I'm trying to understand how broad the definition of "phenotype" is. If someone gets a tattoo, would that tattoo be considered part of their phenotype?
Broadly speaking it most likely is. The exact meaning of phenotype has changed quite a lot over the years. There isn't a clear distinction between what should or should not be included in the definition, and reasonable people can disagree on sound arguments.
Phenotypes were defined by Johanssen in 1911. His definition was "All “types” of organisms, distinguishable by direct inspection or only by finer methods of measuring or description, may be characterized as “phenotypes.”". So in his definition a phenotype is a class of individuals that share a common observable property. Take Mendel's experiment: short and long peas are 2 phenotypes. From heredity we can infer multiple genotypes. An identical phenotype (tall) can result from 2 different genotypes (tall-tall, or short-tall). That is why the distinction is important, because truly understanding genetics will require not only to understand the genetic makeup of individuals (we have more or less solved that), but more importantly how the genes are expressed into a phenotype. Keep in mind all of this was speculated long before we actually found out what genes were. Ironically, Johannsen thought it was an idle pursuit: "The question of chromosomes as the presumed “bearers of hereditary qualities” seems to be an idle one. I am not able to see any reason for localizing “the factors of heredity” (i.e., the genotypical constitution) in the nuclei.".
At the time, genes were really thought to produce strict discrete categories of individuals that were slightly smoothed out through random individual variations. Over time it became obvious that Mendelian traits were more the exception than the rule and that trying to classify individuals in discrete categories was not really helpful. So over time phenotype became to used for traits instead of individual. That's the way I hear people use it today at least (i.e. we would call "being tall" a phenotype of peas, and not "all peas that are tall").
So the modern definition of a phenotype is the observable characteristics of an organism, anatomical, physiological or behavioral. The first 2 points are easily accepted and uncontroversial. But there is debate about what exactly should be included in "behavior". I'm going to ignore behavior for now.
An important point, and I think this is where there is disagreement between Remi.b and me, is that a phenotype must have a genetics basis. I realize this is not explicitly mentioned in any definition, but that is very much the way people mean it. And if not, the definition is totally meaningless. Take the example of a monkey missing a finger congenitally. This is a phenotype, it is due to its genes. If the monkey is missing a finger due to a fight, this is not a phenotype because genes have nothing to do with it (again, ignoring behavior for now).
So, in my opinion, what would not be a phenotype is something that you could conclusively prove to have no genetics basis whatsoever. That's an extremely (impossible) hard case to make. But in retrospect I was perhaps to quick to include behavior in the definition, and this is due to my own biases as a behavioral neuroscientist.
I think this is the narrow definition of phenotype and I don't think anyone would disagree up to that point. You might notice that there has been no mention of evolution so far. It is because genetics and evolution have been developed independently, and even though everybody was quite aware they must be 2 sides of the same coin evolution is not directly relevant to the genotype/phenotype distinction.
This is when the definition of phenotype started to be extended beyond purely physical traits (people included behavior in phenotype but they mostly meant things like gait that require a behaving animal to measure, but it is not behavioral in the broad sense). I see 2 reasons for that change, and R Dawkins was largely involved in popularizing these ideas. First as genetics and evolution started to be integrated together, it became obvious that genes must be selected through evolution, and that phenotype must reflect something related to fitness. W. D. Hamilton did important work on the topic and his ideas were popularized in the "The Selfish Gene" from Dawkins. The second idea is the ethological "revolution" from Tinbergen and Lorenz. Until then animals were seen as automatons (following Descartes) that behave the way they do because of simple conditioning. Ethologists showed that complicated behaviors such as courtship or altruism are instinctive, and therefore must have a genetic basis (Hamilton also worked on altruism a lot). A new generation of scientists pushed things even further with people like E. O. Wilson who studied the social life of insects, J. M. Smith who showed that individually detrimental behaviors can be optimal at the scale of a population etc.
At this point it became clear that the definition of a phenotype was too narrow, because behaviors are (at least partly) related to genes and therefore must have been shaped by evolution. This is the idea put forward by Dawkins in his second book "The extended Phenotype" where he explicitly makes the case that the definition of phenotype should be enlarged to include behaviors such as courtship, nest-making and the like (Dawkins was a student of Tinbergen so it's not entirely surprising he is biased to like animal behavior). Although these 2 books became very popular and widely read in the general public, they were actually meant mostly for biologists. Note that the extension of the definition of phenotypes shifts the focus from pure genetics to evolutionary theory as the rational for extending phenotypes is that more stuff than pure physiology contribute to fitness. That's not entirely relevant to the genotype/phenotype distinction but there is a great interview of Smith by Dawkins where he discusses a lot the history of genetics, evolutionary theory, and the socio-political debates behind it (link below).
So today some behaviors are widely, and I think uncontroversially, thought to have a genetic basis and therefore should be considered a phenotype. This is the second part of the disagreement between Remi.b and me (which derives from the first). If phenotypes must have a genetic basis, then how do you know that a behavior has a genetic basis or not? The gold standard is to show heritability of the trait in a twin-study. If the trait co-occur more in identical twins than fraternal twins, and/or in identical twins raised separately than expected from chance, then you know with confidence that this trait is partly genetically determined. Most famously, for example, this is known to be true for sexual orientation. So I am not saying that a phenotype can only be inherited. Obviously a random mutation is a phenotype but has not been inherited. But to show that a behavior is genetically determined I don't know how you can do that without showing heritability (but I'm not a geneticist so maybe there are other ways). Siddhartha Mukherjee's The Gene is a very good summary of the state of the art in that area.
So where does that leave us concerning phenotypes and tattoos? Well if you agree that behaviors should be considered phenotypes, such as courtship, nest-making in birds, aggressive displays in monkeys... Then I see no reason not to include tattoos in it. Any criterion you apply to "animals" should apply equally to humans. Some birds decorate their nests with flowers as part of courtship. The question left is whether it has a genetics basis or not, and I think you are very likely to find that getting tattooed would co-occur more than expected from chance in twins. Of course the tattoo is not genetically determined. It's probably something like open-mindedness, but a tattoo is observable and open-mindedness is not. This definition of phenotypes means that you can more or less already include quite a lot of things in it. So I can see why someone would disagree with extending the definition of phenotype beyond pure physiology. But clearly phenotypes must be linked to genetics otherwise the definition is purely meaningless.
Johannsen W. The Genotype Conception of Heredity. The American Naturalist 1911;45:129-159.
Dawkins, R. (2016). The selfish gene. Oxford university press.
Dawkins, R. (1982). The extended phenotype (Vol. 8). Oxford: Oxford University Press.
Tinbergen, N. (1951). The study of instinct.
Lorenz, K. (2013). The foundations of ethology. Springer Science & Business Media.
Smith, J. M. (1982). Evolution and the Theory of Games. Cambridge university press.
Wilson, E. O. (2000). Sociobiology: The new synthesis. Harvard University Press.
Siddharta, M. (2016). The gene: An intimate history. Scribner, New York, 9-9.
Standard definition of "phenotype"
The term phenotype is generally not clearly defined. Most often, textbooks will define a phenotype as any anatomical, morphological, physiological or behavioural observable characteristic of an individual. This definition is somewhat vague as it is unclear how far reaching the impact of the behaviour can be. To my knowledge, the only clear definition of a phenotype is its broadest definition as popularized in the book The Extended Phenotype.
The Extended Phenotype definition
In its broad definition, a phenotype is any manifestation in the world of the presence of a given genotype. I am unsure of the detailed presentation of this definition in the book but here is briefly my extended definition.
With the "extended definition", an individual phenotype is not limited to its body. Consider a bird nest as an example. In the "standard definition" only the bird building the nest has a phenotype, the nest itself in not part of that phenotype. With the "extended definition", the nest itself is part of the phenotype of the bird who built it.
With the "extended definition", an individual phenotype is not limited to its temporal existence. An individual can have much consequence to the world after its death. For example, the stoic philosophy of Marc-Aurel Caesar that has affected many people's life after his death is still part of Marc-Aurel Caesar's phenotype.
In short, imagine a world with and without an organism. The difference between these two worlds is the phenotype of that organism. Given the fantastic complexity and connectivity among events in the world, a lot of the world's properties can then be considered part of an individual phenotype.
Which definition is correct
Well, a definition is never correct or wrong. It is just a definition. But let's talk about clarity, uses and fit to theoretical models for both definitions.
It is somewhat unclear what consequences of the genotype in the world are still included in the "standard definition" of phenotype. As such the "standard definition" is somewhat unclear.
For the vast majority of purpose authors use the term phenotype to refer to specific characteristics and therefore, the detailed definition does not matter.
Fitness of theoretical models
In theoretical work, the extended definition is the only definition that fully satisfy our models of ecology and evolution.
Phenotype and heritability
In a comment under the original post @Dirigible's said
No. Phenotypes are qualitative and quantitative characteristics that arise from the interaction of an organism's genetics (genotype) and the environment. There are no genetic determinants for tattoos. This question lacks background research and I'm suggesting that it be closed.
There might be confusion with the concepts of inheritance and of heritability here. Heritability of a given quantitative phenotype is the fraction of variance of this phenotype in the population that is explained by genetic variance (see this post for more info). If a phenotype (that has some non-zero variance) has no underlying genetic variance, then this phenotype is not heritable. But it still remain a phenotype.
There are no genetic determinants for tattoos as a way to justify the
No starting the comment, the comment sounds like (whether or not it was the poster's meaning) a phenotype is a phenotype only if it is heritable. I would strongly disagree with this claim.
More straight forward, @baca in his/her answer said
What would not be a phenotype is something that you could conclusively prove to have no genetics basis whatsoever.
Again, I disagree with this claim.
On top of that, I would argue that tattoos are very likely genetically determined. There are probably genetic determinants for liking tattoo. Probably for liking art in general. There are probably genetic determinants to joining gangs (such as loci affecting aggressiveness for example) that sometimes use tattoos as symbols of membership. There are skin conditions that prevent people from getting tattooed and some of these conditions are likely to be heritable.