The Case Against Teleological Reductionism

Larry Weright

http://scienceandbelief.org/2013/04/04/is-teleology-a-useful-concept-in-biology/

Is Teleology a Useful Concept in Biology?

Everyone uses teleological language – the language of purpose – even when it doesn’t seem to make rational sense. We ascribe motives to everyday processes, so the car won’t start because I don’t want to go to the dentist, your little brother fell down the hole because he was bad yesterday, and the ball keeps rolling away because the dog wants to play with it. We also tend to ascribe purpose to biological processes, so flowers grow because we enjoy them, and the DNA code came into existence in order that life can develop on earth – you might check yourself after saying or thinking such things, but we all think them. A recently study* showed that scientists are naturally ‘promiscuously teleological’, but the current thinking in science is that we shouldn’t use teleological language. The world is the way it is by chance and not by design.

Last month, biologist Harvey McMahon spoke at the Faraday Institute on ‘Teleology and Biology: Friends, Foes, Mutually Exclusive or Complementary?’ McMahon and the members of his research group investigate the curvature of membranes, looking at how the physical shape of cells and their membrane bound ‘organelles’ is suited to function. In his seminar he argued, in a humorous but direct way, that biological language is not pure: it uses language from many fields in the attempt to understand itself. Stripping teleology from our language dehumanises biology, and science suffers for it.

McMahon began his seminar by explaining that biology and teleology seem to be going in different directions. Teleology is about purpose and is a philosophical concept. If something is designed, there was a prior knowledge of what it was for, and perhaps some sort of blueprint. Biology, on the other hand, is about mechanisms and things that are testable by experiment. Naturalistic principles state that there was no prior knowledge of the purpose of any organism, but every living thing evolved within a particular set of constraints.

Despite appearances there is some overlap between teleology and biology. Teleological language is used in biology, despite the best efforts of journal editors. Perhaps there is also a partial overlap in terms of function. Every biological entity has a function that is gained from relationship with its surroundings. Organisms without a function – or purpose – don’t make it through selection. Also discoveries in biology often have strong teleological implications. I think here McMahon was implying that at some point down the line, what biologists find usually has an effect on the way we live. We ourselves gain significance from a sense of order and purpose, and naturalistic mechanisms fit within the overall design and purpose that we see in life.

The purely mechanistic and reductionist view of life that is presented by some scientists has the consequences of isolating it from the totality of life, rather than integrating it. The view that we are simply the consequences of our genes or that given enough information we could be programmed to act in the way we do, isolates the physical world from the totality of meaning. This isolation results in a breakdown of integration. For example, the moral boundaries around our science become blurred and either have to be imposed from outside or in some cases do not exist. What McMahon suggested should be the case is that science should develop in relationship with other aspects of the cosmos, and this will impact on our human relationships, our sense of significance and our well-being.

A number of philosophers also have a problem with isolating parts of life, instead of looking at things as a whole. For example, take the supposed neutrality of science with respect to ethics and politics. Individual scientists are not neutral, so is science neutral? Who we are and what we do are very tightly linked. McMahon also used an example from his own work, where theoretical and experimental biologists work in such separate spheres that they don’t always understand each other, and are at times unable to see the value or relevance of each other’s work.

Science critiques everything that it is capable of exploring, and we can become boxed in by that thinking. In the end there’s very little left to rely on, and we’re only able to fully understand very small and tightly defined parts of biological systems. If this were the only way of thinking about the whole then we would have reductio ad absurdum. McMahon argued that there is a need to emerge from fine focus and allow space for other ways of thinking.

Language acts as a lens or metaphor. Neurobiologists often use the language of computer technology in neurobiology. Switches and central processing units are convenient analogies for nerves and the brain. Does this language limit exploration? Yes to some extent, because neurons are not switches, and the brain is not a computer. Analogies have limits. McMahon suggested that the language of evolution might at times hinder, and the language of teleology might help discovery. Looking through a lens is a necessary part of any intellectual endeavour, but we may need more than one lens in our toolkit.

The Case Against Teleological Reductionism

Larry Weright

“Following the vitalism/mechanism controversy earlier this century, there seemed to prevail among empiricist philosophers an almost paranoid suspicion of teleological concepts. Anything that could cause so deep a confusion and so sharp a division among intelligent men must, it seemed, be intrinsically obfuscatory. Hence, wherever it appeared, the smoke of teleological terminology implied the fire of sloppy thinking.

So the desire to expunge “teleological” explanation was seen a s crucial to developing a unified, mechanistic approace to scientific explanation (of both inorganic and organic reality).

Cites several empiricists who employ a reductionist strategy

They show that the teleological description of a phenomenon (or class of phenomena) is equivalent to some other description which employs only mechanistic terms.

So reduced, the formerly teleological descriptions achieve a “respectable logical status”

These reductions seem to take for granted that a respectable equivalent non-teleological expression is a necessary condition for the legitimacy of a teleological expression.

He looks at three classic reductionist analyses of teleology

2 LOGICAL AND METHODOLOGICAL PRELIMINARIES

The position that every legitimate teleological description is, or must be, equivalent to some non-teleological description of the same phenomenon has been expressed in several different ways.

'translatable into'

'reducible to'

'saying the same thing as'

same 'cognitive content' and differ only in 'emphasis'.

Asserted a logical equivalence

Considering:

1. Whenever it is appropriate to describe a phenomenon teleologically it will, as a matter of empirical fact, always be the case that the same phenomenon can be described non-teleologically.

2. A legitimate teleological description logically entails a non-teleological one.

The position suggests:

· Suggests that teleological descriptions involve nothing new epistemologically.

· Requires that the two accounts be logically equivalent.

Distinction between

1. descriptions which are equivalent in this tight, logical sense

2. descriptions which might be corresponding or correlative in one way or another, but which, if equivalent at all, are so in a weaker sense.

The phenomena which are usually the object of teleological description can be described as a series of events, (e.gg mating of Siamese fighting fish) without in any other way relating the individual events to one another,

Thus such a description does not suggest that there are other objects of similar appearance and behaviour.

“I might say: 'First this happened, then this, and then this'. And the demonstrative pronoun could, in each case, be replaced by a statement in terms of blobs of colour, more or less solid objects, motions of these objects, etc., containing no mention of mating, fish or even life. But the ordinary account of two Siamese fighting fish mating is not logically equivalent to the physico-geometrical account of that phenomenon. If offspring never resulted from the series of events we now refer to as 'the mating of Siamese fighting fish', and if, further, offspring did result from some other series of events, then the original series of events, though in no way changed in themselves, would not be the mating of Siamese fighting fish. The 'mating' account and the physico- geometrical account are saying different things about what is happening, and hence cannot be logically equivalent.

Similar descriptions teleological events are certainly possible, but not logically equivalent.

They are saying something different about what is taking place.

The teleological account says “more” than the physico-geometrical account.

'John went to the store to get some bread' says more than

'John went to the store, and John got some bread'

'the shark was pursuing its prey' says more than

'the shark was travelling behind a certain fish at a certain distance'

‘the missile changed course in order to intercept its target' says more than

'the target changed course and then the missile changed course'

-and even that it says more than 'the missile changed course because the target changed course'.

Generally not disputed that teleological descriptions say more

What is disputed is

1. What is this 'surplus meaning'

2. Whether or not it can be expressed in non- teleological terms.

Boundary separating teleological expressions from non-teleological expressions unclear

Certain terms:

(e.g. 'follow', 'normal', 'important', 'attain') can be “teleologically pregnant.”

Regardless, there are many expressions which are clearly teleological and others that clearly are not.

Most attempts to demonstrate the reducibility of teleological expressions to non-teleological expressions

1. Begin with the a paradigm of teleological phenomena

2. Produce an abstract schema based on that paradigm

3. Produce a non-teleological description of that schema

John Canfield notes:

Attempts to translate teleology out of biology fall into one of two patterns

A) The system of translation is a target schema

The guiding example such instances of teleological behavior as a cat chasing a mouse, or, on the mechanistic side, a homing torpedo proceeding to its goal.

B) The system of translation is a furnace schema.

Different kind of teleological behavior exemplified by a house equipped with an automatic furnace. This schema best fits such biological behavior as homeostasis of temperature and of blood sugar.

Canfield seems to imply that these are mutually exclusive

Wright is unconvinced.

Nevertheless, the distinction provides two usefully different ways of appreciating translation attempts.

3 REDUCTION THROUGH FEEDBACK

Rosenblueth, Wiener and Bigelow

Offers a translation rule which is most easily seen as the product of a target schema.

Teleological behavior is directed towards a goal and involves

'a continuous feedback from the goal that modifies and guides the behaving object'.

Purposeful reactions are those

'which are controlled by the error of the reaction'.

teleological behavior

'becomes synonymous with behaviour controlled by negative feedback ...'

And “feedback” completely describable in non-teleological terminology.

Difficulties:

The problem of specifying a goal object without using the teleologically loaded terms 'goal' and 'error'.

Israel Scheffier: 'The Difficulty of the Missing Goal-Object'

Scheffier points out that there are many kinds of phenomena which are undeniably teleological, but which contain nothing identifiable as a goal-object from which the signals required by the feed- back analysis could be emanating.

“ ... a man's purpose in groping about in the dark may be to find matches that are not there, his purpose in going to the refrigerator may be to obtain a non-existent apple, he may seek the philosopher's stone, the holy grail, the fountain of youth, or a living pulsing unicorn. In every such case his behavior is clearly purposive and yet in none is this behavior guided by signals emitted from a goal-object...”

In reply to the possible objection that these examples all involve human intention and explicit decisions, whereas what Rosenblueth, Wiener and Bigelow were trying to analyze was purposive behavior in non-intended and non-human cases.

Scheffler observes that the same problem arises in the cases in which there is no intention, no explicit decision and even no human.

Examples of such cases are:

1. a standing passenger thrusting his foot outwards suddenly in order to keep his balance in a moving train

2. a rat depressing the lever in his experimental box in order to secure a food pellet

3. a small infant crying in order to attract mother's attention . . ..

“For we cannot plausibly suppose our train-rider to be receiving guiding signals from some region with which his foot is to be correlated (sic). Neither, when the psychologist stops replacing the rat's pellets, can we describe the rat as receiving directive signals from some such pellet. Nor, finally, when mother, expecting baby to sleep, steps out to the corner store is she available for the issuing of signals guiding the infant's behavior toward final correlation with herself”

Feedback mechanisms clearly do not provide the desired translation because in these instances there is no feedback.

4. REDUCTION THROUGH CAUSAL CHAINS

In Scientific Explanation [2], R. B. Braithwaite offers a translation rule which, while also best viewed as based on a target schema, avoids the missing-goal-object difficulty of the previous attempt.

In place of the feedback device Braithwaite employs the notion of the causal chain, nomically determined by initial conditions, to bridge the gulf separating teleology from the more respectable world of ordinary scientific discourse.

On this view, the behaviour of a teleological system is viewed as a causal chain connecting an initial event with a final one.

The important characteristic of a causal chain for this argument is that

“... every event in the system is determined by the whole previous state of the system together with the causally relevant factors in the system's environment or field (which will be called field conditions). Then the causal chain c of events in [system] b throughout a period of time is nomically determined by the initial state e of the system together with the totality of field conditions which affect the system with respect to the events in question during the period.”

Call this set of field conditions f. Then for a given system b with initial state e, c is a one-valued function off; that is, for a given b and e, the causal chain c is uniquely determined by f-the set of field conditions.

What defines a teleological system then is the plasticity of the system’s behaviour towards a goal, its “persistence towards the goal under varying conditions.”

He attempts to represent this plasticity in terms of the undeniably non-teleological concepts of causal chain and field condition.

The first, and easy, step is to translate

'the system is persistent toward the goal under varying conditions' into 'for any given initial state, the system will achieve the goal under a variety of field conditions'.

But that leaves the tough problem of eliminating 'achieve' and 'goal' in favor of non-teleological expressions.

Braithwaite offers the following:

Now consider the property which a causal chain in a system may possess of ending in an event of type gamma without containing any other event of this type. Call this property the gamma-goal-attaining property, and the class of all causal chains having this property the gamma goal attaining class. Every causal chain which is a member of [the gamma goal attaining class] contains one and only one event of type gamma, and contains this as its final event ([2] p. 330).

So we have: 'for any given initial state of the system there are a variety of field conditions which determine a causal chain in the gamma-goal- attaining class.'

According to Braithwaite, all that must be done to justify describing the activity of something as teleologically directed towards a certain goal, is to demonstrate the plasticity of the behaviour of that something from the initial state in question toward that goal.

And the demonstration of plasticity consists in the specification of at least two sets of field conditions under which the goal is reached from the given initial state.

This formulation clearly involves no teleologically loaded concepts, and just as clearly requires no signals to be emitted by the goal, and hence avoids the missing-goal-object difficulty. This formulation does, however, have many difficulties of its own.

Five possibly minor ones, which nevertheless deserve mention, are the following.

1. Problem in determining which is the terminal event of a causal chain. (not clear when it is appropriate to say that a causal chain has ended).

Does the causal chain which began when a driver lost control of his car on a wet road end when all of the pieces settle to the ground after the crash, or when the last hubcap stops rolling and all is quiet, or when the driver is resting quietly in the hospital; or is it when he has finally recovered, or when the car has been repaired or, perhaps, both? Or is it only when the driver eventually dies and removes from the causal stream all those neuroses which resulted from the trauma of the accident?

It would, I think, be rather arbitrary to choose any of these unqualifiedly as the terminal event.

2. Not obvious that the achievement of a goal must constitute the terminal event (when that's clear) of a causal chain.

Eating lunch, though clearly an intended and achieved goal, might be just one event in the larger causal chain constituting one's trip to the zoo.

3. Determining when two events are appropriately of the same type.

Clearly the causal chain which produces a shady resting place for an animal can have other shady moments and yet allow that the shady resting place was the goal of the behavior involved. Hence, shady cannot be the 'type' of event the achievement of the goal is; for if it were, the causal chain would not fit Braithwaite's definition of a gamma-goal-attaining causal chain, because it would then 1 Strictly, more than one is enough, but the more conditions under which the goal is achieved the more plastic is the behaviour.

contain two events of the same 'type.' The same procedure can be repeated with respect to 'restful' and any other potentially relevant predicate, which makes it difficult to imagine just what 'type' of event finding a shady resting place is in Braithwaite's sense.

4. Non-teleological systems could satisfy the Braithwaite criteria for purposive behaviour if achievement of the goal under as few as two distinct sets of field conditions is allowed as a demonstration of plasticity.

For example, a lump of clay would hit the same spot on the floor when dropped from a given place both in still air and in the presence of oppositely directed air currents of the appropriate magnitude and locations.

5. It does not seem unreasonable to suggest that there might be occasions on which we would say a system manifested purposive behaviour even though, given its initial state, it was nomically impossible for it to achieve the goal to which it aspired.

Behaviour might be purposive even though there are no possible gamma-goal-attaining causal chains. The fish struggling to get free of its net would seem to be a case of this type. It is struggling toward its goal of freedom even though achievement is impossible merely in virtue of its initial state and the relevant laws.

More conclusive objection put forth by Scheffler:

'The Difficulty of Multiple Goals'

And the demonstration of plasticity consists in the specification of at least two sets of field conditions under which the goal is reached from the given initial state.

But, given a system's initial state, there may be sets of field conditions which produce goal gamma, and others which produce goal delta. Accordingly when the system is in this initial state, we should be able to say either (or both) that the system is teleologically directed toward gamma or (and) teleologically directed toward delta.

'This', says Scheffler, 'is exactly what we cannot generally say': The cat crouching before the vacant mouse-hole is crouching there in order to catch a mouse. Since no mouse is present there will in fact be no goal-attainment. Nevertheless, the cat's behavior is plastic since there are various hypothetical sets of field conditions, each set including one condition positing a mouse within the cat's range such that, in conjunction with the cat's present behaviour, each set determines a mouse-attaining causal chain. On the other hand, there are also various other hypothetical sets of field-conditions, each set including one positing a bowl of cream within the cat's range, such that, conjoined to the cat's present behavior, each set determines a cream-attaining causal chain. It should therefore be a matter of complete indifference, so far as the present proposal is concerned, whether we describe the cat as crouching before the mouse-hole in order to get some cream.

The fact that we reject the latter teleological description while accepting the former is a fact that the present proposal cannot explain.

This objection is so telling mostly because it attacks the very foundation of the Braithwaite project. It suggests that no analysis will work which takes as its basis a subjunctive statement of goal-attainment (under various possible conditions) of the sort employed in this translation attempt.

5 REDUCTION THROUGH STATE DESCRIPTIONS

Braithwaite contends that his analysis works for furnace-like cases of directively organised behaviour as well as those of the target variety. This presumably would be accomplished by allowing the major variable to be something other than displacement from an object (goal), such as temperature or blood-sugar concentration, and by allowing the plasticity to consist in the maintenance of that variable within a certain range under several sets of field conditions rather than insisting that it tend to zero, as displacement must.

Hence, it need not be surprising that the following translation attempt, even though it takes a furnace schema as its model, offers a solution to the very same problems as do the first two models, and is moreover very like Braithwaite's attempt even in some specific points of detail.

This model, which is offered by Ernest Nagel

Avoids many of the difficulties which beset Braithwaite

Nagel suggests the characteristic feature of teleological systems is the persistence of a system in either manifesting a certain property, condition or mode of behaviour (G) or maintaining itself in a state (G-state) which causally guarantees that it will manifest that property (etc.) at a future time.'

But this persistence is with respect to changes within the system itself, and explicitly excludes environmental effects on the manifestation of G. Hence, the cases of plastic behavior with respect to environmental variables, which Braithwaite considered, are accommodated by including any causally relevant part of the environment as part of the

'directively organized system' in Nagel's terms. So, what is ordinarily considered to be the teleological system in these (target) cases, is merely part of what Nagel takes to be the 'system' in his analysis.

For this analysis, a system must be analysed in terms of a finite number of non-redundant predicates or state variables which are nomically/ physically/causally sufficient to determine whether or not the system manifests G or is in a G-state.

In these terms Nagel's position is that a system 'will be said to be "directively organised during the interval of time T with respect to G" ' if the system begins that interval in a G-state and if, whenever the system undergoes a state-variable change which would ceteris paribus take it out of its G-state, this induces the other state- variables to change in such a way that the system remains in a G-state.

A house is a directively organised system on a certain day (T) with respect to its temperature if it begins that day in a certain temperature range (i.e. in a G-state) and if, whenever (during T) the heat-flow out of the house would, if not compensated for, take the house out of the acceptable temperature range, its furnace comes on long enough to effect the required compensation.

This analysis does admirably achieve one of its goals: it does not contain any terms that even hint of teleology. It is almost a paradigm of non- teleology.

The analysis also clearly avoids the first three of my five 'possibly minor' objections to Braithwaite's model by avoiding talk of terminal events and types of events.

Nagel's has problems:

1. The first and most astonishing is a difficulty so obvious that the previous analyses explicitly avoid it. This is called by Scheffler, in another context,3 'the difficulty of goal-failure”

Most clearly seen by observing that a system can be teleological/goal-directed/directively organised without its ever achieving the state towards which it is 'directed'.

Scheffer's examples are of a dog pawing at a door while trapped in a place from which it will never escape, and of a cat crouching before an empty mouse-hole. Both cases can be legitimately described teleologically even though the respective goals will never be attained: the dog is pawing in order to get out, the cat crouching in order to catch a mouse.

To use the terms of the present analysis: before being consumed, the shark's dinner might have been manifesting behavior which was clearly directively organised toward the goal of not becoming the shark's dinner. Examples can be multiplied indefinitely: trying, which is paradigmatically teleological, does not entail success.

On the other hand, before he will allow a system to be called 'directively organized toward G', Nagel requires that the manifestation of G be causally guaranteed by the system's state description and the relevant laws.

2. A cluster of closely related methodological difficulties. Nagel leaves unanswered a number of questions which it is his business to answer when treating the problem he claims to be treating. He has given us a formula for telling when a system is 'directively organized during interval of time T', but has not said how this relates to important questions couched in less technical terms. For example,

a. when is a system teleological? Only when it is directively organized? Can a teleological system sleep, or does it rather cease to exist for a time? One could offer that the plausible suggestion here is to say that a system is teleological if it is capable of directively organised behaviour. But this suggestion raises a host of new problems, and, more importantly, it is only a suggestion.

b. And until he provides explicit answers to the important questions about teleology, until he tells us when a system is teleological, when it is legitimate to say 'in order to', when 'goal' is the appropriate description of an object, he has not provided us with anything very useful.

6 THE GENERAL CASE

The above criticism of Nagel, like the criticism of the other models, has consisted in bringing to light rather specific difficulties with the way in which the model goes about producing a translation rule. Little of a general nature has so far been said about the programme of reduction itself. The more general argument has been deferred thus far because a crucial point is more easily developed after looking carefully at some serious attempts to carry out the programme.

The teleological reductionist position is almost always described as claiming that, for every legitimate teleological description of a phenomenon, there can be found a non-teleological description which is logically equivalent to it.

Strictly speaking this description is not inaccurate, but it is somewhat misleading. For it seems merely to be maintaining that teleological descriptions can be rewritten, perhaps elliptically, in such a way as to avoid all the traditionally teleological or purposive terms.

But upon examination of what the reductionist philosophers are doing, 'non- teleological' might better be replaced by 'un-teleological' or even 'anti- teleological'.

These men not only want teleological descriptions to be logically equivalent to something else, they want that something else to be part of the presently accepted conceptual framework of science;

they want it to introduce 'nothing new'.

This severely restricts what is to count as a 'non-teleological description', and makes the position more interesting and more important, as well as more difficult to defend, than it at first appeared.

It is their desire to avoid anything conceptually novel, to maintain teleology within the domain of causal principles, which explains something otherwise rather puzzling done by both Braithwaite and Nagel.

The models of teleology offered by these two philosophers might easily have been quite unproblematic behaviourist accounts of teleological systems:

Braithwaite's in terms of plastic behaviour with respect to an environment

Nagel's in terms of a system's maintaining a G-state in spite of a variety of internal changes.

And on the looser, more inclusive interpretation of 'non-teleo- logical' this would have been sufficient.

The translation rule would then be: teleological behaviour is behaviour which is plastic with respect to certain environmental obstacles or, alternatively, behaviour in which a system maintains itself in a G-state in spite of internal changes normally antithetical to maintaining this G-state.

But both authors felt compelled to include something in their models which, from the less restrictive view of the translation programme, is completely gratuitous.

Braithwaite adds to plasticity the further requirement that 'every event in the [teleological] system is determined by the whole previous state of the system together with the causally relevant factors in the system's environment . . .'

Nagel says the same thing in slightly different words when he requires that the 'internal changes', in terms of which his analysis is formulated, be alterations of 'state variables', and that, whatever the nature of the state variables, in respect to the states they represent S is a deterministic system: the states of S change in such a way that, if S is in the same state at any two different moments, the corresponding states of S after equal lapses of time from those moments will also be the same.

The requirement that teleology be maintained within the domain of causal principles, that teleology encompass nothing not already within the science of macro-objects, just is the requirement that teleological systems be causally deterministic.

All three models presented thus far include this requirement

Admits that there is a very good chance that all teleological systems that will ever be discovered to exist, will in fact be causally deterministic systems.

But to require causal determinism, as part of an analysis of what it is to be a teleological system, is to require that a teleological system be causally deterministic by logical necessity.

This “dooms the analysis at the outset.”

For, as Braithwaite and Nagel clearly recognize, what is essential to the teleology of a system is the plasticity of its behaviour and its persistence toward a goal.

And the plasticity and persistence of a system's behaviour is logically consistent with the causal indeterminacy of that behaviour.

When it is said that a system's behaviour is plastic, all that is being asserted is that the system avoids some environmental obstacles. The exact path taken is not important. As long as the goal is reached in a variety of circumstances or, at least, as long as some obstacles are circumvented, the behaviour is plastic.

The exact path taken to reach the goal, or to avoid the obstacles, can be any one of a rather large number.

And if this one were chosen at random, say by a quantum randomiser, the behaviour would be causally indeterminate; it could not be predicted or explained by causal principles.

This has the interesting consequence that behaviour can be teleologically determinate, in the sense that it is subsumed under a true teleological generalisation, while at the same time being causally indeterminate.

The kernel of the essay:

If it is ever useful to think of something as teleological, then it is useful to think of it as something which is not necessarily causally deterministic.

A teleological expression is not just an abbreviation for something complex but essentially causal. Even if all teleological of purposive behaviour is in fact the product of a causal mechanism, and hence is explainable in terms of causal principles, it does not follow that the teleological description of such

behaviour is reducible to a causal description.

Teleological expressions are not 'translatable into' or 'saying the same thing as' some causal expression.

The two do not 'have the same cognitive content', and they differ in more than mere 'emphasis'.

If teleological expressions are ever scientifically useful, they are so because of the way teleological concepts organise an area of scientific interest, not because of their equivalence to causal expressions.

http://aeon.co/magazine/philosophy/steven-poole-teleology/

Can Darwinism Survive without Teleology?

394 39 12 1916

by Gerard M. Verschuuren

Filed under Evolution

152 Comments

DARWIN

From Stephen Graukroger, Descartes: an Intellectual Biography (Oxford, 1995), p. 1:

Since the eighteenth century, there has been in circulation a curious story about Descartes. It is said that in later life he was always accompanied in his travels by a mechanical life-sized female doll which, we are told by one source, he himself had constructed 'to show that animals are only machines and have no souls'. He had named the doll after his illegitimate daughter, Francine, and some versions of events have it that she was so lifelike that the two were indistinguishable. Descartes and the doll were evidently inseparable, and he is said to have slept with her encased in a trunk at his side. Once, during a crossing over the Holland Sea some time in the early 1640s, while Descartes was sleeping, the captain of the ship, suspicious about the contents of the trunk, stole into the cabin and opened it. To his horror, he discovered the mechanical monstrosity, dragged her from the trunk and across the decks, and finally managed to throw her into the water. We are not told whether she put up a struggle.

Those who prefer that their great tales not be ruined by truth will want to stop there. Those who think that fiction survives a dusting of dull fact will want to read the next paragraph as well (pp. 1-2):

The story had a wide currency in the nineteenth and early twentieth centuries, at one stage being taken as a theme for a novel by Anatole France. It exists in a number of versions, some of them explicitly fictional, some purporting to be factual, and the detail varies quite considerably from version to version. So far as I can tell, the story originates no early than the eighteenth century, and it received most attention in an era preoccupied with the theories of La Mettrie, the French Enlightenmentphilosophe who, in his infamous L'Homme Machine (1747), had extended the idea of animals being automata -- developed by Descartes in his L'Homme -- to human beings, offering a materialist account of the mind, and suggesting that Descartes himself had held such a view, but that judicious self-censorship had prevented him making the theory public. There is, in fact, absolutely no evidence that any version of the story is true. Its origins are rather obscure, but by the second half of the eighteenth century it was a propaganda weapon in the fight against La Mettrie's materialism, Descartes himself being seen as the ultimate instigator of this pernicious doctrine. Given this context, the story has all the elements of propaganda, including that favorite propaganda weapon, sexual innuendo, and I have little doubt that it originated as a tool of the eighteenth-century struggle against materialism.

And those with the spirit of a Sub-Sub Librarian, such as myself, will want to read the first footnote, so for your convenience I reproduce it here (from p. 418 of Graukroger):

I first came across this story in print in a recent book on the history of robotics, where it is presented as fact, although no references are given. Investigation showed the story to have had a wide currency between the late eighteenth century and the early decades of this century. For the different versions of the story and their sources see Leonora G. Rosenfield, From Beast-Machine to Man Machine, rev. edn. (New York, 1968), 202-3, and the accompanying notes on p. 236. Descartes is not the first philosopher reputed to have constructed a mechanical companion. Albertus Magnus was said to have had a robot that could move and greet visitors with the salutation Salve! ('How are you!'). Thomas Aquinas, his pupil at the time, is reported to have attacked and broken the gregarious android when he came across it unexpectedly in the night. The story is reported, with references I have not followed up, in G. A. Lindeboom, Descartes and medicine (Amsterdam, 1979), 62.

How like a scholar to hide the Aquinas vs. Robot death match in a footnote like that. It should go in the preview, man: how else do you expect to get buts in the seats eating popcorn?

I'm disappointed that the footnote doesn't mention which of Anatole France's novels is on "the theme" of Descartes's robot (a quick googling indicates that none of his works seem to have been explicitly about Descartes by name). Does anyone happen to know? [Update: A commentator comes through! The novel in question is La Rotisserie de la Reine Pedauque (1893), translated as At the Sign of the Reine Pedauque(1922, available online at the link). Thanks to B. McCarthy; see his comments below for more on this, as well as the Rosenfield book discussed below.]

Incidentally (more Sub-Subness here) the author of From Beast-Machine to Man Machine seems to have been Leonora C. Rosenfield, not G. as reported by Mr. Graukroger. A quick google indicates that seems to have published as Leonora Cohen Rosenfield (although Amazon, for no discernable reason,lists her as Leonora Davidson Rosenfield). Obviously, given the Cohen -> G. mistake, to say nothing of the obvious omission of the story of the golemfrom the aforequoted list of "philosophical robot" stories, the only possibleinterpretation is that Mr. Graukroger is a raving antisemite. These things don't happen by accident, you know.

At any rate, since neither google books nor amazon has a "look inside" feature for Rosenfield's work, my researches must stop here for the moment. Alas. If anyone has a copy of From Beast-Machine to Man-Machine handy, however, please leave any interesting tidbits in the comments.

Ever since Darwin, the concept of teleology has been suspect among biologists. What is so controversial about teleology? Most likely, its history! From the earliest Greek philosophers on, it was widely believed that the world must have a purpose because, as Aristotle would put it, “nature does nothing in vain,” and neither does God, as a Jew or Christian would say. In this often misunderstood view, any change in this world is due to final causes that move things to an ultimate goal, a predetermined end. All things would achieve certain ends or goals because they were designed that way by nature or by God; that’s how hormones, for instance, are supposed to reach their target cells.

However, this belief came under attack when scientists—still called “natural philosophers” at the time—began searching for physical and material explanations, for eternal physical laws that regulate falling bodies and the motion of planets.

Modern philosophy and modern science can be characterized as the rejection of the Aristotelian model of “causation.

Vitalism: the doctrine that "iving organisms are fundamentally different from non-living entities because they contain some non-physical element or are governed by different principles than are inanimate things.

There ensued a Vitalist/Mechanist debate. The Vitalists-Mechanist controversy of the 18th and 19th centuries is often portrayed in terms of the progressive mechanists being opposed by the reactionary vitalists. As far as the mechanists were concerned, the picture is basically correct—the mechanists were charting a new path, one that would prove immensely productive in generating biological knowledge.

They were, in carrying out the program that Descartes had envisaged but only pursued speculatively, showing how many of the phenomena exhibited in living organisms could be explained in terms of the component parts of those organisms carrying out component operations in much the same manner as is the case in human-engineered machines.

Many of the vitalists, however, were astute critics, recognizing the limitations of the mechanistic accounts of their day. The limitations were not just incidental but went to the core of the mechanistic project as it was pursued—they recognized that the mechanist accounts lacked the resources to account for some of the most fundamental features of living organisms. Unlike the human-engineered machines that provide the model for mechanistic accounts, organisms build, sustain, and repair themselves.

In doing this they are not just reactive—they are endogenously active. That is, they behave is ways explained by caused by factors inside the organism or system.

It is at this point in the discussion that Charles Darwin comes into the picture. He replaced teleological explanations with physical explanations in terms of what he called natural selection. According to some, Darwin rephrased teleology from an “a priori drive” to an “a posteriori result.”

Is this then the end of teleology?

Teleology and Function

John Burdon Sanderson Haldane (1892-1964)

[in the `30s] can be found remarking,

“Teleology is like a mistress to a biologist: he cannot live without her but he's unwilling to be seen with her in public.”

“Today the mistress has become a lawfully wedded wife. Biologists no longer feel obligated to apologize for their use of teleological language; they flaunt it. The only concession which they make to its disreputable past is to rename it ‘teleonomy’.” David Hull (1982)

Aristotle: For the sake of what?

"Democritus, however, neglecting the final cause, reduces to necessity all the operations of nature. Now they are necessary, it is true, but yet they are for a final cause and for the sake of what is best in each case. Thus nothing prevents the teeth from being formed and being shed in this way; but it is not on account of these causes but on account of the end; these are causes in the sense of being the moving and efficient instruments and the material. …to say that necessity is the cause is much as if we should think that the water has been drawn off from a dropsical patient on account of the lancet alone, not on account of health, for the sake of which the lancet made the incision."

Aristotle, Generation of Animals V.8, 789a8-b15

The Spookiness of Teleology

For Aristotle, natural phenomena were teleological –

Events happened to produce results

These results explain the events even through they come after the events –

“Nature adapts the organ to the function, and not the function to the organ” (De partib. animal., IV, xii, 694b; 13)

Thus, teleology seems to involve backwards causation—the effects (function) of some causal process are its cause.

In the case of human planning, it is not the actual effect that causes the action, but the thought of the effect –

But there is only prior thought for biology if one is a Creationist

Teleology—Hard to Kill

The scientific revolution seemed to remove purpose from the world.

Events happened solely because of prior mechanistic causes (efficient causation… sort of)

Captured by the quest for mechanisms in biology.

Mechanisms explained their effects in terms of preceding, efficient causes.

Nevertheless, “teleological talk” lives on in the language of functions in biology

“The heart’s function is to pump the blood.”

William Harvey “On the Motion of the Heart and Blood" 1628

“The kidney’s function is to filter and remove waste.”

“The function of the ribosome is to synthesize proteins.”

Evident most clearly in the willingness to say that something is malfunctioning.

Evolutionary Biologist Ernst Walter Mayr ( July 5, 1904 – February 3, 2005) on Teleology

“Consider the following statement: `The Wood Thrush migrates in the fall into warmer countries in order to escape the inclemency of the weather and the food shortages of the northern climates'. If we replace the words ‘in order to’ by ‘and thereby’, we leave the important question unanswered as to why the Wood Thrush migrates. The teleonomic form of the statement implies that the goal-directed migratory activity is governed by a program. By omitting this important message the translated sentence is greatly impoverished as far as information content is concerned, without gaining in causal strength.” Mayr (1974)

Naturalizing Teleology

• Ground (reduce) teleological notions in natural phenomena

• Show under what conditions a natural (mechanical) system has purposes or goals

• Two main naturalizing strategies:

1. Negative Feedback and Cybernetics More recently—language of a program directing activity

2. Products of Natural Selection

Negative Feedback

• First known example of negative feedback:

Water clock designed by Ktsebios

Water clock designed by Ktsebios in the 3rd century BCE – Need to maintain constant water pressure – Employed a float that would start or halt the inflow from the water supply, maintaining a constant level.

It is historically the first documented use of this concept. A device that used a constant flow of water to measure the passage of time.

Water drops falling from orifice of an upper tank accumulate in the lower tank and the increasing water level indicates how much time has expired. However, the accuracy of the clock depends on the constant flow rate of water into the lower-level container. This flow rate is being disturbed by variations in the upper-tank water level, that is, its water pressure varies. To solve the problem in variations in water flow rate,

Ktesibios maintained a constant water level in the upper holding tank by applying a valve with a float. The role of the valve is to open and allow water flow in, when the water level in the upper tank goes down, and to close and cut the water stream off, when the water level in the upper tank reaches the uppermost level. The water level in the upper tank is determined by a float, which operates as part of the valve.

Industrialization and Negative Feedback

James Watt (1736 -1891)

Scottish Inventor

Faced a serious practical challenge – How to control the speed of the steam engine so that all appliances would run at the same rate despite different number being on line at a time

Devised an elegant mechanism for feedback control

Negative Feedback in Biology

• Negative feedback is widespread in biology

Biochemical systems: products of reactions feedback to slow reactions earlier in the pathway.

Physiological systems: when variable deviates from norm, processes initiated to restore it to normal

Motor systems: when action misses the mark, change to guide it to the target

• Walter Cannon: homeostasis

Walter Bradford Cannon, M.D. (1871 – 1945)

Recognizing the Generality of Negative Feedback

Challenge: how to control gun fire targeting aircraft

Use feedback from the first shot to correct the next.

Later, heat seeking missiles and beyond

• Recognizing the commonality between control of anti-aircraft fire and control in biological system, Norbert Wiener created an interdisciplinary movement

–Cybernetics—from the Greek for Helmsperson

The Seeming Insufficiency of Negative Feedback

But…

• Humanly designed negative feedback systems all involve a designer

Who so arranged the parts of the system so that it would reach the target?

• The designer imposed the goal on the system

But where is the designer of biological systems: how did the organism become so organized that it could compensate for deviations?

Teleology and Darwin

• Recall Darwin’s high regard for Paley

Biological organisms are complex systems that are highly adaptive (functional) in their environments

• Darwin offered an (mechanistic) explanation for traits that had seemed to require design

Does natural selection remove the last vestige of teleology from science? or

Does natural selection license teleological discourse in biology

• The function of a trait is that effect of it on which natural selection operated

—that caused ancestors with the trait to reproduce more successfully

Larry Wright “Functions” 1973

Wright’s Distinction Between a Trait’s Function and Other Effects

“Very likely the central distinction of this analysis is that between the function of something and other things it does which are not its function (or one of its functions). . . . The function of the heart is pumping blood, not producing a thumping noise or making wiggly lines on electrocardiograms, which are also things, it does. This is sometimes put as the distinction between a function, and something done merely ‘by accident’.” (Wright, p. 141)

Wright: Functions as Explanatory

“Merely saying of something, X, that it has a certain function, is to offer an important kind of explanation of X.”

The heart beats in order to circulate blood.

To ask “what is the function of X?” is comparable to asking “Why do C’s have X’s (or do X)?”

• The sought for explanatory concerns how X came to be—it came to be because of its function

• But remember the challenge: the function is realized only after X

How could what comes later (evolutionary advantage) explain what came earlier?

Natural Selection as an Explanation

• If an organ has been naturally differentially selected-for by virtue of something it does, we can say that the reason the organ is there is that it did/does that something.

Hence we can say

animals have kidneys because they eliminate metabolic wastes from the bloodstream;

porcupines have quills because they protect them from predatory enemies.

plants have chlorophyll because chlorophyll enables plants to accomplish photosynthesis.

the heart beats because its beating pumps blood.

• The function of X is Z means

(a) X is there because it does (did) Z

(b) Z is a consequence (or result) of X's being there

Thus a two part thesis:

1. We have hearts because of what hearts are for

Hearts are for blood circulation, not the production of a pulse, so hearts are there--animals have them--because their function is to circulate the blood

2. 1 is explained by natural selection: traits spread through populations because of their functions.

Challenges for the Etiological Account

Etiology: a branch of knowledge concerned with causes; specifically : a branch of medical science concerned with the causes and origins of diseases

• Cave fish have remnants of an eye

What is its function?

• It was originally selected for sight

• Is that still its function?

• What is the function of the human appendix?

Darwin: used by other primates to digest leaves – Is that its function in us?

An Alternative to the Etiological Interpretation of Function

Robert C. Cummins

In his paper “Neo-teleology”[1] Cummins argues that the tow part thesis is unsatisfactory. He claims that the presence of a biological trait or structure is not explained by appeal to its function. And to suppose otherwise is to trivialize natural selection.

• Cummins further challenges the principle underlying etiological account, viz.:

“The point of functional characterization in science is to explain the presence of the item (organ, mechanism, process or whatever) that is functionally characterized”

–

As he sees it, the problem is that most functional items are neither necessary nor sufficient for realizing the function and so their occurrence is not explained by citing the function

Relocating the Explanatory Role of Functions

• Cummins claims that it is more plausible that the statement

“The heartbeat in vertebrates has the function of circulating the blood through the organism.”

is appealed to in explaining circulation (the function)

That is, we start with circulation, and identify something as having that function in the context of explaining it.

And may explain the advantage of the heartbeat by identifying the activity it facilitates.

• This is different than explaining the existence of the heartbeat

Cummins’ Account of Explaining How Something Performs a Function

• Functions and dispositions:

“to attribute a function to something is, in part, to attribute a disposition to it. If the function of x in s to Φ, then x has a disposition to Φ in s”

• Dispositions require explanation:

“if x has [disposition] d, then x is subject to a regularity in behavior special to things having d, and such a fact needs to be explained.”

• The appropriate explanatory strategy:

Analytic strategy: – Analyze “d of a into a number of other dispositions d1 . . . dn, had by a or components of a such that programmed manifestation of the d1 results in or amounts to a manifestation of d”

Analytic Strategy in Biology

“The biologically significant capacities of an entire organism are explained by analyzing the organism into a number of ‘systems’—the circulatory system, the digestive system, the nervous system, etc.,— each of which has its characteristic capacities. These capacities are in turn analyzed into capacities of component organs and structures. Ideally, this strategy is pressed until pure physiology takes over, i.e., until the analyzing capacities are amenable to the subsumption strategy.”

• This should seem familiar: mechanism in biology exemplifies this approach

But What Dispositions are Functions?

• Cummins offers a strategy for explaining functions by treating them as dispositions

But which dispositions are functions?

• Considers a condition such as “contributes to the proper working order” of the system of which it is a part

Considers whether these could be cast as “health and life” or “contributing to the survival of the species”

– NO!

Health presumes well-functioning- no malfunctioning.

• Picks out the wrong instances on some occasions

• Doesn’t explain why these are functions

• What’s left is the style of explanation used for functions

BUT, the same strategy is invoked for pathologies

• Does the gene for schizophrenia have the function of producing schizophrenia?

Construals of Function Talk

• The etiological strategy: explain the function of something in terms of what it was selected for

Treat it as an adaptation

Function explained etiologically

• The functional analysis strategy: explain how something is able to perform a function

Treat functions as dispositions of things

Decompose the disposition into sub-dispositions

• A third alternative: explain the function in terms of the contribution something makes to the operation of systems that maintain themselves far-from-equilibrium

Detach function from natural selection

Function in terms of contributions to the maintenance of life in a living system

Autopoiesis and Teleology

Autopoiesis: (from Greek αὐτo- (auto-), meaning "self", and ποίησις (poiesis), meaning "creation, production") refers to a system capable of reproducing and maintaining itself.

• Recall Bichat:

Recall the Vitalists’ Objection

to Mechanism

• Mechanistic accounts were too simple to capture the phenomena found in living systems

• Bichat: Life as the sum of the forces that resist death

– Ordinary mechanisms seem to be subject to death—decompose in the normal course of

ongoing activity

– Living systems maintain themselves in the face of such processes

Living organisms characterized by resistance to death

• Claude Bernard tried to explain such a capacity mechanistically

Components of the system operate so as to maintain the constancy of the internal environment

• Given the degenerating effects of an environment, organisms must continually repair themselves or die

• Maturana and Varela

An autopoietic machine is a machine organized

“. . . as a network of processes of production . . . which: (i) through their interactions and transformations continuously regenerate and realize the network of processes . . .that produced them; and (ii) constitute it (the machine) as a concrete unity . . .

• The processes that maintain an autopoietic machine can be understood as teleological—serving its own maintenance

Biological Organisms as Autonomous Systems

• 2nd law of thermodynamics: In an open system entropy (disorder) always increases

• Maintaining the organization of a living organism requires using energy to resist the increase in entropy

• Moreno characterizes such systems as autonomous

“a far-from-equilibrium system that constitutes and maintains itself establishing an organizational identity of its own, a functionally integrated (homeostatic and active) unit based on a set of endergonic-exergonic couplings between internal self-constructing processes, as well as with other processes of interaction with its environment”

• Kepa Ruiz-Mirazo, Juli Peretó and Alvaro Moreno, A Universal Definition Of Life: Autonomy And Open-ended Evolution • These self-maintaining processes appear as teleological

Darwin may have thought it was, but let’s find out.

Biology Has a Teleological Dimension

What makes biology so inherently teleological? Biological features can be understood in terms of effects—that is, in terms of survival problems that need to be effectively solved. In other words, they serve a function; the green color of a caterpillar has a function, namely, to deceive potential predators; such is their end or goal—or in more neutral, biological terms, such is their function. One could also say that camouflage is “for” deceiving, just like a knife is “for” cutting. So in biology, it remains very common to ask what a feature is “for.” Just like pumps are for pumping (that’s their desired effect), so eye patterns on butterfly wings are “for” protection (that’s the advantageous effect it has on fending enemies off). This is a function of eye patterns, but certainly not a purpose or intention of butterflies.

Prior to any talk of evolutionary theory, William Paley (1743-1805) had argued that something as beautifully designed as the universe must have had a Designer, just like a watch does. In the footsteps of Paley, Darwin also saw a beautiful design in nature, but unlike Paley, he viewed nature as something designed by the trial-and-error test of natural selection during a process of evolutionary change. No matter what, in either case, the results must be design-like (in the sense of well-adapted), because if they were not, they simply would not work in solving problems. If the eye lens, for example, did not function like a physical lens, one would not see very well. There is teleology again.

Apparently, biological features can have and do have effects that are advantageous (or detrimental) to various degrees. But how that is possible in itself is an altogether different story—actually a meta-physical story.

The Metaphysics of Teleology

Somehow our universe has been designed in such a way that specific designs do work, whether it is for better or for worse. It is only due to this metaphysical notion of design and teleology that we can talk about biological designs; all biological designs are “design-based” designs. It is one of the most perplexing things about our universe that it allows for any kind of design to work the way it works.

Did Darwin ignore this part of the story? Or did he really discard teleology? Some keep stressing that he replaced teleology with the causality of natural selection. One of them was George Bernard Shaw who once said that Charles Darwin threw Paley’s “watch” into the ocean. Well, Shaw was wrong. If Darwin did throw something away, it was Paley’s “watchmaker,” but certainly not his famous “watch.” Darwin never threw away the design concept—it was actually essential to his theory.

The artifact analogy of design is as basic to Darwinism as it is to Paley’s natural theology. Since the heart is designed like a pump, it is a successful design “for” circulating blood. After Darwin, the heart still existed “for” circulation; the cause of its existence may have been different, but its teleology was not. However, Darwin ignored, or at least bypassed, the following question: How come that certain biological designs “work,” and are “successful” and “effective” in reaching their “goal”? What is it that makes them “goal-directed”? What carries them through the filter of natural selection?

It’s here that teleology keeps coming back. There is teleology in the biological world because the animate world is design-like—as much so as there is teleology in the technical world of designers because that world is design-like as well. Natural selection may explain that a fine working design has a better chance of being reproduced, but ultimately it cannot explain why such a design is working so well.

And that’s where teleology is needed—even in Darwinism. In that sense, Darwin did not change teleology from an “a priori drive” into an “a posteriori result.” Teleology is not a biological outcome a posteriori but a metaphysical given a priori. Natural selection does not create teleology, but its working is based on teleology.

Where Does Teleology Come From?

The answer may seem mystifying at first sight: Teleology must have been built into nature—as some kind of all-pervasive architecture. It may be so all-pervasive, though, that it easily escapes attention. Natural selection on its own cannot do the “job” unless it works within a framework of purpose and design. Without this “cosmic design,” there couldn’t even be any natural selection. Natural selection can only select those specific designs that are in accordance with the cosmic design (by the way, designers, engineers, and architects must do the same thing).

As it turns out, science does not operate in a vacuum, but it works within a philosophical framework of pre-existing assumptions—and one of them is teleology (some call it cosmic teleology, in distinction from the older idea that biological designs are the product of predetermined goals). Darwin may have thought he could reduce teleology to causality, but his causality mechanism of natural selection can only work on condition that there is teleology in nature. There is “something” in successful biological designs that carries them through the filter of natural selection. To put it briefly, organisms are not teleological because they have survived; on the contrary, their survival is mainly due to the fact that they are teleological. Creation is “loaded” with cosmic design, just like a dice that constantly throws a six must be loaded.

Let me come to a conclusion. The inescapable idea behind all of this is that our universe is ultimately an “intelligent project,” created by an Intelligent Designer. The assumption of a Creator would explain that the universe exists and is what it is; and the assumption of a cosmic design would explain why the universe is this way. In that sense, even Darwinism needs some Divine Help—whether its fans like it or not. There is no way Darwinism could survive without teleology.

[1] In Andre Ariew, Robert E. Cummins & Mark Perlman (eds.), Functions: New Essays in the Philosophy of Psychology and Biology. Oxford University Press (2002)