Skip to content
Navigation
🏠Overview
Knowledge
🔬Scientific Foundation
🧠Critical Thinking
🤖AI and Technology
Debunking
🔮Esotericism and Occultism
🛐Religions
🧪Pseudoscience
💊Pseudomedicine
🕵️Conspiracy Theories
Tools
🧠Cognitive Biases
✅Fact Checks
❓Test Yourself
📄Articles
📚Hubs
Account
📈Statistics
🏆Achievements
⚙️Profile
Deymond Laplasa
  • Home
  • Articles
  • Hubs
  • About
  • Search
  • Profile

Knowledge

  • Scientific Base
  • Critical Thinking
  • AI & Technology

Debunking

  • Esoterica
  • Religions
  • Pseudoscience
  • Pseudomedicine
  • Conspiracy Theories

Tools

  • Fact-Checks
  • Test Yourself
  • Cognitive Biases
  • Articles
  • Hubs

About

  • About Us
  • Fact-Checking Methodology
  • Privacy Policy
  • Terms of Service

Account

  • Profile
  • Achievements
  • Settings

© 2026 Deymond Laplasa. All rights reserved.

Cognitive immunology. Critical thinking. Defense against disinformation.

  1. Home
  2. /Scientific Foundation
  3. /Systematic Reviews and Meta-Analyses
  4. /Evolution and Genetics
  5. /Natural Selection: Mechanism, Phenomenon...
📁 Evolution and Genetics
✅Reliable Data

Natural Selection: Mechanism, Phenomenon, or Philosophical Trap That's Changing Biology

Natural selection is the foundation of evolutionary theory, but debates about its nature persist. Is it a mechanism that causally explains change, or a statistical phenomenon describing patterns? Philosophers of biology in 2024-2025 are engaged in heated discussion: Wei argues that selection is a phenomenon, not a mechanism, while Pérez-González objects. We examine why conceptual clarity is critical for experimental biology, how populations and fitness fit into the mechanistic picture, and what myths about randomness and levels of selection still distort our understanding of evolution.

🔄
UPD: February 27, 2026
📅
Published: February 23, 2026
⏱️
Reading time: 15 min

Neural Analysis

Neural Analysis
  • Topic: Philosophical debate on the nature of natural selection — mechanism or phenomenon, and its impact on biological theory
  • Epistemic status: High confidence in evolutionary facts, moderate in philosophical classification (active discussion 2024-2025)
  • Evidence level: Philosophical analysis + consensus on basic selection principles; disputes concern interpretation, not empirics
  • Verdict: Natural selection is a proven evolutionary process. Classification as "mechanism" or "phenomenon" is a matter of philosophical framework, but affects research design and integration with genetics, development, epigenetics. Both positions have strong arguments.
  • Key anomaly: Confusion between population-level statistics and individual causation; conflating "selection operates at different levels" with "selection is random"
  • 30-second check: Ask yourself: "Does natural selection describe a causal chain (mechanism) or a resulting pattern (phenomenon)?" — if you can't answer, you're in a zone of conceptual ambiguity
Level1
XP0
🖤
Is natural selection a mechanism that explains evolution, or simply a statistical phenomenon describing patterns of change in populations? This question may seem academic, but the answer determines how biologists design experiments, interpret data, and integrate knowledge from genetics, ecology, and developmental biology. In 2024-2025, philosophers of biology launched a sharp debate: Chuanke Wei argues that natural selection should be viewed as a phenomenon rather than a mechanism, while Saúl Pérez-González presents counterarguments defending the mechanistic interpretation (S001, S012). 👁️ This article examines the arguments from both sides, analyzes why conceptual confusion persists a century and a half after Darwin, and reveals which cognitive traps cause even professional biologists to misunderstand the nature of selection.

📌What is natural selection: definitions, boundaries, and conceptual frameworks that define the entire debate

Natural selection is traditionally defined as the process by which organisms with traits better adapted to their environment survive and reproduce more successfully, passing these traits to offspring. This definition traces back to Darwin and remains the foundation of evolutionary biology (S013).

The philosophical problem arises when clarifying ontological status: is selection a causal mechanism that actively produces changes, or a descriptive phenomenon capturing statistical regularities at the population level?

Mechanism
A system of entities and activities organized to produce regular changes (S014, S017). For selection: individuals, populations, genes interact through variation, inheritance, differential reproduction, creating a predictable outcome—change in trait frequencies.
Phenomenon
An observable pattern of change in trait frequencies within populations, separated from the underlying mechanisms (differential survival, reproductive success, genetic inheritance) (S012).

D. Benjamin Barros (2008) developed the concept of stochastic mechanisms, recognizing that evolutionary processes include probabilistic elements but can still be considered mechanisms (S018).

Wei (2024) proposes distinguishing selection as a phenomenon from its underlying mechanisms. According to this position, calling selection itself a mechanism is a category error: we observe the result (population change) but attribute causal power to an abstraction rather than concrete biological processes. More details in the Thermodynamics section.

This distinction is not merely semantic—it affects how researchers formulate hypotheses and interpret experimental data.

Population as entity: the central point of disagreement

The key dispute concerns the role of populations. Skipper and Millstein (2005), along with Barros (2008), argued that populations should be considered one of the constitutive entities of the natural selection mechanism (S004).

Position Argument Consequence
Population as entity Populations have properties (genetic structure, demography) that actively participate in evolutionary dynamics Selection is a population-level mechanism
Population as context Population is not an entity but rather a level of description at which selection effects manifest Selection is a phenomenon; mechanisms lie at the individual level

Pérez-González (2025) defends the first position, that populations are indeed biological entities (S001, S003). This debate reflects a deeper problem: how individual and population levels of explanation relate in biology.

Related materials: irreducible complexity and intelligent design, creationism versus evolution.

Conceptual diagram of differences between mechanistic and phenomenological interpretations of natural selection
Schematic representation of two competing philosophical frameworks: mechanism (entities + activities → causal explanation) versus phenomenon (population patterns + statistical description)

🧩Steel Version of the Arguments: Seven Strongest Cases for the Mechanistic Interpretation of Natural Selection

Before critiquing the mechanistic interpretation, it's necessary to present it in its most convincing form. Defenders of this position advance several powerful arguments that cannot be ignored (S001).

🔬 The Argument from Experimental Manipulation: We Can Intervene in Selection

If natural selection is merely a descriptive phenomenon, how do we explain that biologists can experimentally manipulate its components and predict outcomes? Artificial selection in breeding, experiments with bacterial populations under controlled conditions, studies of adaptation to new environments—all demonstrate that we can intervene in causal processes, not simply observe statistical patterns. More details in the Climate and Geology section.

Mechanisms by definition permit such intervention: we alter components (selection intensity, sources of variation) and observe predictable changes in outcomes.

  1. Artificial selection in breeding of crops and livestock
  2. Controlled experiments with bacterial populations
  3. Studies of organism adaptation to new habitats
  4. Manipulation of selective pressure intensity under laboratory conditions

🧬 The Argument from Integration with Molecular Biology: Genes as Real Entities

Modern evolutionary biology is tightly integrated with genetics and molecular biology. We identify specific genes undergoing selection, track changes in their frequencies, link genotypes with phenotypes and fitness (S001).

This integration works precisely because natural selection is treated as a mechanism operating through concrete biological entities (alleles, genotypes, phenotypes), not as an abstract statistical phenomenon. If selection were only a phenomenon, such integration would be conceptually problematic.

📊 The Argument from Predictive Power: Mechanistic Models Work

Population genetic models based on mechanistic understanding of selection possess impressive predictive power. Fisher's equations, Wright's models, quantitative genetics theory—all treat selection as a causal process with identifiable parameters (selection coefficients, fitness, heritability).

These models successfully predict evolutionary trajectories in natural and laboratory populations. Fisher's fundamental theorem works precisely within the mechanistic framework (S001).

🧠 The Argument from Causal Explanation: Selection Answers "Why?"

Biologists use natural selection for causal explanations: why do giraffes have long necks? Why do bacteria develop antibiotic resistance? Why do peacocks have bright tails?

The answers appeal to selection as a cause, not simply as a description of a pattern. If selection is only a phenomenon, these explanations lose causal force and become mere redescriptions of observed changes. The mechanistic interpretation preserves the explanatory power of evolutionary theory (S001).

⚙️ The Argument from Stochastic Mechanisms: Probability Doesn't Exclude Causality

Critics sometimes point to the stochastic nature of evolutionary processes as a problem for mechanistic interpretation. However, probabilistic processes can be mechanisms if they involve identifiable entities and activities producing regular (though probabilistic) outcomes (S018).

Quantum mechanics, radioactive decay, many biological processes—all are stochastic, yet no one denies their mechanistic nature.

🔁 The Argument from Multilevel Selection: Mechanisms at Different Levels of Organization

Multilevel selection theory shows that selection can operate on genes, organisms, groups, even species. This hierarchical structure fits naturally into a mechanistic framework: at each level there are entities (genes, individuals, groups) and activities (replication, reproduction, extinction) organized into selection mechanisms (S004).

The phenomenological interpretation struggles to explain how the same "phenomenon" can manifest at such different levels of biological organization.

🧰 The Argument from Research Practice: Biologists Think Mechanistically

Analysis of actual research practice shows that evolutionary biologists formulate hypotheses, design experiments, and interpret results while implicitly assuming the mechanistic nature of selection. They search for "mechanisms of adaptation," "selective pressures," "sources of variation"—all these terms reflect mechanistic thinking (S001).

If selection were simply a phenomenon, research practice would look completely different. The connection between theory and practice indicates that mechanistic interpretation is not just philosophical convenience, but a reflection of how biologists actually understand evolution.

🔬Evidence Base: Detailed Analysis of Empirical Data and Philosophical Arguments from 2024-2025 Sources

Systematic analysis of evidence from recent publications shows that the debate about the nature of natural selection rests on incompatible interpretations of the same data. Wei (2024) and Pérez-González (2025) provide material for testing both positions. More details in the Physics and Meta-Analysis section.

📊 Wei's Position: Selection as a Phenomenon Requiring Mechanistic Explanation

Wei (2024) distinguishes between explanandum (what requires explanation) and explanans (what explains). Natural selection is a phenomenon: an observable pattern of change in trait frequencies in populations (S012). Mechanisms—specific biological processes (differential survival, reproductive success, genetic inheritance)—produce this phenomenon.

Wei's central problem: if a population is an entity in a mechanism, what activities does it perform? Populations don't "do" anything in the sense that organisms reproduce or genes replicate (S004). Populations are a level of description at which the effects of individual processes are observed.

🧪 Pérez-González's Counterarguments: Populations as Real Biological Entities

Pérez-González (2025) argues that populations have causal properties: genetic structure (allele frequencies, heterozygosity), demographic characteristics (size, age structure, growth rates), ecological relationships (competition, predation, symbiosis) (S001). These properties are emergent and not reducible to the sum of individual properties.

Fitness only makes sense at the population level. Individual fitness is not an intrinsic property of an organism, but a relative measure of reproductive success in the context of a population and environment (S001). This makes the population a constitutive part of the mechanism, not merely a context.

If fitness is a statistical quantity, how can it be part of a causal mechanism? Answer: statistical quantities are causally relevant if they reflect real differences in biological properties.

🧾 Statistical Interpretation of Fitness

Fulda (2017) analyzes the tension between mechanistic and statistical interpretations (S014). The statistical interpretation treats fitness as a population parameter describing average reproductive success, rather than as a causal property of individuals.

Reconciliation is possible: statistical quantities are causally relevant if based on real differences in biological properties. Average genotype fitness is a statistic, but it's based on causal differences in survival and fecundity that have mechanistic explanations (physiology, behavior, morphology) (S014). Statistical and mechanistic interpretations complement each other at different levels of analysis.

🔎 Fisher's Fundamental Theorem: Mechanism Through Mathematics

Okasha (2008) showed that Fisher's fundamental theorem—the rate of increase in mean fitness equals the genetic variance in fitness—has deep mechanistic significance (S010). It connects population-level changes (phenomenon) with genetic variation and inheritance (mechanisms).

The theorem works only under certain assumptions: absence of mutation, migration, random drift; additive genetic variance. This emphasizes that statistical patterns in evolution depend on specific biological mechanisms, rather than being autonomous (S010).

Interpretation What It Explains Level of Analysis Problem
Phenomenological (Wei) Patterns of change in trait frequencies Population Doesn't explain how populations produce effects
Mechanistic (Pérez-González) Specific biological processes Organismal + population Requires defining population causality
Statistical (Fulda) Mathematical regularities Abstract Connection between statistics and causation

🧬 Integration with Developmental Biology: Adaptive Developmental Bias

Natural selection creates adaptive developmental bias—the tendency of developmental systems to produce certain phenotypic variations more frequently than others (S008). This blurs the classical Darwinian distinction between random variation and directed selection: selection shapes not only the distribution of existing variants, but also the probabilities of new ones appearing.

This integration supports the mechanistic interpretation: selection acts through specific biological processes (genetic networks, epigenetic mechanisms, developmental processes) that can be studied experimentally (S008). The phenomenological interpretation struggles to explain such integration.

⚙️ Self-Organization and Natural Selection: Interaction of Mechanisms

Batten et al. propose the formula: "self-organization proposes, natural selection disposes" (S007). Self-organization creates structures through physicochemical processes independent of selection. Selection then "chooses" among these patterns those that increase fitness.

This shows that natural selection doesn't explain all evolutionary change. Some patterns arise through self-organization. However, this doesn't refute the mechanistic interpretation; rather, it shows that evolution involves multiple mechanisms interacting with each other (S007). The phenomenological interpretation struggles to explain such interactions.

  1. Selection is not the only evolutionary mechanism; self-organization creates initial variations.
  2. The mechanistic interpretation explains how selection interacts with other processes.
  3. The phenomenological interpretation remains at the level of describing patterns without explaining their origin.
  4. The statistical interpretation connects population parameters with biological causes.

Related materials: irreducible complexity and intelligent design, Lamarckism and epigenetics.

Abstract fitness landscape with population trajectories and statistical distributions
Conceptual representation of a fitness landscape: individual organisms (points) distributed across a fitness gradient, population trajectories (lines) showing evolutionary dynamics, statistical distributions (clouds) reflecting variation

🧠Mechanism or Correlation: Distinguishing Causality from Statistical Association in Evolutionary Processes

The central problem in debates about the nature of natural selection is distinguishing causality from correlation. More details in the section Thinking Tools.

🔁 Criteria for Causality in Biology: Manipulation, Mechanism, Counterfactuals

Philosophers of science identify several criteria for establishing causality: (1) manipulative criterion — can we change the proposed cause and observe a change in the effect? (2) mechanistic criterion — can we identify the physical process linking cause and effect? (3) counterfactual criterion — would the effect change if the cause were different? (S008).

Natural selection satisfies all three criteria: we can experimentally manipulate components of selection (changing environment, sources of variation, selection intensity) and observe predictable changes. We identify specific biological processes (differential survival, reproduction, inheritance). Counterfactual scenarios show that selection is necessary for observed evolutionary changes.

If selection is merely a phenomenon (an observed pattern of change), how do we distinguish it from other phenomena? The mechanistic interpretation offers a clear answer: identify specific causal processes and distinguish selection mechanisms from drift or migration mechanisms.

🧩 Confounders in Evolutionary Research: Drift, Migration, Mutation

Evolutionary changes can occur not only through selection. Genetic drift (random changes in allele frequencies), migration (gene flow between populations), mutation (emergence of new variants) — all are confounders that create patterns resembling selection outcomes (S008).

Distinguishing these processes requires careful experimental design and statistical analysis. This creates a problem for phenomenological interpretation: how do we distinguish selection from other phenomena if they all produce visible changes in populations?

Process Mechanism Statistical Pattern Distinction from Selection
Natural selection Differential survival and reproduction Systematic change in adaptive allele frequencies Directional, predictable by phenotype
Genetic drift Random fluctuations in small populations Random walk of frequencies, loss of variation Non-directional, independent of fitness
Migration Gene flow between populations Equalization of allele frequencies between groups Homogenizing, unrelated to local environment
Mutation Emergence of new genetic variants Low frequencies of new alleles, random distribution Source of variation, not factor in frequency change

📊 Statistical Signatures of Selection: How to Recognize Its Action

Population geneticists have developed numerous statistical methods for detecting signatures of selection in genomic data: tests for excess rare alleles, linkage disequilibrium patterns, ratios of synonymous to non-synonymous substitutions. These methods are based on predictions from mechanistic models of selection: selection creates specific patterns of genetic variation distinct from neutral evolution (S008).

The success of these methods supports the mechanistic interpretation: we can predict what statistical patterns selection will create because we understand it as a causal mechanism. If selection were merely a phenomenon without mechanistic content, such predictions would be impossible.

  1. Determine which traits vary in the population and whether they are heritable.
  2. Measure differences in survival and reproduction between carriers of different variants.
  3. Test whether these differences correlate with changes in trait frequencies in subsequent generations.
  4. Exclude alternative explanations (drift, migration, mutation) through statistical analysis and experimental control.
  5. Identify the mechanism through which the trait affects survival or reproduction (e.g., physiological, behavioral, ecological).

The connection between arguments about the complexity of biological systems and the mechanism of selection becomes clear: selection is not merely a statistical pattern, but a causal process that explains how complexity can arise without external design. Mechanistic understanding of selection allows us to predict evolutionary changes and distinguish them from other processes, which is impossible if we view selection as a phenomenon without causal content.

⚠️Conflicts and Uncertainties: Where Sources Diverge and Why Consensus Remains Elusive

A century and a half of research has not produced consensus on the nature of natural selection. Analysis of sources from 2024–2025 reveals several key points of disagreement where even authoritative voices diverge fundamentally. For more details, see the section on Logical Fallacies.

🕳️ Ontological Status of Populations: Entity, Context, or Abstraction?

The sharpest dispute concerns populations. Wei argues that populations are not entities of the selection mechanism in the strict sense—they do not perform activities, lack clear boundaries, and possess no causal powers independent of the individuals composing them (S004).

Pérez-González objects: populations have emergent properties (genetic structure, demography) that are causally relevant and not reducible to individual properties (S001).

This disagreement reflects a fundamental question in philosophy of biology: how do different levels of organization relate? Reductionists see populations as convenient abstractions; anti-reductionists insist on the causal autonomy of the population level.

🧬 Fitness: Property of Individuals or Population Parameter?

The second disagreement concerns the interpretation of fitness. Three main approaches:

  1. Propensity interpretation—fitness as an individual's disposition to reproduce in a specific environment;
  2. Statistical interpretation—fitness as the average reproductive success of a class of organisms;
  3. Contextual interpretation—fitness as a relationship between phenotype and environment, not existing outside a specific population and moment in time.

Each approach has implications for understanding selection causality. If fitness is a property of individuals, then selection acts at the organism level. If it's a population parameter, selection becomes a statistical phenomenon rather than a mechanism in the classical sense.

⚙️ Mechanism or Description: Can Selection Be Both?

The third conflict concerns the very definition of mechanism. Okasha (S008) proposes distinguishing between mechanism in the narrow sense (a system of components with clear causal interactions) and mechanism in the broad sense (any regular process explaining a phenomenon).

Position Selection as mechanism? Consequence
Narrow interpretation No—selection is a description of statistical patterns Selection does not explain but reformulates observations
Broad interpretation Yes—selection is a regular causal process Selection has explanatory power at the population level
Hybrid position Yes, but with caveats—mechanism at population level, description at individual level Selection works as a mechanism only under certain conditions

🔄 Causality and Selection: Does Selection Act or Only Describe?

The fourth conflict concerns causality. Some authors argue that selection is not a cause but a filter: the environment selects, organisms do not choose. Selection does not act actively but passively excludes unfit variants.

Others object: selection is a causal process where differential reproduction of organisms with different traits leads to changes in allele frequencies. Without selection, populations would not change predictably.

Paradox: if selection is only a filter, why is evolution directional? If selection is a cause, why does it not act at the individual level but only at the population level?

📊 Empirical Testability: How to Distinguish Selection from Drift?

The fifth conflict is practical. How do we distinguish natural selection from genetic drift in real populations? Both processes change allele frequencies, but selection is directional, drift is random.

Problem: in small populations drift dominates, in large ones—selection. But population size itself depends on ecological conditions that create selection. How do we separate causes?

Criterion 1: Repeatability
If trait change repeats in independent populations—likely selection. If random—drift. But populations are rarely independent.
Criterion 2: Directionality
If change is directed toward adaptation—selection. But adaptation is defined post hoc, creating circularity.
Criterion 3: Molecular Signatures
Selection leaves signatures in the genome (low variability in coding regions). But signatures may come from other processes.

🌍 Universality of Selection: Does the Mechanism Work Everywhere?

The sixth conflict concerns boundaries of applicability. Selection works in biology, but does it work in economics (S001), culture, technology? If yes—selection is a universal principle. If no—selection is specific to biology.

Problem: in economics and culture there are no clear analogs of genes, replication, inheritance. Applying selection to these domains may be metaphor rather than mechanism.

Consensus is unattainable because each position is logically consistent within its own premises. The choice between them is a choice of philosophical framework, not an empirical question.

Related questions about the boundaries of evolutionary theory can be found in articles on intelligent design and complexity, creationism, and Lamarckism with epigenetics.

⚔️

Counter-Position Analysis

Critical Review

⚖️ Critical Counterpoint

Even rigorous analysis can miss blind spots. Here's where this article is vulnerable to objections — and why they're worth attention.

Overestimating the Philosophical Dispute for Practical Biology

Perhaps the Wei vs Pérez-González debate is scholasticism that doesn't affect real research. Biologists have successfully studied evolution for decades without resolving whether it's a "mechanism" or "phenomenon." However, the history of science shows that conceptual revolutions (e.g., the shift from typological to population thinking) have changed methodology and opened new research directions.

Underestimating the Power of Statistical Interpretation

The article leans toward Pérez-González's mechanistic position, but the statistical interpretation (Fulda 2017) has powerful arguments: population genetics is a mathematically rigorous science that predicts evolution without appealing to "mechanisms" in the philosophical sense. Perhaps selection is precisely a law of nature at the level of thermodynamics (as Fisher believed), not a mechanism at the level of biochemistry, and our article underestimates this position.

Risk of False Dichotomy

Presenting the dispute as "either mechanism or phenomenon" may be an oversimplification. Natural selection can be both simultaneously, depending on the level of analysis and research question. A pluralistic approach (selection as a family of processes, not a single entity) may be more productive than choosing one philosophical framework.

Insufficient Attention to Empirical Tests

The article focuses on philosophical arguments but doesn't propose specific experiments that would distinguish between mechanistic and phenomenal interpretations. If the distinction has no empirical consequences, it may be metaphysical rather than scientific — and we don't provide a clear answer to the question: what observation would refute the mechanistic position?

Obsolescence with New Data

If epigenetics, horizontal gene transfer, or other "non-canonical" inheritance mechanisms prove dominant (rather than marginal), the classical Darwinian framework may require radical revision. Our article is based on the modern synthesis, which itself is under pressure from the extended evolutionary synthesis, and in 10 years the Wei-Pérez-González debates may look like arguments about Ptolemaic epicycles.

Knowledge Access Protocol

FAQ

Frequently Asked Questions

Natural selection is the process by which organisms with traits better suited to their environment survive and reproduce more often, passing those traits to offspring. Darwin described it as the main driver of evolution: variations arise randomly (mutations), but selection is non-random—it "chooses" what works in specific conditions. For example, light-colored moths in polluted industrial areas become more visible to predators and die out, while dark ones survive (industrial melanism). This is not a teleological process—there's no goal or plan, only differential survival and reproduction (S001, S012, S013).
This is an active philosophical debate. Wei (2024) argues that natural selection is a phenomenon (an observable pattern of changing trait frequencies in populations), not a mechanism, since populations are not causal agents. Pérez-González (2025) counters: selection is a mechanism consisting of entities (individuals, populations) and activities (variation, inheritance, differential reproduction) organized to produce evolutionary change. Both positions rely on different philosophical definitions of "mechanism." Consensus: selection is a real process, but its ontological status depends on whether we consider populations constitutive elements or contexts (S001, S003, S012, S004).
Conceptual clarity affects experimental design and theory integration. If selection is a mechanism, biologists look for causal chains at the level of individuals and populations, studying how variation and inheritance interact. If it's a phenomenon, focus shifts to statistical patterns and mathematical models of population genetics. This isn't abstract wordplay: the answer determines how we connect evolution with developmental genetics, epigenetics, and self-organization. For example, if populations aren't part of the mechanism, then population models describe outcomes, not processes, and we must seek mechanisms at other levels (individual, molecular). Philosophical confusion slows interdisciplinary synthesis (S001, S004, S014).
No, this is a common myth. Variation (mutations, recombination) is random relative to an organism's needs—mutations don't arise "because they're needed." But selection itself is non-random: it systematically favors traits that increase survival and reproduction in a given environment. If selection were random, evolution wouldn't produce adaptations—complex structures like eyes or wings. The confusion arises from conflating the source of variation (random mutations) with the direction of selection (determined by environment). Darwin clearly separated these aspects, but popular culture often misses the distinction (S002, S008, S013).
Yes, but it's a controversial area. Classical Darwinian selection works at the individual level: those better adapted leave more offspring. However, group selection is possible if groups differ in composition and compete with each other—for example, groups with cooperative individuals may outcompete selfish groups. This requires specific conditions (low migration, high intergroup competition). Multilevel selection theory recognizes that selection can occur simultaneously on genes, individuals, and groups, but debates about which level dominates continue. Okasha (2008) analyzes the mathematical foundations of these debates through Fisher's theorem (S010, S004).
Fitness is a measure of an organism's or genotype's reproductive success. Two interpretations exist: statistical (fitness as observed offspring frequency in a population) and causal (fitness as a property of an individual causally influencing reproduction). The statistical view sees fitness as a final pattern, the causal view as a mechanistic explanation. Pérez-González (2025) criticizes Wei for conflating these interpretations: if fitness is only statistical, then selection cannot be a mechanism. But if fitness is a real property (e.g., a gazelle's running speed from predators), then selection causally explains evolution. The philosophical choice between interpretations determines whether we consider selection a mechanism (S001, S003, S014).
Darwin didn't know the mechanism of inheritance—this was the main gap in "On the Origin of Species" (1859). Mendelian genetics (rediscovered in 1900) provided the answer: traits are transmitted through discrete units (genes), not blended. The Modern Synthesis (1930s-1940s) united Darwinian selection with Mendelian genetics and population mathematics (Fisher, Haldane, Wright). We now understand selection as changes in allele frequencies in populations: if an allele increases fitness, its frequency rises. Genetics gave selection a material substrate and transformed evolution into a precise science. Without Mendel, selection would have remained a descriptive hypothesis (S009, S013).
Developmental bias is the tendency of developmental systems to produce some phenotypic variations more easily than others. For example, in vertebrates it's easier to change the number of digits than to add a new limb. This constrains the "space of possibilities" for selection. The key question: is this bias created by selection itself (adaptive bias) or independently (constraint)? Source S008 argues that adaptive bias is created by selection—selection "tunes" development so useful variations arise more frequently. This blurs the boundary between Darwinian (random variation + selection) and Lamarckian (directed variation) mechanisms. Modern evo-devo (evolutionary developmental biology) actively investigates this zone (S008, S007).
No, but it complements it. Self-organization is the spontaneous emergence of order in complex systems without external control (e.g., animal coat patterns, tissue formation). Batten et al. (source S007) formulate: "Self-organization proposes what natural selection disposes." Self-organization generates structural possibilities, selection chooses functional ones from them. This isn't competing mechanisms, but division of labor: physicochemical laws create "raw material" for evolution, selection filters it by survival criteria. Ignoring self-organization leads to adaptationism—the belief that everything is explained by selection (S007).
Because populations are statistical aggregates, not physical objects with clear boundaries. Skipper and Millstein (2005), Barros (2008) argued that populations are constitutive elements of the selection mechanism. Wei (2024) objects: populations don't produce causal changes, they're merely the context in which individuals act. If populations aren't entities of the mechanism, then selection isn't a mechanism but a population-level phenomenon. Pérez-González (2025) counters: populations are real as organized systems of interacting individuals, and their structure (size, density, genetic diversity) causally affects selection dynamics. This isn't scholasticism: the answer determines whether we study populations as objects or as epiphenomena (S001, S004).
Fisher's fundamental theorem of natural selection (1930) states: the rate of increase in mean fitness of a population equals the genetic variance in fitness within that population. This is a mathematical formalization of selection. Okasha (2008) provides a philosophical analysis of the theorem: it connects population statistics (variance) with evolutionary dynamics (change in fitness). The theorem demonstrates that selection is a statistical process, but does not negate its causality. Debates surrounding the theorem reflect a broader conflict between statistical and mechanistic interpretations of selection. Fisher himself viewed selection as a law of nature, analogous to the second law of thermodynamics—a population-level principle rather than an individual mechanism (S010).
Yes, selection is not the only evolutionary process. Genetic drift (random changes in allele frequencies in small populations), mutation (source of new alleles), migration (gene flow between populations), and self-organization in development—all influence evolution. Kimura's neutral theory (1968) argues that most molecular changes are selectively neutral and become fixed through drift rather than selection. This does not deny selection's role in adaptations, but shows that evolution is a mosaic of processes. The myth that "selection explains everything" (panselectionism) has been empirically refuted. Modern evolutionary biology integrates multiple mechanisms (S007, S013).
Directly. If selection is a mechanism, experiments must reveal causal relationships: how trait variation affects survival, how inheritance transmits traits, how environment modulates selection. If selection is a phenomenon, experiments focus on measuring population patterns and testing mathematical models. For example, in studying epigenetic inheritance: if selection is a mechanism, one must demonstrate that epigenetic marks causally influence fitness and are transmitted to offspring. If it's a phenomenon, showing correlation between marks and frequency changes in the population suffices. Philosophical confusion leads to methodological uncertainty and hinders integration of data from genetics, ecology, and evo-devo (S001, S004, S014).
Deymond Laplasa
Deymond Laplasa
Cognitive Security Researcher

Author of the Cognitive Immunology Hub project. Researches mechanisms of disinformation, pseudoscience, and cognitive biases. All materials are based on peer-reviewed sources.

★★★★★
Author Profile
Deymond Laplasa
Deymond Laplasa
Cognitive Security Researcher

Author of the Cognitive Immunology Hub project. Researches mechanisms of disinformation, pseudoscience, and cognitive biases. All materials are based on peer-reviewed sources.

★★★★★
Author Profile
// SOURCES
[01] Does the natural selection mechanism still work in severe recessions?[02] In-plane orientation control of (001) <i>κ</i> -Ga <sub>2</sub> O <sub>3</sub> by epitaxial lateral overgrowth through a geometrical natural selection mechanism[03] A comment on Nishimura, Nakajima, and Kiyota’s “Does the natural selection mechanism still work in severe recessions? Examination of the Japanese economy in the 1990s”[04] Bat-inspired algorithms with natural selection mechanisms for global optimization[05] Sex differences in energy metabolism: natural selection, mechanisms and consequences[06] Survival of the Fittest – Utilization of Natural Selection Mechanisms for Improving PLE[07] Does Natural Selection Mechanism Still Work in Severe Recessions? -- ]Examination of the Japanese Economy in the 1990s ---[08] Thinking about evolutionary mechanisms: natural selection

💬Comments(0)

💭

No comments yet