Information

1. Objective and Subjective Information
2. Information and Physics
3. Informational content and informational representation
4. Information, Perception and Causation
5. Information, data and knowledge

Objective and Subjective Information

      

Many people take the word 'information' in a mind-dependent 
sense, i.e in the sense of being informed of something, being 
made aware or cognisant of it of it. However, there is also a 
more objective, scientific view of information, which has its own 
discipline, infomation theory, and which is the sense of the word 
'information' we will be using. Information-theoretical information
 can be  marks on paper, bits and bytes in a computer, tones going 
down a phone line, and so on. It does not need to a human subject 
to be aware of it, nor does it need to have been created with the
 intention to communicate something.

The two conceptions of information can be related by the following 
formula: objective information is potential subjective information.
Objective information is a set of contrasts or differences between the 
properties of a material body or stream of energy, which could in 
principle be the object of study of a human subject, but which still
exists in the absence of a human cogniser. In terms of the old 
conundrum, a tree falling in the forest makes a sound (ie it creates  
sound waves in the air, which carry objective information), but no-one
hears it (the objective information does not enter anyones mind). 
Objective information is really just a recognition of the fact 
that the universe is not a featureless, blank expanse.

Although we have defined OI in terms of SI, that does not mean OI
is independent of SI, and we can even suppose if we wish that
SI is dependent on OI -- that is that SI is just OI which happens
to reside in the human brain.

Information and Physics

  
Is OI a physical quantity ? is it like mass, charge or velocity ? 
It is, in a way more fundamental than that. A difference in charge 
is one kind of difference, a difference in mass another, but all 
differences contain information. On the other hand, without physical 
propertiesthere could be no way for information to be embodied. The 
amount of information a physical system contains is a charateristic 
of the system, but not an independent physical property of the system. 
When we said 'information is marks on paper..' that was not quite right.
It is not the marks themselves, but the difference they make
that is the information.

Information is dependent on the physcial in that it needs some material
or energetic substrate or carrier to embody it -- it is not free-floating.
It is independent in the sense that it does not require any particular
embodiment. The complete works of Shakespeare can exist on paper,
CD-ROM or in the human memory.

Information supervenes on physical properties. Material bodies which are identical will carry identical 
information; changes in information require changes in its physical base; different physical systems do not 
necessarily contain different information.

Informational content and informational representation

Let us dwell on the point that different physical systems do not 
necessarily contain different information.

The strings:-

OLDLANDOFMYFATHERS #1
PMEMBOEPGNZGBUIFST #2
encode the same information, providing
you know that every letter in the second string is the
alphabetical succesor of the corresponding letter in the first string.

Of course, any number of strings can be used to encode the same
information in any number of ways. The string:-

HENWLADFYNHADAU

encodes it using the complicated system of rules known as 
Welsh. String #1 is not the meaning or ocntent or message of string #2
any more than string #3. The content is present in all three 
strings, but is more apparent to most readers in string #1. 
There is always a duality
between informational content and informational form. 
Content always has some form, and form some physical realisation.


Now, the constancy of content accross representations 
depends on rules for transforming one representations
into another, such as the 'add 1 to each letter' rule.
Given a rule, we can generate representation after 
representation of the same content.

Alternatively, given two representations of similar
information capacity we can invent a rule that links
them. 

The first case rules --> representations is the kind of 
situation we have in speech, game-playing and code-breaking.
The second is that of science and code-breaking.

It has already been established that any two representations
can be linked by a concocted rule. So, given two or more strings,
how do we know whether they are linked by a concocted rule or a real
one ? The surest sign is that the the rule is a relatively simple
one thatlinmks a relatively large number of cases, like the rule
that maps string #1 onto string #2. 
Without such limit on the complexity of a rule, no such distinction
can be made. Pseudo-cryptography such as Baconology or the 
'Bible Codes' are expamples of what happens when an attempt
is made to decode a message with rules of arbitrary complexity:
any 'hidden' message whatsoever can be decoded.

The example of science has been given. In science, our informational 
forms are are sets of source-data, and our rules showing how one set
of data evolves into another over the course of time are physical
laws (or rather proposed theories about physical laws). 

The purely informational law of minimal-rule-explaining-maximal-relation is clearly equivalent to the features of citeria of 
minimum complexity (Occam's razor) and maximum generality used as criteria for choosing
scientific explanations.

Information, Perception and Causation

Now, so far we have 3 kinds of rules:-
1. Human-made rules (languages, codes games).
2. Laws of the universe.
3. Theories regarding (2).

Turning to (3), we started by noting that 
informational rules can be seen as causal relations.
We can also see look at cause-effect pairs as information transfers.
This is useful in understanding the operation of the senses, for example.
Since there is a causal chain involved in vision -- photons strike a
surface, are selectively reflected and absorbed, impinge on the retina, 
set in motion a chain of neural firings -- there is automatically a transfer
of information form place to place (and a transformation of
informational content from one form to another).

Since the concept of information we are using is not 'loaded'
with either a physical or mental interpretation, many of the mind-body
dichotomies of traditional philsophy vanish.
Is the world as we see it? Yes, because the content of our inner 
representations (the form the information content has in our brains)
originated, and the effect of the transformational rules tends to 
preserve the content. No, becuase the way it is represented in the
brain is nto the way it is represented in the external world.
The reflectance characteristics of the surfaces of objects are not
the same as our red-green-blue colour vision -- 
but the latter can represent the former.

Perception is direct (because the content is the same allong the causal 
chain), indirect (becuase the form varies), causal (because case-effect 
pairs and information transfers are the same thing) and representational
(because information content is always in some specific form).

Now, we have said that causal relations are information 
transfer. Causal interactions are also said to be transfes
of energy. To complete the triangle, information transfer would
have to be energy transfers. Is that the case ? 
It seems to be. High-frequency photons have more
energy than low-frequency ones, and more information-carrying capacity
as well. If we continue in the vein if taking information
as basic, we would define energy as infomration-carrying capacity,
bandwidth.
Given that the amount of energy at a given place and time are always 
limited, then the amount of information is that can be 'read out'
of a physical situation is limited too. We cannot measure physical
situations to an endless degee of accuracy. There is,in short,
an uncertainty principle.

Peter D. Jones 21/01/02

Information, data and knowledge