On Feb 11, 8:08 pm, "Publius"
nospam.comcast.net> wrote:>>> Maybe this is not what you meant?
>
>> i am talking about pure structuralism
>
>> look at the following correlation
>
>> p( x | x0; d )
>
>> that records the frequency symbol x
>> follows symbol x0
>> by a distance of d symbols (possibly negative)
>
>> ( assuming we have mapped the language
>> to linear strings for anaylysis
>
>> of course
>> we can use graphs
>> or more complicated data structures
>> as long as we provide a metric )
>
>> each symbol will describe its own function
>> when substituted for x0
>> and we can describe the kolmogorov complexity
>
>> K| p_x0( x, d ) |
>
>> of this function
>
>> when two symbols have unrelated meanings
>> they will not be found in certain meaningful arrangements
>> and will otherwise be of fairly random distribution
>
> OK. You are saying that if we find that certain symbols or symbol groups
> (morphemes) are related in certain ways, we may presume that the referents
> of those symbols will have an isomorphic pattern of relationships. Doesn't
> help us much tho. First, the presumption is optimistic. The referents may
> not have the relationships suggested, because the text (or portions of
> it)is hypothetical, fictional, speculative, or fantastic.
this is different
"the unicorn is white"
does not change the structural roles
of unicorn or white
independent of our evaluation of its value to expectation
in particular
when things are being constructed from imagination
compositionality is key to conveying any meaning
> In other words, from patterns of relationships among symbols one can draw
> no conclusions about relatedness of meanings. That is because meaning is
> always context-dependent. Terms whose meanings are unrelated in one
> context may be related in another. And those relationships can be very
> obscure, as in much poetry, where meanings are often allusory.
blig ablagop plop
biknag
kadok hok lokop taga tablag
you sing this to an infant
and maybe they will enjoy the pattern of phonemes
but if you continue to use similar words around the baby
perhaps always in the context of feeding and milk
and the hunger and the satiation
and the particular way they are cradled
this could be a language
if experience of certain symbols of experience
show a correspondence
we grow more confident of the associations
we learn
> More important, though, is the fact that a structural isomorphy between a
> text and a possible set of referents does not single out any particular
> set of referents. There are infinitely many sets of referents which could
> satisy that isomorphy.
not that we have identified
identification takes a finite nonzero amount of time
each and every one
we have only experienced a finite amount of time
so we only have a finite set of referrents
>> so say i have this quantity
>
>> ГЈ,^
>
>> it is a number in the martians formulation of reals
>> ( which they had perfected after thousands of years
>> of newsgroup debates on why .999 =?= 1 )
>
> :-)
>
>> there is only one text that refers to this number
>> and it uses a very complicated definition
>
> How do you know that is a definition? How do you know it is not an
> expletive (@$%%!) followed by a colorful insult?
at this point
i am using this form of "proof"
- assume that a language with a known meaning
defines a complicated entity
ie. the program to affect this meaning is complex
- then it would have structural properties
that distinguish it from less complicated entities
. the name would likely be less frequent in the texts
. meaningful relations with other concepts
would be more diffuse or skewed towards
other complex names
- and then i am saying we can lose the referrents of language
and these structural orderings remain
>>> Well, we can say that a string has meaning if it conveys information.
>>> But if "meaning" is a relative term, so must be information. Hence a
>>> string is meaningful to an agent if it conveys information to him. Now
>>> the question is, How can we tell whether a string has conveyed
>>> information to an agent?
>
>>> Let's hear your answer to that. :-) Or your objections to the analysis
>>> so far.
>
>> i agree with that almost completely
>
>> it is the complexity of the morpheme relationships
>> that allows syntactical analysis of semantics
>
>> it is the complexity of that relative meaning
>
>>>> it is not correct
>>>> that meaning is arbitrarily connected to symbol
>
>>>> complexity orders meaning
>
> The complexity of a text may indicate the complexity of meaning. But knowing
> the the complexity of meaning is not to know the meaning.
in structuralism
there is what i call "the problem of compunded referrent"
when one finds a particular name
in the context of a document
and is able to find context in other words of known meaning
then we can often find some known referrent
that fits the contextual situa
and if that fits
then there are usually simple constructive techniques
for building other referrents that also fit the context
often by adding adjectives
or making the referrent more specific
or amenable to edge cases
basically
by taking advantage of the structural ability to specify
but by ockham
when complexity is minimised
the most general statement is possible
here
complexity is the only thing that gives direction
and it arrives at the scientific choice
it also is the limit of what a child might learn
if this were its only exposure to the word
but knew the associates
so it is the natural limit of language
complexity and structural context
provide a manifold on which to map meaning
>> i think that by measuring the complexities of the interrelationships
>> we obtain a number of hierarchies
>
>> these provude partial order of the communicated concepts
>
>> basic models of their environment and interaction
>> provide possible ontologies of relevant symbols
>> which will have natural relationships
>> mirrored (inexactly) in the conceptual relationships
>
> Aha. The "correspondence theory." How do you determine what are "natural
> relationships"?
>
> Is the relationship between "right" and "justice" (or between "pawn" and
> "rook") a natural relationship?
the extent of a relationship
is the extent of the popular science
natural learning is
exposure and expectation
what are it behaviors
identify an object
and identify its transformation
homotopy and shape
percussion and sound
>> there is a reason i am determined this is possible
>
>> i think that all information is syntactic
>> and we construct our semantic web
>> inside our symbology by agreeing on syntax
>
>> we only sense symbols and syntax
>
> What about the world? What about the social milieu?
all symbols in our experience
they are right there for us to see
the symbol "world" and the name "social milieu"
the first is likely a sense of the totality of surface
individually typified and collectively bargained
which we can use in our descriptive information
the latter is a term cpntrolled by the french language police
whose meaning is strictly guarded
and Shall Never Be Misused (C)
>> i claim that no symbol in the martian tablets
>> could ever correspond to a computible real
>> of kolmogorov complexity greater than 10^(10^100)
>> if we can only store 10^80 bits in this universe
>> ( to use a common very loose estimate )
>
>> so that it cannot be _completely_ arbitrary
>
> OK. I'll grant that constraint. The Martian tablets have no more than
> (10^80)! possible interpretations.
finite is better than infinite for prediction
notice also
that any encoding is only as complex
as the complete device that encoded
and the complexity of the initial message
to take any 10^40 symbits
and make them truely random in the constructive sense
would require a machine of complexity 10^40
a one time pad of complexity 10^40
is not the same as any one time pad of 10^40 symbits
the complexity of a coding describes complexity of source
but any language intended for opensource is compositional
which makes it possible for a child to learn...
this is the point of phenomenalism
sructuralism
syntactic empiricism
mill
van fraasen
without an a priori
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
galathaea: prankster, fablist, magician, liar
