This is Lecture Eight
Causality and Logic in Thinking
A flashback from the Syllabus:
Causality and Logic in Thinking
mental models incorporate tacit knowledge; mental causation goes without
reasoning and inference;
unlimited inconsistency tolerance in the mind; logic as an emerging feature
in mental mechanisms
1. Tacit Knowledge
1.1. What is Tacit Knowledge
a famous concep,t which is everyone's favorite
it's a weak concept as such, but has a strong expressive power
generally known after Kuhn (1062) cited Polanyi (1958); several rounds
of citation since then (most recently Nonaka et al).
reference: POLANYI M (1958) Personal Knowledge: towards a post-critical
philosophy London: Routledge and Kegan Paul
Polanyi, Michael (1891-1976)
"Polanyi`s concept of knowledge is based on three main theses: First, true
discovery, cannot be accounted for by a set of articulated rules or algorithms.
Second, knowledge is public and also to a very great extent personal (i.e.
it is constructed by humans and therefore contains emotions, "passion".).
Third, the knowledge that underlies the explicit knowledge is more fundamental;
all knowledge is either tacit or rooted in tacit knowledge.
"Both Quantum Mechanics and the theory of relativity are very difficult
to understand; it takes only a few minutes to memorize the facts accounted
for by relativity, but years of study may not suffice to master the theory
and see these facts in its context. At all (these) points the act of knowing
includes an appraisal; and this personal coefficient, which shapes all factual
knowledge, bridges in doing so the disjunction between subjectivity and
objectivity."
New experiences are always assimilated through the concepts that the individual
disposes and which the individual has inherited from other users of the
language. Those concepts are tacitly based. All our knowledge therefore
rests in a tacit dimension."
"Tacit knowledge" has been all but hi-jacked by management gurus, who use
it to refer to the stock of expertise within an organisation which is not
written down or even formally expressed, but may nevertheless be essential
to its effective operation.
Originally, Polanyi's interest was in the kind of knowledge which we routinely
use and take for granted, such as the ability to recognise the face of a
friend: it is irreducible to explicit propositional knowledge and cannot
be articulated. It cannot therefore be taught, although of course there is
obvious evidence that it can be learned or acquired."
1.2. An introduction to Polanyi:
Tsoukas, H.: Do We Really Understand Tacit Knowledge? http://is.lse.ac.uk/events/esrcseminars/tsoukas.pdf
tacit knowledge is widely misunderstood as a particular
kind of knowledge
cf. Eliasmisth
in Internet Encyclopedia: tacit vs explicit is
like "knowing how" and "knowing that" (after Ryle)
http://www.artsci.wustl.edu/~philos/MindDict/tacitknowledge.html
"One of the most distinguishing features of Polanyi’s work is his insistence
on overcoming
well established dichotomies such as theoretical vs. practical knowledge,
sciences vs. the
humanities or, to put it differently, his determination to show the common
structure
underlying all kinds of knowledge. Polanyi, a chemist turned philosopher,
was categorical
that all knowing involves skillful action and that the knower necessarily
participates in all acts
of understanding. For him the idea that there is such a thing as “objective”
knowledge, selfcontained,
detached, and independent of human action, was wrong and pernicious. “All
knowing”, he insists, “is personal knowing – participation through indwelling”
(Polanyi and
Prosch, 1975:44; italics in the original)."
“the aim of a skilful performance is achieved by the observance of a set
of rules which are not known as
such to the person following them”
"Skills retain an element of opacity and unspecificity; they cannot be
fully accounted for in terms of their particulars,
since their practitioners do not ordinarily know what those particulars
are; even when they do know
them, as for example in the case of topographic anatomy, they do not know
how to integrate
them (Polanyi, 1962: 88-90). It is one thing to learn a list of bones,
arteries, nerves and
viscara and quite another to know how precisely they are intertwined inside
the body (op.cit., p.89)
"How then do individuals know how to exercise their skills? In a sense
they don’t."
---> remember
this later!!
Polanyi (1969:147) remarks, “the way the body participates in the
act of perception can be generalized further to
include the bodily roots of all knowledge and thought. […] Parts of our
body serve as tools for observing objects
outside and for manipulating them”.
Polanyi, M. (1969) Knowing and
Being, Edited By M. Grene, Chicago: The University of Chicago Press
1.3. How to interpret tacit knowledge
It is totally misconstrued to think of tacit knowledge as something that
"could" be made
explicit; as knowledge waiting for conversion or discovery.
It is more correct (ie closed to Polanyi's and later Kuhn's version) to
think of tacit knowledge
as having to do with the condtions of knowledge - many of them biological,
some social,
and several learned (as e.g. learning a profession).
As it were, many of the conditions of knowledge are themselves knowledge-like.
1.4. A critical assessment of Polanyi
Polanyi mixes several things, and his followers mix up even more.
Reading the texts, it becomes obvious that implicit and tacit are different
components.
Skills are mostly implicit but maybe not tacit.
The tacit dimension however includes skills.
What is the difference?
Rriding a bycicle is a skill; knowing chemistry is (based on) tacit knowledge.
Skill is low-level, tacit knowledge is high-level.
Both are effortless, unmonitored (i.e. non-focal), etc.
In terms of my own terminology, skill is (closer to) the motoric part of
an episode;
whereas tacit knowledge refers to (some part of) a mental
model.
Their separation is a incomplete, because mental models coexist with the
episode
which inlcudes motor parts etc. Yet it is clear that
skill-like and tacit elements are
often different.
In some other aspect, Polanyi's concern is Kantian.
"The way we look cannot be part of what we see". - this is Kant's enigma
of the a priori.
(For, to be able to see at all, one needs something
beforehand, a way to seeing)
Tacit knowledge includes factors hidden behind the open stage performance
of explicit reasoning, utterances, and judgments. In
this regard, the notion of tacit
knowledge imnplies a criticism of the "public-process"
view of the mind, which
focuses on the result of cognition, rather than the
process itself. Exactly as in Kant -
where perception is the end of a story, not the beginning;
it is the output of a complicated
system of interactions, rather than raw input for the
mind.
2. Mental Models Support Tacit Knowledge
2.1. The Structure of Mental Models
There is a deductively available structure in mental models.
In other words, once we have accepted the notion of mental model as outlined
earlier, there will be several consequences: these bones
show the place where the
flesh should go, and in this way we can deductively
clothe up the entire skeleton.
One immedate consequence is the multi-layered nature of mental models by
which we
can intepret logical properties of cognition as well
as phenomena of tacit knowledge and
much else.
We are now going to discuss the relationship between explicit and implicit
parts of
mental models, but need to take a byway first.
2.2. The Nature of Consciousness
and what it bears for mental models
etc.
Consciousness is, in general terms, somebody else's problem, but there
are some disturbing facts.
There is a slight interference between the consciousness problem (1st person)
and the scientific
problem of the mind (3rd person).
For instance, experience is the basis of embodied concepts - and experience
is 1st person.
Then, consciousness is usually associated with focal attention and awareness,
and hence it
sneaks into the discussion when talking about the relationship
between overt (explicit) and
covert (tacit) knowledge, which is the question we will
consider in the context of mental models.
So, some reflection is necessary. Here is a brief outline of a view of
consciousness compatible with mental models.
Take it as it is - as a speculative, superficial account provided just
for the sake of clarity.
- Consciusness may be like a flashlight that points
at certain things in the mind and not others;
the place where it points changes
dynamically.
- Mental models operate by their own causal powers;
ie. consciousness has no specific causal
effect. (The question whether it
has other, non-specific effects, such as increasing mental activity
or activating certain mechanisms
in the mind, is left open.)
- Conscious attention concerns not entire mental models
but their selected aspects or attributes
(such as propositions associated
with them - hence the linguistic experience).
- Attributes made conscious serve like handles or hooks
by which entire mental models are
activated. (Remember that we did
not decide whether activation is by consiousness or otherwise)
The situation should be familiar from artificial intelligence and applied
computer science.
"Objects" and "frames" contain structured information in "slots" and variables;
instantiation of a variable or a slot
means instantiating an entire class to which the object
belongs.
"Underlying the majority of these is the concept of object-orientation,
namely the recognition that the decoupling of data and the code
that acts upon them, is based on an artificial distinction, and that models
which combine the data and code into distinct "objects" offer
both more intuitive and a functionally richer conceptual entities. This
paradigm shift can be particularly seen in three areas: programming
languages, databases and user-interfaces. Within programming, the
evolution has been from procedural based languages where the code
was encapsulated within procedures and kept separate from the data, to
object-oriented languages where the program is built up of self
contained "objects" which encapsulate both the data and the actions of
the items being modeled."
It can be no accident that the pyschological theories of remembering have
long given up the videotape or storage room
notion of memory and have focussed, since F. Bartlett (1932), on "schemes",
or more recently, "scripts" etc.
All that this model of consciousnes implies is that the relevant mental
unit of processing is not some selected element of focal
attention but an entire mental model - an entire "scheme", or "script",
if you wish.
It might be frightening that "someone else" is doing the work inside, and
"we are not masters in our mind" - but that's not a
scientific argument. And it can be reversed - it is
at the same time relaxing to see that the system does not depend
on willful operations in a fragile way (I, for one,
could not sleep if this was the case...)
2.3. How Mental Models Support Tacit Knowledge
A mental model is like an iceberg, the largest part of it is "invisible".
A mental model is a complex material entity that models actors and constraints
of an episode; such a
mental model can be activated by any element of the
episode or any reference to the actor(s).
An active mental model is identified for (and by) the self by means of
some (few) marked attributes, which are related to
the activation enforcing elements.
Consequently, most (in fact almost all) of the mental model is unavailable
for inspection and introspection, most of the time.
Mental models are ineffable.
In other words, mental models are "deep";
there is always still more in them to get out.
This parallels the notion of causal depth and underlines our earler
remark that mental models are material systems.
Explicit knowledge is transparent ("it can be overviewed"); therefore it
is tempting to believe that transparency must
be a property of any knowledge.
- see the old alliance of transparency with rationalism
and enlightenment
- see the misinterpretation of Polanyi in knowledge
management; where "tacit" means "waiting to be discovered"
Is now a "deep" mental model indeed knowledge, if it's not even
available for operation?
The answer is yes.
Mental operation (as we remarked re consciousness) is
autonomous, and does not depend on some transparent,
directly experienced, overt mental
entities and their conscious manipulation by means of a central will.
If that picture is rejected (as the notion of mental
model quasi forces us to reject it), the "availability for operation"
also obtains a new form. What is
not available for an overt and conscious operation may still be available
for
operations of a more fundamental
kind.
2.4. Thinking and Mental Processes
Mental processes are
- not public (like the functioning
of the heart ---> why should the mind be any different? just because
of
consciousness? that's cheating - all the evidence is otherwise, and now
we use the royal road to have
direct access? That can be - and is in fact - a very mischievous strategy)
- not transparent (contra Chomsky
and the entire tradition of "mind is thought is language")
- requires no effort - it just happens
(Neural Networks and "relaxation" go in this direction!)
3. Logic and the Mind
3.1. Causality and Reasoning
Rather than being based on reasoning and inference, mental processes are
autonomous and causal.
Mental models possess causal powers and work spontaneously.
Thinking is not the world of logical consequence but that of causal effect.
In particular, results of thinking do not depend on some isolated properties
alone, such as the truth of
propositions.
It is a question, therefore, whether - and how - logical properties are
maintained in thought.
3.2. Logical Properties
3.2.1 Deduction Follows Pre-Established Routes of Causality
it is wrong to look at deduction as a way to reach a conclusion
inference is a readout of a completed mechanism
deduction, in fact, is just public a posteriori justification of
something that has already happened inside
deduction of this kind is not independent of "semantic" properties -
in the world of mental models (in the mind, in language
etc) NOTHING ever follows from P (a proposition);
logical consequence is based on what P is.
Is there no room for formal logic then? Is there no domain-independence?
How come that logic is so effective?
3.2.2. The Laws of Thought
Boole, De Morgan etc. 19. Century: truth tables and connectives; the birth
of formal logic (Lullus' program completed)
truth tables are (mostly) valid but irrelevant (-----> cf. Johnson-Laird:
in my words, the solution of a logic puzzle is the
building of a narrative - of what happened -
and not a truth table)
(several other remarks here. "logic" never goes without verb!! which
makes it
suitable for telling a story. Te copula - is - is a special
verb, difficult!!)
-----> empirical tests? just replace "runs faster
than" by "is", "is not"
irrelevant because they are just summaries
"connectives" (and, if, not, or) do connect - for the mental mechanisms,
they are like flow pipes, or the wires in a network
that transfer activity; in fact they describe the
routes by which the activity can flow. They are like (meta) constraints
in
a mechanism consisting of elementary mechanisms linked
by the connectives.
3.2.3. Logic as an Emerging Feature
Sometimes (or often - depending on ceteris paribus conditions) activities
follow identical patterns under identical constraints.
Combinations that tend to follow the same identical pattern can be recognized
as having an "abstract" validity.
This means that a (partial) generalization accross domains is possible.
Here is the double nature of logic:
writing logic is biulding a machine;
so we can know (ie explain, predict) how works
but thinking means using this machine
(as a flow of activities, a causal mechanism) rather than reasoning about
it
4. Inconsistency and the Mind
Inconsistency? What inconsistency?
-- Logician caught
by his students trying to prove not A from A.
Logic produces consistency; the mind can use systems recognized in logic
as formal systems but there
is more to the mind than just that.
The mind can also produce incosistency.
Much as mental models are mostly tacit, mental processing is mostly inconsistent.
- we know this as a phenomenon
- accordingly, there are several attempts do deal with inconsistency
paraconsisstent logic (Polish logic
in particular)
non-monotonous logic (e.g. changing
set of propositons, such as in time) - McDermott
- but mostly the concern is with consistency; inconsistency is considered
as a deviation,
to be repaired (n-m-n) or accommodated
(p-c-l).
Here the situation is different.
the basic situation of a system of metnal models is
that of arbitrary inconsistency
imagine the mind as a multi-agent system, each making
truth-claims
Why pursue consistency at all?
I believe this has to do with the extension of mental
models towards universality.
Linkage of mental models with "and"?
From sitation dependent to situation independent knowledge
Matching the pieces into a larger picture.
Consistent subset = intensively used subset
Is consistency only an artifact in language? Probably yes....?
I don't like to hear that... Sounds plausible though.
This requires further research.