What is function in biology?
Abstract: In this paper, I am going to argue Prior’s account of functional analysis is most promising theory of function in biology based on the selected important desiderata for a theory of biological function. I will then defense Prior’s view against Neander’s attack from the prima facie problem with teleological explanations.
In the second half of the course, we have surveyed many accounts on theory of function in biology. Namely we have studied a few rejected classical theory of function in  and , the standard etiological approach proposed by Wright , the system approach proposed by Cummins , the goal oriented approach proposed by Boorse , the forward-looking approach proposed by Bigelow  and the functional analysis approach proposed by Prior . Neander’s paper does not make any new proposal but rather fortify the Wright’s argument. Godfrey-smith’s paper spoil the fun of the party by proposing that the two major camps in this debate, Wright and Cummins respectively, are talking about two different notation of functions. Out of the all the functional theory, I think Prior’s account that is based on a modified Cummins approach is the most promising approach to explain what is function in biology. I will first summarize the desiderata for a theory of biological function by collecting all the functional criteria from all the papers.
An acceptable account of theory of biological function must able to answer the following questions:
- An acceptable functional theory must distinguish between function and merely effect. For example, the heart’s function is to pump blood; its function is not to make throbbing sound. This example is used by both Wright and Cummins. This is pretty much one of the most fundamental acid test for any functional theory.
An acceptable functional theory must explain both artificial (conscious) and natural (biological) function. Actually in my opinion, throwing away this criterion will open up all the limitation of biology functional theory. If we can treat artificial function and natural function as two distinguish and non-intersecting categories, most of the argument in functional theory will simply disappear. Biological function will become much like physical function and chemical function that is just a descriptive account of the causation effects and what happen as it is. However Wright, Cummins and pretty much everyone else choose to keep the criteria, I have to include it in this paper to keep the premise consistence with all the papers
An acceptable functional theory must distinguish function from malfunction, non-function, vestige, statistically normal function, accident and unintended function. This criterion is where the different functional theories disagree with each others, in another word this is the heated battle ground. The Cummins camp attacks the Wright camp that the etiological approach cannot deal with any one of those distinctions or all of the above. The Wright camp defenses the etiological theory by counter argue that their account can handle the distinctions. One problem I observe is that the actual meaning of function seems twisted one way or another by each camp to fit their functional theory. Here I will propose an acid test for each type of distinction to keep the criterion simple. A heart that cannot pump blood, say there is some genetic defect and the heart is born with a missing a valve, is a malfunction heart. An appendix is a non-functioning and vestige organ in human. If a virus causes 99% of the population to go blind, the function of an eye is still to see. The function of a belt bucket is not stopping bullet, although sometimes accident does happen and it does stop bullet and saves life. The function of Prior’s hypothetical grape-prawn peeler is to shell prawn. Alternatively, we can substitute a real world example for the grape-prawn peeler. When laser was invented, it was intended be used as laser weapon, but it turns out laser also is very useful in reading data from DVDs.
An acceptable functional theory must take into account of super function. Here is an example of super function. Normally the function of an eye is to see. However in the comics X-Men, Cyclops, a mutant superhero, can shoot laser from his eyes. Therefore function of his eye is both seeing and shooting laser. Yet we do not know whether shooting laser from eyes has any evolutionary advantage or even an inheritable genetic trait. The comics do not mention he has any offspring and just the fact of being a mutant always puts his life in great danger. In the next section, I am going to outline Prior’s functional theory and contrast it with Wright’s etiological theory. I will examine both theories base on the desiderata defined above. Prior’s define function statement as the output of functional analysis. “According to this account the ascription of a function is both theory and interest relative – the function of a particular organ will be that of its effects which features in our best account of some interesting (higher-level) capacity of the organism in question.” He also said that his account is familiar similar to Cummins’ functional analysis that “functions are appealed to in order to explain the capacities of their containing systems, and not the existences of the item to which the function is attributed.” In contrary, Wright’s etiological account states that “The function of X is Z means: (a) X is there because it does Z, (b) Z is a consequence (or result) of X’s being there”  In short, Prior’s theory does not care about how the organ comes into place, it only cares about the effect and relationship between the organ and the organism or the organism and the ecosystem. On the other hand, Wright’s theory requires the function to explain how the organ comes into place.
Both Prior and Wright’s theory easily pass the first criteria although they have different explanation. In Prior’s account, he can follow Cummin’s explain the function of heart under the context of the blood circulatory system. Since the pumping of blood but not making the heart beating sound is a desired effect of the heart in the circulatory system in explaining its capacity, we can distinguish the function of the heart as pumping blood against making noise is merely an effect. Wright would explain the function of heart by stating that the heart pumping blood gives evolutionary advantage to the species hence increase the fitness to survival for those who population within the species who process this genotype, as a result a blood pumping heart exists because that is what was selected for. The heart beating sound does not give any evolutionary advantage, it is a selected of, a trait merely got inherited tagging the tagged along with the pumping function of heart. Since the first criterion is the most basic requirement of any functional theory, I am not surprised both theories can explain it well.
The second criterion is does not have much controversy since both Prior and Wright deliberately took the effort to make their theories compatible with artificial selection. Prior’s approach actually originates from functional account of artificial function and extend it to cover natural function. The function of a mechanical part is determined by its design intention within the mechanical system. In the pre-Darwin where creationism rules the study of nature theory, natural function pretty much share the same account as artificial function. The only difference is the conscious agent for artificial function is human and the conscious agent for natural function is God. Medieval doctors explain biological function in term of physiological function which a result of the functional analysis procedure for the whole body. However this account is not without problem. By definition no one know the real intention of God, so the natural function is really only based on what people’s guesses about God’s intention or on powerful religion figure’s claim on knowing the intention of God. If the intention is not known or incorrect, the function could be totally wrong. For a long time in the history, people think generating emotion is also a function of the heart, which is obviously wrong under the light of modern science. Wright’s approach is the opposite; it extends the biological account of natural function to cover artificial function by replacing natural selection with the selection by the designer or user. In this sense, it actually simplifies the problem of the etiological account. The biggest problem of etiological account is base on the natural selection theory which it is somewhat questionable as we learn in previous lectures. Replacing natural selection with an artificial selection originated from a conscious agent eliminate the uncertainty in the equation. If there is any doubt, we can always get clarification by asking the agent for his selection criteria.
Prior’s theory does not have much problem with the third criterion. Once the nominal function is identified through functional analysis in the context of the system and its capacity, any derivation from the nominal function can treat as malfunction, non-function or accident. If the system changes and the nominal function are no longer required, it becomes a vestige. If the context of the system and its capacity is indentified correctly in related to a higher level system, we can deduce the intention of the function and figure out the difference between the statistical normal function and the nominal function. Unintended function can be deal with by a running a new functional analysis after updating the context and capacity of the system. Wright account has a harder time answering the question posed by the third criterion. Malfunction and non-function can be understand in the relationship of token and type, deviations from of token from the type that lower the fitness can be seen as a malfunction or non-function. Vestige is the easiest to deal with since the etiological account keeps track of the historical record of the function. Statistical normal function is a bit harder to deal with using the backward looking account, but if we patch the etiological theory with Bigelow’s  forward looking account, we can still distinguish between function and statistical normal function. For example, a virus that blind 99% of the population would not change the function of the eye, because having an eye that can see still have an evolution advantage. In the etiological account if a function is not inherited then we can treat it as an accident instead of a function. Backward looking account cannot deal with unintended function, but it has no problem in the forward looking account since unintended function is still a function as long as it has a selective advantage.
I believe the forth criterion really sets the two theories apart. According to the comics, Cyclops acquires the ability of laser eye due to first generation genetic mutation. Clearly there is no previous selection history Wright can use in the backward etiological account. Having a laser eye does not make him more attractive to women, so it does not give him any reproductive advantage. In the world of X-men, the government hunts down mutants, the evil mutants launch a war on the good mutants for world domination and occasionally the good mutants have to defense the Earth from alien invasion. Mutants seem to have a worse survivability compare to the normal people. If there is no forward selection advantage, we cannot use Bigelow’s forward etiological account. However, it would be odd to say shooting out laser is not a function of Cyclops’ eyes because that’s what they really do. Prior’s account has no such problem, we can examine Cyclops as mutant, which is a different system compare to normal people. We can easily perform functional analysis on his eyes and concludes they have the function of seeing thing and shooting lasers. This approach is exactly how the medieval doctors determine the function of human body parts and the X-File scientist in Area 51 figure out functions of alien body parts when they dissect the alien.
In Neander’s paper , she argues against Cummin’s account following a different strategy. She employed the prima facie general problem concerning all teleological explanations which explains the means by the ends; a development or trait is explained by reference to goals, purposes of functions. Teleological explanations appeals to the effects but it is often that the purposes, goal and functional effects are never realized. In the case of artificial function, instead of appeal to the actual effect of the function, the teleological explanations appeal to the intended effect of the agent. When the purposes go unfulfilled, the agent still has a purpose. Clearly, the explanatory power of purposive explanations does not derive from their explicit reference to future effects so much as their implicit reference thereby to past intentional attitudes to those future effects. In short, teleological explanations are just a species of ordinary casual explanation. The problem with biological function is there is no intentional agent. Prior cannot appeal to God nor Mother Nature’s intention or evolutionary design, since those misguided pseudo-explanation are not scientifically respectable.
In Prior’s paper, he sort of response to the teleological problem by stating that it is just a “fundamental fact about human beings (or animals, plants, insects) that they are highly interested in survival and reproduction.”  He dodged the theological question of the ultimate intention by simply assuming it is a fact. He also stated that the ultimate purpose of biological life from is contagious by citing an example of an imaginary world with human life form wish to die happily instead of survive and procreate. I am not very please with Prior’s defense since his defense is pretty weak. He is just making a groundless assumption about human nature or biological life form in general. I can already contradict his idea about human purpose with any major religion beliefs. For example: to Christianity, the purpose of human being is not to survive or to reproduce; rather it is to go to heaven; to Buddhist, the purpose of human being is to reincarnate and reach nirvana.
I am going to response Neander teleological problem by using the human centric view that stating it is us, the human being, gives purpose to the nature. Nature may not have any purpose by itself; but that does not mean human cannot assign some purpose to the nature. At the end of the day, it is human, neither God nor Mother Nature, is where the conscious agent that all the teleological intention ends. With the advance of modern science, like genetic engineering, geo-engineering, climate control, space colony, human will eventually has the power to shape the nature to the way we want. There will be no natural intention, everything is artificial intention. For the time being, natural intentions are merely artificial intention that we haven’t figure out a way to control yet. To summarize, the function of an egg to a chicken is a mean to reproduce. The function of an egg to human is breakfast, and the function of a chicken to human is lunch or dinner. Prior’s account is the most possible explanation for function, my minor twist to Prior’s account is human is in the center of all system, so we are the conscious agent and eventually nature function will be mostly equivalent to artificial function, except that nature function is generated by evolution engineering.
1. Wright, Larry (1973) “Functions” Philosophical Review. Vol. 82, pp. 139-68
2. Cummins, Robert (1975) “Functional Analysis” Journal of Philosophy. Vol. 72, pp. 741-64
3. Boorse, Christopher (1976) “Wright on Functions” Philosophical Review. Vol. 85, pp, 70-86.
4. Bigelow, John, and Robert Pargetter (1987) “Functions” Journal of Philosophy. Vol. 84 (4), pp. 181-96.
5. Neander (1991) “The Teleological Notion of ‘Function’” Australasian Journal of Philosophy.
6. Prior (1985) “What is Wrong with Etiological Accounts of Biological Function?” Pacific Philosophical Quarterly.
7. Godfrey-Smith (1993) “Functions: Consensus Without Unity” Pacific Philosophical Quarterly.