The Graur–Leibowitz Thesis:
Development, Function, and the Limits of Biological Explanation
A methodological inquiry into development, biological function, and the risk of turning uncertainty into ontology
Core distinction: Not every change is development. Not every activity is function. Not every mutation is information. Not every reconstruction is proof.
Introduction: The Methodological Problem
This article presents the Graur-Leibowitz Thesis as a test of conceptual discipline in biological explanation. Its concern is not whether biological change occurs, whether genomes exhibit measurable activity, or whether evolutionary biology has produced real scientific knowledge. Its concern is more exact: when do terms such as development, function, mutation, and information remain bound to the evidence that justifies them, and when do they begin to carry more explanatory and ontological weight than the evidence can support? By reading Yeshayahu Leibowitz and Dan Graur as complementary boundary-figures, the thesis asks how science preserves its strength not only by expanding knowledge, but also by limiting the language through which knowledge is converted into claims about reality.
The Graur–Leibowitz Thesis is a methodological thesis about the limits of biological explanation. It examines how scientific concepts such as development, biological function, mutation, and information may expand beyond the evidential conditions that originally justify them.
The central risk examined here is conceptual inflation: the transformation of limited scientific findings into broader claims about the structure of biological reality. In this article, ontology refers to claims about what exists or how reality is structured. Epistemology refers to the conditions under which such claims are justified.
The thesis reads Yeshayahu Leibowitz and Dan Graur as two complementary boundary-figures within scientific reasoning. Leibowitz was an Israeli scientist, philosopher, and public intellectual associated with biochemistry, neurophysiology, philosophy, and the history of science. Graur is a molecular evolutionist whose work concerns evolutionary bioinformatics, genes, genomes, genome composition, and the distinction between functional and nonfunctional genomic fractions.
Their importance does not come from belonging to the same school of thought, holding the same worldview, or advancing the same scientific program. Their importance comes from the fact that both expose a similar problem from different directions: scientific language can expand beyond what scientific evidence is able to justify.
The first problem concerns the word development. In its strong sense, development refers to a process in which an internal organization unfolds through ordered stages. Ontogenetic development is the clearest example. An organism develops because an internally regulated biological process is already present. The process is not merely a sequence of changes. It is the organized realization of a biological structure.
When the same word is extended to historical biological change, the status of the word changes. Species may change. Populations may vary. Genotype frequencies may shift. Phenotypes may become more or less common under environmental pressure. These are legitimate biological observations. The methodological question is whether such observations are sufficient to justify the strong use of the word development.
The second problem concerns the word function. In modern genomics, biological activity is measured in many forms: transcription, binding, regulation, interaction, expression, chromatin structure, and biochemical response. These are real findings. The question is not whether activity exists. The question is whether activity is automatically function.
The Graur–Leibowitz Thesis therefore rests on a double distinction:
Not every change is development.
Not every activity is function.
From these two distinctions follows a broader methodological concern. A scientific paradigm may remain productive, institutionally successful, and rhetorically stable while its central concepts gradually expand beyond their conditions of justification. This does not mean that the paradigm is simply false. It means that the relation between evidence, interpretation, and ontology must be examined carefully.
The issue is not anti-scientific. It is the opposite. The Graur–Leibowitz Thesis argues that science remains strongest when it distinguishes between what has been observed, what has been inferred, what has been assumed, and what has been converted into a general picture of reality.
The central question is therefore not whether biological change exists. It clearly does. The question is how far scientific language may go in interpreting biological change. When does change become development? When does activity become function? When does mutation become information? When does a coherent evolutionary reconstruction become proof?
The Graur–Leibowitz Thesis does not offer an alternative biological theory. It offers a test of conceptual discipline. It asks whether the language of biological explanation remains bound to the evidential standards that make scientific reasoning valid in the first place.
What Is the Graur–Leibowitz Thesis?
The Graur–Leibowitz Thesis is not a biographical comparison between Yeshayahu Leibowitz and Dan Graur. It does not claim that Graur continued Leibowitz, that Leibowitz anticipated Graur, or that the two thinkers belonged to the same intellectual school. The thesis is methodological, not biographical.
Its central problem is the expansion of scientific concepts beyond the conditions that justify them.
Leibowitz defines the limits of the concept of development. Graur defines the limits of the concept of biological function. Together, they reveal a general danger in scientific explanation: a concept may remain useful, familiar, and institutionally accepted even after its explanatory status has become unclear.
In the case of Leibowitz, the central issue is the word development. Development, in its strong sense, refers to an internally organized process. Ontogenetic development is the clear case: an organism develops through an internal biological program, structured stages, and regulated realization. When the same word is applied to the origin of species or to large-scale biological history, the term becomes more complex. Species change, populations vary, and biological forms shift over time. But the methodological question remains: does this justify the strong concept of development, or only a broader metaphor of historical change?
In the case of Graur, the central issue is the word function. Modern genomics can detect many forms of biochemical activity. A genomic sequence may be transcribed, bound, regulated, expressed, or involved in a measurable interaction. But Graur’s importance lies in the distinction between activity and function. Activity is something measured. Function is something inferred. To call an element functional, one must show that it has a relevant biological role.
Development is not identical with change.
Function is not identical with activity.
These distinctions are simple, but their implications are substantial. If every change is called development, then the concept of development loses its strict meaning. If every activity is called function, then the concept of function loses its evidential discipline. In both cases, scientific language becomes vulnerable to conceptual inflation.
Terminological caution:
This article does not claim that evolutionary biology technically equates ontogenetic development with evolution. The issue is different: whether the explanatory language used to describe historical biological change sometimes imports implications of organized development, direction, functional emergence, or structural realization that require additional evidential support.
The thesis also clarifies the relation between ontology and epistemology in biological explanation. Ontology concerns what exists or how reality is structured. Epistemology concerns what may be claimed as known. A scientific theory becomes vulnerable when it moves too quickly from epistemology to ontology: from “this is how we interpret the evidence” to “this is how biological reality itself is structured.”
The Graur–Leibowitz Thesis does not deny biological change. It does not deny mutation. It does not deny natural selection. It does not deny that genomic elements may have functions. Its claim is narrower and more precise: scientific reasoning must distinguish between what is observed, what is inferred, what is assumed, and what is projected onto reality as a general explanatory structure.
This is why Leibowitz and Graur are complementary. Leibowitz protects the concept of development from uncontrolled metaphorical expansion. Graur protects the concept of biological function from uncontrolled evidential expansion. One guards the boundary of the concept. The other guards the boundary of evidential warrant.
One-sentence formulation:
A scientific paradigm becomes methodologically vulnerable when it continues to use concepts such as development, function, mutation, and information beyond the evidential conditions that justify them.
This vulnerability does not necessarily make the paradigm false. It makes the paradigm dependent on constant conceptual discipline. Without that discipline, uncertainty may harden into ontology, metaphor may become mechanism, and interpretation may begin to function as proof.
Yeshayahu Leibowitz and the Concept of Development
Yeshayahu Leibowitz is important to the Graur–Leibowitz Thesis because he clarifies the conceptual status of the word development. His relevance does not depend on treating him as an evolutionary biologist, nor on presenting him as the author of an alternative biological theory. His relevance lies elsewhere: Leibowitz helps expose the difference between development as a strong biological concept and development as a metaphor for change over time.
In the strong sense, development refers to an internally organized process. The clearest case is ontogenetic development: the development of an individual organism. In this case, development involves a biological program, ordered stages, internal regulation, and the realization of a structure that is already encoded in some form. The developing organism does not merely change. It unfolds according to an internal organization.
This distinction matters because the same word, development, is often extended far beyond this strong case. People speak of the development of species, the development of society, the development of culture, the development of institutions, and even the development of the universe. These uses may be linguistically convenient. They may also be intellectually useful. But they do not automatically carry the same explanatory status as ontogenetic development.
The Graur–Leibowitz Thesis treats this as a methodological warning. If development means any sequence of change, then the concept becomes too broad. It loses the distinction between a process that unfolds from an internal organization and a process that is interpreted retrospectively as if it had direction, order, or inner necessity.
Leibowitz’s contribution is therefore not a denial of change. Biological populations change. Phenotypes vary. Genotype frequencies shift. Organisms adapt to environments. Species may differ from earlier forms. None of this is denied by the conceptual distinction. The question is more exact:
When does change justify the strong concept of development?
This question becomes especially important in the transition from ontogeny to the origin of species. Ontogeny is directly observable in individual organisms. Its stages can be studied, described, and related to internal biological mechanisms. The origin of species, by contrast, is reconstructed through historical inference. It involves evidence from genetics, morphology, physiology, paleontology, population biology, and comparative biology. That reconstruction may be scientific, but it is not the same kind of process as observing the development of an embryo.
The methodological risk appears when the language of ontogenetic development is transferred too easily to the history of species. In such a transfer, the word development may begin to imply more than the evidence has established. It may suggest direction where only sequence has been shown. It may suggest internal order where only retrospective pattern has been inferred. It may suggest a process of realization where only accumulated change has been described.
This is the problem of conceptual inflation. A concept that is valid in one explanatory setting expands into another setting without carrying the same evidential conditions with it. In the case of development, the strong concept belongs first to processes with internal organization. When applied to large-scale biological history, the concept requires additional justification.
For the Graur–Leibowitz Thesis, the central lesson is simple:
Not every change is development.
This statement does not close inquiry. It opens it. It asks what kind of evidence is needed before a sequence of changes can be treated as development in the strong sense. It asks whether the term development is being used as a mechanism, a metaphor, a historical description, or an ontological claim about the structure of biological reality.
From Ontogeny to the Origin of Species
The distinction between development and change becomes most important when the term development is transferred from ontogeny to the origin of species. Ontogeny refers to the development of an individual organism. The origin of species refers to the historical emergence of biological forms across generations, populations, and lineages. These two cases are related to biology, but they do not have the same epistemic status.
In ontogeny, development is observed within the life of an individual organism. The organism develops through internally regulated stages. Embryonic development, tissue differentiation, organ formation, and maturation all occur within a biological system whose organization is already present in encoded, regulated, and inherited form. The process can be studied directly. Its phases can be observed, compared, disrupted, and experimentally examined.
For this reason, ontogenetic development is the strongest case of development. It involves more than change. It involves the realization of an internally organized biological structure.
The origin of species is different. It is not observed in the same way. It is reconstructed through historical inference. The evidence may include comparative anatomy, genetics, paleontology, embryology, population biology, molecular similarity, and patterns of distribution. These forms of evidence may be scientifically significant. But they do not function in the same way as the observation of an individual organism developing through ordered stages.
The thesis does not deny that species change. It does not deny population-level variation. It does not deny mutation, selection, adaptation, drift, or inheritance. It asks a more precise methodological question:
When does evidence of biological change justify the strong concept of development?
At the population level, biological change may appear as shifts in genotype frequencies, allele frequencies, phenotype distributions, or adaptive traits under environmental pressure. These are real and measurable phenomena. A population may become better adapted to a local environment. A trait may become more common. A genetic variant may spread. A phenotype may increase or decrease in frequency.
But population-level change is not identical with the origin of new biological structures.
This sentence is methodologically decisive. It separates two levels of explanation. On one level, one observes or models change within populations. On another level, one infers the historical emergence of organs, body plans, regulatory systems, genes, genomes, and species. The first level does not automatically warrant the second. The transition requires additional evidence and additional argument.
The risk appears when the observed dynamics of population change are treated as if they already contain the full explanation for the origin of biological novelty. In that case, a local and measurable process is extended into a large-scale historical explanation. The extension may be legitimate, but it must be argued. It cannot be assumed merely because both levels belong to the same evolutionary vocabulary.
The thesis therefore distinguishes between three claims:
-
Biological populations change.
-
Population-level changes can be measured, modeled, and interpreted.
-
Such changes explain the origin of new biological structures.
The first claim is observational. The second claim is methodological. The third claim is historical and ontological. The third claim is the strongest. It requires the strongest evidential warrant.
This is where the concept of development becomes vulnerable to conceptual inflation. If population-level change is immediately described as development, then the term may carry more explanatory weight than the evidence itself supplies. Change becomes development. Adaptation becomes origin. Variation becomes innovation. Retrospective pattern becomes internal direction.
The issue is not whether evolutionary biology may study the origin of species. It clearly may. The issue is whether the language used in that study preserves the distinction between what is observed and what is inferred.
The Modern Synthesis as a Paradigmatic Integration
The Modern Synthesis is a central point in the methodological background of the Graur–Leibowitz Thesis. Its importance is not only historical. It is also conceptual. The Modern Synthesis brought Darwinian evolution through natural selection together with Mendelian genetics and helped stabilize a unified evolutionary framework in the twentieth century.
This integration was powerful. It gave evolutionary biology a common language. It connected genetics with natural selection. It allowed biologists to describe evolution through changes in gene frequencies, inheritance, variation, adaptation, and differential survival. It helped transform evolutionary biology into a more formal, more quantitative, and more institutionally stable scientific field.
The issue examined here is not whether the Modern Synthesis was historically important. It clearly was. The issue is methodological:
What kind of problem did the Modern Synthesis solve, and what kind of problem did it leave unresolved?
Before the Modern Synthesis, evolutionary explanation was divided between different explanatory languages. One language emphasized gradual adaptive change in populations. Another language emphasized heredity, genes, mutation, and discrete variation. The Modern Synthesis presented these languages as parts of one framework. Population-level change could be understood through genetics. Mutation could supply variation. Natural selection could act on that variation. Evolution could be modeled as change in the genetic composition of populations over time.
This was a major synthesis. But synthesis is not the same as conceptual resolution.
The Graur–Leibowitz Thesis asks whether the integration of these explanatory languages fully explains the transition from population-level change to the origin of new biological structures. A shift in gene frequencies is one level of explanation. The emergence of new genes, new regulatory systems, new organs, new body plans, and new species is another level of explanation. The methodological question is whether the first level, by itself, justifies the second.
This is where the earlier distinction becomes important:
Population-level change is not identical with the origin of new biological structures.
The Modern Synthesis made it possible to speak about evolutionary change in a unified way. But the unification of vocabulary does not automatically remove the burden of justification. If mutation supplies variation, and selection filters variation, then the explanatory burden becomes precise: one must show how such processes account for the emergence of stable, heritable, functional biological organization.
This does not mean that the Modern Synthesis is scientifically empty. It means that its conceptual scope must be examined carefully. A framework may successfully organize research while still leaving open questions about the status of its strongest claims. It may explain many cases of adaptation, variation, and population change, while the origin of complex biological novelty remains a higher-level explanatory problem.
The risk is conceptual inflation. A framework that explains measurable population change may begin to be treated as if it has already explained the full origin of biological form. Mutation becomes the source of novelty. Selection becomes the organizer of that novelty. Population change becomes development. Retrospective reconstruction becomes mechanism. The language becomes unified before every level of explanation has been independently secured.
This is not a rejection of the Modern Synthesis. It is a request to distinguish between the achievements of the synthesis and the interpretations built upon it.
The Modern Synthesis can be described as a paradigmatic integration. It unified previously distinct explanatory tools into a common framework. It gave evolutionary biology a stable research program. It helped make mutation, heredity, variation, selection, and population change parts of one scientific language.
But the Graur–Leibowitz Thesis asks a further question:
Does a unified scientific language show that the underlying explanatory gap has been closed?
The answer cannot be assumed. It must be examined.
Dan Graur and the Concept of Biological Function
Dan Graur is important to the Graur–Leibowitz Thesis because he clarifies the methodological status of the word function in biology. His role in the thesis is not that of an external critic of evolutionary biology. It is the opposite. Graur is useful precisely because he works from inside molecular evolution, genome evolution, and evolutionary reasoning.
Graur’s academic profile identifies him as a professor associated with Tel Aviv University and the University of Houston, with research focused on evolutionary bioinformatics of genes and genomes. His work includes extensive engagement with genome evolution, genomic composition, molecular evolution, and the classification of functional and nonfunctional genomic material.
Biochemical activity is not automatically biological function.
A genomic sequence may be transcribed. It may bind proteins. It may participate in a measurable biochemical interaction. It may appear in regulatory contexts. It may generate experimental signals. These findings may be real and scientifically important. But none of them automatically establishes that the sequence has a biological function in the strong sense.
This distinction matters because activity and function do not have the same epistemic status. Activity is something detected. Function is something inferred. Activity may show that something happens. Function claims that what happens has a relevant biological role.
Graur’s methodological importance lies in his refusal to let biological activity become biological function without sufficient evidential warrant. This refusal is not merely technical. It is a boundary principle. It protects biological explanation from conceptual inflation.
In this respect, Graur performs for the concept of function what Leibowitz performs for the concept of development. Leibowitz asks when change may legitimately be called development. Graur asks when activity may legitimately be called function. Both questions are questions about the limits of scientific language.
The issue is not whether genomic elements may have functions. They may. The issue is whether function can be assigned before the relevant warrant has been supplied. A genomic element should not be called functional merely because a measurement has detected activity associated with it. The stronger claim requires a stronger standard of justification.
Graur’s position also clarifies the relation between mutation, information, and function. A mutation may introduce genetic change. But genetic change is not automatically biological information in the strong explanatory sense. To become information in that sense, the change must be shown to contribute to a stable, heritable, biologically relevant organization.
This is the same methodological structure again:
Not every mutation is information.
Not every activity is function.
Not every change is development.
The point is not to close biological inquiry. The point is to protect it. If every detected activity becomes function, then the concept of function loses its evidential force. If every mutation becomes information, then the concept of information loses its explanatory discipline. If every change becomes development, then the concept of development loses its strict meaning.
Graur’s importance, then, is not only empirical. It is conceptual. He forces a scientific distinction between what has been observed and what has been justified. In doing so, he provides the second half of the Graur–Leibowitz Thesis: the boundary of evidential warrant.
ENCODE and the Difference Between Activity and Function
The ENCODE debate is not only a dispute about the human genome. It is a dispute about the conditions under which biological function is justified.
The ENCODE Project, the Encyclopedia of DNA Elements, was designed to map functional elements in the human genome. Its 2012 integrated Nature paper reported that the project had systematically mapped regions of transcription, transcription-factor association, chromatin structure, and histone modification, and stated that these data enabled the assignment of biochemical functions for 80% of the genome. The same paper reported that 80.4% of the human genome participates in at least one biochemical RNA- or chromatin-associated event in at least one cell type.
This finding was scientifically important. It showed that large parts of the genome could be associated with measurable biochemical events. It expanded the empirical map of genomic activity. It gave researchers new data about transcription, regulation, chromatin states, and genomic interaction.
But the methodological issue is not whether ENCODE detected activity. It did. The methodological issue is whether such activity is sufficient to establish function.
This is the point at which Dan Graur becomes central. Graur and colleagues criticized ENCODE’s movement from biochemical activity to biological function in their 2013 article, “On the Immortality of Television Sets: ‘Function’ in the Human Genome According to the Evolution-Free Gospel of ENCODE.” Their criticism targeted the way a broad use of function could detach the concept from an evolutionary standard of justification and inflate estimates of genome functionality.
Biochemical activity is not identical with biological function.
A genomic region may be transcribed. It may bind proteins. It may be associated with chromatin marks. It may produce a measurable signal in a particular cell type. These observations may be meaningful starting points for research. But they do not automatically establish that the region has a selected, necessary, stable, or systemically relevant biological role.
In this sense, the ENCODE debate provides a concrete example of conceptual inflation. A measured event becomes activity. Activity becomes function. Function becomes biological significance. The transition may be justified in some cases, but it cannot be assumed in all cases merely because the activity has been detected.
The Graur–Leibowitz Thesis uses ENCODE as a methodological case study. ENCODE helps show how a scientific finding can be real while its interpretation remains contested. The data may be valuable, and the inference may still require scrutiny. A genome-wide map of biochemical events is not the same thing as a genome-wide warrant for biological function.
This distinction matters because the word function carries explanatory weight. To say that a genomic element has a function is not only to say that something happens near it or through it. It is to say that the element plays a role in the biological system. That stronger claim requires criteria. It may require evidence of conservation, selective constraint, causal contribution, regulatory necessity, phenotypic effect, or other forms of biological relevance.
Graur’s role is to force this question:
What exactly must be shown before activity can be called function?
That question parallels the question raised by Leibowitz about development:
What exactly must be shown before change can be called development?
The structure is the same. Leibowitz protects the concept of development from uncontrolled metaphorical expansion. Graur protects the concept of function from uncontrolled evidential expansion. ENCODE is the concrete site where the second danger becomes visible.
Mutation, Information, and the Limits of Inference
A mutation may introduce genetic change, but genetic change is not automatically biological information in the strong explanatory sense. This distinction is central to the Graur–Leibowitz Thesis because it extends the same methodological discipline from the concept of function to the concept of information.
Definition of biological information:
In this article, biological information does not mean any genetic difference. It means a genetic difference that is integrated into stable, heritable, functional organization within a biological system.
The term information requires special caution. In ordinary language, information may mean any difference, signal, sequence, pattern, or stored content. In biological explanation, however, the term can carry a stronger burden. It may imply functional organization, regulatory meaning, heritable stability, or contribution to biological form. The thesis is concerned with this stronger use.
In the previous section, the distinction was between biochemical activity and biological function. A genomic sequence may be active without being functional in the strong biological sense. The present section applies the same structure to mutation: a hereditary mutation may occur, may be transmitted, and may alter a genetic sequence, but that does not by itself settle whether it introduces biologically meaningful information.
The issue is not whether mutations exist. They do. The issue is not whether mutations can be inherited. They can. The issue is not whether mutations may sometimes contribute to biological change. They may. The methodological question is narrower and more precise:
When does genetic change become biological information?
A mutation is first a change in genetic sequence. Biological information, in the strong explanatory sense used here, is a stronger claim. To describe a mutation as information, one must show more than alteration. One must show that the alteration contributes to stable, heritable, functional biological organization. Without such justification, the word information risks becoming another case of conceptual inflation.
The distinction can be stated directly:
Mutation is not identical with information.
Genetic change is not identical with functional novelty.
This matters because the modern evolutionary framework often treats mutation as the source of variation on which selection acts. At one level, this is a valid and useful formulation. Mutation can generate variation. Variation can be inherited. Selection can change the distribution of variants in populations. But a further claim is often added, explicitly or implicitly: mutation is treated as the sufficient source of new biological information required for the emergence of new structures, systems, and forms.
That further claim requires independent justification.
A population may contain variation. A genotype may become more common. A trait may spread under environmental conditions. These are population-level facts or models. They do not, by themselves, establish the origin of new biological information in the strong sense. To make that transition, one must show how genetic changes produce new, stable, functional organization rather than merely altering existing variation.
This is where the problem returns to the earlier distinction between ontogeny and the origin of species. In ontogeny, information is already embedded in an internally organized developmental system. The organism develops because the biological system already contains regulatory structure, inherited organization, and mechanisms of realization. In the origin of species, by contrast, the question is how such organization arises in the first place.
The thesis therefore separates three claims:
-
Mutations occur.
-
Mutations may contribute to genetic variation.
-
Mutations explain the origin of new biological information and organization.
The first claim is biological. The second claim is evolutionary and population-based. The third claim is stronger. It is historical, functional, and ontological. The third claim requires the greatest evidential burden.
This does not mean that mutation cannot contribute to biological information. It means that the contribution must be demonstrated rather than assumed. A mutation becomes biologically informative in the strong sense only if it can be shown to participate in a functional system, contribute to stable organization, and become integrated into a heritable biological structure.
The danger appears when mutation is treated as explanatory by default. In that case, the argument moves too quickly:
mutation → variation → selection → adaptation → novelty → information
Each step may be legitimate in specific cases. But the chain as a whole cannot be treated as automatically demonstrated merely because the first step occurs. A mutation is not a complete explanation. It is an event that may or may not acquire functional significance.
This is the same methodological structure seen in the ENCODE debate:
activity → function
mutation → information
change → development
In each case, the first term is weaker than the second. Activity is weaker than function. Mutation is weaker than information. Change is weaker than development. Scientific reasoning becomes vulnerable when the stronger term is granted before the required warrant has been supplied.
The Coincidence Paradigm: Retrospective Sufficiency and Biological Order
In this article, the term Coincidence Paradigm is used as an analytic term. It is not presented as the name of an established scientific theory. It is not intended as a caricature of evolutionary biology. More precisely, it refers to a problem of retrospective sufficiency: the tendency to treat a reconstructed sequence of non-teleological events as sufficient to explain organized biological order once the outcome is already known.
The Coincidence Paradigm does not refer simply to randomness. Random events occur in nature. Mutation, variation, drift, environmental pressure, and population-level change may all involve contingent elements. Natural selection itself is not random in the same sense as mutation. The methodological question is different. The question is whether a sequence of contingent or non-directed events can be treated as an explanation for organized biological novelty merely because the sequence can be reconstructed after the fact.
Non-directed events are treated retrospectively as if their accumulation were sufficient to explain organized biological order.
This structure has two dimensions.
The first dimension is ontological. It concerns what is claimed to have happened in biological reality. On this level, the claim is that mutations, variation, selection, drift, inheritance, and environmental pressures accumulate across generations until new biological structures, functions, organs, systems, and species emerge.
The second dimension is epistemological. It concerns what can be known or justified from the available evidence. On this level, observed patterns in genotype frequencies, phenotype distributions, morphology, physiology, and molecular similarity are interpreted as evidence for a historical process that is not directly observed in its entirety.
The distinction between these two dimensions is essential. The ontological claim concerns the actual origin of biological organization. The epistemological claim concerns the evidence used to infer that origin. A scientific explanation becomes vulnerable when the epistemological reconstruction is treated as if it were already identical with the ontological process itself.
This is the methodological concern behind the Coincidence Paradigm. It does not deny that biological change occurs. It does not deny that mutations occur. It does not deny that selection can act on variation. It does not deny that populations change over time. Its concern is the transition from these claims to a stronger claim: that accumulated non-directed events have been shown to generate the organized biological novelty required to explain the origin of complex structures and species.
That transition requires justification. It cannot be secured by terminology alone.
The Coincidence Paradigm therefore names a risk of explanatory compression. Several weaker claims are compressed into one stronger claim:
mutation occurs → variation exists → selection acts → populations change → biological novelty emerges → development has been explained
Each step may be legitimate in particular cases. But the chain as a whole cannot be treated as automatically established merely because the earlier steps are real. A mutation is not yet information. Variation is not yet innovation. Selection is not yet the origin of organized form. Change is not yet development.
This is why the Coincidence Paradigm belongs inside the Graur–Leibowitz Thesis. Leibowitz warns that the word development may expand beyond its strong meaning. Graur warns that the word function may expand beyond its conditions of justification. The Coincidence Paradigm identifies the larger explanatory pattern in which such expansions can become mutually reinforcing.
The paradigm works by retrospective coherence. Once a biological form exists, its history can be reconstructed as a sequence of changes that led to it. That reconstruction may be useful. It may even be scientifically plausible. But retrospective coherence is not identical with demonstrated mechanism. A sequence that can be narrated after the fact is not automatically a process that has been shown to generate the structure being explained.
This is the difference between a historical reconstruction and a complete explanatory mechanism.
A historical reconstruction asks: how might this structure have emerged?
A complete explanatory mechanism asks: what has been demonstrated to generate such structures under the relevant conditions?
The Coincidence Paradigm becomes problematic when the first is treated as if it had already satisfied the second.
This point is especially important in discussions of biological novelty. If a small change affects an existing trait, that is one level of explanation. If a sequence of changes is said to produce a new biological system, that is a stronger claim. If such a sequence is then used to explain the origin of organs, regulatory architectures, developmental systems, or species, the explanatory burden becomes stronger still.
The thesis does not claim that such explanations are impossible. It claims that they must remain bound to their conditions of justification. The stronger the explanatory claim, the stronger the required demonstration.
Ontology and Epistemology in Biological Explanation
The Graur–Leibowitz Thesis turns on the relation between ontology and epistemology in biological explanation. Ontology concerns what exists or how reality is structured. Epistemology concerns what may be claimed as known, and under what conditions such claims are justified.
The central danger is the conversion of epistemological reconstruction into ontological claim.
A scientific explanation often begins from limited evidence. It observes patterns, compares forms, measures activity, identifies correlations, reconstructs histories, and proposes mechanisms. This is normal scientific work. The problem begins when an inferred reconstruction is treated as if it were already identical with the structure of reality itself.
In biological explanation, this danger appears in several forms:
change becomes development.
activity becomes function.
mutation becomes information.
retrospective coherence becomes proof.
Each transformation moves from a weaker claim to a stronger claim. Each may be justified in particular cases. But none is justified automatically.
The distinction between ontology and epistemology is therefore not abstract. It is a practical safeguard. It asks whether a claim about biological reality is supported by the kind of evidence required for that claim. If a population changes, that is one claim. If the change is called development, that is a stronger claim. If a genomic element is active, that is one claim. If the element is called functional, that is a stronger claim. If a mutation occurs, that is one claim. If the mutation is called biological information, that is a stronger claim.
The stronger claim always requires the stronger justification.
This is where Leibowitz and Graur become complementary. Leibowitz protects the concept of development from becoming an ontological claim when it may still be a metaphor or retrospective interpretation. Graur protects the concept of function from becoming an ontological claim when it may still be only a measured activity.
The same structure appears in the Coincidence Paradigm. A historical sequence may be reconstructed after the fact. A set of changes may be arranged into a coherent explanatory narrative. A biological structure may be interpreted as the result of accumulated non-directed events. But the reconstruction remains epistemological. It is a way of organizing evidence and inference. It becomes ontological only when it is treated as a demonstrated account of how biological reality itself was produced.
This movement is not always invalid. Science often moves from evidence to inference, and from inference to claims about reality. But the movement must remain disciplined. A claim about reality must not outrun the conditions of justification that support it.
The methodological problem is especially important in evolutionary reasoning because many central claims concern processes that are not directly observable in their complete historical form. The origin of species, the emergence of organs, the formation of regulatory systems, and the rise of complex biological novelty are reconstructed from evidence. Such reconstruction may be legitimate, but it remains dependent on the strength of the inferential bridge.
The Graur–Leibowitz Thesis asks whether that bridge is always strong enough to support the concepts placed upon it.
This question does not deny evolutionary change. It does not deny common patterns, molecular similarity, adaptation, mutation, selection, or genomic activity. It asks how these findings are interpreted. It asks whether the concepts used to interpret them remain proportionate to the evidence.
A scientific paradigm becomes vulnerable when what is known only through inference is treated as if it were already known as the structure of reality itself.
This vulnerability is not a failure of science as such. It is a failure of conceptual discipline. Science requires inference. But science also requires the ability to distinguish between observation, inference, hypothesis, mechanism, and ontology. When these levels collapse into one another, the scientific explanation becomes rhetorically stronger than its evidential basis.
The risk can be described as premature ontologization.
Premature ontologization occurs when a limited epistemic claim is converted into a broad claim about reality before the necessary justification has been supplied. In biological explanation, this may happen when observed change is treated as development, biochemical activity is treated as function, mutation is treated as information, historical reconstruction is treated as demonstrated mechanism, or a coherent explanatory story is treated as proof.
The thesis treats these transitions as points requiring discipline. It does not forbid them. It requires that they be earned.
The Paradigmatic Trap
A paradigmatic trap emerges when a scientific framework continues to produce knowledge while gradually expanding its concepts beyond their conditions of justification. The framework may remain productive. It may generate research, organize data, produce models, and maintain institutional authority. Yet its central concepts may begin to carry more explanatory weight than the evidence can support.
This is the central risk identified by the Graur–Leibowitz Thesis.
A scientific paradigm does not become vulnerable only when it stops working. It may become vulnerable while it is still working. The danger is more subtle. The paradigm may continue to explain many things, but it may also begin to protect its own explanatory language from scrutiny.
The paradigmatic trap begins with a small movement:
a useful concept becomes a general concept.
a general concept becomes a background assumption.
the assumption becomes part of the ontology of the field.
At that point, the paradigm no longer merely interprets evidence. It begins to define the terms in which evidence itself is understood.
In the case examined by the Graur–Leibowitz Thesis, the trap appears through a series of conceptual expansions:
change becomes development.
activity becomes function.
mutation becomes information.
retrospective reconstruction becomes proof.
Each expansion may appear reasonable in isolation. Biological populations do change. Genomic regions do show activity. Mutations do occur. Evolutionary histories can be reconstructed. None of these points is denied.
The problem appears when the weaker claim is treated as if it already contains the stronger claim.
Change may support an account of development, but it is not identical with development. Activity may support an account of function, but it is not identical with function. Mutation may support an account of information, but it is not identical with information. Reconstruction may support an account of origin, but it is not identical with proof of mechanism.
The paradigmatic trap is therefore not the presence of error alone. It is the accumulation of small transitions that are not examined carefully enough. A term becomes broader. A distinction becomes less visible. A metaphor becomes technical language. A model becomes a default explanation. A default explanation becomes the language of the field.
This is why successful paradigms can be especially vulnerable to conceptual inflation. Their success gives their language authority. Once a paradigm becomes institutionally stable, its concepts may be treated as if their meaning were already settled. Researchers may continue to investigate details inside the paradigm while the central conceptual transitions remain unexamined.
The trap is not that the paradigm produces no knowledge. The trap is that the knowledge it produces may be organized by concepts whose conditions of justification have changed silently.
A paradigm remains scientifically disciplined when it can still ask:
What has been observed?
What has been inferred?
What has been assumed?
What has been named?
What has been demonstrated?
Without these distinctions, uncertainty begins to harden into ontology. A gap in evidential support becomes a research program. A research program becomes a background assumption. A background assumption becomes a picture of reality.
This process is gradual. It rarely announces itself. It does not require bad faith. It may occur through normal scientific language, institutional repetition, textbook simplification, research specialization, and the need for coherent explanatory narratives.
The paradigmatic trap is therefore not primarily psychological. It is structural. It emerges when a framework becomes strong enough to interpret every difficulty as an internal problem to be absorbed, rather than as a possible sign that the framework has reached a conceptual boundary.
A difficulty becomes “complexity.” A gap becomes “future research.” An unresolved transition becomes “accepted framework.” A conceptual weakness becomes “standard terminology.”
This is how a paradigm can remain productive and still become epistemically fragile.
Definition:
The Graur–Leibowitz Thesis defines the paradigmatic trap as a condition in which a scientific framework continues to produce knowledge while its expanded concepts increasingly protect the framework from examining the gap between what it demonstrates and what it assumes.
This definition is intentionally narrow. It does not imply that the paradigm is false. It does not imply that its research is useless. It does not imply that its central phenomena are unreal. It implies that its strongest concepts require renewed examination.
A scientific paradigm is strongest not when it explains everything, but when it can still distinguish between what it explains, what it models, what it infers, and what it has not yet demonstrated.
What the Graur–Leibowitz Thesis Claims and Does Not Claim
The Graur–Leibowitz Thesis is not a rejection of biological science. It is a demand that biological explanation remain answerable to its own conditions of justification.
This section is necessary because the thesis can easily be misunderstood. A critique of conceptual inflation may be mistaken for a critique of biology itself. A critique of the misuse of the word development may be mistaken for a denial of biological change. A critique of broad definitions of function may be mistaken for a denial that genomic elements can be functional. These are not the claims being made.
The thesis is methodological. It concerns the relation between scientific concepts and the evidence required to justify them.
The Graur–Leibowitz Thesis claims that the concepts development, function, mutation, and information must be used with conceptual discipline. It asks when each concept is justified, what kind of evidence it requires, and whether the concept has expanded beyond its original conditions of justification.
It claims that not every change is development. Biological populations change. Traits vary. Genetic frequencies shift. Species may differ from earlier forms. But the strong concept of development requires more than change over time. It requires a demonstrated form of internal organization, direction, or structured realization. Ontogenetic development is the strongest case. Historical biological change is a different explanatory category.
The thesis claims that not every activity is function. A genomic region may be transcribed, bound, expressed, regulated, or associated with a measurable biochemical event. Such activity may be important. It may also become a starting point for further research. But biological function is a stronger claim. Function requires evidence of biological role, not merely detectable activity.
The thesis claims that not every mutation is information. A mutation is a genetic change. It may be inherited. It may contribute to variation. But biological information, in the strong explanatory sense, requires more than alteration. It requires a demonstrated contribution to stable, functional, heritable biological organization.
The thesis claims that retrospective coherence is not proof. A historical sequence may be reconstructed after the fact. A biological structure may be placed inside a plausible evolutionary narrative. Such reconstruction may be useful and scientifically meaningful. But a coherent narrative is not automatically a demonstrated mechanism of origin.
The thesis also claims that a successful scientific paradigm may become vulnerable to conceptual inflation. This does not mean the paradigm is false. It means that its central concepts may expand until they begin to carry more explanatory weight than the evidence directly supports. A paradigm may remain productive while still requiring renewed examination of its strongest terms.
At the same time, the Graur–Leibowitz Thesis does not claim to offer an alternative biological theory. It does not present a rival mechanism for the origin of species. It does not replace evolutionary biology with another scientific framework. Its function is not constructive in that sense. It is diagnostic and methodological.
The thesis does not deny biological change. It does not deny variation, heredity, mutation, adaptation, selection, drift, or population-level dynamics. These are legitimate biological phenomena. The question is how far they may be extended in explanation and what kind of evidence is required when they are used to explain biological novelty.
The thesis does not deny that genomic elements may have functions. Many genomic elements clearly do. The issue is not whether function exists. The issue is whether function may be inferred from activity alone. A criticism of inflated definitions of function is not a denial of function. It is a demand that function be justified.
The thesis does not deny that mutations may sometimes contribute to biological novelty. It denies only the automatic elevation of mutation into information. The claim is not that mutation is irrelevant. The claim is that mutation must be shown to contribute to functional organization before it is treated as biological information in the strong sense.
The thesis does not claim that evolutionary biology is simply wrong. Its argument is narrower. It asks whether certain explanatory transitions inside evolutionary reasoning are always justified with sufficient care. The transitions from change to development, activity to function, mutation to information, and reconstruction to proof must remain visible.
The thesis does not make a religious argument. It does not depend on theology, revelation, creationism, intelligent design, or any metaphysical alternative to science. It is concerned with scientific language, scientific justification, and the limits of inference. Its demand is internal to scientific reasoning: a claim must remain proportionate to the evidence that supports it.
This is why the thesis must also apply its own discipline to itself. It must not overstate what it demonstrates. It does not demonstrate that the Modern Synthesis is false. It does not demonstrate that evolutionary reconstruction is invalid. It does not demonstrate that genomic activity is meaningless. It does not demonstrate that mutation cannot generate functional novelty.
It shows something more limited and more precise:
biological explanation becomes vulnerable when its central concepts are used beyond the conditions that justify them.
The Graur–Leibowitz Thesis therefore stands on four methodological distinctions:
change is not automatically development.
activity is not automatically function.
mutation is not automatically information.
reconstruction is not automatically proof.
These distinctions are not anti-scientific. They are the conditions under which scientific explanation remains disciplined.
Why the Graur–Leibowitz Thesis Matters for Science
The importance of the Graur–Leibowitz Thesis lies in its defense of the boundary between scientific explanation and conceptual overextension. Its purpose is not to weaken biology, but to strengthen the discipline by which biological claims are made, interpreted, and converted into broader explanations.
Science depends on concepts. It cannot function by data alone. Concepts such as development, function, mutation, and information allow biological research to organize findings, compare cases, build models, and formulate explanations. But the same concepts can become vulnerable when they are extended beyond the conditions that justify them.
A scientific concept must remain answerable to its conditions of justification.
The thesis matters because it identifies a recurring danger inside successful science. A concept may begin as a precise tool, then become a general explanatory term, then become part of the background ontology of a field. Once that happens, the concept may stop being examined with the same rigor that originally made it useful.
This danger is not external to science. It is internal to science.
It does not require fraud, ideology, bad faith, or ignorance. It may arise through normal scientific work: specialization, institutional success, textbook simplification, research momentum, technical vocabulary, and the need for coherent explanatory frameworks. A paradigm may become so productive that its central terms appear self-evident.
The Graur–Leibowitz Thesis resists that process.
It asks that biological explanation preserve four distinctions:
change is not automatically development.
activity is not automatically function.
mutation is not automatically information.
reconstruction is not automatically proof.
These distinctions are not minor terminological corrections. They are safeguards against premature ontology. They prevent scientific language from converting limited inference into a general picture of reality before the necessary justification has been supplied.
This is why Leibowitz and Graur matter together.
Leibowitz matters because he clarifies the boundary of development. Development, in the strong sense, cannot simply mean any change across time. It refers most clearly to an internally organized process, such as ontogenetic development. When the term is extended to historical biological change, the extension requires justification.
Graur matters because he clarifies the boundary of function. Function cannot simply mean any detectable biochemical activity. A genomic element may be active without having a demonstrated biological role. To call it functional requires justification beyond detection.
Together, they show that the strength of science does not lie only in discovery. It also lies in restraint.
Scientific restraint is not hesitation for its own sake. It is the refusal to let a concept do more explanatory work than the evidence permits. A disciplined science does not merely ask whether a claim is useful, elegant, explanatory, or institutionally accepted. It asks whether the claim remains proportionate to what has been shown.
This is especially important in large scientific paradigms. A paradigm gives science power. It unifies research, organizes problems, stabilizes vocabulary, and allows cumulative work. But for the same reason, a paradigm can also hide conceptual inflation. Once a vocabulary becomes dominant, it may begin to absorb unresolved questions into its own terms.
Central vulnerability:
science may continue to produce knowledge while losing clarity about the limits of its own concepts.
The Graur–Leibowitz Thesis matters because it gives this vulnerability a clear structure. It does not say that a paradigm is false because it is powerful. It says that a powerful paradigm must be especially careful about the concepts through which its power operates.
A scientific paradigm should therefore be able to ask:
What has been measured?
What has been inferred?
What has been demonstrated?
What has been assumed?
What has been converted into ontology?
These questions are not hostile to science. They are part of science’s self-correction.
The thesis also matters because it protects the difference between explanation and naming. To name something is not necessarily to explain it. To call a sequence of changes “development” does not yet explain development. To call a biochemical event “function” does not yet explain function. To call a mutation “information” does not yet explain biological novelty.
A scientific explanation must do more than assign a term. It must show the mechanism, the evidence, the limits, the conditions, and the criteria by which the explanation could be tested or challenged.
This is why the thesis is relevant beyond the specific figures of Leibowitz and Graur. It offers a general methodological rule:
Science protects itself by limiting its own language.
This does not mean that science should avoid bold theories. It means that bold theories must preserve the distinction between what they propose, what they infer, what they model, and what they demonstrate. A theory may be powerful without being complete. A model may be useful without being ontological. A reconstruction may be plausible without being proof.
The thesis stands against two opposite errors.
The first error is anti-scientific rejection: the claim that because scientific concepts have limits, science itself is unreliable. The thesis rejects this. Conceptual limits do not weaken science. They define the conditions under which science remains reliable.
The second error is scientific overextension: the claim that because a paradigm is successful, its strongest concepts may be treated as already secured. The thesis rejects this as well. Scientific success does not cancel the need for conditions of justification.
The thesis occupies the disciplined middle position:
Science is strongest when it can criticize its own conceptual inflation.
This is the final importance of the Graur–Leibowitz Thesis. It shows that the defense of science is not identical with the defense of every expanded use of scientific language. Sometimes the defense of science requires drawing a boundary.
Development must remain tied to the conditions under which development is demonstrated. Function must remain tied to the conditions under which function is established. Information must remain tied to the conditions under which biological information is shown. Ontology must remain tied to the epistemology that justifies it.
The thesis therefore ends where it began: with a demand for conceptual discipline.
Not every change is development.
Not every activity is function.
Not every mutation is information.
Not every reconstruction is proof.
These statements do not close biological inquiry. They keep it open. They prevent inquiry from being prematurely settled by inflated concepts. They require biological explanation to remain answerable to evidence, inference, mechanism, and justification.
That is why the Graur–Leibowitz Thesis matters for science.
The Graur-Leibowitz Thesis therefore proposes a sober rule for scientific reasoning: explanatory power must not be purchased by conceptual inflation. A scientific paradigm may be productive, coherent, and institutionally successful, yet still require renewed scrutiny at the points where its language moves from observation to inference, from inference to mechanism, and from mechanism to ontology. The purpose of the thesis is not to weaken biological science, but to strengthen the boundary conditions under which biological explanation remains scientific. In that sense, its final demand is simple: concepts must not be allowed to outrun the evidence that gives them meaning.
Selected Sources
Stanford Encyclopedia of Philosophy — “Yeshayahu Leibowitz”
Tel Aviv University — Dan Graur Academic Profile
University of Houston — Dan Graur Academic Profile
Graur et al., “On the Immortality of Television Sets,” Genome Biology and Evolution, 2013
PNAS — Historical and conceptual discussion of the Modern Synthesis
Related Source and Reference Pages
This article belongs to the public essay layer of RATIUM.AI. For readers who want to move from this article into the broader source, technical, and orientation layers of the project, the following pages provide the relevant entry points.
Articles
The articles page gathers the public essay layer of RATIUM.AI, including arguments on stable AI governance, decision-control architecture, visible governance versus real authority, universal reason, technical competence, purpose governance, and the doctoral-scale framing of CEP.
Foundational Source Dossier
The foundational source dossier presents the deeper intellectual corpus behind CEP, LoopGuard-AI, and the broader RATIUM.AI research structure.
Technical & Reference Dossiers
The technical and reference dossier page collects architecture, visual explanation, methodological context, FAQ material, and technical source pages related to LoopGuard-AI and CEP.
RATIUM.AI / LoopGuard-AI / CEP FAQ
The FAQ page provides a structured orientation layer for readers who need concise explanations of RATIUM.AI, Benny Dunavich, CEP, LoopGuard-AI, AI governance, evidence boundaries, and the relationship between the project’s source dossiers, technical materials, and public articles.