Chapter 12: Integrating Action and Though

From: Harnad, Stevan (
Date: Tue May 20 1997 - 19:05:30 BST

Chapter 12: Control of Thought and Action

Karl Lashley (1951), one of the forefathers of contemporary cognitive
neuropsychology, wrote a famous paper about the problem of serial order
in behaviour. This problem may not sound like a very cognitive one, but
it turns out to be at the core of many aspects of cognition.

We can have an inkling (only an inkling) about what it's like to have
various kinds of brain damage from things that happen to us even if our
brain is intact: The tip-of-the-tongue phenomenon gives us a glimpse of
what it's like to have anomia (inability to remember names); the
occasional trip-up with our tongues gives a taste of dysarthria
(problems controlling speech); not remembering a film or an event
is like an episodic memory deficit; all forms of forgetting are mild
hints about amnesia; clumsiness is a taste of motor disorder, and so on.

This Chapter is concerned with movement at a higher level than muscle
control. In all the cases discussed here there is nothing wrong with
movement per se; there is a problem with how movements are put together.

All of us have had the following experiences now and then:

Doing something with one object that we should have done with another,
either because we are no paying attention, or because we are rushing
under stress: binning a letter instead of the envelope, fetching a knife
when we needed a fork, etc.

Sometimes we fetch the right thing but put it in the wrong place, or we
leave out an important step in a sequence, or we do something with
an object when we had not meant to, such as turning off the light while
there is still someone in the room: This is called "capture" or
"utilisation" behaviour. It occurs when you absent-mindedly do the
accustomed thing with an object, like a light-switch, even though you
had not intended to. The object "captures" your behaviour by just being
there. When we are attentive and calm, this kind of automatic response
is inhibited, but when we are tired, or thinking of something else, or
stressed, the inhibition may weaken, and we do it.

"Perseveration" is another kind of lapse we sometimes get into: We may
repeat a repetitious action more often than we had intended to, as in
putting more spoonfuls of sugar into our coffee than we had meant to.
To show that perseveration need not happen only when you are distracted,
think of when you have lost something, and keep coming back to the same
place in search of it even though you have searched it several times
already; or if you're working on a problem and you keep trying the same
strategy even though it has not worked.

It turns out that all of these normal lapses have their much more
severe counterparts in brain damage. The general category of
neuropsychological deficit that we are concerned with here is related
to the "apraxias" (disorders of voluntary movement [rather than of
automatic reflexes]). The tendency to do the wrong thing with objects
(such as trying to smoke a spoon) or leaving out important steps in
a sequence (such as extinguishing a match after lighting a cigarette)
is called "ideational apraxia." And just as it was important to note
that visual agnosia (not recognising things you see) is only agnosia
when the patient can see (i.e., is not blind or otherwise visually
handicapped), ideational apraxia applies only to people who can move
perfectly well, so the deficit is in the ORGANISATION of movements
rather than in making them.

Similarly, if someone smokes a spoon because they mistake it for a
cigarette, then that is not an ideational apraxia, but a visual agnosia.
If asked, they would say they had not meant to pick up the spoon and
smoke; it happened automatically, and they know perfectly well that it
was a spoon, and that they had made a mistake; it's just that they keep
making such mistakes.

"Perseveration" -- doing the same thing or using the same
(unsuccessful) strategy over and over again -- is another problem in
the organisation of actions. Some forms of perseveration are caused by
injury in the frontal lobes of the cerebral cortex, areas that are
concerned with, roughly speaking, executing planned actions.

"Capture Behaviour" or "Utilisation Behaviour" occurs when a stimulus
triggers a response that would be normal if the patient had meant to do
it, but the patient did not mean to do it. Such patients, when they see
a glass of water, reach over and drink it, even though they are not
thirsty and didn't mean to drink then. They just do what they are
accustomed to doing with objects -- again, a tendency that is normally
inhibited in the brain, so that when we DO want to do something, the
inhibition is released. In these kinds of patients (again with injury to
their frontal lobes) the inhibition is not there, so the stimuli
"capture" the actions that are normally performed only voluntarily.

This kind of "stimulus-bound" behaviour also has counterparts in normal
behaviour: It is thought that certain forms of obesity are caused by
"stimulus-bound" eating: Such people we overeat by huge amounts as long
as they are in the presence of food; they can't seem to inhibit the
behaviour that food "captures." But if they are not around food -- can
neither see or smell it -- they can actually go for food much longer
than normal eaters. The hypothesis was that their eating, instead of
being regulated by the internal symptoms of appetite and satiation, is
instead captured by the sight and smell of food.

One of the ways that theories are tested and diagnoses are made in
clinical neuropsychology is through the study of "dissociations" and
"double dissociations." A single dissociation occurs if a brain
injury causes a deficit in one capacity, but not in a related one:
Say, an agnosia that causes difficulty in recognising faces
(prosopagnosia) but no difficulty in recognising other kinds of things.
This means that not the same mechanism underlies object recognition and
face recognition.

An even stronger conclusions can be drawn from "double dissociation,"
which occurs when injury in one area causes one deficit -- say, face
recognition -- and not a second -- object recognition -- whereas injury
in another area causes a deficit in the second but not the first. This
is strong evidence for some independence in underlying function, but
not yet enough evidence to conclude that two completely independent
"modules" are involved. (The criteria for concluding that independent
modules exist are discussed in Chapter 3.)

Some hard to imagine dissociations are: "alexia without agraphia"
(loss of the ability to read, but sparing of the ability to write),
"capgras syndrome" (loss of the ability to recognise the "essence" of
familiar people with sparing of the ability to identify their
exact imposters"), "discopia without disgraphia", and of course the
strangest of all dissociations: the split brain.

A good concept to have in mind when you consider disorders in the
integration of thought and action is that of a "motor programme."
If you recall ethology (the study of behaviour in its biological
context), Lorenz & Tinbergen had the concept of the "fixed action
pattern" -- a pattern of movements that the animal executes in
full. Such action patterns are seen in courtship behaviours in animals;
the way birds build nests and carnivores capture prey are all examples
of fixed action patterns. The evidence for them has been strengthened by
work in neuroethology, where the entire nerve circuit that generates
certain behaviour patterns -- from simple reflexes to complicated
species-specific behaviour has been found. The human counterparts
of these fixed action patterns are the action sequences -- some
inborn, like crying or nursing, and some learnt, as in tying your
shoes or taking of your coat -- that are likewise based on motor
programmes that need to be integrated with our conscious thoughts and
goals when we act in the world.

These motor programmes, or at least the ones that have been learned,
are first very deliberate and need a lot of attention to get them right.
But once they have been learned -- and overlearned, to the point
where they are error free and you don't need to think about them any
more -- they become automatic. The process by which deliberate conscious
actions such as driving turn into familiar tasks that we no longer have
to think about to perform is analogous to the process by which short
term memories -- which first have to be rehearsed -- become long-term
memories, retrievable with no effort at all. But if there is one form of
memory that these motor programmes resemble the most, it is procedural

Unlike basic motor programmes, which tend to be fixed, automatic and
stereotyped, voluntary sequences of action, at least new ones that have
not yet been overlearned, require attention and thought, taking into
account the goals for the action, and how to achieve those goals.

Two models for the organisation of behaviour are described in this
chapter: Norman & Shallice's model is based on neuropsychological
evidence, and is meant to explain the difference between routine,
automatic behaviour and new, nonroutine behaviour that requires attention
and conscious control. The model accounts for a lot of the findings
with movement control disorder, especially of the routine kind, but
to handle more complicated, nonroutine behaviour a higher cognitive
model is required, one that an deal with all problems we confront, and
not just the action-sequencing ones. Newell's SOAR model is a candidate
for this "unified theory of cognition." The model is formulated at a
much more abstract level than motor action. It is a "production system,"
where "productions" can be physical movements or more abstract
"actions", such as chess moves, or logical inferences.

The general principle underlying SOAR is that there is an initial state,
plus a goal state, and the task of the model is to reduce the distance
between the current state and the goal state until the goal is reached.

Newell & Simon were the inventors of the first chess-playing computer
programmes. They have also written programmes that can prove
mathematical theorems and discover scientific laws. SOAR, however, is a
symbol system, and that means it has the familiar weak spot of all
symbol systems: the symbol grounding problem and the frame problem. The
chapter puts Norman and Shallice's neuropsychology-based model
alongside Newell's SOAR model, and suggests that these two models, one
bottom-up from sensorimotor behaviour and the other topdown from
knowledge and thought, into one unified theory of cognition.

For the details of the progress toward this goal, you will have to wait
for your 2nd and 3rd year courses.

Lashley, K.S., (1951). The problem of serial order in behavior. In
L.A. Jeffress (Ed.), Cerebral mechanisms in behavior. New York: Wiley.

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:52 GMT