Archive for ‘intentionality’

March 6, 2011

The trouble with epistemology

by Neil Rickert

My view of epistemology is probably colored by the fact that I am a mathematician.  We mathematicians seem to think differently about such things.  For example, nothing could be clearer to a mathematician, than that it is possible to know the axiom of choice, yet at the same time to not believe it.  Thus mathematicians are likely to see something wrong in the “knowledge is justified true belief” that philosophers often assume.

If we go by what the epistemology literature tells us, we might conclude that “truth” is the name of an immaterial magical substance that permeates the universe and that people search for.  We might conclude that the universe is filled with immaterial objects known as propositions.  We might also conclude that perception is a magical system for picking up such propositions, and that it has a builtin magical filter that allows it to mainly pick up true propositions.

Well, okay, that was a bit overstated.  The main point is that epistemology gives a very artificial account.  It comes across as an account of what would constitute knowledge for an ideal rational agent living in an imaginary Platonic universe.  Knowledge is defined in terms true beliefs, but “true” and “belief” are never really defined.  So we are left with true beliefs as something like abstract objects.  This leaves epistemology as a kind of logical calculus of abstract objects, so it has something of the appearance of mathematical Platonism.  Because the knowledge is in terms of abstract objects, it does not relate to reality.  But never mind — the epistemologist comes up with a property of intentionality which is supposed to provide that otherwise missing connection.  But intentionality is left as unexplained, so either mysterious or magical.

By contrast, when I look at science, the scientists are very concerned with connecting their scientific statements with reality.  Epistemology has a problem with intentionality, and that problem carries over to scientific epistemology (the philosophy of science).  But science itself does not seem to have that intentionality problem, for it carefully defines its terms in ways that connects it to reality.  It should be obvious from this description, that I see a serious mismatch between scientific epistemology and the science that it is supposed to explain.

The real problem of knowledge is expressed in the question “How is it possible to have knowledge at all?”  This can be further elaborated as “How is it possible for a sequence of letters to say something about reality?” and “How is observation even possible?”  In short, the real problem of knowledge is the problem of intentionality.  That is the problem that drives science.  Epistemology massively avoids dealing with that problem.

February 15, 2011

Purpose (7) – summary and index

by Neil Rickert

I have been pondering whether to continue this series with more posts.  But I think I’ll end it here for now.  I will later have some independent posts that are related.


One of the arguments that we repeatedly see from creationists, is that there is something missing from a purely mechanistic view of the world, and that is where they want to put their deity or intelligent designer.  What they see missing, is an explanation of an apparent purpose.

Part of that missing purpose is a backward construct from their theology, and their wish that they (or humankind) be a product of purpose.  I cannot find any basis for that.  However, there does seem to be a basis for seeing apparently purposeful behavior in biological systems.  And that’s what I have been discussing.

My main emphasis has been to show that there is an adequate natural account for this apparently purposeful behavior, so there is no need to call on theology for a pseudo-explanation.


For ease of future reference, here’s an index to the posts in this series:

February 7, 2011

Purpose (6) – background

by Neil Rickert

For this post, I want to go over how I became interested in purpose.

It all started with a personal interest in understanding how humans learn, so I spent some time studying the problem of learning.  I tried to combine two approaches.  One of those was to use the methods of AI, which would require modeling learning as a computational problem.  The other was to look at natural learning such as occurs in biological systems.  The aim was that ideas I might glean from looking at natural learning systems could perhaps provide guidance for computation learning.

Computers are versatile, so when you have a specific problem, you can usually come up with a way of solving it.  But the general problem for learning is that we don’t start with specific problems; we somehow just learn without being told what to learn.  And the general problem that I ran into was one of setting direction.

One of my grad school professor, S. Kakutani, would sometimes ask “Pick a number. Square it.  Is that a theorem?”  The idea was that if I pick a large number x, and square it (multiply it by itself) to yield y, then the statement “y is a perfect square” is a true statement that has probably never been stated before.  But no mathematician would consider that a worthy result.  We don’t just come up with true beliefs; we come up with interesting true beliefs.  And that leaves the difficulty of deciding what is interesting.  Hubert Dreyfus, a sometime critic of AI, expressed the general problem this way.

Using Heidegger as a guide, I began to look for signs that the whole AI research program was degenerating. I was particularly struck by the fact that, among other troubles, researchers were running up against the problem of representing significance and relevance – a problem that Heidegger saw was implicit in Descartes’ understanding of the world as a set of meaningless facts to which the mind assigned what Descartes called values and John Searle now calls function predicates.

So I went looking for sources of such direction.  I remember some particularly useful online discussions with Chris Malcolm of Edinburgh University, and those discussions were part of what set me looking for a good account of purpose, eventually leading me to the position that I have been presenting in this series of posts.

January 31, 2011

Purpose (5) – nature and purpose

by Neil Rickert

It seems obvious enough that purpose, as I have been describing it in this series of posts, is entirely natural.  However, some ID (intelligent design) proponents probably disagree.  In this post, I shall explain why I believe it to be natural.  I have already provided much of the basis for seeing that purpose is natural, in that I have connected it with measurement.  However, the examples that I have used, such as the thermostat, are man made.  In particular, the measurement aspect of the thermostat is the result of human design.  Thus we might expect some ID proponents to claim that they show the need for an intelligent designer.

In order to complete the picture, I need to provide examples of natural measurement.  For that, I want to turn attention to homeostatic processes.  To say that a process is homeostatic is to say that keeps itself in some sort of equilibrium.  Such homeostasis works on the basis of feedback, where the process is reacting to its own current state and modifying its behavior in ways that tend to keep it in a reasonably stable range.  That feedback is a form of self-measurement.  Thus a naturally occurring homeostatic process already exhibits natural measurement.

read more »

January 24, 2011

Purpose (4) – chaos

by Neil Rickert

In earlier posts in this series, I have pointed out places where there is apparently chaotic behavior.  In this post I shall further comment on chaos.

I’ll start by indicating why I have been mentioning it.  Many people seem to take the view that we could manage with only using mechanistic explanations, and that the vocabulary of intentions is unnecessary, though perhaps convenient.  For example, that appears to be the view of that Dennett is suggesting in his The Intentional Stance.  I mentioned examples of chaotic behavior because they are cases where mechanistic explanation breaks down, and thus fails to be adequate.  Thus they show that mechanistic explanation is not adequate, and that there are cases where we need teleological explanations.

Now it may be that the world is still entirely mechanistic, and that mechanistic explanation fails for chaotic behavior because as finite beings we are limited in the amount of information we can have, and we cannot have enough for a full mechanistic explanation.  Or it could be that the world is not entirely mechanistic, and a mechanical explanation could not be completed even in principle.  You, the reader, will have to decide which of those positions you want to take.  I will present my current tentative view at the end of this post.

read more »

January 17, 2011

Purpose (3) – measurement and purpose

by Neil Rickert

In this post, I plan to connect the notion of purpose to that of measurement.  Here, I am using “measurement” to refer to the process of measuring, rather than to an individual item of data.  I could not find much in the way of a philosophical account as to what measurement is, though a google search for “philosophy of measurement” (with the quotes) turned up some references to metrology and some discussion of the problems of measurement in psychology.

Measurement and chaotic behavior

I will be interpreting “measurement” quite broadly, so that much of what we consider to be ordinary observation can be considered to be measurement.  Our ordinary experience in measuring things is generally a good guide to what counts as measurement.  In particular, when we are measuring something, we normally expect that the result of that measurement will be a decimal number.  While scientists mathematically model measurements as being real numbers from a continuum, actual practical measurement gives discrete values of limited precision.  So a measuring process is a discretizing process, or something similar to a digitizing process.  Because of this discretizing and because of limits of measuring equipment, measuring typically gives approximate but inexact results.

The process of measuring is a purposeful one.  Or at least that is how I see it.  However, it seems to be a common view that perception is passive (see the Wikipedia entry), so those who see perception as passive might also consider observation to be passive.  And that could lead to disagreement over whether measurement is a purposeful activity.  My main aim, that of connecting measurement with purpose, does not depend on assuming that measurement is a purposeful activity.

When measurement is carried out by people, such as when using a ruler to measure the width of a window, it isn’t clear that it makes sense to talk about that measuring being a chaotic process.  We would, instead, tend to think of it as inherently ambiguous.  For example, if that window is just between 31.6 inches and 31.7 inches, then we might be unsure which of those two measurements to choose.  When we automate measurement, as with a digital thermometer or a digital voltmeter, that same problem does show up as a chaotic process.  A tiny change in the input causes the digital measurement reading to jump from one value to the next.  When we idealize measurement mathematically, the ideal measurement value is a discontinuous function (a step function), and it is that kind of discontinuity in the idealized measurement that leads to chaotic behavior in an automated measuring system.

The same problem of ambiguity or chaos shows up for simple observations.  A cat enters the room, walks over to the mat in front of the fireplace, and settles there.  At the end of this, “the cat is on the mat” will be true, but when the cat just entered the room it would have been false.  There will be a stage where it is ambiguous as to whether we should say that “the cat is on the mat” is true or false.  If we were to setup an automated “cat on the mat” detector”, then it would behave chaotically at the stage where it is making the transition from false to true.

Purpose and measurement

Here is how you can construct an automated system or robotic system with apparently purposeful behavior.  First identify what you consider the purpose, and then devise a way of measuring the extent to which that purpose has been achieved.  We then program the automated system so that it periodically measures its degree of purpose achievement, and then takes action intended to move it closer to that goal.

The programming could amount to using an algorithm that moves the system closer to its goal.  Or it could even be a trial and error procedure, that tries something and if that doesn’t work, tries something else.  The use of randomness could be part of that trial and error procedure.  As it finds something that seems to work (as determined by the periodic measurements, the trial and error program can attempt modifying what works to find an even better way of solving the problem.

The house thermostat is an example of just this.  The thermostat measures whether the room is warm enough (i.e. meets the intended purpose of the thermostat), and if not it turns on the heating system so as to heat up the house.  As another example, consider the heat seeming anti-aircraft missiles.  Once the missile has been targeted on a heat source (typically the jet engine of an enemy aircraft), it attempts to reduce its distance from that heat source to zero by modifying its own motion.  We can see that as having a purpose of colliding with the enemy aircraft (the heat source).

A theory of purpose

As a tentative theory of purpose I want to suggest that, at least within scientific discussions, we should take “purpose” to mean just such a measurement controlled program as just considered.  In ordinary non-scientific conversation, “purpose” is sometimes used in other ways.  However, we should be more careful about usage when using “purpose” scientifically.  Specifically, we should use the terminology of purpose, only when we have reason to believe that there is some kind of measurement going on, with a program of behavior that is controlled by the measurement in a way that is likely to achieve the indicated purpose.

To illustrate this, I would like to analyze some examples, including those mention by Eric Thomson in a comment to my previous post in this series.  One of the examples he mentioned was “a (naturally formed) system of gulleys in a mountain side that tended to somehow sort small and large stones into two piles.”  In that case, I see no measurement going on, so I see no basis for considering that to be the following of a purpose, except in a metaphorical sense.  Contrast that, however, with the apple orchardist who has a system of moving belts and diverters to sort the apples by size.  In that case, the orchardist would be checking on the sorted apples, and adjusting his apparatus so that it sorts them as wanted.  That checking is a kind of measurement of whether the purpose is being achieved, so it would be appropriate to describe apple sorting as the purpose of the apparatus.

For another example, consider the heart (also mentioned by Eric Thomson).  It is often described as having a purpose of pumping blood through the body, though that is not how Eric describes it.  However, I cannot see anything measuring whether that purpose is achieved.  However, there are biological feedback system that are, in effect, measuring whether the heart is rhythmically pushing blood out into the arteries, so we can reasonably describe that as a purpose.  And note that what Eric says is the purpose fits that well enough.

January 10, 2011

Purpose (2) – Teleological explanation

by Neil Rickert

Briefly, a teleological explanation is an explanation based on a purpose.  The thermostat, for example, has the purpose of maintaining a suitable temperature, and we often explain its operation in terms of how it meets that purpose.  Biologists sometimes talk of teleonomic explanations.  An explanation is said to be teleonomic if it is based on apparently purposeful behavior, and teleological if it is based on a purpose arising from a conscious agent.  Since I am not particularly concerned with the role of consciousness here, I shall not make that distinction and will use “teleology” to describe both cases.

Mechanistic explanation

Teleological explanation can be contrasted with mechanical explanation.  A mechanical explanation is one based on the idea of physical matter in motion, as described by laws of physics.  When we describe the behavior of objects using Newton’s laws of motion and Newton’s law of gravity, we are providing a mechanistic explanation.  No purpose is assumed by such an explanation, so the mechanistic explanation is entirely non-teleological.  Scientists usually prefer mechanistic explanations, where they are available.

Chaotic behavior

Mechanistic explanation can break down, when there is chaotic behavior.  Chaos, as I am using the term, is well described by:

Mathematically, chaos refers to a very specific kind of unpredictability: deterministic behaviour that is very sensitive to its initial conditions. In other words, infinitesimal variations in initial conditions for a chaotic dynamic system lead to large variations in behaviour.

When a mechanistic explanation deals with chaotic behavior, the explanation is of limited use.  In particular, it is unable to make reliable predictions.

Example – the thermostat

We can think of the thermostat typically used as part of a heating system for a house.  The thermostatically controlled system keeps the house at a near uniform temperature.  The thermostat, as a simple device, is part of the controlling functionality.  If we ignore the electrical aspects, then a thermostat has a simple mechanistic description.  In the traditional version, a bimetallic strip bends as it is heated, due to the differential expansion of the two metals.  And this bending brings two surfaces (usually part of a switch) into contact.  Once we include the electrical characteristics, things become a bit more complicated.  The switch does not instantly go from open to closed.  Rather, the resistance of the circuit goes from very high (open circuit) to very low (closed circuit).  The transition in resistance is chaotic, and that limits the accuracy of a mechanistic account of what happens.  However, the fact that the electrical transition is chaotic does not interfere with the intended purpose of the thermostat.  So we can give a good teleological account without getting into the details of the chaotic behavior.

Pseudo-mechanistic explanation

Typically, the full thermostatically controlled system is explained by describing the thermostat as switching from open to closed (or from off to on) when it reaches the set temperature.  But when we describe it that way, we are not giving a mechanistic account of the thermostat, for we are not talking about parts in motion.  We have, in effect, replaced the thermostat in our description with an abstract ideal machine that just switches.  The full explanation of the thermostatically controlled system, when given that way, has the general form of a mechanistic explanation, except for our substitution of the ideal abstract machine for the actual mechanism of the thermostat.  I shall use the term “pseudo-mechanistic” for an account that has the general form of a mechanistic explanation, but is based on the “mechanism” of an abstract ideal machine.  Such an explanation is implicitly teleological, for we have constructed that abstract ideal machine based on the intended purpose of the actual thermostat.

When it comes to pseudo-mechanistic explanation, the big example is the digital computer.  We typically explain its operation in terms of logic gates, flip-flops, latches, etc.  The flip-flop is used as a one-bit memory device.  Electrically, it is a transistor like component, but with the transistors operating in a non-linear region.  We normally describe it as having two stable electrical states, and being switchable between the two.  The states might be indicated by a voltage or a current flow, depending on chip design.  The transition from one electrical state to the other is chaotic.  We often describe this as a memory cell which can have the value 0 or 1, and in using that description we do not mention the electrical values.  Likewise, a logic gate is usually described as having an output value of 0 or 1, depending on the inputs.  Again, the actual physical device has output voltages or currents (depending on chip design), and the transition between the output levels that we label “0” and “1” is chaotic.  In typical computer explanations, we describe the logic gate as an idealized abstract machine that can have an output of 0 or 1.  Our explanation of computer operations is in the form of a mechanistic explanation, except that it is based on these abstract ideal machines such as logic gates.  And our use of abstract ideal machines is based on the intended purpose of the actual electrical circuits, so is implicitly teleological.  There’s a bit of an irony here, for AI (artificial intelligence) proponents are often outspoken in their favoring a mechanistic view of everything, yet they rely on a teleological account of their computers.

Purpose in biology

When the inputs to a neuron reach a sufficient level, the neuron “fires” and transmits a signal.  This is usually described as a threshold event, with the neuron activating (or firing) when the input reaches a threshold.  With threshold events, there is a large output change from a small input change (that last little bit of input that pushed to the threshold).  And because of that, we should consider the operation of the neuron to be chaotic.  This is an example of an apparently purposeful action of a biological cell, though it is hard to be precise about what we should consider the purpose, since the operations of the brain are not yet fully understood.

When we talk of a struggle for survival, we are using teleological language, and assuming some sort of intrinsic survival purpose in the biological organism.  If we say that the purpose of flowers is to decorate our living rooms or our gardens, then we are imposing our own purposes on the plants.  If, however, we say that the flowers have the purpose of increasing the likelihood of successful pollination, then we are ascribing a purpose which is a better fit to what the plant appears to be doing.  It is difficult to discuss biology without the use of teleological language, because the appearance of purposeful action is so common.


I have illustrated how widespread is our use of teleological language.  At the same time, I have suggested we often run into situations where a purely mechanistic explanation is unsatisfactory, often because there are chaotic aspects to the behavior we are describing.

January 3, 2011

Purpose (1) – Introduction

by Neil Rickert

In a recent blog post at BioLogos, Dave Ussery wrote:

I do believe that life’s history is infused with purpose and that this process is God’s process.

I agree with the first part of that.  The second part is why scientists often try to avoid talk about “purpose.”  The problem that science sees in the second part, is that it is an attempt to explain purpose that is not evidence based.  And yet people find “purpose” and other intentional words a very useful part of their vocabulary.

This post is intended to be the first of a series where I discuss “purpose” and similar intentional words, and where I attempt to provide a basis for intentional language that is entirely natural and is consistent with the scientist’s requirement of evidence.  If there are aspects of our use of purpose that are inherently mystical or religious, I won’t be attempting to deal with those.

Science usually attempts to give what we might consider to be mechanistic descriptions of what it is studying.  However, in our non-scientific lives, many of our descriptions are based on purpose, rather than on mechanism.  Take this blog, for example.  A mechanical description would be about the physical events that take place to cause a pattern of illuminated pixels on your screen.  That’s fine if we are interested in the physics of digital displays, but it is not satisfying if we are interested in what the blog is about, in what meaning it is attempting to convey.  To convey those meanings and interests, we need an account that uses our intentional vocabulary, our words that talk of purpose, goal, aim, intention, meaning, etc.

The attitude of many scientists with regard to purpose is well illustrated in a recent post to an internet forum:

Purpose is a human construct implying intent, which is another human construct. Science has nothing to say about purpose and intent. Once you’ve started talking about purpose you’ve left the realm of science.

I am sure that expresses a rather common view.  However, if to talk about purpose is to leave the realm of science, then study areas such as Psychology and Cognitive Science will have to be abandoned or at best to be recognized as being outside the realm of science.  I do no see any reason to exclude those fields from science.

There is an alternative to excluding purpose from science.  And that alternative is to find a natural basis for purpose such as will make purpose itself potentially something that science can study.  And that is the direction that I will be taking in this series of posts.


We use “purpose” and other intentional words in a number of different ways.  A brief discussion is appropriate.

I might say that the purpose of my automobile is to get me to work and then home again.  However, when I use “purpose” in that way, I am not really talking about the automobile having a purpose.  Rather, I am talking about me having a purpose, and using that automobile as part of how I fulfill that purpose.  So when I use “purpose” in that way, it really means the same as “use”, as in “I have a particular use for my automobile.”  We can discount that particular way of using “purpose” as not being about what we ordinarily understand as purpose.

The other extreme would be to limit the word “purpose” to the case where a person has a conscious purpose.  I want to avoid that particular usage, because it is rather difficult to characterize what we mean by “conscious.”  In between, there are many systems that exhibit what we might describe as “apparently purposeful behavior.”  And that is where I wish to target this series of posts.  That kind of behavior can be seen in many biological organisms, including those such as plants that we would never consider to be conscious.  But it is not restricted to biology.  We can describe a thermostatically controlled system as having apparently purposeful behavior.  Because of its relative simplicity, the thermostat is reasonably well understood, so I shall occasionally use it as an example in the series.  Here, and throughout this series, I shall sometimes use “thermostat” as shorthand for “thermostatically controlled system.”

September 16, 2010

How science works

by Neil Rickert

John Wilkins has asked about how scientists think.  This is intended, in part, as a response, though it also fits into my series on epistemology.

The received view of science is that it is based on analysis of facts.  But that is too simple.  We express our facts in terms of concepts.  An understanding of science has to begin by looking at concepts.  The history of the scientific investigation of electricity and magnetism illustrates this particularly well.  Today we may heavy use of electricity and electronics.  Many of the facts that we use are expressed in terms of voltage (or electromotive force), current, resistance, inductance, capacitance.  None of those concepts were in use at the time intensive investigation of electricity and magnetism began in the 18th century.  The electrical facts that we commonly use today were inexpressible at the start of that historic research program.  Coming up with a suitable conceptualization was an important part of that research.

The received view is that science is mainly concerned with discovering regularities in its data, and that scientific laws are presentations of those discovered regularities.  But that is not at all what I see.  Rather, I see a primary concern as one of finding ways to actually have facts (or symbolic representations).  Many of the scientific laws are abstracted from measuring conventions.  In some sense, the scientists are really solving the intentionality problem, the problem of having representations that are actually about something in the real world.  What is usually considered to be the intentionality problem is sometimes called “The Symbol Grounding Problem”, an expression used by Stevan Harnad in a 1990 paper.  Scientists, being the pragmatists that they are, go about this the other way.  They might be said to be solving the “symbolizing the ground” problem.  Instead of starting with symbols and asking how they can be about something, the scientists start with the something and come up with a systematic way of expressing factual information about that something.

The newly formed concepts that emerge from scientific research allow us to represent facts that could not previously have been expressed.  That is, the new concepts are the basis for a representation system, which we can think of as something like a coordinate system.  As mathematicians are well aware, when you are constructing a coordinate system there is some flexibility in how you do that.  Where possible, scientists use that flexibility to create a coordinate system that is mathematically nice, and this is part of why mathematics is so useful in the sciences.  Physics produces more sui generis concepts than do other sciences, so physics has more flexibility than other sciences in constructing their coordinate systems in a way that allow the use of mathematics.  In the other sciences, many of their new concepts are in some way derived from the more basic concepts of physics, so that those sciences have less flexibility to mathematically structure their systems of concepts.

In the universities, it is traditional that science majors are expected to take laboratory classes.  It is in these lab classes, that they master the intentionality of the scientific concepts.  That is to say, it is in these classes that they learn to connect their technical concepts to reality.

And that’s my heretical view on part of how science works.

July 28, 2010

Free will and randomness

by Neil Rickert

Free will and randomness

While we are on the “free will” topic, a few more thoughts.

Much of the belief that we don’t have free will stems from a view that the universe operates in a completely deterministic manner, so the future is already decided.  Personally, I have never seen any convincing evidence of determinism.  That the laws of physics are deterministic is not sufficient to conclude determinism.  Some of those laws are only approximate, and they might not be a complete set of laws.

One of the arguments for free will is based on the thesis of compatibilism, the view that free will is compatible with determinism.  It’s a clever argument, but many people do not find it persuasive.

In any case, the current evidence from physics, is against determinism.  Some phenomena, such as radioactive decay, seem to be random.  That shifts the argument over free will from one of whether determinism allows it, to one of whether randomness can account for the appearance of free will.

The skeptical response is that no, randomness does not have what is required.  When we talk of people having free will, we expect that to indicate that can determine their future actions.  That seems to require something like determinism and to be contrary to randomness.  A common comment is that random behavior is not what we mean by “free will.”

That skeptical reaction is, I think, mistaken.  It is true that by “free will” we do not mean random behavior.  However, randomness in the world need not result in random behavior.  If our response to randomness in the world is linear, then our response will be as random as the events that we are responding to.  However, there is the possibility of a non-linear response, perhaps a highly non-linear response.

Take yachting, as an example.  If the wind is blowing in the direction that the yachtsman wants to go, then he just sets his sails and the wind carries him along.  If the wind blows in the opposite direction, then a linear response would he to allow the wind to carry him away from his destination.  But that is not what the yachtsman does.  Instead, he reconfigures his sails for tacking.  He cannot sail directly into the wind, but he can adjust his sails for a direction across the wind and partially into the wind.  Then he switches to going across the wind in the other direction, but still partially into the wind.  This way he sails a zigzag course that takes him, overall, in the direction of his destination.

In electronics, we can use a diode to give a non-linear response to possibly random electrical currents.  In biology, we see semipermeable membranes which can presumably give non-linear responses to random variation in the chemical constituents of the environment.

My suggestion is that randomness could indeed be sufficient to account for our apparent free will.  It’s just that we have non-linear ways of responding to randomness, and we can use those to exploit the available randomness in order to achieve our goals.