavioural disposition. It does not imply that actual behaviour has any effect on DNA or 'gets written to DNA'. The assumption is that the DNA sequence arises by random mutation in the context of natural selection. The Lamarkian case of behaviour altering DNA is almost certainly possible - in fact post Craig Ventnor it is definitely possible for us but as I mentioned there are no rules in genetics, everything is what is plausible in the context. Giraffes getting long necks by trying to reach up is totally implausible because the mechanism needed is outside anything we have come across and would have to be so complex it ought to be hard to miss.

I mention AID only in that prior to its elucidation we had no real understanding of how random mutation rates might be programmed. At least now we see how that can be done. AID seems to punch holes mostly in antibody genes but it also punches holes in other genes at lower rates. Whether this is important is not clear, but it gives us a model for how that might work.

I am not clear of the argument in the wasp case. However, it may be relevant that homogeneous behaviour is a very potent driver of speciation. If half of a colony of wasps started building longer shaped nests because of a mutation they would almost certainly survive better in a different part of the forest and inbreed for enough generations for their DNA macroformatting to drift sufficiently for them to be unable to backcross with the short nest wasps. That was the bit Darwin should have tweaked to because of the well known infertility of mules of even very similar phenotypes.

Best wishes

Jo

Dear Haines,

You ask: "You mention examples of imprinted complex behavior that is learned. "Presumably any innate behaviour is in some sense or other DNA encoded, or is so to some degree". The issue is,  how does this imprinted behavior become encoded in DNA?"  

My point was that imprinted behaviour is not encoded in DNA. What is encoded is the imprinting procedure: follow-whatever-duck-you-see-first; speak-whatever-language-you-hear-in-your-first-few-years, etc. For this reason, though my children speak English by imprinting, their children may speak some other language. Lamarckianism predicts that my children's English-speaking will have an impact on their DNA in such a way that makes it more likely that their children will speak English. That's why it's wrong.

So here's my first point: both deterministic (or sphexish, to use Dennett's term) and flexible behaviours can be encoded in DNA because conditional commands can be DNA encoded: i.e., you can have a DNA-encoded command "If F, then do X". (I think this is what Jo was getting at when he wrote of the DNA determining the disposition -- rather than the behaviour.) Where the antecedent of the conditional is environmentally sensitive, this demands the functional analogue of a sensor to determine when the antecedent of the conditional is satisfied.

You say: "the emergent whole contains more information than that held by its constituents."

I agree. A and B occurring together carries more information than A and B each carry. But here's my second point: I am not sure what you mean when you say that this can't be explained by natural selection. For instance, a rabbit may be deterministically programmed to run north if it hears simultaneous loud noises east and west, but to run east from a loud noise westward, and west from an eastward sound. You could say this is a case of non-reductionistic emergence, but why does natural selection have a problem with it?

Finally: "can the behavior of the parent during the growth of the embro somehow influence or select which variant possible replication is favored, so that the resulting cells are a function of the parent's
behavior?"

I think that in a literal sense, the answer is yes. For instance, the mother can take thalidomide. But Lamarckianism says that the parents' behaviour will replicate itself in the offspring through genetic changes in the DNA. Unfortunately, you can't by ingesting thalidomide produce a baby who is genetically immune from morning sickness (because she is genetically prone to take thalidomide).

best,

Mohan

Jo, you say, "I think we have some fairly simple wire crossing here". Were the truth told,
this blind person simply stumbles over the wires.

"For the molecular biologist 'encoding of behaviour in DNA' simply means that the DNA
sequences provides the code that determines behavioural disposition. It does not imply that
actual behaviour has any effect on DNA or 'gets written to DNA'." Well, yes, the issue is
sequencing.

"The Lamarkian case of behaviour altering DNA is almost certainly possible". Are you here
merely distinguishing DNA from its sequencing? So are you saying that behavior might affect
DNA, but not its sequencing?

"...behaviour altering DNA is almost certainly possible ..., but as I mentioned there are no
rules in genetics, everything is what is plausible in the context." Do I infer correctly that
behavior can alter DNA in particular cases, but we can't generalize those cases and assume
this is a general fact? 

Your example of the AIDS virus (?) suggests that random mutation rates can be "programmed".
I'm unsure if this means chemically in the lab or if, say, environmental change or diversity might
in some cases increase mutation rate.

But I have trouble following your discussion of the hypothetical long-nest wasp. It seems that
you are only saying that genetic isolation lends itself to speciation. But not in that context,
do we have any reason to assume that hypothetical longer nests will be better adapted to a different niche? Surely all the specifics of nest design can't be assumed to be adaptively
functional. Maybe a simpler crude example: small birds, similar in structure and habits, in
the same environment, built quite different nests. To argue these differences are there because
of a subtle adaptive advantage seems a stretch. More likely, it seems a genetically encoded
habitus. 

If I can summarize what I be to take your point, there is nothing to disprove that behavior
might not alter genes, but in the absence of a generally present mechanism, whether it
actually happens will depends on particular instances. 

Haines

Mohan, thanks for the reply.

"My point was that imprinted behaviour is not encoded in DNA. What is encoded is the
imprinting procedure: follow-whatever-duck-you-see-first; speak-whatever-language-you-hear-
in-your-first-few-years, etc."

Sorry, but not sure I understand. You seem to say that what is encoded are hard-wired
procedural rules, rather than behavior. If "behavior" simply means patterns of action, then
the difference with "procedural rules" is that they bring in a cause for behavior rather than
just describe a pattern of action. To use the silly example of genetically passing knowledge
of English along to your children, the difference would be that while linguistic competence
can be passed along, English is a learned behavior that takes advantage of the linguistic competence.

Humans, of course, have evolved to have a very generalized competence, particularly in
language, and some animals do likewise. Other animals have an intelligence that is hard-
wired. People argue that in both cases, it was adaptively advantageous and the
mechanism of this development was natural selection. But of course, such an argument
is an inference that cannot be proven or disproven. One criteria for the merits of an
hypothesis is whether it saves the data. The problem is, that there is data that is hard
to explain in these terms. This is widely taken to imply the need for a more universal
explanatory hypothesis.

My question about other mechanisms for genetic change was in part stimulated by
arguments that are raised about the development of functional organs, particularly the eye.
The development of the parts of the eye has to be guided by their functionality in the eye
as a whole, and those parts are not in themselves adaptive in relation to the environment,
but only their relation to each other. The eye as a whole probably developed as an adaptive response to circumstance, but the eye is a functional relation of its parts which cannot
explain their origin.

More specifically, you suggest "behaviours can be encoded in DNA because conditional
commands can be DNA encoded." OK, and then you say, "Where the antecedent of the conditional is environmentally sensitive, this demands the functional analogue of a sensor
to determine when the antecedent of  the conditional is satisfied." But in the elements of
the eye, I don't see that they can satisfy that condition, for when the parts of the eye
evolved, there was no criteria for selection except the function they would only later
acquire. This has led to some adventurous speculation for which I find little appeal.

My point about information was that evolution is an emergent process, and an emergent
process cannot be reduced to causal determinants. People evade this problem by saying
that the emergent whole "supervenes" on its base, but this is actually only a description,
a kind of Humean constant conjuncture of observables at the base and the emergent levels.
I have tried to cook up an answer to this conundrum, which I won't burden you with, but
an implication is that natural selection is a constraint on possibility, and as such novel
behavior or phylogeny is not explained by genetic "encoding", but by constraints on the
expression of genetic possibilities.

This might suggest that your distinction of procedural rules and behavior end up being
the same thing. Learning English arises from the constraint on mental and social
possibilities that are ultimately genetic possibilities, but so too are hard wired responses
to stimuli. An introduction of randomness in explanation might be operational in
thermodynamics, but philosophers find the concept rather troublesome. Instead, a mutation
would represent a different genetic probability distribution, and so the engine of evolution
would not be environmental causal determination, but actualization of possibilities.

Haines  
 

Dear Haines,
You might want to look into the "Baldwin Effect".   It is not Lamarckianism, but it is a mechanism whereby learned behaviour might become encoded in DNA.  The basic idea is that learned behaviour can alter the selection pressures on an organism in a way that provides a kind of positive feedback. So, for example, if a certain learned behaviour improves fitness, then the ability to learn that behaviour might be selected for.  But it might also be that hardcoding some of the behaviour in DNA makes the behaviour easier to learn.  In this case hardcoding of the behaviour could be selected for. 

There is some controversy about whether the conditions are ever right for this kind of thing, but as far as I know, it is generally accepted as a theoretical possibility. Sorry if I haven't explained this well, but there is plenty of literature on the topic.

Richard

Dear Haines, We do not seem to be getting much clearer. I guess it would help to know what the philosophical import of your questions might be? 

In my comment on the possibility of behaviour altering DNA, I was just referring to altering the base sequence. My point was that I could almost certainly alter the sequence of my germ cell DNA in a laboratory in a way that would vaguely approximate to what people think of as Lamarckian and there are probably other examples in bacteria or viruses that vaguely approximate in some other way but it is way outside reasonable plausibility that these are relevant to the evolution of vertebrate behaviour on a wide front. When I referred to AID I was meaning activation induced deaminase, as in my previous post, not the AIDS virus. Sorry for not being clear.

In the case of the wasps I think it is very likely that longer nests would have an effect on niche. The length might affect how often nests were blown down on southwest facing slopes or how easy it was for honey buzzards to break off chunks to carry away. If half a colony of wasps were building long and half short the nest might easily end up useless, whereas separate pure colonies would make good nests. 'Adaptive functionality' is a hugely complex phenomenon.
 
Maybe what you are getting at in the small bird example is an interesting point that struck me when visiting the Galapagos Islands (no, really, it did). It is another example of how Darwin was actually wrong about many things but deserves the credit for getting the basic idea of evolution by mutation and speciation right. (There is no suggestion that his mistakes merit a return to Lamarck.) The famous finches of Glalpagos come in about nine species - large, medium and small ground, woodpecker, vegetarian, warbler, tree etc. It is quite likely that all these birds derive from a single pair of mainland finches that got blown to Galapagos in the same storm by a freak chance. (There are only about thirty species of small non-seabird on the islands, indicating just how rarely pairs of birds of the same species arrived in the same lifetime.) If, as it seems, conditions were favourable, the pair would have had hundreds of offspring. These would have spread sparsely around the islands. Because there is normally significant allelic variation in genes for physical phenotypic traits (it may have the advantage that individuals can be recognised) the offspring will have varied a bit. Those with slightly fatter beaks would tend to congregate in areas with seeds that are very nutritious but hard to open. Those with thinner beaks would tend to congregate where there were more bark inhabiting insects that were hard to tweak out. As soon as these congregations started to interbreed locally there would be a major enhancement in the difference between groups. If there are ten genes that contribute to beak thickness, which in an outbred population are randomly mixed to give an average beak, preferential interbreeding of birds with e.g. fatter beaks (because of territory) will rapidly lead to a group with all ten genes being of the pro-fatter alleles. This may be quite enough for them to look like a new species. No natural selection in the sense of variable survival needs to be suggested. All the birds at this stage can survive. However, in order for true speciation to occur the macro-formatting of the birds' DNA must drift enough for them to become back-cross infertile. This will happen naturally by DNA structural drift (again no survival pressure) as long as the groups live in different places and have the behavioural trait of sticking to their patch. Once speciation has occurred a new force will come in. Birds with DNA that encodes the behaviour of 'chasing girls with the same size beak as you' will procreate better than birds who chase any old girl and at territorial overlaps too often pair up with a genetically incompatible mate. So from this point on mutations in behaviour encoding genes will be selected for on a survival of offspring basis (not survival of individual).

But to come back to your point "To argue these differences are there because of a subtle adaptive advantage seems a stretch. More likely, it seems a genetically encoded habitus." I am not sure why you see these as alternatives rather than both true. In conventional theory both are true. Except that differences are not there because of adaptive advantage in the sense that their existence is due to adaptive advantage. Their existence is due to mutation and their persistence is due to adaptive advantage. Feathers did not evolve because they were good for flying. They evolved by mutation and then continued because they were good for flying. Successful speciation is dependent on the random occurrence of mutations that ensure new behaviour patterns that tend to ensure selective interbreeding within a group with the same DNA macro-format (by that I mean things like the number of chromosomes the DNA is split up into).
Best
Jo

Haines:

Like Jo, I am not clear where you are going with this. Here is a menu of thoughts.

1. Posing the conundrum about the eye is a bit of a red flag, since this is a favourite example of creationists; yet, I thought it had long been solved. (I don't know the details of eye-development or evolution, though; so I won't stick my oar in about it.) But there is no general problem about the evolution of complexity, which is why I don't understand the puzzle about the evolution of flexible behaviour patterns. Maybe you could spell it out with an example.

2. Again, I don't myself believe in "emergence" unless it is a name for something quite non-mysterious like my example of the rabbit in my last post. Broad thought that water "emerged" from hydrogen and oxygen, but his quite unconvincing reason was that the properties of water cannot be predicted from those of hydrogen and oxygen. On that understanding of emergence, natural selection is emergent all the time. (The supervenience talk can be taken as legislating emergence away -- so your take on it is right, I think, though I draw the opposite conclusion from it.)

3. On the other hand, some of the things you say sound very like evo-devo: the evolution of developmental patterns. Some of the evo-devo people, Richard Goodwin, for example, think that it appeals to influences other than natural selection -- Chomsky used to be in this camp as well. Others, like Sean Carroll, think that these sorts of constraints are part of natural selection -- his book is marvelous on the topic of gene-regulation, homeobox genes, etc. -- that's my view as well, development is the product of selection.

4. On the Baldwin effect, it is not what you are looking for. Here's the idea. Take a learned behaviour B. B is advantageous, but not innate. Sometimes, though, B will pop up in the population as the result of a mutation. Because it is advantageous the newly mutated gene for B is selected for. This selected-for behaviour displaces the learned behaviour because it is more reliably expressed.

Lamarckianism is where a gene for behaviour B arises as a result of an organism evincing B. The Baldwin effect is not an example of this.

best,

Mohan

Richard, your dropping the phrase "Baldwin Effect" was what I was looking for. Thanks.

By coincidence, the next book I'll shortly be reading is Jablonka and Lamb, Evolution in Four Dimensions. In pursuing this Baldwin Effect I discovered that this book discusses it at length. I've downloaded a lot of material, going back to Waddington (whom I often don't find agreeable) and including Corning (whom I find disagreeable), but many articles are promising. As a result of your suggestion, my plate is now piled so high that I hestitate to pursue things any further until I've digested some of it.

Your point that adaptation is not passive, but active, so that behavior and circumsance are not readily separated is noted. Traditional (social) action theory presumed a mind-body (intent-world) ontological contradiction that is clearly ideological and characterizes the modern West. It has long been subject to criticism. An important point is that our action changes or defines the world in which mind operates, or as I'd rather put it, mind and action frame the world and therefore constitute the (observable aspect of) the world in which action takes place. I'm not here adopting a Wittgensteinian position, but speaking in terms of the philosophy of science, and I'm also not much drawn to "evolutionary epistemology" (such as Wuketits) because of its Kantianism.

I suspect this point is relevant to your "whether the conditions are ever right for this kind of thing". That is, if I understand correctly, you doubt whether external circumstance are often such that choices determine the environment to which the organism adapts. In contrast, the perspective I s, persuing is that observables are not intrinsic properties, but extrinsic, a function of sensory apparatus, axioms, etc. Or, more accurately, the sensory apparatus (and brain) constrain the object's probability distribution for giving rise to observable effects. This is not just pulling a philosophical fog over is what is really a simple affair, for there's a difference: the relation of organism and enviornment is not causal and mechanistically determinant, but one of processes.

To put this in more specific terms, Darwin combined Buffon (speciation) with Malthus (mechanism of selection) to conclude that the causal relation of organism and environment determines relative reproductive success. The Scottish Enlightenment (as in Adam Ferguson and Adam Smith)  reduced society to autonomous individuals who choose to establish a relation with resources accessible in (the economic) environment. Indeed, they invented a new word, "rational", to characterize those relations that happen to result in an increase in one's "talents" (wealth). This conception was contested right from the beginning, but it nevertheless came to prevail, particularly in the sciences.

I might also note as an aside that today there is no consensus of just what causation is; there is a range of competing theories, none of whichis entirely satisfactory. The traditional (positivist) notion was that explanation ultimately relied on the force of universal laws. Today this view has collapsed in principle because of the reification, and singular causality has taken its place by default. 

For example, the Foucaultian notion of Self being alien to the Other and standing in a relation of relative power, does not seem quite right either in terms of ethology or human psychology. For example, animal consciousness apparently sees the Self as another Other, not as something alien or in competition. Bacterial quorums seem to parallel representing the human individual as a social being. So what we are and how we behave is not solely an intrinsic matter (a mix of accident and possibilities arising from genes), but is a constraint of personality on social possibilities. Not only is the world to which we adapt socially constructed, but the "we" is a constraint on population possibilities. I hope I'm not appearing to argue an extreme case here, for the constraint of the individual on environmental possiblities is actually a constraint on an environmental probability distribution, and an environment that does not well support metabolism will probably in fact have an effect on reproductive success. However, the point here is the active individual which is inseparable from population, and which does its best to cope; not a passive individual being causally determined by circumstance.

Sorry for all this philosophical commentary. Some people find such speculation fun and others find it a distraction. All I can say to justify myself is that we today seem in the midst of a major "paradigm" shift in which the ontological presuppositions of the modern West seem to be yielding  to a new conception. If so it calls for being imaginative and bold. I'm reminded here of Henri Bergson, who argued that  physics and chemistry (which he wrongly saw as necessarily employing a mechanistic method appropriate to closed systems) had to be supplemented by "philosophy", by which he meant an intuition inferred from the bio-evolutionary record that living systems are emergent and escape reductionist explanation. In fact, there has always been challenges to the Newtonian mechanistic world view, from Rousseau, German Romanticism, Heidegger, etc. But again, I wander at your expense.

Haines      

Mohan,

That the evolution of the eye is appealed to by creationists does put discussion under a political cloud, but the puzzle remains, at least in my mind. You note that it has long been solved, and I hope that in my readings I'll encounter that solution. Until then I can only stick with the conventional criticism of functionalism in that function is a result rather than cause of development of the constituents of an organ.

I didn't realize that my reference to the evolution of flexible behavioral patterns was at all problematic. For example, humans inherit an extraordinary linguistic potential that can be expressed in a wide variety of ways though social learning. Humans also adapt to the most extreme kinds of niches. To judge from the case of the famous Alex the Parrot, parrots also have a capacity to use language similar to that of humans (but without much extrasomatic transmission). Alex can tell you when you are wrong, can pose questions, can create vocabulary and insist that you use it, etc. and so parrots could teach each other an expanded use of language, but they choose not to. Most parrots have no need to actualize this potential beyond a very elementary level. Often this inherited flexibility is called "adaptability". Perhaps I don't quite get the point of your question.

As for "emergence", this notion has evolved in unhelpful ways. According to my OED, it only means to become visible, to appear, become known. In other words, the word roughly refers to novelties. But as radical empiricism came to prevail, it was defined in epistemological rather than ontological terms. Pepper (1926) was not the first to shift its meaning to refer only to what is unpredictable and in contrast to categorize mere novelty as unequivocally determinant. This division between ontological mechanism and epistemological unpredictability has bedeviled the discussion ever since. My perverse view is that all novel outcomes are emergent in the sense they are improbable (except in the hypothetical limiting case of closed systems) and therefore cannot be absolutely reduced to initial conditions (which is why we learn about standard deviation in school). Unpredictability is not just an artifact of ignorance, but is in principle de re natura (yes, statistical causality boils down to causality, but the issue is really one of modal realism). So I do agree that natural selection is emergent all the time, but I thought the bone of contention is often over whether environmental conditions cause the genetic distribution in a population.

At the moment I happen to be reading François Jacob, The Possible and Actual (1982) I disagree with his defining the "possible" in epistemological terms as a mental counterfactual, rather than in ontological (scientific realist) terms. He says that natural selection "does not act merely as a seive, eliminating detrimenal mutations... It integrates mutations and orders them into adaptively coherent patterns... It is natural selection that gives direction to changes." Perhaps this is what you meant regarding eye formation. In any case, I need to explore the mechanism involved here.

In my own field this issue takes the form of a debate over historicism. 19th-century historicism not only held that an explanation of a particular situation requires that it be understand as part of a greater whole (which in itself remains a causally mechanistic view), but that the history by which the whole came about has a determinative effect on the situation. In our post-modernist era, this is dismissed as "grand narrative" and "meta-theory". This objection seems right to the extent that the past (what is separated from the present in time) does not exist in the present and so can't have any causal influence in the present (our mental representations of the past, are not a reflection of the past, but a construction that is to a degree compatible with the framed traces that happen to exist in the present and are shorn of unobservables). However, there is this to be said about historicism: the effect of the past is the traces that survive into the present, and these do constrain action in the present.

In other words, Jacob seems to be pointing out that the genetic code is not intelligible simply as an isolated constraint, but to understand it we need to know how it arrived in its present condition and therefore its history. I find this unpersuasive, in that the past does not exist in the present except as a mental inference, and like a Markov process, the present is ignorant of its past.

However, it seems to me, on the contrary, that history is in fact a determinant of the present in terms of the depth of its contradictions. Another tricky word, but I take "contradiction" to be the unity and interdendence of processes that are opposite with respect to the direction of their change in probability, or as it is more commonly put, change in their entropy. Mario Bunge is one among many who pursue this line. The depth of a contradiction exists in a present as the relative probability, speed and magnitude of possible change, and this cannot be directly experienced in the present, but only known from an inference from past change. One might not find this objectionalble in epistemological terms, but as a realist, I assume that depth of contradiction is real, which is to say, makes a difference that is independent of our consciousness. In the early stage of early civilizations, there is typically greater inventiveness but not much sophistication, for it actualizes a broad probability distribution; with maturation, its inventiveness declines but it finds itself in a very sophisticated (improbable) state. This is called "locking in" in general systems theory. However, knowledge based on evidence for one moment in time, cannot convey whether the civilization is immature or mature, but one has to infer it from the degree to which the civilization historically was able to cope.

To return to your point, for "emergence" to be a useful concept, it must be explained in naturalistic terms, and whether some novelty is predictable or not is really an issue of epistemology. There is much interest in how nature gives rise to novelites that are not reducible to a prior state of affairs. To mention some popular works, David Layzer, Cosmogenesis (1990), David Bohm, Wholeness and the Implicate Order (1980), Stuart Kaufmann, Investigations (2000), Eric Chaisson, Cosmic Evolution (2001), etc. I'm not suggesting I entirely agree with these authors, but they serve to hint that a new way of representing our  world seems in formation.

You bring up evo-devo. If this is taken to mean that genes are constantly being reconstructed, I didn't realize it was a controversial issue. I'll do some reading on evo-devo, Richard Goodwin and Sean Carroll. I get the feeling that Goodwin may be hung up on the dynamics of far-from-equilibrium systems involving negative and postive feedback to give rise to extraordinary trajectories, and,  if so, I might find myself a bit critical (I've never understood the point of chaos theory, and I distance myself from general systems theory to the extent it implies closed systems). But you have pointed me in the direction of some useful reading, and I thank you.

On your characterization of the Baldwin Effect, what it seems to amount to, to put it in the language of the philosophy of causality, is that we can't  infer a causal relation between regularly proximate events because both may be effects of a different cause (what Wesley Salmon calls pseudo-causality in The Causal Structure of the World, among many others who discusses this point). If so, then you are right that it is not what I'm looking for.

But I also hesitate to characerize what I am looking for as support for Lamarckianism, given the undertainty of its definition. Jean Marx: a change acquired in response to an altered environment becomes hereditary. Gordon Kaplan: the essence of Lamarckianism is often imprecisely summarized as the inheritance of acquired characteristics, the stable inheritance of adaptive changes induced in an individual organism by an altered environment. Pietro Corsi: (I've not read his book) Lamarck felt that change in circumstance induces change in habits and thence change in form. It strikes me that the ambiguity of a) environment causing change in an individual's genes through some yet unknown mechanism, b) individual adaptation of organism and its environment causes change in that individual's genes, c) the habitus or behavior of the individual, regardless of whether it is adaptively advantageous, brings about that change. So is Lamarckianism addressing behavior in general, only adaptive behavior, or merely circumstance as a causal factor?

But it strikes me that the basic issue here is really over the unit of selection. Lamarckianism seems to imply an individual organism, but the neo-Darwinian synthesis sees population as the unit. In the latter, it seems one can legitimately speak of population genetic change as an effect of adaptation. However, the issue remains that reproductive sucess may have other determinants, such as income, or China's one-child policy, or geographic distribution. Then genetic drift would not simply reduce to relative environmental adaptability. That the economy is ultimately determinant means that it is the ultimate souce of all possibilities, not that it is mechanistically causal for all aspects of life.
 
Do you think I am right to infer that the basic issue is over the unit of selection? If the unit of the individual organism happens to be the focus, I suspect the question boils down to the influence of behavior, and to suggest that individual behavior (short of exposing oneself to dangerous radiation, etc.) can affect genetic structure seems a big stretch.

But assumed here so far is a static individual. In fact the individual is a process that originated from cell reproduction and continues to replace cells, and so my original question was whether this reproduction within the individual's life span might not be subject to internal selective constraints that have nothing to do with the environment or adaptation to maximize metabolism? If this is so, are then the sperm and ovum also subject to such constraints? (my ignorance) If so, the genetic code passed on to descendents can be affected by intrinsic constraints. Whether such intrinsic constraint can be stretched to include learned behavior, though, seems a bit much (such as behavior spawning a protein that somehow finds it way to block gene expression in the reproduction of spem or ovum cells).

Haines

Haines,

Though some wires are getting uncrossed, I think that you and I may still disagree about the evolution of flexible behaviour. I don't think that, in and of itself, it raises any problems for straight natural selection.

I may have been a bit harsh in my earlier post with regard to the Baldwin effect. The idea is that the prevalence of a particular learned behaviour makes it advantageous for parts of that behaviour to become innate.  Here's David Papineau 2005 on the subject:

"By way of an example, consider the woodpecker finches again, and suppose that there was a time when their tool-using behaviour was not innate but socially learned.  That is, young woodpecker finches would learn how to use tools from their parents and other adepts.  Now, this socially transmitted culture of tool use would give a selective advantage to genes that made young finches better at learning the trick.  For example, it would have created pressure for a gene that disposed finches to grab suitable tools if they saw them, since this would give them a head start in learning the rest of the grub-catching behaviour from their elders.  But this gene wouldn’t have been advantageous on its own, in the absence of the tool-using culture, since even finches with that gene wouldn’t have been able to learn the rest of the tool-using behaviour, without anyone to teach them."

Papineau (who notes, by the way, that Baldwin was no Lamarckian) reasons that in the above case, the acquisition of "a gene that disposed finches to grab suitable tools if they saw them" would have no selective advantage if this didn't streamline the learning/teaching process -- so the selective advantage depends on the existence of the teaching/learning process.  

My earlier claim does hold good of the Papineau analysis. The learned behaviour does not causally affect the genome. That's why it is not Lamarckian. What I overlooked is that the existence of the learned behaviour affects the fitness landscape.

I am not sure I understand your idea about units of selection. It's certainly true that in the neo-Darwinian synthesis, the equations deal with gene-frequencies, and for this reason the population is the target of analysis. Nevertheless, causation is at the individual level even in the Darwinian model. However, this claim has been thought controversial. For a brief polemic see my 2010.

Finally, evo-devo. In my understanding, this is not the idea that "genes are constantly being reconstructed," but the idea that major evolutionary transitions occur when selection works on genes that regulate development.

best,

Mohan

Dear Haines,  
I don't believe that adaptive behaviour can ever be 'assimilated' directly. This is the so-called Baldwin Effect. However, I have suggested that adaptive behaviour can  'prime' the genome. It then needs a simple trigger from the environment of the organism to be manifested. Please note that this is very different from epigenetic inheritance which is on the 'easy' (ie genetic as distinct from somatic)  side of the Weismann Barrier.

Please see http://tinyurl.com/l6bure for a little more detail about Genetic Priming.

John Jacob Lyons

Dear Haines,
You seem to be returning to feet on earth in the last response to Mohan, thank goodness.


The basic issue is certainly the unit. Lamarck proposed that adaptive evolution occurred within individuals and post-Darwinism puts it in populations. The problem with Lamarckism is twofold. Firstly, germ cell (sperm,ova) DNA is pretty immune to effects of the individual's behaviour or resultant environmental shifts. Secondly, and more importantly, any effect of environmental shift on germ cells has a vanishingly small probability of inducing a genetic change of any relevance at all to that environmental shift. Heating drosophila may make their wings change shape and this may be passed on to progeny. However, it is inconceivable that a giraffe stretching its neck to a new type of leaf will lead to chemicals in the leaf altering germ cell DNA in just such a way that the progeny have necks just that big longer enough to eat that type of leaf. It is this second problem that makes Lamarck so implausible.


I would just add that the issue is probably not the unit of selection. In Lamarckism there is no mention of selection at all. Things just get better. And as I intimated in relation to the finches, even the post-Darwinian view does not really need selection.  Evolution could still occur without any selection if there was an infinitely favourable environment. There would just be a lot of simple species around as well as complex ones. Nevertheless, in the context that you have revealed as your interest, selection of individuals as a driver of population change clearly is the crux. (I am not sure what selection of population units would mean.)
 
Best 
Jo

Mohan,

"I think that you and I may still disagree about the evolution of flexible behaviour." Goodness! I wouldn't venture to disagree in a field in which I'm totally ignorant. I'm just exploring possibilities. In fact, I don't see how a resolution of the issue is likely to have any effect on my own work. However, your citation of Pappineau hits a bit closer to home, for I've followed his work on naturalism and physicalism and generally find his views sympathetic.

His characterization of the Baldwin Effect is what I had assumed it to be, and it does not seem particularly troublesome (or helpful on some core issues). On the other hand, I'm still in the dark as to what is implied by saying he is not Lamarckian, for according to Corsi, I gather, there can be no definitive definition to support a categorical statement.

My impression of Enlightenment thinking is that it breaks with the traditional presumption of static forms to posit their progressive plasticity. I suspect the new science of geology was involved here.  Buffon seems to fall within this view in that he felt that external circumstances had some influence on the evolution of physiology, and certainly there at the time a strong element of geographic determinism regarding human nature. On the other hand, I'm not sure about Enlightenment embryology. If I remember correctly, it held that that the emergence of an organism is an actualization of a hidden miniature form, a humunculus, hidden in the seed. Although this humunculus was by then a physical rather than ideal form, it seems to be static.

So at the time the issue in a sense was a combination of embrology and geographic determinism by which cicumstance might influence a plastic inner germ. If I understand correctly, Lamarck merely injected the mediation of habitus into this as an influence on the plastic germ, although I gather he still felt that held that the germ was shaped by internal factors as well. Mendelian genetics, I assume, saw "germ plasticity" as a combination of randomness and direction in that the possiblities of genetic randomness are constrained by the past state of the gene. (I'm summarizing just to be sure you don't have to correct me). It seems the gene gives directionality to evolution by conveying the constraint of the past as well as opens possibly by due to an random contingency of the present (mutations). I guess my concern right from the beginning is an attribution of possibility to a reified randomness.

The Baldwin Effect seems related to a population unit because the extrasomatic tool use of finches, to use Pappineau's example, is a social artifact and natural selection operates on the relative reproductive success within a population. So this is not the old view (Lamarck, for example), where the habitus affects the individual organism. Thanks for straightening me out on evo-devo.

I'm on alien turf when it comes to your 2010 polemic over drift, but I believe I understand Millstein's distinction of drift as outcome and drift as process. Usually, statistical outcomes are descriptions after the fact (such as the half-life of radio-isotopes) rather than explanations. The notion of "process" can (and usually is) also defined in epistemological terms as a mental inference after the fact: a state of affairs represents an identity that undergoes accidental changes in time. In thought a process is an observable state of affairs represented in memory as a persistent identity that differs from another observable state of that same identity, which the mind projects onto an imaginary time dimension to avoid logical contradiction. There is, of course, also a philosophical problem with the essentialism implied by persistent identity (I'm inclined rather to what is called a structural explanation of persistence, but that's not a popular position). In contrast, there is a long tradition of an ontological definition of process that in the West goes back to Artistotle, is present in Bergson, implicit in the American pragmatists, and abstractly in Heidegger, but not so clearly in Whitehead and Hartshorne, and definitely not in the current standard work of Rescher.

This ontic notion of process is a unity of real possibility and actuality, and so it it entails modal realism (modal realism has become quite the fashion these days, but primarily in the form of David Lewis' plurality of possible worlds, which is not much help here). My belief is that we have direct experience of process in our action, but we cannot represent it adequately in thought due to limitations of mind (like Augustine speaking of time). So to make process operational in thought, we have to break it down into modal "aspects". I'm inclined to do this in terms of three aspects: exogenous possibilities, extrinsic potencies (the combination of which is a real probability distribution), and local actuality, and I define each aspect as anchored on the others. With this in mind, I was not clear about your difference with Millstein on drift. You seem to define drift in epistemic terms, and perhaps Millstein wants to lend it ontic meaning. I have an intuition here that the word drift is necessarily epistemic, and if so Millstein needs to find another word for what becomes manifest after the fact as drift.
 
Well, again, I must apologize for submitting you to my biological ignorance and philosophical speculation.

Haines

John,

Thanks for your interesting comments on genetic assimilation. I have the intuition that "assimilation" is a more cautious term than causal determination, and that your putting assimilation in quotes reflects an even greater caution.

The basic issue seems to be the Weismann Barrier. In the absence of any explanatory mechanism, I get the impression the Weismann Barrier was originally only a hypothetical generalization, perhaps an ideological implication of Western social atomism. But since the 19'70s, microbiologists better understand the physics involved, and some have concluded from experimental results that the Weismann generalization may only be approximately true or even open to question.

You note a correlation between behavior and genetic propensities. While the word "propensities" is these days a philosophical swamp, I'm sympathetic to it in principle because it implies that a propensity, although an unobservable, is a real property of the genome. One cannot infer a propensity from a finite number of observations. That is, It is incompatible with traditional/positivist/radical empiricism (which these days might give it a certain credibility). However, you don't explore the implications of this, but instead turn to what seems to be the Baldwin Effect, which seems not to violate the Weismann Barrier generalization. Is this so?

However, in your subsequent comment, I am left uncertain. Are you objecting to Waddington's use of the Baldwin Effect rather than the Effect itself as generally understood (merely as a correlation of behavior and genetic code in a population)?  But then I'm unclear about your citation of Waddington's observation about Drosophila, where environmental change apparently have an impact on genes. Are you saying this violates the Weismann Barrier, which you had previously honored? Or is your point instead that environemtal change is not the same as behavior? (a point I would be inclined to contest). 

I gather most discussion these days of penetrating the Weissman Barrier has to do with reverse transcription. You don't mention it, and I don't know if this is because you see these cases where the soma changes the gene as being irrelevant to the issue of behavior or if (except with virii) it is not somatic.

I'm unfortunately ignorant about all these things. I get the impression that a protein that blocks
gene expression is in principle irrelevant because it does not change the gene itself, and so does not, I suppose, have implications for inheritance. But can an enzyme like reverse transcriptase that serves as a template for the construction of DNA, be considered a somatic effect? If so, then the next question would be whether it is conceivable that the enzyme might be affected by behavior or environment. I've never see a suggestion that it can (outside the exceptional case of AIDS).
  
Haines

Dear Haines,

1. You "confess" (that's in quote marks, because when philosophers say they are in the dark, they mean that their interlocutor has not been sufficiently illuminating!): "I'm still in the dark as to what is implied by saying he is not Lamarckian".

Here's what Papineau says on the subject:

When James Mark Baldwin and others first posited the Baldwin effect over a hundred years ago, their concern was precisely to uncover a respectable Darwinian mechanism for the Baldwin effect. The great German cytologist Augustus Weismann had already persuaded them that there is no automatic genetic inheritance of acquired characteristics: the ontogenetic acquisition of a phenotypic trait cannot in itself alter the genetic material of the lineage that has acquired it. The thought behind the Baldwin effect is in effect that an alternative Darwinian mechanism might nevertheless mimic Lamarckism, in allowing learning to influence genetic evolution, but without requiring Lamarck’s own discredited hypothesis that learning directly affects the genome. (my emphasis)

The understanding of Lamarckianism implied in the italicized sentence above -- that it is directly contradicted by Weismann -- is what I have been working with in this thread.

2. You say: "The Baldwin Effect seems related to a population unit because the extrasomatic tool use of finches, to use Pappineau's example, is a social artifact and natural selection operates on the relative reproductive success within a population"

Two points:

First, whether learning is social or not -- i.e., whether there is teacher and learner, as opposed to just a single individual learning from her own experience -- is not relevant to the issue of units of selection. As far as the Baldwin effect is concerned it is the result of the learning process -- individuals who can use a twig as a probe -- that is important. For it is the individual pupil and the individual teacher who are advantaged by the development of the capacity to recognize twigs

Second, in my view, the fact that population genetics describes reproduction rates in a population does not imply that it is committed to population level causation. In fact, the theory of natural selection is inimical to population level causes. This is the point that I was making in my last post, and against Millstein.  Broadly speaking, I agree with Jo when he speaks of selection of individuals as a driver of population change. (I would, however, demur when he uses the causal metaphor 'driver': in the theory of natural selection population change is a mathematical, not a causal, consequence of the births and deaths of individuals.)

(Sorry for all the italics in the last para. I wasn't shouting: I just couldn't seem to normalize the font using PhilPaper tools!)

cheers,

Mohan

Mohan, thanks for trying, but I'm still uncertain.

Your discussion of the Baldwin Effect seems to be what I thought it to be: an observed correlation between behavior and genotype. You point out that the aim of Baldwin was to explain this correlation of behavior and genotype in Darwinian terms, even though it might superficially seem to mimic Lamarckianism. I guess he did pretty well at it, but this seems to lend support to Darwinism rather than to disprove Lamarckianism. I don't think much of Popper, but he would ask, can Lamarckianism be falsified?

Now, as for the Weismann Barrier, which contradicts Lamarckianism, I don't know how Weismann  arrived at it. For example, did he repeatedly cut off the tails of mice to see if that might have a long-term effect on heredity? If he did something like this, his Barrier only represents a generalization of experience. This is how "law" seems now understood in the philosophy of science, where explanation increasingly relies on singular causality, a desciption of the operative mechanism in particular cases. Without at the time having any idea of the physical mechanisms that might be involved, Weismann's Barrier would seem only a generalization of experience, not a real universal determinant. That Weismann's inferences contradict those of Lamarck does not seem relevant to discussions today, which instead is over the mechanisms involved. 

I suspect the issue of unit is perhaps only a matter of perspective and so won't belabor it here, for I'm agreeable to what you say. I also put aside the question of the extent to which adaptation is the sole or principle determinant of reproductive success.

Instead, what I'd like to do is to try to clear up a bit a comment I made earlier. It seems that genotype is traditionally seen as an "entity", a closed system, a self-sufficient structure. However, it is also said that genotype has an effect on phenotype. If so, and if genotype is self-contained, then surely it must have free energy. Maybe Morowitz gets into this, but for the moment I'll assume that the genotype, once formed, is not entirely self-contained. Instead, it could be represented as one aspect of a broader process. That is, a probability distribution that is not actualized until it enters a relation with its environment (phenotype). This also seems to be what people often imply. However, if the genotype is a probability distribution that is actualized by the constraint of circumstantial structures, then the inside-outside ontological contradiction seems to break down, and the  Weismann Barrier issue might be mute, for genetic possibilities are always actualized by external constraints.

But the question remains why genotypes changes. Here random mutation steps in, but I raised some objection to it, and perhaps I should elaborate. If we take randomness to be the equivalent of non-determination, it then implies there are things or properties that are not contingent. This seriously runs the danger of an ontological dualism, of an objective idealism. However, modern Western science depends on its assumption of a coherent universe (where "coherence" might be variously defined). For something to be random seems to separate it from the mutual relations that define coherence. So I suppose that "random mutation" is only a way of speaking, and we lack the divine omniscience necessary to understand the determinations (not necessarily causal) at work. This abstruse point might have some relevance, for if mutations are really random, then by definition they are not influenced by phenotype; if the determination of their change is simply beyond our ken, then there is no objection in principle over the possibility of phenotypical influence on mutation. In other words, I suspect the whole issue of Lamarck vs. Darwin may only be epistemological, an effect of how we frame the world or ideology.

Haines  


 

Haines,
If you are going to do some reading on the Baldwin Effect, can I suggest you look at
Hinton and Nowlan (1987) "How Learning Can Guide Evolution" Complex Systems 1:495-502

This article is short and sweet and it is responsible for rekindling contemporary interest in the Baldwin Effect. It has a nice discussion of exactly how the effect works, and includes a computer simulation which shows that the strength of the effect could be quite strong. I believe the article has been re-printed in numerous collections.

Cheers,
Richard.

Richard, your citation much appreciated, but I can't find it either in WorldCat or in the article databases accessible to me. But I do see Weber and Depew's, Evolution and Learning: The Baldwin Effect Reconsidered (2007), which might serve me just as well because it includes the history and changes in the idea, including its revival and interaction with the modern synthesis.

However, the rather abstract and philosophical question I've been raising does not speak for or against the Baldwin Effect. Roughly, it is a methodology that assumes that a) in principle all things are processes, b) (more contentiously) explanation must be in terms of processes rather than, say, a description of a causal mechanism uniting entities. To represent things as processes I have argued that it requires a unification of three modal aspects: a) structure (the local and observable effect of the past), b) real exogenous (all levels are seen as constraints of a structure on the possibilities of a more universal level) possiblities that are constrained by structure as a probability distribution, and (Morowitz would insist) potency, which I define as a probability gradient to anchor it to structure.

For example, a gene can be described, but to use it in explanation requires that it be seen as part of a broader process that engages unobservables: to see it as a structure (the local and observable outcome of the past), coupled with a probability gradient (to drive change), coupled with possibilities (inherited from the more universal level on which the living organism represents a constraint). Now, all this is obviously obscure and may seem gratuitous, but the implication is not so unconventional: the gene is only a (local, observable) structure that is only one aspect of a broader process, and this broader process is the unit of analysis necessary for explanation.

For example, this suggests that whether "information" can pass both ways across the Weissman Boundary is a false question, for it reduces a process to the causal or functional relation of the (static, local, observable, closed) entities of genotype and phenotype.

I admit that I've been imposing some vague and general conceptual garbage on rather specific empirical questions, and this does tend to invite trouble. However, I beg everyone's indulgence, for I'm not trying to reconstruct microbiology, but merely exercising a methodology by trying it out in a field entirely unrelated to the field for which it was developed.

Haines

Hinton and Nowlan can be found here: http://tinyurl.com/247ojew

The take-home message:

"Most biologists now accept that the Lamarckian hypothesis is not substantiated; some then infer that learning cannot guide the evolutionary search. We use a simple combinatorial argument to show that this inference is incorrect and that learning can be very effective in guiding the search, even when the specific adaptations that are learned are not communicated to the genotype."

Mohan

Thanks, I've got it (but not yet read).

I'll read Hinton and Nowland as I proceed with Jablonka and Lamb, but at first glance, they may not disagree. Hinton and Nowland define "Lamarckianism" to mean a causal effect on genotype by phenotype, but I get the feeling that a reduction of phenotype to merely an expression of genotype is now longer much held and may not be exactly what Lamarck had in mind. I get the feeling there is wide agreement that there is a range of structures, both inherited and acquired, that constrain and actualize the possibilities offered by genotype. That is, the issue is not the Weismann Boundary, but rather a reduction of a complex emergent process to just genotypic determinism. But at this point in my reading, I'm just guessing.

Thanks for the link.

Haines

Some of my friends like Jablonka and Lamb. You'll have to make your own mind up. Proceed at your own risk.

As for HInton and Nowland, their point is interestingly like Papineau's (a couple of decades later). Suppose you have to put two things -- a and b -- together in order to achieve a functioning unit. This faces the notorious pre-adaptation problem -- a and b have no adaptive value by themselves.  Suppose that an organism mutates to a, and learns b.  Then it has the a-b advantage, waiting for a mutation to b.

So learning b changes the adaptive landscape, and thus paves the road for selection subsequent to a b-mutation.

best,

Mohan

Mohan,

I infer from your comment that Jablonika and Lamb may be a bit controversial and not everyone is particularly sympathetic to them. I've just started J&L, but they seem to begin with an argument that what is involved is an emergent process that does not lend itself to genotype reduction, and various external constraints actualize genotypic possibilities in various ways. I get the off-hand impression that the evo-devo folks tried to address this issue in terms of general systems theory, such as with non-linear feedback--an approach that over the years has lost its appeal for me.

The pre-adaptation problem is one I clumsily tried to bring up in connection with the evolution of the eye.

I've got some good reading ahead (can't dedicate too much time because it distracts from my research and writing in an entirely different field), and I want to thank you for helping me toward defining the landscape.

Haines

I get the feeling that we have exhausted this topic (or are exhausted by it), but I'm left uncomfortable because I allowed philosophical issues in by the back door and hence in a rather incoherent way. If my insistence might is excused, I'd like to summarize my philosophical stand lest it be misunderstood.

Basically I am criticizing a conventional notion of change in terms of the causal relation of entities (defined by their intrinsic properties such as genotype and phenotype). I won't belabor that criticism, but only offer an alternative to it. This is to see change as arising, not from a relation of entities, but as a relation of processes.

While this is hardly an unconventional suggestion, the usual definition of process is rather problematic: a given state of affairs that changes in time. This "state of affairs" is simply an entity defined in terms of its intrinsic properties, and so it does not really offer anything new. I instead define process as a relation of real modalities, of which only one (structure)  is defined by its intrinsic properties. The two other modal aspects of process (potency and possibility) are non-local. This, too, has sometimes been argued, but I try to anchored them on structure. I'll not elaborate this point here.

A relation of processes therefore becomes fundamentally different than a causal relation of entities, which becomes a hypothetical (absolutely closed) limiting case. To emphasize the distinction, I use the term processual superposition (with no implications of QM) for a relation of processess. In terms of it, change arises from the constraint of both structures on the probability distribution of the other that is actualized through their mutual framing.

Simply put, change can't be reduced to an expression of intrinsic properties or even a dialectic of intrinsic properties, and so therefore the question of whether or not traffic across the Wiesmann Boundary is two-way is a false question, for boundary implies closure, and no process is absolutely closed.. Now, I suppose that in most cases, a natural selection of random mutations is a close enough approximation, but its success does not preclude the constraint of phenotypical structures on the outcome, and to judge by some recent experimental work, these other constraints can sometimes play a significant role.  So the neo-Darwinian synthesis in these terms is not incorrect, but needs to be seen instead within a framework more universal than biological evolution as a typical case near the limit. I can't presume to take a stand on either side of the debate, but I believe I can suggest an alternative "ontology".

My comments are not meant to drag out the discussion, but merely to help avoid misunderstanding what I have already said.

Haines Brown

I'm well out of my depth here but I'd like to just mention Schroedinger's 'faux-Lamarckism' of which I'm a fan. 'Use it or lose it' seems to sum it up. The mutations may be 'random' but their survival would be behaviourally determined. He proposes that our ability to walk is the result of our wanting to walk, not simply the result of our being given the tools to do so. I'm not sure if this adds anything to the discussion.   

He also says some interesting things about genetic mutations, pointing out that the locus of the mutation is too confined for the statistical laws of physics to apply. I've never heard another physcist comment on this and wonder whether he's right. 

In my own field of biology, "epigenetics" has a precise definition, specifically variation in DNA methylation, DNA oxidation, or histone acetylation. All three of these can modify gene expression. Differences in DNA methylation can alter histone acetylation and acquired changes in DNA oxidation can influence DNA methylation. It has been recently determined that postnatally aquired modifications of DNA methylations can be inherited on a "permanent" basis (they do not "re-set" during the imprintation stage of development but retain the "acquired" state). In addition to chemical or other "physical" exposures altering DNA methylation and oxidation, in some cases, "behavioral" differences, such as postnatal maternal care, can alter DNA methylation.

Therefore, if a "behavioral" condition, such as maternal care, can alter DNA methylation, postnatally acquired changes in DNA methylation can be inherited and do not automatically re-set, the possibility of inheritence of characteristics due to "behavioral" situations is not implausible, althought it is likely to be rare.

Hi Haines,

Please see my recent posts in the discussion section of this blog. If you wish to, please reply to my email address; [email protected]

http://scienceblogs.com/evolution/2011/05/whats_evolution_got_to_do_with_1.php

https://www.academia.edu/2551188/GENETIC_PRIMING_HOW_ADAPTIVE_BEHAVIOUR_SHAPES_THE_GENOME
This paper that I wrote for The Linnean Society relates to the issue you raised. Best wishes, John

I consider myself a neo-lamarckian, as stated in my book "Darwinland".