On Feb 14, 9:08 pm, "Publius"
nospam.comcast.net> wrote:>>> It is only a verbal quibble, but it has philosophical implications: it
>>> tempts us to imagine that information can exist without a recipient or
>>> a context. But without those all strings of bits are just random
>>> strings, no matter how interesting or complex their internal structure.
>> but all communication is just bits of this information
>
>> from our very first experience of the world
>> to our most elaborate expressions of meaning
>
>> all of it is just this neutral mathematical information
>
>> there is no context outside syntax
>
> That is inconsistent, g. If it relates somehow to our experience of the
> world, then it is no longer neutral. The world (as experienced by us)
> supplies the context. What were previously random bits with intriguing
> patterns are now signals with information content. We can relate that string
> of bits to features of our experience.
all of it is symbols
whether you are looking at words on a screen
or rocks in a field
our perception is symbolic
at all stages of perception
our way of understanding the world
is symbolic abstraction
when we relate hearing symbols
with olfactory symbols
there is meaning composed
when we relate visual recognitions of natural scenes
and visual recognition of self-drawn designs
meaning is composed
but in the latter we can manipulate the meaning
>> when someone tries to teach a refferent
>> all they really do is present two symbols
>> in a relation that they want understand as meaning "referrent"
>
>> pointing to or something similar
>
> Yes. And that pointing is what makes that symbol a carrier of information.
> It doesn't carry any without it.
it too is a symbol
"fingerpointing"
it is not something an infant is born knowing
it is a syntactic context
between geometric direction and object location
fingerpointing, object, and symbol
are arranged in the syntax of our visual symbols
and we learn an association of object and symbol
of reference
even pan troglodytes can learn this syntax
anim cogn ( 2002 ) i5 p107-14
>> communication already assumes the ontology of a channel
>> ie. two endpoints and a transport of mathematical information
>
> Yes. And information theorists don't pay enough attention to what happens at
> the two endpoints (though that really isn't their job). It is encoded on one
> end and decoded on the other. What is encoded is some complex state of
> affairs, as experienced by someone. That code is then sent over the channel,
> where it is decoded. The recipient then has information about the state of
> affairs experienced by the sender, without having to experience it himself.
> The coding scheme is entirely arbitrary, and need only be agreed upon by
> sender and recipient for the communication to work.
i think it is all one theory
information can be transmitted along channels
we can build networks of channels
and model information dynamics across the graph
the vertices of the network
can be given computational complexity
which transforms its inputs to outputs
through possibly stateful transitions
if these computational nodes
are only equipped with learning algorithms
for the most basic and biologically motivated
A = B
one symbol is associated another
then protocols may form semantics
intentions can grow into language
and network patterns can form
larger scale structures through language
~~***********..&&&
i was probably wrong to choose your martian tablets
as an exemplar
without anything but planetary context
the tablets can hold many secrets
i know this is what you are pressing strongly on
and you are right
we don't even know if we are extracting the right information
or if there are infinitessimal grooves
or patterns in magnetic doping
spectral reflectance
or any new and strange technology of communication
or even if they are footprints in the mud
from birdlike ancient motorlife
here the information content
still increases our knowledge
because it tells us a certain level of complexity
existed once
if we saw them
and were able to rule out simpler known forces
we could propose theories of complex dynamics
try to predict were to find more
with even this type of information
science and models are possible
but i did intend more
what i really wanted to show
was that meaning really is syntactically contextual
and that meaning can be learned
only if understood compositionally
but of course that understanding can be coded
to the complexity reachable by the nodes
and made uncompositional
in transmission
as long as a translation occurs to make it compositional again...
>> now if all context is syntactic
>> Ra'a' for instance
>
>> how can we learn to predict meaning
>> without their being information in your sense?
>
>> when do we exceed mere strings of symbols?
>
> Why do you claim that all context is syntactic? We assume the sender and
> recipient share a world and a coding scheme (arrived at by pointing, in one
> way or another). Is that syntactic?
syntax is merely relations of symbols in the sensespace
visual syntax consists of visual connectives
strings of symbols
graphs of symbols
visuogeometric relations of symbols
and their temporal evolution
auditory syntax
is separation of sources
and their temporal evolution
hunger syntax
might only be temporal
or may consist of distinct natures
auidovisual syntax
relates auditory symbols
to visual symbols
in the collective experience processing
and so on
speech is audio-related
perhaps to visual or other experience
writing is visual-related
...
>>> I agree. There are infinitely many possible interpretations of those
>>> tablets.
>
>> hey!
>
>> you gave me 10^80 (or 10^80! or something similar)
>
> Heh. That was just a throwaway concession. The number of possible
> interpretations is infinite. But for any given interpretation, we only have
> 2^(10^80) ways to encode it.
>
> The number of bits or complexity of the message doesn't predict the number
> of bits required to encode that message in some other way, or the complexity
> of the referent set (provided we have some means other than bits to measure
> it). Complex messages can be encoded with but a few bits. Think of Paul
> Revere's message:
>
> 01 = "The British are arriving by land."
>
> 10 = "The British are arriving by sea."
>
> Think you could extract that information from "01"?
no
those are very complex referrents
on a low complexity signal
because that is a ridiculously small amount of information
so very little science is possible
i concede this point
and thank you for forcing me to recognise
something i do think is crucial here
>> complexity arguments conquer arbitrariness to infinity
>> if we exist in a finitistic universe
>> (or if it has at least a finitistically observed past)
>
> Only the physical universe is finite. So we have only a finite number of
> ways to physically encode a message. But the universe of referents, which
> includes imaginary and hypothetical worlds as well as the physical world,
> seems to be infinite. And any of those finite strings can be mapped to any
> of those possible referents.
i challenge this
because i don't know what it means otherwise
here is the situation i picture:
a string is a communication
if it is meant to convey meaning
both parties must agree on how to interpret
to negotiate this interpretation also requires symbols
and only symbols in my understanding
because all of our sensation is symbols
in my understanding
so something has to have been pattern recognised commonly
some type of symbolic connection made
and the number of these
whether constructed in the imagination of symbolic connectives
or developed in the natural symbols of experience
is still finite
it is potentially infinite
in the future sense of something that
if continued
could continue
but the past of an individual is finite
and this is what matters here
>>> Someone mentioned that there is a bit of common context, i.e., that we
>>> and the Martians inhabit a common solar system. But we cannot know
>>> whether the Martians perceived the solar system in the same way we do,
>>> or classified their percepts the same way. Nor do we know whether there
>>> is any astronomical material in the text at all.
>
>> we have much more common context
>
>> every symbol is a record of an event
>
>> this includes every piece of scientific data
>> we can extract from study of the planet
>
> It is a record of something. Not necessarily an event. And not necessarily a
> spatiotemporal event for which we might recover a clue from the Martian
> rocks.
but this is how all science progresses
from the assumption
that a model might predict
from whatever information present
these are events in that sense
>> we cannot guarantee any other individual
>> perceives the world the way we do
>> but through complexity and compositionality
>> we do come to learn to represent
>> the same symbol-referrent relationship over time
>
> Only if we are in an ongoing dialogue.
>
>> the corresponding inverse problem
>> appears isomorphic to science in general
>
> We are immersed in a sea of data, at one end of a communication channel. We
> have no access to the other (sending) end, and have no key, agreed to
> between the sender and us, which can unambiguously decode the data received.
> So we construct a model universe consistent with the data, and which allows
> us to predict some future data. The more successful those predictions, the
> more confidence we have in our model. But that data set, like any other,
> could encode infinitely many "realities." We grasp at the one that works
> best, for the time being, of those we've thought of.
>
> "Wie Schiffer sind wir, die ihr Schiff auf offener See umbauen müssen, ohne
> es jemals in einem Dock zerlegen und aus besten Bestandteilen neu errichten
> zu können."
> ---Otto Neurath
this i think
is an important sentiment
but i would add
" ..and still our predictions grow finer
as long as we remember "
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
galathaea: prankster, fablist, magician, liar
