In the previous volume of "Folia Baeriana", one can find a statement: "Development, not evolution, could be considered as the central theoretical framework for biology. In this case Baer and not Darwin would become the central historical figure in theoretical biology" (Salthe, 1993:247). This may mean that the arguments put forward by Karl Ernst von Baer, a contemporary of Charles Darwin, who happened to be both, a forerunner and critic of the theory of natural selection, have attained a growing interest in the period of the formulation of a post-Darwinian approach in biology more than a century later.

The term post-Darwinism, in the sense of 'biology which comes after the period of Darwinism'¹, was first used not later than in 1986, at the meeting on structuralism in biology in Osaka (Ho, 1989). That meeting, which recognised a clear opposition to the 'mainstream' theoretical biology, also admitted its historical continuity as concerns the 'marginal' tradition in theoretical biology coming from J. H. Woodger, C. H. Waddington and other members of the Cambridge Club of Theoretical Biology (Goodwin et al., 1989:vii; Abir-Am, 1987). The conference in Osaka promoted the setting up of the 'Osaka Group for the study of dynamic structures' in 1987 (Sermonti & Sibatani, 1998). On the other hand, it is clearly noticeable that the nomogenetic approach in Russian biology, represented by L. Berg, A. A. Lyubischev, S. Meyen and others, which has taken its tradition back to K. E. v. Baer and his criticism of Darwinism, in many aspects resembles above-mentioned structuralistic biology (Øðåéäåð 1977; Brauckmann & Kull, 1997). Via L. Beloussov, an embryologist from the Moscow University, who participated in both, there is also a personal link between the Russian-Estonian nomogenetic conferences and Osaka group meetings. G. Sermonti, the editor-in-chief of the Italian journal 'Rivista di Biologia', is also one who has attempted to tie together the Russian, Japanese and West-European structuralistic trends in biology (cf. Sermonti, 1994; Beloussov, 1994). The long tradition of Estonian Spring Schools in theoretical biology, which began in the 1970s, and a series of Winter Schools organised together with colleagues from St Petersburg and Moscow, represent a similar approach in the theory of biology (Kull, 1998b).² During the previous decade,

Thus, a set of principles for an alternative biology was worked out during quite a long period of time. Particularly, these principles concern a different view on morphology and taxonomy. As an illustration of this view may serve a radical statement by G. Webster (1989:13): "/.../ the structure of taxonomies should be immediately deducible from a theory of transformations. If such a theory is developed then knowledge of what actually happened in the history would be of no great interest and phylogenetic speculation would become a purely antiquarian activity".

However, nomogenetic-structuralistic biology is still a quite 'marginal' tradition, while the selectionist or adaptationist view (or "panglossian paradigm" - Gould & Lewontin, 1979) has reconciled some of that criticism with introducing the view of the constraints of evolutionary process. The main selectionist mechanism of evolution leading to adaptations has not found a sound alternative. Within the latest years, the situation seemed to be changing. Several important ideas have appeared, which allow to give a sketch for understanding the basic mechanism of adaptive biological evolution in a considerably different way than that of selectionism. If so, the announcing of post-Darwinian³ biology could become grounded. Nevertheless, as always in science, this is not anything new, but the continuation of an old tradition.

The main point of structuralistic biology has never consisted in the study of evolution. However, when relating the approach to Darwinian biology, the analysis of evolutionary mechanisms becomes inevitable. Thus, in this paper I am going to describe some aspects of explaining biological evolution, comparing the neo-Darwinian and post-Darwinian views. The possibility of doing so comes from the understanding that a series of alternative schools of thinking in the theory of biology (neo-Darwinism, nomogenetics, biological structuralism, biosemiotics, the concept of autopoiesis, etc.) allow to form a comprehensive picture.

For Darwin, an important unexplained problem was that of missing links. The gradualistic view of evolutionary process did not correspond well to the absence or rarity of intermediate forms. The punctualistic theory of evolution, as developed by N. Eldredge, S. J. Gould and S. M. Stanley (Eldredge, 1989), partly resolves this difficulty accepting H. Paterson's (1993) recognition concept of species and treating species as having long periods of stasis and short

Among the early anti-Darwinian theories developed at the end of the 19th century, one of the most important was the autogenetic theory of evolution (C. Nägeli, T. Eimer, et al.). According to its main statement, there exists an immanent (intrinsic) source of evolutionary change, the intrinsic trend towards the diversification of structure and behaviour. T. Eimer stressed also the predictability of that trend, using the term orthogenesis. That view was considered to be close to the tradition of Lamarck. Indeed, its supporters usually considered the inheritance of acquired characters essential.

After the works of August Weismann, it became generally accepted that acquired characters cannot be inherited. In modern terms, the genetic memory works as read-only, as the ROM. As a result of this discovery, the autogenetic theories, together with all that they included, were thrown into the dustbin of history. The nomogenetic theory of evolution (L. Berg, A. Lyubischev, S. Meyen) has been one of the rare (and quite unknown for the majority of biologists) approaches which still holds and develops the autogenetic view.

A prevailing approach in evolutionary theory after A. Weismann has been the attempt to explain how a 'blind watchmaker' (the term used by R. Dawkins for the mechanism of natural selection) could create adaptive evolution in all its forms. The description of evolution through the description of characters was replaced by the description of the evolution of genes and genomes. Progress was remarkable, not to say - astonishing. However, some paradoxes have been quite hard to explain, among them seemingly one of the most difficult - the paradox of time, or the speed of evolution, which became even more exposed after the discovery of punctualistic evolution leaving quite little time for important rearrangements. The paradox states that the number of possible combinations of the structure of DNA is so vast, and the number of all existing events of selective choice (the minimum time step being one generation, and the minimum unit of fixation being a population) is relatively small, so that the finding of some close to optimal solutions is highly improbable.

Under autogenetic evolution I mean the concept according to which these are not DNA and environment which determine the course of evolution, but organism (and population) itself as a self-organising system (usually comprising a collective together with other specimens of the population and symbiotic organisms which it communicates with) has a certain (semiotic) freedom to choose the way of development and the environment to live in, and, as a result, to direct its evolution.

It is not reasonable and correct to consider the genome (or a gene, or a population of them) to be the element of evolution, since the genome per se does not have any meaning outside an organism. DNA as such is not a text, but it turns to be a text together with an interpreter, together with the one, which can read it, which can recognise something in it - a cell, or an organism, which includes the mechanisms of transcription and translation, at least.

DNA is never replicating or functioning in a pure form, but always together with an apparatus, which uses the information it contains for its restoration and rebuilding.

The reader (the mechanism, which uses DNA) is, in fact, the same age as the organism's DNA, since DNA cannot be replicated without the replication mechanism already being there. As a matter of fact, the reader renews itself much more frequently than the genome, but still it never disappears totally.

The reading mechanism as a whole (as a cell) is an active adaptive system, which is able to (automatic) search for more suitable structure and conditions; in some cases it can also learn, whereas the changes in its structure, i.e. the findings of the organism, could not be stored in its ROM, i.e. genome.

To illustrate the view according to which the genome is only a dead memory which cannot guide the organism's behaviour - on the contrary, it is the cell that uses the genome in one or another way - let us pay attention to some important well-known facts:

(a) as a rule, the cell uses only a part, quite often a minor part of its genome, the other parts of the genome being totally inactive;

(b) one and the same genome could be used in very different ways, as appears in case of the cells of different tissues of a multicellular organism;

(c) some cells can live a considerable period without having any genome in them (e.g., erythrocytes, and cells from which DNA has been removed experimentally);

(d) genome may undergo considerable changes (rearrangements of the positions of genes, some duplications, etc.) without any remarkable influence on the behaviour of the cell.

Borrowing a musical metaphor from K. E. von Baer, we may say that the genome is like a keyboard on which the cell can play different melodies. Or, more linguistically, the genome is like a vocabulary, on the basis of which a speaker can compile different expressions or create different texts.

Measured values for the coefficient of inheritance show that usually phenotypic characters are partially (for instance, by 60%⁴) determined by the genotype. However, several characters are inherited very strongly. Nevertheless, it should always be considered that offspring grow up and live in conditions, which are very similar to those under which their parents lived. This

It is generally accepted and known that information about a structural or behavioural change can be transmitted from one generation to the next also via other channels than DNA structure, but that change is reversible. There are several ways and examples to illustrate that:

(a) parents are able to select and find a suitable environment, the environment could influence the development, and as a result of this, the offspring happens to be in a new environment and morphologically or behaviourally changed;

(b) at reproduction cells of one tissue give cells of the same tissue, i.e. the developmental choice is inheritable without changes in the genome;

(c) genes could be marked via DNA-methylation, the marks being inherited sometimes over several generations (an example of epigenetic inheritance); the marks may be related to gene expression or gene usage.

E. Jablonka et al. (1998) distinguish between four inheritance systems: epigenetic (EIS), genetic (GIS), behavioural (BIS), and linguistic (LIS). The means of information transmission include, correspondingly, regeneration of cell structures and metabolic circuits (EIS), DNA replication (GIS), and social learning (BIS, LIS). These inheritance systems transmit variations from generation to generation, whereby the variations include cellular morphology (EIS), DNA base sequences (GIS), patterns of behaviour (BIS), and language structures (LIS). Accordingly, only one of these four inheritance systems concerns the "memory" of DNA.

As opposed to the genocentric view of biological evolution, the distinction between several independent inheritance systems makes it clear that GIS cannot explain all that is going on in evolution. We should also consider that the change or stability of the environment (i.e. the environmental information) is an obligatory component of inheritance. Changes in any of these inheritance systems may have evolutionary consequences.

This is the statement known as the concept of organic selection or 'Baldwin effect' (Baldwin 1896), which was proposed a century ago by the palaeontologist Henry Fairfield Osborn (1857-1953) and the psychologists James Mark Baldwin (1861-1934) and Conway Lloyd Morgan (1852-1936), and was later analysed by G. G. Simpson (1953) and C. H. Waddington (1953a; 1953b). Waddington used it to develop his model of genetic assimilation. Recently the Baldwin effect attracted attention again, which is expressed by the book of R. K. Belew & M. Mitchell (1996), and several other publications (Emmeche, 1994; Deacon, 1997; Jablonka et al., 1998; Kull, 1998a). My view can be interpreted as a possible mechanism for the 'Baldwin effect', which uses the

P. J. Bowler (1992:81, 131-132) described the 'Baldwin effect' as follows: "The mechanism was based on the assumption that a deliberately chosen new behaviour pattern could influence the evolution of a species, but not by means of the inheritance of acquired characters. The body of each animal would adapt itself to the new situation, but instead of this adaptation being directly inherited, it would give the species time during which random variation could come up with truly heritable equivalents, which would then be favoured by selection. Thus, selection need not be the purely mechanistic process condemned by most Lamarckians, but could include a role for the active participation of the organisms themselves. /.../ This presented habit as the guiding force of evolution - not via Lamarckism, but through its ability to define a trend in bodily structure that would be followed up by natural selection". However, what is said here about natural selection may be unnecessary (cf. the paragraph about differential reproduction below).

There are at least two important ways how a changed phenotype or a different usage of the genome may influence further stochastic changes in the genome.

(1) After the discovery of the evolutionary molecular clock by M. Kimura, it became clear that there is an important difference in the speed of the fixation of mutations between the used and non-used sites of the genome. E.g., the regions of the genome, which are coding the enzyme active sites, have different speeds of evolution as compared to the other regions. It means that a reversible change in the usage of the genome may further lead to irreversible changes in its structure. Mutations appearing in the used sites will be fixed differently from those appearing in the non-used sites.

(2) The mate choice in the populations of biparentally reproducing organisms is dependent primarily on the characteristics of the phenotype (specific mate recognition system, SMRS, according to H. Paterson), which could have reversible modifications. As a consequence, the mutations appearing in the sites, which influence the SMRS, fix differently from the mutations not influencing the SMRS.

It is also important to add that there are several ways of influencing ROM by organism, without writing anything in it, but changing it.

(1) It is possible to forget some information stored in ROM. The role of sexual reproduction happens to be, in this context, an enhanced mean of forgetting.

(2) Through organism's behaviour, it is possible to change the relation and composition of endosymbionts. These endosymbionts also have their ROM, which is in this way selected by the organism.

In case of genetic mutation, it is quite improbable that a concrete mutation could appear simultaneously in many organisms of a population. For phenotypic changes, vice versa, it could be quite usual that a particular change takes place simultaneously in many specimens since the genome and

New evolutionary findings could be rapid since these are primarily a result of the functioning of organisms, a new or changed way of using the ROM by organisms. Corresponding genetic changes could be treated as after-effects of the morphological and behavioural change.

Rapid morphological changes in speciation, as described by punctualists, and gradual genetic changes, as described by molecular evolutionists, are thus found to be in correspondence, since the latter follow the former.

In the existing models of Darwinian theory of evolution (synthetic theory of evolution), the organism is not considered to have a multi-level structure with independent activity and a possibility to use its genome in various ways. Assuming the organisms to have activity, we find the autogenetic theory of evolution work. Darwinian theory of evolution happens to be a special case of the autogenetic theory of evolution, assuming the organism to be very simple and passive.

The main material for evolution is phenotypic variability. If phenotype and genotype are strongly connected (i.e., the phenotypic variability corresponds to the genotypic one), then evolution is Darwinian. If phenotype and genotype are more or less uncoupled due to plasticity, then the directional changes are phenotypic, and genetic variability is of minor importance.

An evolutionary change is like finding of a new melody by a player - the organism. It has a number of ways to keep this melody so long that it could be fixed by the stochastic changes of the genome.

Another important aspect (besides the Baldwin effect and its consequences described above) which consequences are rarely considered, is the individuality of almost every single genome. The size of any genome is so large, and the modifications (particularly in case of sexual reproduction), although rare, are nevertheless frequent enough to make each genome individual. The interconnectedness of the processes in the cells results in the individuality of the context in which the 'products' of genes work; i.e., the context of a particular allele can be different in every single individual (cf. Holdrege, 1996). Thus, no gene can have a constant meaning. Therefore, what has been selected for in one generation may not have the same meaning in the next one, which, in an extreme case, may turn selection itself quite senseless.

In Fisher's theorems of natural selection (Fisher, 1930) one can notice an unspoken assumption that reproduction is transitive. According to transitivity, if A gives B, and B gives C, then A gives C. However, because of the

Explaining the phenomenon of 'phenocopies' (i.e., identical phenotypes in organisms having different genotypes), geneticists speak about 'plasticity', or the 'norm of reaction', as the amplitude of the phenotypic variability of a genotype (or "the range of phenotypes that may possibly develop", Dobzhansky et al., 1977:29). However, what we have here is much more than plasticity. The reason is that 'norm of reaction' is a term, which principally cannot have a finite description for any genotype. Expressing this in a more physical language: the boundary conditions for an organism in natural circumstances are indeterminable. A description of plasticity would be analogous to listing of all the possible meanings, which a particular word can have in all possible conditions. Thus, it cannot be viewed as an operational term. This also means that plasticity, generally, cannot be used as a measurable property of a genotype.

According to the neo-Darwinian mechanism of evolution, the complex structure of organism does not play a direct role in the mechanism of an evolutionary change, i.e. in the replacement of genotypes. In most population-genetic models, a change in genotype and a change in phenotype are considered as being simultaneous. Hierarchy is thus still not represented in the models of evolution. However, if organism itself is regarded as the subject of evolution (Weingarten, 1993), and the organism's search turns out to be a direct factor of evolution, the hierarchical and multi-level structure of organism becomes inevitably considered in the mechanisms of evolution (Salthe, 1985; 1993). This view is leading to a possible 'cognitive turn in biology' (cf. Hoffmeyer, 1997). The possibility of solving several existing profound controversies in the theory of biology through considering the hierarchical structure of organism was proposed by E. Oldekop (1930) already. The recent work by E. Jablonka & M. Lamb (1995; Jablonka et al., 1998) gives an insight into a possibility of integrating the Lamarckian phenomenology and epigenetic mechanisms into Darwinian evolution just by using the multi-level description of processes.

In order to explain the relationship between natural selection and the mechanism described above, we need a very clear definition of natural selection. Assuming that natural selection necessarily requires difference in the survival (i.e. differential reproduction) of the offspring of two genotypically distinguishable sets of organisms (estimated through the difference in the number of surviving adults per adult in these two groups), we may conclude that according to the Baldwin mechanism (described above) natural selection is not necessary for evolution (whereas the notion of evolution is still accepted in its traditional definition as the irreversible change in the genetic structure of population). Thus, if the offspring of two genotypically distinguishable sets of organisms have the same percentage of non-viable organisms (e.g., if these sets are indistinguishable according to their mortality rate), there is no natural selection in this case according to the definition accepted.

Natural selection can certainly be effective if the selective factor (e.g., an antibiotic) is very strong, so that the population number temporarily decreases down to few surviving individuals only. But if the population is permanently large, then natural selection cannot be very effective; Baldwinian mechanism, however, may still effectively work and, consequently, evolution may continue.

If a broader definition for natural selection is used, then Baldwinian mechanism still includes natural selection (since some genotypes die anyway, e.g., already as zygotes; cf. footnote 1).

It is quite difficult to find a direct experimental or empirical evidence which could distinguish between the selectionist and autogenetic (in the version presented here) theory of evolution. Namely:

1. paleontological dating is not precise enough and it is very difficult to obtain information about the concurrent changes of genome, thus, it is almost impossible to decide which of the changes was first - the genetic or the morphological one;
2. if the mutation process, which follows a change in the usage of the genome, is slow, then the experiment, which could show it, will take more time than accessible for any laboratory or field experiments.

The main idea is that the neo-Darwinian theory of evolution (including the theory of population genetics and behavioural ecology) formally considers phenotype and genotype to be existing and functioning simultaneously. It means that no important theoretical conclusions (as far as I know, of course) have been drawn, which could result from the time difference of phenotypic and genotypic changes.

In this sense, what I propose is actually not a refutation of the neo-Darwinian theory, but a more general view to which a possible time difference is added. When the time difference is becoming equal to zero, we can get exactly the existing theory as a special case.

Even research carried out into the role of the plasticity of phenotype in evolution has not assumed the difference in time, i.e. the pre-existence of the phenotype, and has not considered the possibility of maintaining the phenotypic difference without the genetic one throughout several generations.

The plasticity, if existing, is usually assumed to be a result of particular genes. It could be analysed in this way, but then we have to accept the possibility that the genome is used in several different ways by the phenotype.

The other point is that the existing neo-Darwinian theory assumes the reproduction of organisms and genes to be structurally transitive. According to contemporary knowledge, changes in context make reproduction intransitive. It follows that we need a broader theory. However, this in its turn means that we can get the neo-Darwinian theory as a special case, namely, for the situations when transitivity is assumed.

There are also several paradoxes, which were not easy to solve in the framework of the classical approach.

Only a very small piece of genome could be produced as an optimal one. This means that at the given time of the evolution of life on the Earth, the number of generations and the number of organisms born give the maximum estimation of the different genotypes ever built and checked for their survival. This number is seemingly lower than 4⁵⁰, which means that no longer than one 50-bases DNA-sequence could have been checked (by natural selection) for all its possible combinations of primary structure.⁶

In case of sexual reproduction, almost every individual has its own genotype which has never existed before. Nevertheless, a remarkable part of offspring is viable. This paradox may not be easy to solve from the point of view of strong neo-Darwinism, since, strictly speaking, the viability of a unique genotype is unpredictable. Also, if many of new genotypes are viable, it means that natural selection is, at least, not intensive. The offspring, which is genetically different from the parent, does not, evidently, contribute to the fitness of the parent's genotype.⁷

The views described here are close or complementary to the views of Eva Jablonka, Richard C. Strohman (1993), and many others. Below, some principal differences between neo-Darwinism and post-Darwinism are shortly pointed out:

(1) the main process for post-Darwinism is symbiosis and coherence (from which, in some conditions, competition may follow), whereas for neo-Darwinism it is competition (from which sometimes symbiosis follows);

(2) the first evolutionary event for neo-Darwinism is the mutation of DNA and the distribution of the new mutant (allele) in population, whereas for post-Darwinism it is an ontogenetic change (a change in the usage of genetic memory), which is later followed by stochastic fixation in memory (mainly due to 'forgetting of un-used');

(3) the one, which makes the choice, is environment for neo-Darwinism, and organism itself for post-Darwinism;

(4) for neo-Darwinism, DNA (together with environment) is the determinant of all the structure and through that also of the behaviour of organism, whereas for post-Darwinism DNA is like a thesaurus, or vocabulary from which the organism uses the entries it needs.

(5) for neo-Darwinism, the main role of sexual reproduction is to provide new genetic variants, whereas for post-Darwinism the importance of sexual reproduction comprises (a) the creation of species, and (b) forgetting of the unnecessary, i.e., making of the genetic memory dynamic;

(6) generally, neo-Darwinism can be regarded as a restricted special case of post-Darwinism.

According to Baer (1886:188): "Ohne Zweifel ist auch der Organismus ein mechanischer Apparat, eine Maschine, die sich selbst aufbaut. Der Lebensprocess verläuft unter ununterbrochenen chemischen Vorgängen; deswegen könnte man einen Organismus auch ein chemisches Laboratorium nennen; allein er ist zugleich auch der Laborant, indem er die für den Fortgang der chemischen Operationen notwendigen Stoffe aus der Aussenwelt aufnimmt; kann er sie nicht haben, so hört der Lebensprocess auf."

The evolution of organisms, as well as of language, is autogenetic evolution. The statement that the evolution of organisms (or language - Bichakjian, 1994) can be Darwinian is apparently true when applying only very general features of the Darwinian mechanism, when only diversification, genealogy and historical replacement are considered. A closer consideration indicates that the mechanisms of dialogue (biparentality or mutual recognition) and autogenetics become applicable, and this means that the mechanism is post-Darwinian.

However, according to E. Mayr (1988:535) "Darwinism is not a simple theory that is either true or false but is rather a highly complex research

Acknowledgements. I thank the colleagues from the Estonian Schools of Theoretical Biology, S. V. Chebanov, and late S. V. Meyen for many creative talks.