In the ‘Moderation’ thread, William J Murray tried to make a case for ideological bias among evolutionary scientists by referencing a 2006 Gil Dodgen post, in which numerous authors emphasise the lack of teleology within the evolutionary process. I thought this might merit its own OP.
I disagree that authors are showing a metaphysical bias by arguing against teleology. I wrote
Evolutionary processes, conventionally defined (ie, variations and their changes in frequency due to differential survival and reproduction), do not have goals. If there IS an entity with goals that is also directing, that’s as may be, but the processes of evolution carry on regardless when it isn’t. It is important to erase the notion of teleology from a student’s mind in respect of evolutionary mechanisms of adaptation, and most of those quotes appear to have that aim. Organisms don’t, on the best evidence available, direct their own evolution.
To which WJM made the somewhat surprising rejoinder: “how do you know this”? Of course the simple answer is that I qualified my statement ‘on the best evidence available’ – I didn’t claim to know it. But there is a broader question. Is there any sense in which evolutionary processes could, even in principle, be teleological? I’d say not. You have a disparate collection of competing entities. Regardless whether there is a supervening entity doing some directing, the process of differential survival/reproduction/migration cannot itself have goals.
An example of evolution in action: the Chemostat.
The operator of a chemostat has a goal – often, to create a pure cell line. The process by which this is achieved is by simultaneous addition and removal of medium, which causes purification by random sampling, which is evolution (a form of genetic drift). How can that process have a goal? There is no collusion between the cells in the original medium to vote one to be the sole ancestor of all survivors. How do I know this? That would be a pretty daft question. I think it would be incumbent on the proponent to rule it in, rather than for me to rule it out.
Strong and weak nuclear forces. Gravity. Electro-magnetic forces. Each of of these forces moves matter by applying motion through a change in energy potentional.
Not necessarily. I would agree that formulating them as “rules” is a way of modeling nature (helping us understand it). But without them, how do we even have “mechanism”?
Yes.
Neil,
The fact that the lower-level elements of a system follow strict rules does not mean that the system itself is incapable of flexibility.
Think of a neural network being simulated on a digital computer. The (virtual) neurons are deterministic, the software implementing them is deterministic, the processor is deterministic, the logic gates are deterministic. Yet the neural network can evolve new rules for classifying its input. Flexibility built on top of lower-level inflexibility.
Yes, it would, but the models at the upper end would give you information not derivable from models lower down (and vice versa). To make this (I hope) clear: it is my belief that we will, eventually (and in someways we have already) make robots – or man-made entities anyway – that can intend things, and bring them about. And those man-made entities can be built from “the bottom up” from parts and interacting systems. So that is the sense in which I am a “materialist” – I think that the right physical bits and pieces brought together in the right configuration results in an intending, executive agent. I think that’s what happens when we grow from a one-celled conceptus to a fully functioning human being – a physical combination of parts has been put together in a configuration that constitutes an intentional agent. But that does NOT mean that I think that if we knew where all the bits were, and the direction they were travelling in, we’d have a model of a person. We wouldn’t. It would be vast, but it would tell us nothing. I think to understand people we need to model at a much higher level.
caused by, certainly; “predictable” not so much, and I’m not sure what to make of “subsumed”. I think we have to be very careful of words like “predictable” – predictable by whom, based on what information? If I want to predict what will happen next when you say “My name is….” a model based on molecules, or sub-atomic particles, won’t work. But a person-level model works pretty well. You might say “sun-glasses”, but the odds are you will say “William”. I can therefore predict that outcome. I see no reason for that to be at odds with the idea that the system I can predict at william-level is nonetheless a configuration of mechanistically interacting parts.
Yes, I guess. And I’m not attaching any magic to the word “emergent” – it’s just a word to describe a property of a system that isn’t possessed by its contituent parts. A tornado is an emergent entity. In fact most things are, but we tend to reserve the word for large-scale systems.
Neil:
Lizzie:
You can get them even without the stochastic noise, as the neural network example shows.
In that scenario there is noise at the transistor level and below, but the logic gates are designed to be digital, which keeps the noise from “infecting” the higher levels of the system.
If the training input is also deterministic, then the network’s rules evolve deterministically.
So the answer to Neil’s question is “Rules come from mechanism, and it can even be deterministic mechanism.”
I re-read the Mossimo and Bich paper (“What Makes Biological Organization Teleological?”, 2014), and still think that they have a deeply compelling view. In particular I appreciate their criticisms of the idea that evolutionary considerations are sufficient for naturalizing teleology. Their use of Rosen and Kaufman is quite important.
As a minor aside, they criticize Varela for denying that autopoietic systems are teleological. This was indeed Varela’s view for most of his life, but in Weber and Varela (“Life After Kant: Natural purposes and the autopoietic foundations of biological individuality”, 2002) Varela changed his mind once he realized that autopoiesis is a theory of what Kant calls the purposiveness of organisms.
The Kantian view is, however, undermined by Kant’s own insistence that teleological judgments are merely expressions of how the mind organizes its experiences, and lack the objective validity of mechanistic judgments. (The details of Kant’s view of the difference between judgments of deterministic mechanism and judgments of teleology need not concern us; suffice it to say, for now, that I think Kant’s way of drawing this distinction is indefensible.)
A related point worth emphasizing is that even mechanism is grounded in a particular mode or way of embodied coping with reality. We adopt a “mechanistic stance” when we choose to isolate some part of a system, subject that part to spatial or temporal manipulation (actual or simulated), and observe the consequences. In other words, mechanisms are co-constituted by the kinds of investigative technique we bring to our inquiry into reality, and are by no means disclosive of how reality is in itself.
On this basis — among many others — we can conclude that “the dead world of mechanism”, as Brian Ellis calls it, is not the metaphysics of science. All of our scientific theories — physical, chemical, biological, psychological, etc. — show that reality is fundamentally active and interactive. The dynamics of real things have both active and passive aspects, and this is because there are real powers (which are themselves — and here I am much more of a Heraclitean than an Aristotelian — temporary configurations or constellation of processes or forces).
I would also like to add that if “materialism” or “mechanism” is not the ontology of nature according to our best contemporary empirical science, then a great many arguments against naturalism (e.g. Hart in Experience of God) are simply wrong.
I have spent plenty of time thinking about that. And I don’t see how it can ever be an adequate simulation.
Neil,
Such simulations don’t have to be of actual biological neurons. An abstract neural network implemented in software has the ability to evolve its rules whether or not it is intended or designed to mimic biology.
So once again, you have higher-level rules evolving flexibly in a system built of lower-level components that strictly follow rigid rules. Rules arising from mechanism.
All of those things are descriptions of behaviors reified as causes. In other words, gravity doesn’t push atoms around; atoms moving around a certain way in certain conditions is called “gravity”. Nobody knows what causes the effect known as gravity. All we have is a model.
I would say a hammer has a goal. But it’s one that is imposed upon whatever material is used to construct a hammer.
A hammer needs to meet certain criteria in order to fulfill the purpose of a hammer, and the fulfillment of those criteria is the goal of a hammer.
You won’t find many hammers made of jello, for example, or many one time use hammers.
So I’ve been catching up on a lot of reading wrt teleology and what I have seen so far is that there is widespread acceptance of teleology in biology, one might say it’s a fact of life. The debate is over how to account for the existence of that teleology.
The problem with any computational theory is that it is necessarily based upon switching, and that ignores the actual dynamics of living organisms.
Neil, can you say more about rules? At what point in the history of life do you see the appearance of rules? Human conventions? The genetic code?
Thanks
I’m not sure that is correct, though I suppose it depends on what you mean by “switching”. Our computers use switching. But did an abacus?
We use “rules” for several different things. Formal rules, as used in logic, are different from social rules, and those in turn are different from mechanical rules.
We ascribe mechanical rules. I’m not sure that they can exist, except as ascriptions. And we ascribe them back in time, even to the early cosmos. But I’m inclined to say that ascribed rules didn’t really appear until there were ascribers. So that would seem to require some sort of agency as a prerequisite.
William J. Murray,
‘All we have is a model’ could be applied to the matter itself. As Einstein showed, it is not a different thing from the energy that results from setting it in motion. It’s just that some forms of energy are more solid than others.
Nonetheless, you don’t need to know how ’cause’ is implemented to be able to use it, and investigate it. It isn’t ‘just a model’; it is an observed regularity. If two objects have mass, they will exert force on each other. How? By distorting spacetime. How do they do that? Er … There is probably a limit to our capacity to do the investigation, or even to understand the answer. But none of this creates an opportunity for ‘non-material causation’. First, find a phenomenon, then we can worry about mechanism. That’s what the fundamental forces are: phenomena.
Yes indeed, but the important thing about the description is that that it is also a prediction. Scientific models are essentially predictive rather than explanatory, although they can be cast as explanations. For instance, we can say: “The iron filings moved because they entered a magnetic field”. But that is just a manner of speaking – what we are really saying is that a magnetic field is a field that has the property of moving iron filings.
They are, however, to be any use, more than just descriptive. It’s no use just to be able to say “The iron filings moved when they entered the magnetic field”. A scientific model has to be able to say, with high degree of probability, that “The iron filings will move when they enter the magnetic field”.
Did you ever listen to that clip by Richard Feynman about magnets?
Here it is.
Yes, all we have are models, but what we have are predictive models, not mere descriptions. And it is the replicability of those predictions that gives us confidence in our models – as well as confidence that the world is “well behaved” – that models that work today will also work tomorrow. Which is itself interesting, and evidence, I’d say, that reality is real! Which is sort of cool.
Exactly. And it may turn out that “intention” is one of them. But I think the evidence is against it. It’s testable though, or could be.
Why do you say this? Firstly, not all computing need be based on switching – analog computers are possible. Secondly, actual neurons do “switch” – action potentials are discrete events, and the downstream neurons summate action potential inputs over time. So computerised neural networks really are pretty good models of actual neural networks, and “learn” by means of “Hebbian” mechanisms. The implementation in silica is a bit different from implementation in protein, but not that different. One difference is that in actual neural systems the currents are ionic rather than electronic, but that’s scarcely a difference in principle.
Why would you think that mechanisms are not dynamic or interactive?
A key defining feature of mechanisms that is their operation depends on the context : hence they are interactive. Further a mechanism responds to the context through the operation of the organized components. Hence they are dynamic: that is, their state and outputs vary with time based on context.
Coming from an admitted scientific realist, that statement puzzles me.
Mechanisms are a way of formulating scientific theories, so if the entities or structure of such a scientific theory tell us something about the world as per scientific realism then mechanisms tell us something about the world. I accept that other theories from different sciences may tell us something else about the world. But seems wrong for a scientific realist to say mechanisms in a successful scientific theory tell us nothing about the world.
In the paper you cited earlier, Kukla takes the stance-relativist position as well. At one point she says something to the effect that physics may tell us the world works this way, but only from that stance. I find that problematic for physics in particular.
I think at a minimum one has to accept that physics constrains what happens in the world. So I don’t see how one could claim stances are independent in a way that allows us to choose a viewpoint and so ignore physics.
Stances applied to sciences seem to be another way of thinking of the separate domains of science. I agree that, in general, each science yields a valid set of explanations within its domain, and that, in general, we cannot expect to reduce explanations (and maybe ontology) from one domain to lower one.
But even if a complete reduction is impossible, there are many examples of inter-domain explanations, such as aspects of chemistry using quantum mechanics and memory (psychology) reduced to biochemistry and neuroscience. Often mechanisms provide the inter-domain structure: the emergent causal properties of the whole mechanism operate on phenomenon at the higher domain and the causal organization of the components is expressed and explained in the lower domain.
So it seems wrong to me to assert that domains or stances are separate, and so concepts and models from one cannot bear on another.
I would prefer to say predictive and explanatory power are each standards used to assess scientific models.
David Deutsch talks about a copper atom on the end of the nose of a statue of Churchill. Even if we had some huge physics model which perfectly predicted the history and position of that atom, would we consider that a useful scientific model of why that atom was there? No, we would want an explanation in terms of history of England and why people honor their heroes.
I interpret WJM’s concerns with scientific explanation as based on an intuition that they should make sense in terms what people see day-to-day. The Feynman video has an example. At one point, Feynman says perhaps the questioner hopes that Feynman will be able to say that electromagnetic attraction is like elastic bands. But the opposite is true: elastic bands work because of electromagnetic, molecular forces.
If we want to go beyond human artifacts like statues to understand all aspects of the whole world, then physics is the science that must be used. And its explanations are mathematical and can only be understood through the mathematics. Any expression in everyday language using everyday concepts is going to be partly wrong and incomplete.
Better explanations will come from unifying and generalizing the mathematics, as in attempts at a GUT. Not in trying to make the science accord with our everyday intuitions about causality and explanation.
Elizabeth,
Predictive descriptions are still not causes, which was the point I was making to Robin. To say that gravity is a material cause is erroneous thinking; gravity is not a material cause, it is a description of a mechanistic effect. We don’t know what causes it.
Thus, to say that all known effects (berhaviors) are “caused by” material causes is erroneous all the way down. because all scientific causes are models of behavior all the way down.
It might be justifiable to say that all known effects are mechanistic – but then, even that is not known to be true. Since all we have, scientifically speaking, are models, then what does it mean when we talk about intentional effects? These effects or patterns would not be plausibly described as deterministically lawful, stochastically distributed or chaotic. The patterns of behavior would conform, in some way, to an intention.
We certainly see matter behaving in ways that conform to intention – we experience it, more or less, directly with our own bodies as we intend to move, and then move accordingly. Also, our bodies manipulate other matter in accordance with an intention, matter moves around in ways that we currently cannot sufficiently model without the use of intention as a descriptive commodity. While all the details of the motions of the matter involved appear to individually follow all of the mechanistic predictions, they appear to have been set in an overall, cumulative motion not reducible to those models alone.
IOW, without intentional agency, the words you are reading cannot be sufficiently modeled. They are not deterministic, probabilistic or chaotic. They do, however, conform to a general pattern of what we can expect by an intentional agency; they appear to be arranged specifically and directed towards a teleological purpose that corresponds to the intention involved.
IOW, we wouldn’t expect matter to behave in certain ways without without an additional reified cause: an intentional agency. Like with gravity and forces and energy transfers, whatever is causing that behavior is essentially unknown; all we can do is observe and model the effects and not the distinction between intentional effects and other kinds of effects.
Some arrangements/behaviors of matter, while additionally conforming to other known laws and forces, appear to be incomplete descriptions without intention.
I’m surprised that Ruth Garrett Millikan hasn’t come up in the context of this discussion of function and purpose in biology. I’d expect Dennett to draw heavily upon her thesis in his upcoming book.
I don’t actually see any basis for that.
Physics constrains physicists, and other people (particularly engineers) find it useful to follow some of those constraints. But I don’t see that it otherwise constrains anything.
Yes, I almost agree with that, except that they aren’t standards. They are more like guidelines.
And I am agreeing with you!
I was, however, pointing out the difference between a “descriptive” and a “predictive” model. As Feynman so rightly says, once you get down to the “bottom” of the explanatory chain, you get to where “that’s the way the world is”. That doesn’t we can’t produce higher-level explanations, including teleological ones, and Feynman gives some good examples. But if “reductionism” means anything sensible, it means that what happens in the world is a result of a few fundamental properties that it has, and what physicists do is try to find out what those are. What people like me do is study the systems that result.
Right. Because it is fundamental. And, again, as Feynman points out – the interesting question about gravity is not “what causes it” but how it relates to the other fundamental forces. When I last looked it still seemed to be out on a limb. A “Theory of Everything” wouldn’t be an explanation – it would just be a simpler fundamental model.
Well, it depends what you mean by “reducible”. I am fairly confident that intentional systems arise from the fundamental properties of the world as (imperfectly) understood by physicist. But I’m even more confident that we won’t understand intention better by understanding more about those fundamentals (pace Penrose). I think we will understand it more by understanding the higher level systems that we observe in animals capable of intentional behaviour. And I think we have a pretty good handle on that right now, actually.
Right: they can only be understood at the level of the system. This is what I understand that Dennett was getting at with his three “stances”: Physical, Design, and Intentional.
And some phenomena can only be understood at a “higher” level stance.
Example: bird flight
We can understand why birds fly in terms of physics – the “physical stance” – in terms of muscle activation and aerodynamics. We could also understand it in terms of the genes and homologies that enabled wings to evolve.
We can understand why they fly in terms of “design” – being able to fly serves the continuation of their species by allowing them to find food and escape danger.
We can also understand why the fly in terms of intention – to find food, to distract a predator from their offspring, to attack a predator, to attract a mate.
The Physical stance doesn’t allow us to answer questions relevant to the design stance or the intentional stance, but it isn’t (I would argue) that some new explanatory feature is introduced at each level – it’s simply that we are asking a different question – we are, to use your term “reifying” a property that emerges from a lower level of analysis. And if we don’t do so, our models will be predictively useless. And I would argue that the test of a model is its usefulness.
Well, I would put it slightly differently: Some arrangements/behaviours of matter can only be predicted usefully at the Intentional level. I do not think that entails there being an additional ingredient called “intention”, other than the configuration itself.
oops I see that others have said something similar.
Allan Miler said:
Same thing, for all intents and purposes. Indeed, observations themselves are models – representations of sensory stimulation modeled in our mind/brain.
That phrase doesn’t even have any coherent meaning given the conversation to this point. Perhaps you mean there is no opening for a non-mechanistic model? I disagree. We already have a non-mechanistic model: intentional effects. Some people, however, seem to insist that intentional effects are ultimately subsumed by mechanistic models. I think that’s an ontological commitment and has no scientific import when it comes to model making.
The question is if science should be insisted on as a methodology committed to only making models of behaviors that can be reduced to mechanistic patterns, and/or if science should insist that all behaviors are ultimately produced by combinations of mechanistic behaviors. It seems to me that these are ontological commitments and do not necessarily represent any natural boundary for scientific investigation.
For example, if we find an unknown artifact, would it necessarily be impossible to recognize the patterns of matter as intentionally arranged, and would it be impossible to scientifically extract a good idea of its purpose?
In what sense is “intentional effects” a “non-mechanistic model”, William?
Serious question – I am still not clear on what you are calling “non-mechanistic” nor what you are calling a “model”.
It wouldn’t be “necessarily impossible”, no. Anthropologists do this all the time.
You really need to catch up with physics reseach in the latter 20th century William. The majority of physicists today agree that the evidence shows that mass causes gravity. They’ve even used the support of said evidence to locate both quasars and blackholes – both mere theoretical objects not 15 years ago.
Contrary to your beliefs, quite a lot is actually known these days about the causes of material forces. And, as has been noted before, the more we understand, the less evidence we find of anything other than the material having any effects in this universe.
William J. Murray,
Rendering the whole discussion a mere bit of Sunday-afternoon time-passage. Everything, including observed regularity, is, in some sense, ‘just a model’. Therefore … ?
I’m glad you think so, which is why I put it in scare-quotes. You are the one who keeps banging on about ‘materialist ontology’ as if there is something else.
It would be perfectly possible to recognise that kind of intent. But one would also need to be aware of the recently-discovered (160 years) principle of Natural Selection. Because this provides an environmental input to what survives, it can look as though it was intended that what survives was tuned to the environment. But in fact, in the case of NS, what survives is tuned by the environment.
If the artefact does not replicate, its arrangement is more likely due to intent. If it replicates, that probability goes down.
Stepping back a bit:
It seems to me that if we want to talk about intention we need to talk about an intender. And an intender, I suggest, has to be some kind of sapient entity capable of both imagining alternative futures and of executing actions that will preferentially tend to bring one, rather than the other, about.
There might be definitions of “teleology” that don’t require an intender (and I suggest that those definitions are covered by “teleonomy”) but surely it’s intention we are interested in here?
So one issue is: can an intender be the outcome of physical interactions, where those physical interactions can be “reduced” to fundamental properties of the universe? I suggest yes.
And the next is: can there be an intender thus defined that is NOT the outcome of such physical interactions? I suggest no.
EL said:
I don’t know what you mean by “higher-level”. To me, it seems like you are saying “Mechanistic models don’t actually sufficiently predict intentional effects, but since I’m ontologically committed to the idea that intention is not a fundamental model, it must be something that is produced by mechanistic effects., so I call it a “higher-level” or “system-level description of aggregate mechanistic effects.” IOW, you can’t make the direct prediction from any collection of known mechanistic models, but you are assuming it is an aggregate effect of mechanistic models, ergo “higher level” or “system-level description”. Am I correct?
Is it reasonable to say that matter is behaving as mechanistically expected, if the behavior of the matter in the system cannot be sufficiently modeled without reference to a non-mechanistic descriptor? Let’s say that at what you call the “system” or “higher” level an intentional agency (reified cause) must be evoked or else the description is flawed/insufficient. Doesn’t this factually mean the system is not (as far as we know) reducible to known mechanism or interactions thereof?
You seem to be in agreement here, but what follows doesn’t appear to me to be rational or scientific: why insist that the added, necessary descriptor is a “higher” level or “system” level descriptor, and not a fundamental descriptor? Aren’t all descriptors – gravity, strong and weak nuclear forces, entropy, inertia, electro-magnetic fields, dependent upon observations of some sort of system? Can a solar system be described without gravity being utilized as a system-wide descriptor?
It seems to me that you are unnecessarily qualifying “intention” as a different kind of descriptor that has been produced by an aggregation of other effects (“higher”, “system”) for no sound reason other than an a priori commitment that intention is not as fundamental as any other characteristic necessary in describing a system.
Well, that may be the interesting scientific question, but “what causes gravity” is, at least for me, a very interesting philosophical question.
Given the following statement:
… I dont’ see how you can possibly justify this statement which preceded it:
If you are confident a more thorough understanding of the fundamental mechanistic properties of the world will not provide a better modeling of intention (which is supposed produced by those properties), how can you possibly be confident that intention arises from those mechanistic properties?
Pardon me, but that sounds like metaphysical bias.
I’ve selected this quote because it ties together two issues: whether my acceptance of the stance/pattern distinction (or something like it) is inconsistent with scientific realism, and whether stances overlap or bleed into each other.
I can accept quite easily that stances can overlap or bleed into each other, and in many cases we can and do adopt different stances towards the same object at the same time. For example, a physician treating her patient adopts both a discursive stance — treating the patient as a person with beliefs, desires, anxieties, etc. — and also a teleological stance — treating the patient as a diseased organism.
Apart from the discursive stance (taking others as capable of playing the game of giving and asking for reasons) and the teleological stance (taking others as organisms-in-environment, with relevant norms of proper functioning) there is also what I would like to call the nomonological-mechanistic stance. This is like Dennett’s machine stance, but I want to stress the role that laws play in selecting what parts of the system to count as invariants and what parts of the system to count as variables.
Now, in one important sense — and I think this follows from what Kukla says, but she doesn’t stress it as much as I want do — all the stances are real, because a stance is a kind of embodied coping, not an intellectual schema. By this I mean that a stance, in order to really count as a stance at all, must do two things: it must be generally successful at classifying and describing the modally robust properties of the processes it detects and tracks, and it must be able to provide an explanation as to why it is generally successful.
I want to add the second qualification because of a very interesting problem that the Dennettian can be asked: why isn’t there a religious stance? After all, if pragmatic workability is the only criterion of a stance being acceptable, and stance-relativism implies that we need to reject metaphysical realism (which might be true, and might not be), then the stance theorist can’t reject religious claims on the grounds that they fail to correspond to reality.
The difference here is that religion is basically magic, and magic is not a genuine stance. It is instead what I call a pose. A pose does, of course, generally detect, track, and classify modally robust properties. It’s not like a psychotic episode or sustained hallucination! But the conceptual resources deployed in a pose do not account for why the pose is as generally successful as it is. Religious practices are poses because when it does genuinely engage with reality in terms of managing group dynamics and enforcing shared norms of conduct, not in terms of untestable pseudo-explanations.
As for physics: I’m inclined to follow Ladyman and Ross here in thinking that the only kind of physics which constrains all stance-relative ontological commitments is fundamental physics, which today consists of quantum mechanics, general relativity, and thermodynamics. Possibly dynamical systems theory could be considered a fundamental physical theory, but I simply don’t know.The “fundamental” means that any measurement taken anywhere in space or in time can confirm or disconfirm a hypothesis entailed by the theory.
But, that’s actually a very loose constraint, and beyond that, “rainforest realism” reigns: there are many different real phenomena, and different stances track and classify real phenomena in different ways, and so we have a plurality of characterizations of reality. What makes them characterizations of reality is simply that they can support testable counterfactuals. Testable counterfactuals at the propositional level is our criterion for tracking modal properties of processes at the ontological level.
It might be objected that at this point my scientific realism has been watered down to the claim that
I firmly agree with Thompson here: all processes are equally real and none has absolute ontological primacy. (Incidentally, this is why I am no longer consider myself an “emergentist”, since “emergentism” does presume some “base” that has ontological primacy over whatever “emerges” from it.) This process ontology is (a) firmly grounded in our best empirical science, so it is a scientific metaphysics, but also (b) since it denies that any level has any ontological primacy, it is antithetical to both “materialism” (“matter has ontological primacy”) and to “idealism” (“mind has ontological primacy”).
To this I would only add that different stances are more or less reliable ways of tracking real processes by characterizing those processes in mechanistic, teleological, and/or discursive vocabularies. It’s a very milquetoast scientific realism, to be sure, but remember that my basic reason for accepting scientific realism in the first place was to acknowledge the no-miracles argument about scientific progress. So a very milquetoast scientific realism will do quite fine!
It is a further question as to why we have the stances we do. My best conjecture these days is that the discursive, teleological, and nomological-mechanistic stances are all social practices that exploit the cognitive modules underlying folk psychology, folk biology, and folk physics (respectively). Those modules in turn are kinds of cognitive representation that are partly innate and partly acquired through socialization — but they wouldn’t be evolutionarily stable and developmentally canalized if they weren’t also tracking real features!
(If Karniloff-Smith is right in thinking that human cognition involves a lot of mapping between modules, that would certainly support the idea that the stances can bleed into each other and that one can take different stances towards the same object at the same time.)
I also strongly suspect that, for most of human history, the different stances bled into each other far more than they do for us Westerners today. What we call “animism” can be described as “the original stance” prior to the differentiation of the discursive, teleological, and nomological-mechanistic stances — a differentiation that is enforced by our social practices and their institutionalized histories.
Can you say more? Link?
EL said:
What generates such suggestions, other than a priori metaphysical bias?
Robin said:
How does mass cause gravity, Robin? You’re doing nothing here but begging the question>
Robin, I think you are not comprehending the nature of the point. Others here, like EL, do understand and agree with me, so please consider that you may actually not be understanding the point.
The mass-causing-a-indention-in-space-time is a model used to predict gravitation effects. How does mass attract matter? How does mass make an “indention” in space-time? These are not explanations, Robin: they are descriptions of what occurs, and models that aid in predictions. They are not explanations of how mass does what they describe it as doing.
I don’t think that makes sense.
If we ever tried to model intentional effects, we would do that mechanistically. The mechanistic model might not be very good, but it is still what we would do. Because mechanism is how we model.
The mistake that I see, is that many people believe that because our models are mechanistic, therefore the world itself is mechanistic. I don’t see a basis for that belief.
Allan Miller said:
I consider the terminology, and the epistemological and ontological concepts they imply, very important. When one asserts that all causes have been shown to be material, what are they actually asserting? In fact, no cause has been shown to be “material”. Many behaviors have been shown to be sufficiently described in mechanistic terms; but many are also known not t be sufficiently described in mechanistic terms. So, no, not all “causes” are known to be “material” – meaning, mechanistic.
No, you are. I’ve specifically included other kinds of ontologies in my dialogue and have even agreed that it might be time to drop the term “materialist” altogether due to lack of useful meaning.
Translation: Neil Rickert is a mechanistic ideologist, in addition to being an atheist and a reductionist.
Mechanistically, mechanistic, mechanism!!!
I wouldn’t trust a myopic person like that talking about ‘intentionality.’
Right. And this is absolutely key, I think, and the point I was trying to make (but you make better) in an early OP (can’t remember which one): that science isn’t biased against non-mechanistic models, it’s rather that they are the only kind that science can do. Whether or not you think there are more things in heaven and earth horatio than are dreamt of in your scientific models, or not, is a matter of belief, not of science.
But I would certainly add that it is perfectly possible to make a mechanistic model (if I’m understanding that term right) of intention. That’s why robotics is actually viable.
Gregory, why did you omit the second part of Neil’s post? It totally negates your point.
Excellent!!!!
I concur.
The first: evidence for.
The second: lack of evidence for, or reason to posit.
Elizabeth,
The point stands because he has no alternative. What alternative did he name, Lizzie?
Btw, you’re not a woman, Lizzie, you’re a clone.
(Sorry missed this one)
No. I’m saying that systems are hierarchical (systems and systems of systems) and that at a “higher level” (i.e. systems of systems of systems….) the system has properties not necessarily shared by the subsystems of which it is comprised.
I’m not assuming it – it is supported by evidence. For instance, take that robotic dog above. The reaction of the robotic dog to being kicked can’t be modelled in terms of its circuits or physics – that wouldn’t tell you anything. The system is chaotic, and the only way of predicting what the dog will do to a particular kick would be to run entire simulation model, and the output might be quite different in response to a different kick. So not only would the only mechanistic “model” be as large as the phenomenon itself, it wouldn’t generalise at all.
What does generalise, however, is the simple model “when you kick the dog it tries to get up again”. That will work every time (even if it is lots of different kicks rather than one repeating gif). I assume you agree (there seems no good reason to disagree) that that robotic dog is “mechanical”. Yet its behaviour can only be sensibly model at the “intentional level” – at the level of the whole system, and in terms of teleology – what the “dog” is “trying” to do (stay on its feet, even if kicked).
You could put it that way, which is why I don’t find “reductionist” a good descriptor for my own position. But the issue is not whether the description is flawed/insufficient, I would argue – this is why I keep saying that scientific models are NOT merely “descriptive” to be successful models they must be predictive And a mechanistic model, say of that “dog” would be a terrible predictive model. But an intentional model is highly predictive – it will work for all kinds of kicks and hazards the “dog” meets: the simple model is: “the dog tries to stay on its feet”.
I think we have confusion of terms here – I’m not talking about “system-wide” fundamental descriptors (or, rather predictive models), when I talk about “system-level” models. I make the assumption that our fundamental predictive models (laws about fundamental forces) are universally applicable, i.e. not suspended sometimes – it’s why we call them “laws”. By “system-level” model, I mean a predictive model that works efficiently. For many complex phenomena, a higher-level model is far more parsimonious because the “entitites” it invokes are systems in themselves. As with the dog – instead of trying to model the dog as a vast piece of computer code (which could be done, and probably was, prior to the prototype), but which would have little predictive efficiency, you can do it much more effectively by modelling it as a system, as a “dog trying to stay on its feet”. It’s a “higher level” model because its units are systems of systems, not fundamental particles. Even though I have no problem in agreeing (or assuming) that the dog is composed of fundamental particles and nothing else.
Gotta run for now….
What mysterious force sets matter in motion?
That would actually be a bad mistranslation.
Is it really the case there are no non-mechanistic but empirically testable models in scientific explanation? What about in the social sciences? Are economics and sociology non-mechanistic?
I’m pretty sure that there’s a big disconnect between what different people here mean by “mechanism.” My reading of Lizzie and Neil is that they are using “mechanism” as a synonym for “empirically testable model of some phenomena”.
For others — Murray, perhaps? — the term “mechanism” seems to mean something that’s ontologically or epistemologically far more demanding — perhaps along the lines suggested by Laplace:
Hope this helps a bit!