Transcript Causality and information

History, Theory, and
Philosophy of Science
(In SMAC + RT)
7th smester -Fall 2005
Institute of Media Technology
and Engineering Science
Aalborg University Copenhagen
2nd Module
Causality and information
Luis E. Bruni
Causality as an ontological question
Arthur Peacocke (Chapter 2)
“…the succession of events which form causal chains is independent
of the choice of frame of reference and, indeed, the concept of
causality is affected by this initial theory of Einstein only to the extent
that we now have to recognize that causal influences can never be
transmitted through the universe at a speed greater than that of light”.
Why?
Think about the implications of these statement.
What is causality? What are our presuppositions about causality?
Causality
Causality  or Causation  a process linking two or more events or
states of affairs so that one brings about or produces the other.
One event is the cause of another if:

(a) the event occurs prior to the effect

(b) there is an invariant conjunction of the two events

(c) there is an underlying mechanism or physical structure attesting
to the necessity of the conjunction.
Since (c) is not always demonstrable in empirical data the requirement
may be replaced by tests assuring that no third variable controls both
or mediates between the two events. Without this weaker test, a cause
may be termed spurious and genuine otherwise.
Aristotelian causality
1) Material  the direct physical corelate.
2) Efficient  mechanical workings  the agent.
3) Formal  abstract forms towards which developing
entities naturally progress  the model.
4) Final  finality  intelligibility.
The hierarchical nature
of Aristotelian causality (I)
Ex: the causality of a battle.
Material causes  soldiers and guns  affect only a subfield of the
overall action.
Efficient causes  officers  their scale of involvement is most
commensurate with that of the battle itself.
Formal cause  the battle’s strategy.
Final cause  the reasons of the State  a head of state influences
events that extent well beyond the time and place of battle.
The hierarchical nature
of Aristotelian causality (II)
Ex: the causality of building a house.
Material causes  bricks, cement, tools.
Efficient causes  bricklayer.
Formal cause  the blueprirint  the architect’s idea.
Final cause  the family that lives in the house.
The hierarchical nature of Aristotelian causality  irreconcilable
with the Newtonian reductionistic and universal picture.
Deterministic causality
The systems’ behavior is specified without probabilities (other than
zero or one)  is predictable without uncertainty once the relevant
conditions are known.
Deterministic systems leave nothing to chance and are of necessity
lawful  there are no options.
Deterministic systems conform to the ideal of a machine in which wear
and tear, mechanical failures and unreliabilities are absent.
Modern computers are conceived as deterministic machines.
Newtonian systems
Five conceptual presuppositions of the Newtonian
approach  Newtonian systems are:
1) Deterministic  given the initial position of any entity in
the system, a set of forces operating on it, and stable
closure conditions  every subsequent position of each
particle or entity in the system is in principle specifiable
and predictable.
2) Closed  they admit of no outside influences other than
those prescribed as forces by Newton’s theory.
3) Reversible  the laws specifying motion can be calculated
in both temporal directions.
Newtonian systems
4) Atomistic (strongly decomposable)  reversibility
presupposes that larger units must be regarded as
decomposable aggregates of stable least units  that what
can be built up can be taken apart again  increments of
the variables of the theory can be measured by addition
and subtraction.
5) Universal  they apply everywhere, at all times, and over
all scales.
[Depew and Weber (1994), Ulanowicz (1997)]
Posibilistic causality
The systems’ behavior includes options without
specification of probabilities within that system.
In contrast to deterministic systems  possibilistic
systems leave some uncertainty in the specification of
future states and behavior, even if all relevant conditions
are known.
Possibilistic systems  also called non-deterministic.
Probability theory and statistics
Probability theory and statistics  had been created
expressly to circumvent an observer’s ignorance about
detailed events that everyone assumed were amenable to
classical deterministic mechanics.
The same mathematics could be applied to events that were
inherently stochastic (provide one accepts indeterminacy in
a world of Newton’s law).
Since the advent of quantum physics  growing
credibility has been accorded to indeterminacy over
ignorance as the proper object of statistical considerations.
The Laplaceam Demon
"We may regard the present state of the universe as the
effect of its past and the cause of its future. An intellect
which at a certain moment would know all forces that set
nature in motion, and all positions of all items of which
nature is composed, if this intellect were also vast enough
to submit these data to analysis, it would embrace in a
single formula the movements of the greatest bodies of the
universe and those of the tiniest atom; for such an intellect
nothing would be uncertain and the future just like the past
would be present before its eyes."
Pierre-Simon Laplace
Different kinds of systems
Deterministic systems.
Non-deterministic systems.
Stochastic systems  combined a random component with
a selective process.
Teleonomic  teleological systems  goal-oriented
behaviour  “self-organised” systems.
Predictibility
Social and cultural events  informational, semiotic, mental, or
cognitive processes  are rarely uni-causal phenomena and as
deterministic as in the natural sciences.
Causality in the social sciences therefore tends to be multi-causal and
probabilistic  as in information theory.
Predictability  the theoretical importance of causal explanations is
that one can apply them to explain what happened and predict what
will happen.
Their practical importance is that they lead one to produce or to
prevent causally related events by direct or indirect intervention.
What is it that probabilities measure?
There has been a shift in attitude  from “probabilities measure our
ignorance about a deterministic situation”  quantitative epistemology
 to “probabilities reflect an indeterminacy inherent in the process
itself  it bears also upon the ontological character of events.
This shift has not permeated Information Theory  the central concept
“uncertainty”  a state of knowledge, not a state of nature.
Information Theory  quantifies changes in probabilities.
Information  anything that causes a change in probability
assignment.
Shannon’s information
If an event changes its probability assignment from 50-50 to 70-30 
from more to less indeterminate  there is information  engendered
by whatever caused the bias (the change in probability assignment).
Remember  information = anything that causes
probability assignment.
a change in
Shannon’s formula may be useful for quantifying those factors that
help constrain flows along certain preferred pathways.
Whatever constraints partition the flows (of events) in the observed
proportions  they reduce the indeterminacy  they inform the
system.
Information  imparts order
We always begin work on a problem with some degree of uncertainty
 through repeated observations under different conditions we reduce
that uncertainty  gain information  however under all possible
circumstances a residual “uncertainty” will persist due to the inherent
indeterminacy in the process and its context.
The term “uncertainty” is frequently replaced by “indeterminacy”.
Information  refers to the effects of that which imparts order and
pattern to a system.
Uncertainty = information  has been very confusing.
The confusion comes from a failure to distinguish between
“information” and “information capacity”  Capacity of a system for
either information or indeterminacy  order or disorder.
Cybernetic information
Some think it always necessary to identify a
sender, a receiver, and a channel over which
information flows.
Information theory transcends communication 
probabilities are its fundamental elements.
What about the semantic value of information?
Material  Mechanical
Decartes  mechanical aspects of nature.
Thomas Hobbes  all reality is in essence material 
including God and the human soul.
End of XVII century  material + mechanical
We have on the one side material-mechanical causality.
… and beyond?
It is normally (but not universally) assumed that events at
any hierarchical level are contingent upon (but not
necessarily determined by) material elements at lower
levels.
What kind of causality is implied in informational,
semiotic, mental, or cognitive processes?  as in culture?
If we base culture on digital media  does that make
cultural processes more deterministic? Or more
probabilistic in the sense of Information Theory?
News of a difference
The smallest unit of information is a difference or
distinction, or news of a difference.
A sign  an idea  a complex aggregate of differences or
distinctions
More elaborate signs and ideas can be formed by complex
aggregates of differences  emerging codes.
Information vs. Impacts
Information  a difference that makes a difference to a
system capable of picking it up and reacting to it  for
there to be a “difference” - news of a distinction - there has
to be a biological system that senses it.
Otherwise they would not be differences, they would be
just impacts  think of a receptor.
So information means a difference that makes a difference
to some system with interpretative capacity
What is a sign?
A sign is something that stands for something to some
system with capacity for interpretation
Sign-vehicle
Object
Interpretant
Smoke
Fire
Let´s get
out of here
Differences and purpose
“The number of potential differences in our surroundings
... is infinite. Therefore, for differences to become
information they must first be selected ...” and categorised
by an interpretative system with such capability of pattern
recognition.
Differences are not intelligible in the absence of a purpose.
Informational, Semiotic, Mental
and Cognitive Processes
Two types of causal links
1) “pleroma” (Bateson)
• the world of non living billiard balls and galaxies
• the material world
• where forces and impacts are the “causes” of events
2) “creatura”
• the world of the living
• where distinctions are drawn and a difference can be a cause
• the equivalent of cause is information or a difference
Information is always contextual,
and context is always hierarchical.
History, Theory, and
Philosophy of Science
(In SMAC + RT)
7th smester -Fall 2005
Institute of Media Technology
and Engineering Science
Aalborg University Copenhagen
2nd Module
Causality and information
Luis E. Bruni