NOTE-114: SEP — Biological individuals

Contribution

This is a survey of the philosophy of biological individuality. The one structure it adds of its own is a Venn-style framework built across Figs 2, 3, 5 and 6. First (§4) it divides biological individuals into organisms, some parts of organisms and some groups of organisms. Next (§6.1) it overlays a living agents (physiological) region, then (§7) a Darwinian (evolutionary) individuals region, which yields nine numbered regions with one example each. §7 and Fig. 6 argue that all nine are plausibly occupied. The entry also reports two quantitative-style frameworks from others: Godfrey-Smith’s B/G/I space for collective reproducers (Fig. 4) and Queller & Strassmann’s cooperation/conflict plane (Fig. 7). To the second it adds its own informal “meta ratio” extension (§8).

Key insight

“Organism” and “biological individual” have long been used interchangeably (Lidgard & Nyhart 2017a: 18, quoted in §1), and the entry denies that they are the same. The two main kinds of individual, evolutionary and physiological, each run across the part / organism / group axis. So parts and groups of organisms can be Darwinian individuals without being living agents (genes; eusocial colonies), or living agents without being Darwinian individuals (hearts; coral reefs). Organisms are “exemplars of (but not strictly identical to) both” (§3.2).

Assumptions

These are the survey’s premises and the stances it takes, with where each appears.

• Scope. The entry concerns biological individuals only. What makes anything an individual of any kind is “bracketed off” (§1, citing Strawson 1959, van Inwagen 1990, Lowe 2016, Wiggins 2016 and others).
• Individuals versus processes. Biological individuals “have three-dimensional spatial boundaries, endure for some period of time, are composed of physical matter, bear properties, and participate in processes and events”. Processes (photosynthesis) and events (speciation) lack that suite (§1). The process ontology of Dupré, Nicholson and Meincke 2019 is noted as a challenge and not adopted.
• One superordinate category. “Biological individual” names a superordinate category whose relation to “organism” is worth exploring (following Lidgard & Nyhart, Dupré & O’Malley 2009, Nicholson 2014, Pradeu 2016a/b). Okasha’s (forthcoming, BJPS, doi:10.1086/728048) scepticism that any such category exists is noted and set aside (§1).
• The framework’s scope. Biological individuals “include exactly” organisms, some parts of organisms and some groups of organisms (§4). “Some” is never cashed out: no criterion is given for which parts or groups qualify.
• All organisms are living agents (§6.1: “supposing that all organisms are living agents but that there may be both parts and groups of organisms that are not”). This is a stipulation that drives Fig. 3.
• Darwinian individual follows Godfrey-Smith 2009: a member of a population with variation, heritability and differential reproductive advantage, which is both a bearer and an active reproducer of heritable traits (§3.2).
• Physiological individual follows Pradeu 2016b: 807, “a functionally integrated and cohesive metabolic whole, made of interdependent and interconnected parts” (§3.2).
• Hierarchy. The evolution-of-individuality literature (§9) assumes that “many or all biological individuals are hierarchically organized: earlier individuals provide the material basis for later individuals”.

Key results

What the entry argues or reports, section by section.

• §1 (Focal Question). Lidgard & Nyhart 2017a compiled 23 historical criteria for “individual or its contained subset organism”. “Causal integration, cohesion, collaboration, or agency of parts” is the most-cited criterion (reported in §2).

• §2 (puzzle cases). The Michigan fungus is reported at more than 10 tons and more than 1500 years old (Smith, Bruhn & Anderson 1992: 431). The coral reef is a polyp–zooxanthellae–calcite complex. Either it is an organism with better integration than its constituents, or it is an ecosystem “that contains several distinct organisms … as proper parts”. The empirical facts do not settle which; antecedent concepts are also needed.

• §3.1 (Beyond Organisms). Parts of organisms are “often enough, themselves biological individuals”, and so are the populations and lineages that organisms constitute. Hence, if internal compositional complexity or forming populations is necessary for organisms, “given that there are organisms, there must be some biological individuals that are not organisms”.

  • Microbialism (Dupré & O’Malley) and Eliminativism (Haber 2013, Okasha 2011) are stated.
  • Dupré’s promiscuous individualism permits counting a lichen as one, two or three individuals (lichen; fungus; cyanobacterium) depending on purpose. The entry objects that it “seems even to suggest that whether there are any biological individuals at all is ‘a matter of biological judgment’”.
  • Holobiont follows Gilbert & Tauber 2016: 842, “the multicellular eukaryote plus its colonies of symbionts”. Lloyd & Wade’s 2019 euholobiont/demiholobiont distinction, Doolittle & Booth’s “holobiosis” process view, and Suárez & Stencel’s 2020 “part-dependent” account (holobionts are individuals and ecosystems at once) are reported without adjudication.

• §3.2 (Distinctions).

  • Evolutionary versus physiological (metabolic) individuals is “the most commonly recognized distinction” (Pradeu 2016a, b).
  • Godfrey-Smith 2013 places squid–Vibrio as an organism that is not a Darwinian individual, and aphid–Buchnera as both (Fig. 1). Fig. 1 also places the fruit fly in both, and a chromosome as Darwinian but not an organism.
  • Pradeu’s immunological criterion (2012: 244) makes “true organisms” holobionts: the macrobe plus “all and only those microbes at least tolerated by its immune system”.
  • Godfrey-Smith 2016c: 782 names the “tension in seeing symbionts as both organisms themselves and also parts of larger organisms”; the entry calls it a live issue.

• §3.3 (Conceptual Space). Pluralisms are listed: J. Wilson’s genetic/functional/developmental/evolutionary individuals; DiFrisco’s 2019 kind pluralism; Harper’s ramet/genet (an aspen grove is many ramets but typically one genet); and practice-based pluralism (Kovaka, Chen, Love 2018’s “strong individuality pluralism”; Kaiser & Trappes 2021).

• §4 (framework). Organisms, some parts and some groups are “not mutually exclusive”. A bacterium can be an organism and part of a corporate organism; a eusocial colony or integrated multispecies community “may be true organisms”. The entry says this “may also capture what truth lies behind” proposals to extend “organism” to some parts and groups (Queller 1997; Okasha 2011).

• §5 (groups).

  • Superorganisms (Wheeler 1911/1920) are distinguished from trait groups (D.S. Wilson). The common response (Sterelny 1996; Okasha 2006, 2018) is that eusocial colonies may be evolutionary “(and even physiological)” individuals but trait groups are neither.
  • Species and clade selection (Stanley, Gould, Jablonski) is set against Vrba’s species sorting: on that view species and clades are products of selection, not agents in it.
  • The species-as-individuals thesis (Ghiselin 1974; Hull 1976, 1978) treats organisms as the species’ physical parts. The “now widely accepted insight” is that organisms belong to species “(as parts or members)” by extrinsic, relational properties (Barker 2010).
  • Mariscal & Doolittle 2020 treat all of life (LUCA and its descendants) as one individual.

• §6.

  • Living agents have metabolism plus structural and dispositional properties (§6.1). Organisms are singled out among physiological individuals by an immune system or a life cycle.
  • Life cycles (Bonner 1993) and bottlenecks: Dawkins’s (1989: 264) claim that a single-cell bottleneck is “the essential, defining feature” of an organism is called “hyperbolic” (§6.2). Neither reproduction nor the capacity to reproduce belongs to the generic life cycle, because sterile castes exist; “all organisms have life cycles that allow them to form reproductive lineages”.
  • Godfrey-Smith’s 2009 simple, scaffolded and collective reproducers are presented with the Fig. 4 B/G/I cube. Placements: buffalo herd (0,0,0); sponge (0,0,.5); slime mold (0,.5,.5); aspen ramet (.5,.5,1); Gonium (1,0,.5); oak from acorn (1,.5,1); Volvox carteri (1,1,.5); “us” (1,1,1).
  • Autonomous agency (§6.3) combines relative autonomy from the environment with control over parts (R.A. Wilson 2005: 62–65; Moreno & Mossio 2015). Four accounts of why cognitive metaphors pervade biology are listed: Wilson’s crystallization, Okasha 2018, Figdor’s 2018 literalism, and Godfrey-Smith 2019.

• §7 (nine regions, Figs 5–6).

  1. Fruit fly: organism, living, Darwinian.
  2. Gene: part, Darwinian, not living.
  3. Eusocial insect colony: group, Darwinian, not living.
  4. Lysosome: part, neither.
  5. Clade such as Bryophyta (about 10,000 moss species): group, neither.
  6. Heart: part, living, not Darwinian.
  7. Coral reef: group, living, not Darwinian.
  8. Virus: part, living, Darwinian, via “a scaffolded form of reproduction”.
  9. Squid+Vibrio: organism, living, not Darwinian.

  The authors say some regions “may turn out … unoccupied” or may collapse together.

• §8. Queller & Strassmann’s “organismality” is high cooperation and low conflict among parts (Fig. 7A: mouse high, yeast floc low). The authors call this the internal ratio. They propose a meta ratio of internal to external cooperation/conflict ratios as an extension, and link it to Godfrey-Smith’s “movement of individuality”.

  • Clarke 2013: 427 holds that something is an evolutionary individual iff it has policing mechanisms (which inhibit within-object selection) and demarcating mechanisms (which maintain between-object selection). Her view is functionalist, so evolutionary individuals are multiply realizable and the bottleneck condition is rejected.
  • Clarke 2016a nonetheless denies that biofilms are evolutionary individuals, against Ereshefsky & Pedroso.

• §9. Established facts: prokaryote to single-celled eukaryote about 2 Gya; to multicellular eukaryotes within the last 600–800 Mya; “no counter-examples to this evolutionary trend”. Five contested defaults follow:

  1. Fitness decoupling (Michod), challenged by Bourrat and colleagues’ tradeoff-breaking and ecological-scaffolding/endogenization accounts.
  2. Individuality equated with complexity (McShea’s nestedness hierarchy versus Maynard Smith & Szathmáry’s eight transitions; O’Malley & Powell 2016’s “turns”).
  3. Which events count: Herron 2021 would narrow major transitions to those creating a new population of evolutionary individuals, excluding the genetic code and language.
  4. Directional bias versus Gould’s contingency.
  5. Sociality as under-theorized (Erwin’s public and club goods).

Claims

id	claim	strength	support
C1	Equating biological individuals with organisms is a mistake, because parts and groups of organisms are often individuals	moderate	informal argument from examples, §3.1; the stronger “must be” version is explicitly conditional on compositional complexity being necessary for organisms
C2	Evolutionary vs physiological is the most commonly recognized distinction among kinds of individual	moderate	assertion attributed to Pradeu 2016a, b; a survey judgement (§3.2)
C3	Biological individuals include exactly organisms, some parts and some groups of organisms, in categories that overlap	weak	stipulated framework (§4); “some” gets no criterion, and the overlap example given (Buchnera in a “human” gut) is a factual slip
C4	All nine regions of Fig. 5 are occupied	weak	the authors’ tentative placements, graded “less contentious to more contentious” (§7); regions 3, 7, 8 and 9 are each hedged (“perhaps”, “more controversially”, “plausible candidate”)
C5	Genes are Darwinian individuals but not living agents (region 2)	assertion	§7, following replicator and Darwinian-individual views; contested elsewhere in this record (LIT-157 treats genes in multicellular organisms as mere reproducees)
C6	Honey-bee colonies are Darwinian individuals and living agents only by the cognitive metaphor	informal argument	§7, resting on the §6 account of living agency
C7	Squid+Vibrio is an organism and living agent without being a Darwinian individual	weak	Godfrey-Smith 2013 via Fig. 1, reported in §3.2; in §7 the authors themselves raise the rival reading that it is no organism at all
C8	All organisms have life cycles that let them form reproductive lineages, though neither reproduction nor the capacity for it is part of the generic life cycle	informal argument	§6.2, from sterile castes; sits uneasily with C7’s non-reproducing organism, a tension the entry does not note
C9	Individuality can be measured as a meta ratio of internal to external cooperation/conflict ratios	weak	the authors’ own proposed extension of Queller & Strassmann (§8); no operationalization or case is given
C10	Something is an evolutionary individual iff it has policing and demarcating mechanisms	reported	Clarke 2013: 427, stated in §8 without endorsement
C11	Transitions to single-celled eukaryotes (about 2 Gya) and multicellular eukaryotes (600–800 Mya) are established, with no counter-examples to the trend	assertion	§9, stated as established fact with no citation at that sentence

Concepts

• Focal Question — “what are biological individuals?”, with four paraphrases (§1).
• Biological individual — the superordinate category: spatially bounded, enduring, material bearers of properties that participate in processes (§1).
• Evolutionary / Darwinian individual — a member of a Darwinian population that bears and actively reproduces heritable traits (Godfrey-Smith 2009, 2013).
• Physiological individual — a functionally integrated, cohesive metabolic whole of interdependent parts (Pradeu 2016b: 807). The entry treats these as “living agents”.
• Living agent — has a metabolism, heterogeneous specialized parts, capacities for growth, reproduction and self-repair, and a locus of control. Organs are living agents but “obligately-dependent” (§6.1, §6.3).
• Corporate organism — a multispecies organism formed by symbiosis, e.g. a lichen or squid+Vibrio (R.A. Wilson 2005).
• Holobiont — roughly “the multicellular eukaryote plus its colonies of symbionts” (Gilbert & Tauber 2016). Euholobiont versus demiholobiont follows Lloyd & Wade 2019.
• Scaffolded / simple / collective reproducers — Godfrey-Smith 2009: 87–88.
• Internal ratio / meta ratio — the authors’ reconstruction and extension of Queller & Strassmann (§8).
• Policing / demarcating mechanisms — Clarke 2013.

Connections

How the entry treats parts versus individuals. This is the point of the mereology tag.

• Parthood is informal. The entry uses “part”, “proper part” (§2), “compose” (§3.1), “physical parts” and “parts or members” (§5) without definition. It explicitly sets aside the general metaphysics of individuality, van Inwagen 1990 included (§1). So it takes no position on any of the principles laid out in LIT-124 (the M ⊂ MM ladder, supplementation, the Special Composition Question). Several of its puzzles are nonetheless composition and counting questions in LIT-124’s sense:
  • the lichen counted as 1, 2 or 3;
  • the reef as organism versus as an ecosystem with organisms as proper parts;
  • ramet versus genet.
• Organisms. Organisms are the middle region. They are all living agents by stipulation (§6.1), and all but region 9 are also Darwinian.
• Parts of organisms. Parts land in four regions: gene (Darwinian only), lysosome (neither), heart (living only) and virus (both). “Most parts of the cellular machinery … such as lysosomes … or ribosomes, are biological individuals” (§7), so the class is broad, but “some” in “some parts” is never bounded. The virus placement treats an obligate intracellular replicator as a part of its host. The only stated ground is its scaffolded reproduction on “the replicative machinery of their host”; the entry does not discuss free virions.
• Overlap: parts that are organisms. An organism may be both an organism and a part of one (the Buchnera example, misattributed to a human gut) or a group (superorganisms) (§4). Figs 2–6 nonetheless draw the three categories as disjoint regions and give each example a single region. The overlap the text insists on is therefore not represented in the diagram.
• Groups and holobionts.
  • Groups span eusocial colonies (Darwinian, not living), coral reefs (living, not Darwinian) and clades (neither). Trait groups get no region.
  • Holobionts are treated in §3.1 as a candidate replacement paradigm and in §3.2 through Pradeu’s immunological criterion. Only squid+Vibrio enters the framework, as an organism (region 9), and the placement is immediately contested.
  • Suárez & Stencel’s part-dependent account, on which holobionts are both individuals and ecosystems, is the one explicitly part-relative theory cited; it is reported, not assessed.
• Evolutionary versus physiological. The two distinctions cut across the part/organism/group axis rather than along it: each kind of individual has members among parts, organisms and groups. Organisms are the region where the two usually coincide. §6.1’s heading, “Physiological Individuals as Living Agents”, identifies the physiological axis with living agency. Clarke’s policing mechanisms (§8) make evolutionary individuality turn on suppressing selection among parts, and Queller & Strassmann’s organismality is a cooperation/conflict ratio among parts. So on the evolutionary side, the relation of parts to the whole is the criterion itself.

Links to held works.

• LIT-049 (Bourrat’s coarse-graining account) is cited directly in §9 (as Bourrat 2023). LIT-005 (Bourrat on information-theoretic individuality) is not cited. Bourrat 2015, 2019, 2022 (with Doulcier and others), forthcoming (endogenization of scaffolded properties) and Black, Bourrat & Rainey 2020 are cited.
• LIT-157 (Bourrat, reproducers) postdates the revision. Its reproducee / reproducer / Darwinian-individual split directly revisits region 2: it puts genes with hearts as reproducees.
• LIT-188 (Griesemer, scaffolding individuality) postdates it too. It continues Griesemer’s progeneration and material-overlap view, which the entry cites (Griesemer 2000, 2014, 2016).
• LIT-167 (Baedke, The Organism) postdates it. It takes up the entry’s organism-centred and agency material (§6.3: Walsh, Fulda 2023).
• LIT-142 (Meincke on pregnancy) is not cited; the entry cites Meincke 2019 on autopoiesis instead. It is a direct case of the §4 overlap question, whether one organism can be a part of another, argued from a process ontology the entry notes but does not adopt.
• LIT-210 (SEP Philosophy of Microbiology, by O’Malley) matches the entry’s Microbialism thread (§3.1), though the entry cites O’Malley’s papers rather than that SEP entry.
• LIT-160 (SEP Units and Levels of Selection) appears under Related Entries as “natural selection: units and levels of”.
• LIT-201 (Damuth, multilevel selection) bears on the §5 trait-group material; it is not cited.

Bearing on the record

For the nucleation record, this is the organising reference for the biological-individuality cluster. Its most usable content is the nine-region placement table and the vocabulary (Darwinian / physiological / living agent; simple / scaffolded / collective reproducer; policing / demarcation). For the mereology tag it supplies cases of disputed parthood and counting (lichen, reef, holobiont, virus-in-host, organism-in-species), not a theory of parthood. Pairing it with LIT-124 is where the formal questions get asked.

It carries nothing for ML practice. No ANTH- document bears on it, and it should not be cited from the anthology for anything beyond, at most, loose analogy about “individuality” in multi-agent or modular systems, which the entry itself never draws.

Limitations

• It is a survey whose one original framework (Figs 2–6) is openly tentative:
  • the region placements are judgement calls;
  • the key category qualifiers (“some parts”, “some groups”) have no criterion;
  • the diagram is disjoint where the text says the categories overlap;
  • the “all organisms are living agents” stipulation and the “all organisms … form reproductive lineages” claim (§6.2) are not reconciled with a non-Darwinian organism in region 9.
• The §8 meta-ratio proposal has no operationalization.
• The bibliography stops at 2023. Okasha’s sceptical BJPS paper is still “forthcoming” and is set aside, not answered.
• Factual and editorial slips: Buchnera placed in a “human” gut (§4); anabolic and catabolic swapped (§6.1); “Armillaris” (§2); a wrong cross-reference to §4 (§7); inconsistent Clarke years (§1 versus §8).

Open questions

• Which parts and which groups of organisms count as “some”? A criterion would be needed, for instance Clarke’s policing and demarcation or Pradeu’s immune tolerance, applied across the part/group axis. The entry offers none.
• Is squid+Vibrio an organism that is not a Darwinian individual (region 9), or not an organism at all? The entry says this turns on how kinds of reproduction are distinguished (Godfrey-Smith, Griesemer, O’Malley). LIT-157 and LIT-188 are later contributions to exactly that.
• Is a virus a part of its host (region 8), and how does that square with free virions? This is unaddressed.
• Does “organism” survive Okasha’s scepticism about a superordinate category of biological individual? The entry defers the question; reading Okasha’s BJPS paper (doi:10.1086/728048) would settle how much of the framework depends on it.

Corrections to the seeded skim

• The skim says the three-way framework holds “following Pepper & Herron”. In the text (§4), Pepper & Herron only pose the question (is an individual an organism, a part of one, or a group of them?). The framework “that holds that biological individuals include exactly” organisms, some parts and some groups is offered by Wilson & Barker (“Consider then a framework…”).
• The skim’s overlap example, “Buchnera within its host”, quietly fixes a slip in the entry. §4 actually says a Buchnera bacterium may be “part of a corporate organism, such as the human whose gut it is integral to (Andersson 2000)”. Buchnera is the aphid endosymbiont: that is how the entry itself treats it in §3.1 and Fig. 1, and it is the subject of Andersson 2000. The example is right in the skim and wrong in the source.
• The skim’s open question, “does the nine-region framework take a stand on holobionts?”, has an answer. The word “holobiont” appears only in §§3.1–3.2. The framework places one multispecies complex, squid+Vibrio, in region 9 (an organism and a living agent but not Darwinian), explicitly as “perhaps”, and immediately allows that the same fact “disqualifies it as an organism, making it no different in kind from coral reefs” (§7). Whether holobionts are organisms or evolutionary individuals is left as a “live issue” (§3.2).
• Honey-bee colonies: the skim is right that they are Darwinian individuals but living agents only via “the cognitive metaphor” (§7). The Fig. 6 table’s region-3 entry is the general “eusocial insect colony”; honey bees are the in-text illustration.
• The skim does not mention that the general metaphysics of individuality (Strawson, van Inwagen 1990, Lowe, Wiggins, French) is “bracketed off” in §1. That is the key fact for the new mereology tag: the entry uses “part”, “proper part”, “compose” and “member” informally and never engages parthood as such. The tag is justified by §4 and §7, which sort parts and groups of organisms, not by any mereological analysis.
• Smaller slips in the source, not in the skim:
  • §2 prints “Armillaris bulbosa”; the bibliography’s Smith, Bruhn & Anderson 1992 title has Armillaria.
  • §6.1 swaps anabolic and catabolic (“an anabolic dimension in the breakdown of chemical molecules … a catabolic dimension in intracellular synthesis”).
  • §7 sends species and clades to “section 4” when they are discussed in §5.
  • §1 cites “Clarke 2013” and §8 cites “Clarke 2012” for the same identification of individuals with organisms.