Now turn to the other main Lamarckian principle, the idea that such acquired characteristics are then inherited by future generations. All evidence suggests that this idea is simply false, but through most of history it has been believed to be true. Lamarck did not invent it, but simply incorporated the folk wisdom of his time. In some circles it is still believed. My mother had a dog who occasionally affected a limp, holding up one hind leg and hobbling on the other three. A neighbour had an older dog who had unfortunately lost one hind leg in a car accident. She was convinced that her dog must be the father of my mother’s dog, the evidence being that he had obviously inherited his limp. Folk wisdom and fairy tales are filled with similar legends. Many people either believe, or would like to believe, in the inheritance of acquired characteristics. Until this century it was the dominant theory of heredity among serious biologists too. Darwin himself believed in it, but it was not a part of his theory of evolution so his name is not linked to it in our minds.
If you put the inheritance of acquired characteristics together with the principle of use and disuse, you have what looks like a good recipe for evolutionary improvement. It is this recipe that is commonly labelled the Lamarckian theory of evolution. If successive generations toughen their feet by walking barefoot over rough ground, each generation, so the theory goes, will have a slightly tougher skin than the generation before. Each generation gets an advantage over its predecessor. In the end, babies will be born with already toughened feet (which as a matter of fact they are, though for a different reason as we shall see). If successive generations bask in the tropical sun, they will go browner and browner as, according to the Lamarckian theory, each generation will inherit some of the previous generation’s tan. In time they will be born black (again as a matter of fact they are, but not for the Lamarckian reason).
The legendary examples are the blacksmith’s arms and the giraffe’s neck. In villages where the blacksmith inherited his trade from his father, grandfather and great grandfather before him, he was thought to inherit the well-trained muscles of his ancestors too. Not just inherit them but add to them through his own exercise, and pass on the improvements to his son. Ancestral giraffes with short necks desperately needed to reach high leaves on trees. They strove mightily upwards, thereby stretching neck muscles and bones. Each generation ended up with a slightly longer neck than its predecessor, and it passed its head start on to the next generation. All evolutionary advancement, according to the pure Lamarckian theory, follows this pattern. The animal strives for something that it needs. As a result the parts of the body used in the striving grow larger, or otherwise change in an appropriate direction. The change is inherited by the next generation and so the process goes on. This theory has the advantage that it is cumulative — an essential ingredient of any theory of evolution if it is to fulfil its role in our world view, as we have seen.
The Lamarckian theory seems to have great emotional appeal, for certain types of intellectual as well as for laymen. I was once approached by a colleague, a celebrated Marxist historian and a most cultivated and well-read man. He understood, he said, that the facts all seemed to be against the Lamarckian theory, but was there really no hope that it might be true? I told him that in my opinion there was none, and he accepted this with sincere regret, saying that for ideological reasons he had wanted Lamarckism to be true. It seemed to offer such positive hopes for the betterment of humanity. George Bernard Shaw devoted one of his enormous Prefaces (in Back to Methuselah) to a passionate advocacy of the inheritance of acquired characteristics. His case was based not upon biological knowledge, of which he would cheerfully have admitted he had none. It was based upon an emotional loathing of the implications of Darwinism, that ‘chapter of accidents’:
it seems simple, because you do not at first realize all that it involves. But when its whole significance dawns on you, your heart sinks into a heap of sand within you. There is a hideous fatalism about it, a ghastly and damnable reduction of beauty and intelligence, of strength and purpose, of honor and aspiration.
Arthur Koestler was another distinguished man of letters who could not abide what he saw as the implications of Darwinism. As Stephen Gould has sardonically but correctly put it, throughout his last six books Koestler conducted ‘a campaign against his own misunderstanding of Darwinism’. He sought refuge in an alternative which was never entirely clear to me but which can be interpreted as an obscure version of Lamarckism.
Koestler and Shaw were individualists who thought for themselves. Their eccentric views on evolution have probably not been very influential although I do remember, to my shame, that my own appreciation of Darwinism as a teenager was held back for at least a year by Shaw’s bewitching rhetoric in Back to Methuselah. The emotional appeal of Lamarckism, and the accompanying emotional hostility to Darwinism, has at times had a more sinister impact, via powerful ideologies used as a substitute for thought. T. D. Lysenko was a second-rate agricultural plant breeder of no distinction other than in the field of politics. His anti-Mendelian fanaticism, and his fervent, dogmatic belief in the inheritance of acquired characteristics, would have been harmlessly ignored in most civilized countries. Unfortunately he happened to live in a country where ideology mattered more than scientific truth. In 1940 he was appointed director of the Institute of Genetics of the Soviet Union, and he became immensely influential. His ignorant views on genetics became the only ones allowed to be taught in Soviet schools for a generation. Incalculable damage was done to Soviet agriculture. Many distinguished Soviet geneticists were banished, exiled or imprisoned. For example, N. I. Vavilov, a geneticist of worldwide reputation, died of malnutrition in a windowless prison cell after a prolonged trial on ludicrously trumped up charges such as ‘spying for the British’.
It is not possible to prove that acquired characteristics are never inherited. For the same reason we can never prove that fairies do not exist. All we can say is that no sightings of fairies have ever been confirmed, and that such alleged photographs of them as have been produced are palpable fakes. The same is true of alleged human footprints in Texan dinosaur beds. Any categorical statement I make that fairies don’t exist is vulnerable to the possibility that, one day, I may see a gossamer-winged little person at the bottom of my garden. The status of the theory of the inheritance of acquired characteristics is similar. Almost all attempts to demonstrate the effect have simply failed. Of those that have apparently succeeded, some have turned out to be fakes; for example, the notorious injection of Indian ink under the skin of the midwife toad, recounted by Arthur Koestler in his book of that name. The rest have failed to be replicated by other workers. Nevertheless, just as somebody may one day see a fairy at the bottom of the garden when sober and in possession of a camera, somebody may one day prove that acquired characteristics can be inherited.
There is a little more that can be said, however. Some things that have never been reliably seen are, nevertheless, believable insofar as they do not call in question everything else that we know. I have seen no good evidence for the theory that plesiosaurs live today in Loch Ness, but my world view would not be shattered if one were found. I should just be surprised (and delighted), because no plesiosaur fossils are known for the last 60 million years and that seems a long time for a small relict population to survive. But no great scientific principles are at stake. It is simply a matter of fact. On the other hand, science has amassed a good understanding of how the universe ticks, an understanding that works well for an enormous range of phenomena, and certain allegations would be incompatible, or at least very hard to reconcile, with this understanding. For example, this is true of the allegation, sometimes made on spurious biblical grounds, that the universe was created only about 6,000 years ago. This theory is not just unauthenticated. It is incompatible, not only with orthodox biology and geology, but with the physical theory of radioactivity and with cosmology (heavenly bodies more than 6,000 light-years away shouldn’t be visible if nothing older than 6,000 years exis
ts; the Milky Way shouldn’t be detectable, nor should any of the 100,000 million other galaxies whose existence modern cosmology acknowledges).
There have been times in the history of science when the whole of orthodox science has been rightly thrown over because of a single awkward fact. It would be arrogant to assert that such overthrows will never happen again. But we naturally, and rightly, demand a higher standard of authentication before accepting a fact that would turn a major and successful scientific edifice upside down, than before accepting a fact which, even if surprising, is readily accommodated by existing science. For a plesiosaur in Loch Ness, I would accept the evidence of my own eyes. If I saw a man levitating himself, before rejecting the whole of physics I would suspect that I was the victim of a hallucination or a conjuring trick. There is a continuum, from theories that probably are not true but easily could be, to theories that could only be true at the cost of overthrowing large edifices of successful orthodox science.
Now, where does Lamarckism stand in this continuum? It is usually presented as well over on the ‘not true but easily could be’ end of the continuum. I want to make a case that, while not in the same class as levitation by the power of prayer, Lamarckism, or more specifically the inheritance of acquired characteristics, is closer to the ‘levitation’ end of the continuum than to the ‘Loch Ness monster’ end. The inheritance of acquired characteristics is not one of those things that easily could be true but probably isn’t. I shall argue that it could only be true if one of our most cherished and successful principles of embryology is overthrown. Lamarckism therefore needs to be subjected to more than the usual ‘Loch Ness monster’ level of scepticism. What, then, is this widely accepted and successful embryological principle that would have to be overthrown before Lamarckism could be accepted? That is going to take a little explaining. The explanation will seem like a digression, but its relevance will become clear eventually. And remember that this is all before we start the argument that Lamarckism, even if it were true, would still be incapable of explaining the evolution of adaptive complexity.
The field of discourse, then, is embryology. There has traditionally been a deep divide between two different attitudes to the way single cells turn into adult creatures. The official names for them are preformationism and epigenesis, but in their modern forms I shall call them the blueprint theory and the recipe theory. The early preformationists believed that the adult body was preformed in the single cell from which it was to develop. One of them imagined that he could see in his microscope a little miniature human — a ‘homunculus’ — curled up inside a sperm (not egg!). Embryonic development, for him, was simply a process of growth. All the bits of the adult body were already there, preformed. Presumably each male homunculus had his own ultra-miniature sperms in which his own children were coiled up, and each of them contained his coiled up grandchildren … Quite apart from this problem of infinite regress, naïve preformationism neglects the fact, which was hardly less obvious in the seventeenth century than now, that children inherit attributes from the mother as well as the father. To be fair, there were other preformationists called ovists, rather more numerous than the ‘spermists’, who believed that the adult was preformed in the egg rather than the sperm. But ovism suffers from the same two problems as spermism.
Modern preformationism does not suffer from either of these problems, but it is still wrong. Modern preformationism — the blueprint theory — holds that the DNA in a fertilized egg is equivalent to a blueprint of the adult body. A blueprint is a scaled-down miniature of the real thing. The real thing — house, car, or whatever it is — is a three-dimensional object, while a blueprint is twodimensional. You can represent a three-dimensional object such as a building by means of a set of two-dimensional slices: a ground plan of every floor, various elevation views, and so on. This reduction in dimensions is a matter of convenience. Architects could provide builders with matchstick and balsa-wood scale models of buildings in three dimensions, but a set of two-dimensional models on flat paper — blueprints — is easier to carry around in a briefcase, easier to amend, and easier to work from.
A further reduction to one dimension is necessary if blueprints are to be stored in a computer’s pulse code and, for example, transmitted by telephone line to another part of the country. This is easily done by recoding each two-dimensional blueprint as a one-dimensional ‘scan’. Television pictures are coded in this way for transmission over the airwaves. Again, the dimensional compression is an essentially trivial coding device. The important point is that there is still one-to-one correspondence between blueprint and building. Each bit of the blueprint corresponds to a matching bit of the building. There is a sense in which the blueprint is a miniaturized ‘preformed’ building, albeit the miniature may be recoded into fewer dimensions than the building has.
The reason for mentioning the reduction of blueprints to one dimension is, of course, that DNA is a one-dimensional code. Just as it is theoretically possible to transmit a scale model of a building via a one-dimensional telephone line — a digitized set of blueprints — so it is theoretically possible to transmit a scaled-down body via the one-dimensional digital DNA code. This doesn’t happen but, if it did, it would be fair to say that modern molecular biology had vindicated the ancient theory of preformationism. Now to consider the other great theory of embryology, epigenesis, the recipe or ‘cookery book’ theory.
A recipe in a cookery book is not, in any sense, a blueprint for the cake that will finally emerge from the oven. This is not because the recipe is a one-dimensional string of words whereas the cake is a three-dimensional object. As we have seen, it is perfectly possible, by a scanning procedure, to render a scale model into a one-dimensional code. But a recipe is not a scale model, not a description of a finished cake, not in any sense a point-for-point representation. It is a set of instructions which, if obeyed in the right order, will result in a cake. A true one-dimensionally coded blueprint of a cake would consist of a series of scans through the cake, as though a skewer were passed through it repeatedly in an orderly sequence across and down the cake. At millimetre intervals the immediate surroundings of the skewer’s point would be recorded in code; for instance, the exact coordinates of every currant and crumb would be retrievable from the serial data. There would be strict one-to-one mapping between each bit of the cake and a corresponding bit of the blueprint. Obviously this is nothing like a real recipe. There is no one-to-one mapping between ‘bits’ of cake and words or letters of the recipe. If the words of the recipe map onto anything, it is not single bits of the finished cake but single steps in the procedure for making a cake.
Now, we don’t yet understand everything, or even most things, about how animals develop from fertilized eggs. Nevertheless, the indications are very strong that the genes are much more like a recipe than like a blueprint. Indeed, the recipe analogy is really rather a good one, while the blueprint analogy, although it is often unthinkingly used in elementary textbooks, especially recent ones, is wrong in almost every particular. Embryonic development is a process. It is an orderly sequence of events, like the procedure for making a cake, except that there are millions more steps in the process and different steps are going on simultaneously in many different parts of the ‘dish’. Most of the steps involve cell multiplication, generating prodigious numbers of cells, some of which die, others of which join up with each other to form organs, tissues and other many-celled structures. As we saw in an earlier chapter, how a particular cell behaves depends not on the genes that it contains — for all the cells in a body contain the same set of genes — but on which subset of the genes is turned on in that cell. In any one place in the developing body, at any one time during development, only a minority of the genes will be switched on. In different parts of the embryo, and at different times during development, other sets of genes will be turned on. Precisely which genes are switched on in any one cell at any one time depends on chemical conditions in that cell. This, in turn, depends
upon past conditions in that part of the embryo.
Moreover, the effect that a gene has when it is turned on depends upon what there is, in the local part of the embryo, to have an effect on. A gene turned on in cells at the base of the spinal cord in the third week of development will have a totally different effect from the same gene turned on in cells of the shoulder in the sixteenth week of development. So, the effect, if any, that a gene has is not a simple property of the gene itself, but is a property of the gene in interaction with the recent history of its local surroundings in the embryo. This makes nonsense of the idea that the genes are anything like a blueprint for a body. The same thing was true, you will remember, of the computer biomorphs.
There is no simple one-to-one mapping, then, between genes and bits of body, any more than there is mapping between words of recipe and crumbs of cake. The genes, taken together, can be seen as a set of instructions for carrying out a process, just as the words of a recipe, taken together, are a set of instructions for carrying out a process. The reader may be left asking how, in that case, it is possible for geneticists to make a living. How is it possible ever to speak of, let alone do research on, a gene ‘for’ blue eyes, or a gene ‘for’ colour blindness? Doesn’t the very fact that geneticists can study such single-gene effects suggest that there really is some sort of onegene/one-bit-of-body mapping? Doesn’t it disprove everything I have been saying about the set of genes being a recipe for developing a body? Well no, it certainly doesn’t, and it is important to understand why.