Evolving Thoughts

What is an individual?

Rob Wilson has a new entry up at the Stanford Encyclopedia of Philosophy, entitled “The Biological notion of an individual”. It discusses an interesting problem, one that goes back to discussions by Julian Huxley in 1911. What is an individual in biology?

The term “individual” means, etymologically, that which is not divisible. Of course we can divide up organisms, but if we do this physically, they immediately thereafter cease to be the organism. Except… there are colonial organisms that can be so divided – sponges, hydras, slime molds, and so on. To make matters worse (much worse, as we shall see) there are mutual symbionts, such as lichens, formed by a symbiosis of algae and fungi. The notion of an individual is not so solidly based as we might think on the prototypes of vertebrates.

The problems are these:

1. What is an organism? The general notion of an organism is that it is independent of its surroundings, it has coherent and mutually interacting parts, and it cannot be divided without damaging the dynamical relations of those parts. Considerations like Huxley’s, or those adduced in the groundbreaking book by Leo Buss, The Evolution of Individuality, have led to a philosophical conundrum.

2. What is an individual in the metaphysics of biology? and particularly in the context of evolution and ecology? The term “individual” has many meanings. I’ll sort through these in a minute.

3. What else in biology is an individual? Are genes individuals? Species? Eusocial insect hives?

Let’s consider this in a bit of detail.

The notion that there are things in biology that tend to reproduce, interact with the environment as a whole, and have internal coherence, is hardly at issue. Obviously most of the observable biological world is like this. But sometimes reproduction is not coextensive with the other two. Consider a clonal stand of quaking aspen: each “tree” is an individual capable of doing all that any similar organism does, but they are derived by propagation from an underground net of roots, and in genetic terms they form a single individual genome. Likewise when a sponge is divided, the parts will spontaneously reorganise themselves into coherent sponges that can breed in the normal ways.

Then there are cases in which entire populations of what should be seen as individual organisms do not breed except for a single queen, in the eusocial insects. In evolutionary terms, this means that the entire hive is a single organism (a view Darwin first proposed, I believe), which is sometimes referred to as a “superorganism“.

So it appears that biology is confounding our received concepts yet again.

The term “individual” is ambiguous. It has at least three meanings that are relevant in this context. The first is the one most people think of when they hear the term in the context of biology: an individual is a functionally integrated system. So a group of people in a room is not necessarily an individual, but a football team is. This is the prototypical sense in biology, derived, as Huxley noted, from our using human beings and things very like them as the exemplar. It is in this sense that the species-as-individuals thesis of Ghiselin and Hull receives criticism (and Ghiselin has tried to defend). Species don’t appear to be functionally integrated objects (although oddly, Mayr, who used this criticism against the S-A-I thesis, thought that species were in fact functionally integrated gene pools).

The second sense is of phenomenal identity. An individual is something (like the single quaking aspen tree) that presents itself in a way that is individuated from the rest of the environment. In this sense, Mt Blanc is an individual. Things like “a” slime mold are individuals in the sense that when they are fruiting, they form a single body of cells, differentiated out in parts. A problem in biology is that a great many things, such as bacterial communities (“biofilms”) that exist solely as such, are “individuals”, contrary to our intuitions based on the vertebrate model.

The third sense is the metaphysical sense. An individual is, according to the traditional metaphysics of the west deriving from Aristotle, something that cannot be further divided without it ceasing to be whatever it is an instance of. Hence a human being, divided down the middle, ceases to be a human being. But there are two distinct senses of a metaphysical individual. One is that of formal individuality – there are no further differentiae in the definition that will divide the things spoken of into something smaller. So although species can be divided into varieties or races, these smaller groups are not of the same kind as species. The other is material divisibility – a human being can be divided into parts, or a population can be divided into classes or firms or whatever.

The proper (traditional) term for the metaphysical sense of individuality is particular. A particular is something here and now, while the class concept, or natural kind in the sense of science, can be anywhere at any time or place (or none). For example, a gold atom exists whenever the right atomic number is instantiated, whether it comes from earth or Andromeda. We can even imagine a universe that has no gold atoms (like the early universe shortly after the big bang). But to be a species is very restricted – only in one place at one time period do species exist. If there are, qua Star Trek, humanoids that are biologically identical to humans on Arcturus V, whatever they are, they are not human beings – Homo sapiens – because that species evolved once, on earth, in one period, the Recent.

So let us consider what a biological individual might be: it might be functionally coherent (but then, like the sponge, it might not), it might be phenomenally salient (but then, like the bee hive or ant colony, it might not), and it might be a particular (but then again, like the Arcturan “humans”, it might not). It seems like to be an individual we need to specify clearly what kind of individual we are talking about.

Attempts have been made to bring together the particular sense and the phenomenal and functional senses of individual. Richard Boyd has argued for a notion of “natural kind” that includes there being a cluster of properties that maintain some kind of homeostasis. Rob Wilson has argued that this explains species. It founders a little on the cases where individual populations have no contact with the rest of their species for long periods, but which, when brought back into contact, interbreed and are demographically replaceable (that is, they do the same ecological and populational things as their conspecfics). It fails to account at all for higher taxa like genera or families, which are not maintained by any shared functional set of properties. But where there is something that functionally maintains homeostasis, we have a “kind” that is an “individual”.

Another attempt was made by Paul Griffiths, who argued that species (and by implication, all biological individuals) have a historical essence, in which being a member of that species was something based on shared historical origins. This accounts for the separated populations being the “same” species, but lacks a general account of what sorts of things do this.

Back to organisms. An organism may be something that is maintained by a HPC – in fact most if not the near-totality of things we call organisms will be. They have developmental “programs”, lifecycles, constant environmental habitats, and so on. But this won’t help us in the case of the quaking aspens to individuate organisms. Or we might say that something is an organism if it derives from a single propagation event, like Griffiths’ historical essences does for species. In this case, the entire stand of aspens is an individual, intuitions be damned.

But I think the problem is set up because there is a mismatch between our phenomenal differentiation of biological activities, and the consequent functional and metaphysical accounts that we deal with. To make matters worse, it has become clear that crucial biological and evolutionary mechanisms like natural selection can operate on a multiplicity of levels, ranging from the molecular to the populational level, so individuating organisms in terms of their being subject to selective pressures fails, and this happens also for other generic processes or properties like reproduction.

So I want to suggest the following alternative. “Individual” is always something that is contextual. There is no class of all and only individual organisms sharing some general property that holds true for all life. Instead, each “kind” of individual is an evolved trait. There are sexual individuals and asexual ones. There are mammalian individuals and angiosperm ones. It is not a metaphysical notion, nor even an observational one, but a property that occurs differently in differently evolved groups. There’s no reason then to expect that an aspen individual, a slime mold individual, and a mammalian individual will be of the same kind of thing.

Rob proposes a notion of organisms as agents – well and good so far as it goes, but it seems question-begging to me. Agency, which here means something like causal autonomy, is a property of biological systems and processes at all scales, and the question of which scale or kind to privilege is a matter of the interest of the biologist, not the facts of the biology. To say that there is a kind of agency that matters is to force fit biology to our preconceptions, in my book.


  1. #1 TomS
    August 11, 2007

    An individual is, according to the traditional metaphysics of the west deriving from Aristotle, something that cannot be further divided without it ceasing to be whatever it is an instance of.

    Is this a forerunner of the concept of “irreducible complexity”?

  2. #2 RyanG
    August 11, 2007

    Why not call sponges Dividuals?

  3. #3 potentilla
    August 11, 2007

    I agree with your conclusion, and add that there are lot of words like this. Ones which are popularly considered to have an obvious meaning but which in fact thave no perfectly definable referent but which must be carefully defined within context. Person. Personal identity. Right/wrong. Love. Freewill. In general, human concepts which are the product of human evolution.

  4. #4 John Pieret
    August 11, 2007

    Hence a human being, divided down the middle, ceases to be a human being.

    But a human who has, oh, say, lost a finger remains a human being while the finger no longer is. Given modern medicine, we can go pretty far in that process. Would a brain a tank still be a human being while Terri Schiavo wasn’t? If so, doesn’t that give us at least a narrower definition of what we mean by an individual “human being”?

  5. #5 Matt Barker
    August 11, 2007

    Nice blog post. I’d like to make one major point, then a series of minor ones.

    First, the major point. The blog post criticizes Rob Wilson’s Stanford Encyclopedia of Philosophy article on biological individuals. To do this, it implies that Rob offers a monistic account of ‘organism’, and then argues by way of a few examples that any monistic account seems problematic because of the seemingly ununifiable diversity of things that people have called ‘organisms’. In short, the blog post thinks ‘organism’ is hopelessly ambiguous, while Rob thinks there is some hope for a unified account of ‘organism’. The blog post sums this up by saying ‘there is no class of all and only individual organisms sharing some general property that holds true for all life’. This is to say, I think, that Rob attempts to delimit necessary and sufficient conditions for a thing’s being an organism, and yet there are numerous counterexamples in which something we’d call an organism doesn’t satisfy one or more necessary conditions, or doesn’t satisfy sufficient conditions.

    Now, I grant that there may be no set of necessary and sufficient conditions that all and only organisms satisfy. Indeed, I think the search for such conditions when analyzing biological concepts is often misguided. Given this, however, the major problem with the post’s critique is two-fold: 1) the monistic view of organisms that the critique targets is NOT committed to biological kinds (including ‘organism’) being individuated by necessary and sufficient conditions–indeed the monistic view on offer proposes a novel way of DOING JUSTICE TO the diversity among organisms that the blog post illuminates and bases its pluralistic alternative upon; 2) ironically, the post’s pluralistic alternative seems itself to invest in necessary and sufficient conditions. In short, if upon reading the blog post you thought (as one commentator did), ‘yah, that’s right, there is too much diversity in the living world to think the monistic view is right, so I like the pluralistic alternative’, then actually I’ll suggest that your intuitions about diversity may be better served by the monistic view that the post criticizes. To do so, I’ll say a bit about the two sides of the problem I’m posing, i.e. (1) and (2).

    The monistic view of organisms that the post criticizes is based in part on Richard Boyd’s HPC view of natural kinds. One attraction of this view of natural kinds is that it abandons the philosopher’s penchant for necessary and sufficient conditions in favor of taking the work of biologists seriously. The basic idea is that we revise our philosophy to fit the biology (strangely, the blog post implies exactly the opposite). To see this, consider how the monistic view of organisms applies the HPC view (which is described in Rob’s SEP article). First, the idea is not (as a quick reading of the blog post might lead you to believe) that ‘organism’ is an HPC kind. Rather, the monistic view is a tripartite view, where an organism is said to be a) a living agent, that b) has a life cycle and c) is functionally autonomous. The HPC view is only applied to part (a) of this view, that is, ‘living agent’ is thought to be an HPC kind. Now, HPC kinds are individuated by property clusters (e.g., living things tend to have a cluster of definitive properties that includes ‘metabolizes’, ‘grows’, ‘reproduces’, etc). Crucially, no one of the properties that compose the definitive cluster is necessary for being a member of the kind that the cluster defines. So a living thing may, for example, not reproduce, and yet still be a living thing (e.g., drone bees). Some n-tuple of the properties that are TYPICAL of the cluster are sufficient for kind-membership in each case. The properties in the cluster tend to go together because underlying mechanisms tend to co-instantiate them; hence the property cluster is in some non-technical sense homeostatic. But sometimes co-instantiation fails, so that some of the properties typical of the cluster are missing; no big deal, so long as a sufficient number still obtain. What counts as sufficient in each case is primarily up to biologists.

    Because the HPC view is a part of the monistic view of organisms, the monistic view can absorb the putative counterexamples that the blog post offers. Those examples do not threaten the monistic view’s list of necessary and sufficient conditions because that view never offered such a list.

    On the other hand, the pluralistic view does seem committed to necessary and sufficient conditions. This commitment is what motivates the abandonment of the more general kind ‘organism’ in favor of more particular kinds ‘sexual individuals’, ‘asexual ones’, ‘mammalian individuals’ and ‘angiosperm ones’ (as the post puts it). Presumably the idea is that there ARE necessary and sufficient conditions that all and only ‘mammalian individuals’ satisfy. Likewise for the other more particular kinds. Hence we’re supposed to recognize only those more particular kinds and not the more general kind ‘organism’. I doubt that there are such conditions even for the more particular kinds because I embrace the very diversity that biology has uncovered, and to which the blog post appeals.

    Before leaving this major point, it is curious to note that the blog post is unwittingly written in a way that shows just how hard it is to abandon the monistic view of organisms. The post claims that individual is ‘a property that occurs differently in different groups.’ But by appeal to ‘a property’ (even if THAT property can ‘occur differently’) it is sneaking in a monistic view of organisms. What is THE property to which the passage could be referring? ‘Being an organism’, of course. And as the ‘a’ in ‘a property’ implies, this is a single property, i.e., one that is picked out by a unified (even if loosely unified in HPC style) concept. To be clear then, the monistic view of organisms does NOT say that there aren’t distinct sub-kinds of organisms, such as ‘mammalian individuals’ and ‘angiosperm ones’. Sure, there may be such kinds. The substantive claim of the view is that these sub-kinds fall under a broader kind, ‘organism’, that is scientifically interesting. This broader kind is frequently appealed to in biological explanation. Indeed, the monistic view of organisms appeals to biological practice to make this case, rather than fitting biology to the philosophy. Along the same lines, the view can accept that ‘individual’ often seems to have the three meanings (functional, observational and metaphysical) that the blog post implies, but it can then note that it is very interesting how often these three notions of ‘individual’ apply to one and the same thing. This interesting fact cries out for explanation, and the monistic (and partially HPC) view of organisms helps on this score.

    Okay, so much for the major point, now for the series of minor ones.

    -The post notes that it is puzzling that Mayr criticizes the view that species are individuals by claiming that species aren’t functionally integrated, when yet Mayr is famous for the view that species are functionally integrated gene pools. This isn’t so puzzling given that Mayr’s claim was actually that species aren’t functionally integrated ENOUGH or IN THE RIGHT KIND OF WAY to be considered individuals. He commits no puzzling inconsistency here. Just as the blog post is pluralistic about ‘organism’, Mayr and many others are pluralistic about ‘functional integration’. There are lots of different kinds of functional integration. Many kinds are not sufficient for individuality, while yet they are sufficient to make the entity in question a scientifically interesting agent (e.g., the immune system is an agent even if it is not an individual).

    -The following is important and may not be clear from the blog post: the monistic view of organisms doesn’t claim that ‘organism’ and ‘agent’ are identical. Indeed, the view explicitly rejects that. Rather, organisms are a very special sub-kind of agent (so lots of agents aren’t organisms). Neither is ‘organism’ identical with ‘biological individual’. Rather, ‘organism’ is again a sub-kind of ‘biological individual,’ which is in turn a sub-kind of ‘individual’.

    -In passing the post makes a few criticisms of the attempt to apply the HPC view to species. It mentions that there are counterexamples in which certain species consist of populations that aren’t proximal. As the above implies, the HPC view is designed to accommodate these very counterexamples. For similar reasons the view is far more adept at capturing higher taxa than the blog post implies as well. See Boyd’s and Wilson’s original works for thorough discussion here.

  6. #6 John Wilkins
    August 11, 2007

    Wow, Matt, nice response (you’ve worked on this recently, haven’t you?). One of the limitations of a blog is that you cannot write the sort of measured treatment and retain your readers, most of whom are nonspecialists barely if at all aware of the issues. Dividing positions into things like the monist view and the pluralist view is something I simply cannot do here. But I take all your points to heart. I have, of course, my own opinions about this.

    Mayr’s point is contradictory because he himself makes a functional claim for species cohesion, so he offers a way to integrate species as individuals. It’s not so much a self-contradiction as an oddity, which is why I framed it the way I did. He could have gone with the S-A-I view if he’d wanted to with no damage to his own view of species. That he didn’t indicates that his prototype of an individual was solidly founded on an orgamismal view – which, as a field biologist, is understandable. But it is not philosophically compelling.

    One of the things about this debate is how much the paradigm concepts [NB: Non-Kuhnian use of the p-word] are based on metazoan, vertebrate biases. My suggestion is that these are evolutionarily derived is something that the “monist” view (as if there were such a beast amongst biologists, which I doubt) simply doesn’t deal with. Microbiologists need to appeal to unit concepts that are radically different to those of metazoan biology, and metaphytic concepts are something else again. But the debate continues to be framed as if there were a rank that was universal and generally, problem cases aside, applicable in all domains of life. As I have elsewhere discussed, there is a prima facie case against there being fixed unit concepts in biology at all, only prototypical ones or exemplars that have limited applicability, and the evolutionary story is a reason why.

  7. #7 Matt Barker
    August 12, 2007

    Hey John. On Mayr: I’m not sure why he actually stopped short of calling functionally integrated species ‘individuals’. He may have had an organismal view of individuals in mind, one that made species seem to far from individuals. But there are other ways that he could have defended his view that would be philosophically compelling. Hull often advanced S-A-I by saying that species are relevantly like paradigm individuals such as organisms in that they, like organisms, are a) spatio-temporally bounded, b) contiuous, and c) cohesive. He implied that these are the marks of individuality. I think many metaphysicians (e.g., Ayers, Armstrong, Wiggins in places) agree with parts or all of this. Mayr’s view can then be seen as accepting this, and then holding to the idea that (a) through (c) comes in degrees or kinds; a certain kind of each is needed for individuality, and species don’t exhibit these kinds. I’ve argued for this elsewhere myself and Rob and I have a paper on there being distinct kinds of cohesion, such that species (even in Mayr’s sense) often have one kind of cohesion but not the kind that in part makes organisms individuals. Sure, BSC-species may exhibit a kind of functional integration, but it just isn’t the integration that many metaphysicians think is necessary for individuality.

    This doesn’t, of course, imply that we can’t view species as agents. But agent-hood is not sufficient for individuality. Individuals are just special kinds (i.e., bounded, continous, cohesive) of agents. A think, like a species, can be an agent without being an individual. Though I think it is also an interesting question whether species are even agents. Certainly Mayr is committed to saying ‘yes’. And Hull said something like ‘species are the things that evolve, split, bud-off new species.’ But I’m not sold on this. I sometimes think species may just be epiphenomena that, perhaps, need explaining. Populations are another matter.

    I haven’t read your case against fixed unit concepts in biology. But I’m sympathetic to parts of what you said above. Although I think there may be something special about organisms that marks them off as a kind, biological agents are a broader kind admitting of many sub-kinds, many of which (like microbial communities in which there is lateral gene transfer) are agents in part because of the context in which they’re in.

New comments have been temporarily disabled. Please check back soon.