Meaning
Meaning... what
does that mean ? (A comic drawing from Joe KS
website.
Last revised 9/6/2014. Return to Major
concepts. See perception, conscience.
The first meaning of meaning is the "performative" value
of a message. By doing what is commanded, the receiver shows that it has
"understood". The dog sits, the lift goes to the fifth. And, millions
if not billions of times per second, a computer understands what the program
commands.
In other words (see form), the
meaning is got from a raw input (raster) by an interpretative process. And the
resulting data, or sometimes code, will be interpreted as the receiver as
actions to do. Something like:
if the input is I, then
produce the output O. That would be the "behaviorist"
standpoint.
The effect may be invisible from outside. Be a simple storing of
data, for instance. Or "evocation" of data from long termed memory.
Then the automata expression E = f(I,E) and 0 = g(I,O),
is better.
Not every part of a message is meaningful for a given receiver.
And we can separate two cases :
- the relevant information is formally written In/on the being, directly
readable by a machine (barcode, for instance)
- the relevant information must be sorted out of the "natural"
features of the being, by processes which may be rather simple (size or weight
of an apple) or highly complex (artificial vision and pattern recognition).
In the 1950's-1970's, meaning was seen as
information, as an extension of Shannon (even if Shannon had excluded meaning
from his formulas). You find an expression of that in books like [Osgood, 1957] or [Moles,
1971 and 1972].
But this approach has its limits. We have tried to go farther in
our Laval Virtual 2013 paper. There we open
the way of another measurement of meaning: the quantity of meaningful data (or
code) that a given receiver (notably a program) can extract from a document
(notably a work received as a raster). This would be somehow symmetrical to
Kolmogorov complexity (see provisional notes in our ontology).
Different evaluation systems will give different appreciations
results. A powerful recognition system, with a large library of pattern and a
large stock of features, will extract more meaning than a basic one. But this
is precisely a confirmation to this approach validity: two different persons
have different tastes, and an expert like Panofsky [46] can write 25 pages on
"Arcadia Ego" by Nicolas Poussin, in which a non-specialist visiting
the museum will only see some undefined characters talking in front of a rather
dull stone monument?
Since Google is investing in these
evaluation tools – and many others, too, including military and medical
professionals - we can expect great things in the years to come. And this kind
of Watson-type cognitive machine will give them the technological resources
they need (if they don’t have them yet).
Quantitative
aspects of meaning
To be "meaning", a message of sign must have a lenght somewhere between
- at least minimal informational weigh to bring the intended meaning
- at most the bandpass of the receiver.
https://www.technologyreview.com/s/604031/this-robot-knows-when-its-confused-and-asks-for-help/
Some references
<Image,
meaning and discovery, by Felic Frankel. 12 pages in [Grau 2007 ].
< Variation,
construction et instrumentation du sens. by
Maryse Siksou, ed.
Hermès/Lavoisier 2003.
< La structure
absente. Introduction à la recherche sémiotique, by Umberto Eco. Mercure
de France, Paris, 1972 (Italian Original Milano 1968).
< The measurement of
meaning by Osgood Charles E., SUCI George J and Tannenbaum Percy H.
University of Illinois Press, 1957.
< Meaning in the
Visual Arts. by Erwin Panofsky, University of Chicago Press.
Chicago, Illinois, USA. 1955.
< General systems
theory (Wikipedia).
Greimas
Mouton
Riss (book of pari)
Ulysses
German Tuutle
(1930) Dictionary or symbols
Logo
Varia to te
developed and edited
Let us try a presentation by levels, from the physical to the
transcendent.
Starting with the general definition, meaning is the effect of a perception on
a receiver
Meaning of anything is the behavior I'll have when I perceive it
1. Physical. Just
forces
Measurement is something in the physics laws (acceleraion,
forces, movement conservation
And heath. Mass.
The signal emerges from the crude physical forces. A kind of prediction on the
receiver. How can it be learned in non-couscious,
Darwin etc.
Structure of the matter.
2. Emergence of the
signal
The signal emerges not as a category of objects, but from the
perception of an observer applying causality. There must be sone ltime lag, a lo
Meaning is the replacement of an object, action or phenomenon, by
a signal "immaterial" signal
Semantics begin where physical determinism ends
and the meaning differs from a physical consequence
The signal is PMA on energy levels, but also on information
quantity
Simplificatin. Retina remanency.
Smoothing. Measurement
A permanent objet. Naming.
Big data as white noise
Part only of the reception (and then expression) can be said in
words; or intenally logically
Perception
audo
touch
smell
temperature
cenaesthesis
Recognition. Higher and higher level: recognize a feature, a
being, an emotion, a behavior, a threat or affordance.
Recognize the human, the Turing test.
Recognise a definite person. Biometry.
Sesame, passwords.
Auto-completion : get the meaning with incomplete
data. Can be done by a machine (Google)
Meaning is an aspect of PAM (Principle of action minimization) : a sign is ligher that its
signified. In the general case.
Théorie statistique de la décision.
And reusability.
It has to be interpred by
the receiver to have any effect. Organs
But to be properly so, the input must be perceived and heard as a signal.
Intermediary case : signal enzymes
"Ciel pommelé, eau
peu éloignée
A natural phenomenon, object, scene
idenficition, detection, recogition,
description, .... decision
e.g. it rains, i'll take my raincoat, I'll not go in
Normandy this weekend, I have not to water my garden..
normally not intentional
Maybe also signals from oher communities, but we can detext (birds songs)
3. The signal as an purposeful emission
General L purpose. Impossible; Subgoals.
Rabhi/
implies some convention between emitter and receiver
Implies memory. Learning
Pierce trilogy : index, symbol, icon
Psycchology and socioloy
Fodor,
semantics
- He considers semantics as a variety of psychology. Description, not
construction. Hence, enhancement and if/when possible, optimization is an
issue. How to produce/get the maximum of meaning with the mininal
message.
OK. : study the effects
- He deals with natural language, and humans. I am on any language, natural or
not, form
Meaning = complexity of the message .
complexity of the receiver
But what the "result" of the application. from the set E
of all expresisions onto the set F of all
"meanings" ??? What is the set of meanings?
Can be defined formalluy on a finite state automaton.
Or more "naturally" empirically on a human receiver.
This "effect" can be an "action" by the
receiver, or only a change in its "knowledge". The message is only an
information and, at least at that instant, has no consequence on the receivers action, behavior or mood.
Emotion as a consequence of perception.
Emotion on the spectator is the meaning of the text, or of the expression of
the character.
Perception takes time. Even if we are not conscous
of it. Cinema.
If two signs have the same meaning... an equivalence relation on
the set of signs.
Note that there is a sort infinity, with meaningless signs like page in Roxame or zeroes before and after the meaningful figures
... 000000017,0000....
Same meaning : the two signs can be
replaced by each other. Typical of equivalence relation.
But a given sign can/will have different meanings in different
contexts
3. Articulated
languages
First there is the cognition move. From reality, or more exactly
from the bits we detect in reality, through which we see reality, and from
those bits the building of high level objects and
ultimately of the world.
Systems of oppositions. Saussure.
Intension is the features included in an expression. E.g. a horse
has four legs. A square has four sides/angles.
Extension is the (list of) set of occurrences, the population of horses. The
set of all squares ...
Two expressions have the same meaning (the same intension) if they
have the same structure (type of compositon of
elements) and each correspondiing element has the
same intension.
What is the "set of meanings" ?
The real ? Then only expressions referring to the same
real object of fearutre have the same meaning.
Language is a poweerful means of etting data to reach aims.
De quoi s'agit-il
La décision.
Meaning, the possibility of error (not dealt with, in genral, by semanticisis) and
transgression (Leech). Tristan, *Thesee
Built of icons by analogy:
- the succession of samples from a sound sensor
Built symbols by convention: codes, standards. .
Thought they have optimization rules.
Coding. The linguistic schema. From the phrase to the surface.
But below this surface there are
- letters/phonemes
- bits
There are infinitely many ways of coding into bits.
Implied in a words are its parts/features (a horse has
four legs, a head and a tail)
Intension is the features included in an expression. E.g. a horse
has four legs. A square has four sides/angles.
Extension is the (list of) set of occurrences, the population of horses. The
set of all squares ...
Two expressions have the same meaning (the same intension) if they
have the same structure (type of compositon of
elements) and each correspondiing element has the
same intension.
From surface. Surface Vs deep structures
But in our perception it is not properly the surface itself we see, but its interpreation by our senses.
Somehow, surface is misuse of language. Surface doesnt
exist by itself. We see the surface with our eyes. Our ears do not hear
surface.
But, with high level beings, they build their surface (membrane, skin,
expression).
But, if you know the sensor, a bit can be an index
Genetics. Nothing but the reception of the male code in the female
womb. Decoding the DNA. We do not perceive it at all, or only indirectly
through.
Animals and humans; the senses. Vision, hearing, Mainly in the brain. But
preceding ways in flowers, jellyfish..
Synonymy. The set of meanings can be an equivalence relation on
the set of expressions. Frequently (generally) there will be an infinite class,
since some minimal specimen of the class may be completed and followdd by any meaningless antecedents, typically, 17,
0000017, 17,000000 etc.
But on a definite machine, there will be limits (number of bits for an integer,
float, double).
More precisely, there will be all the "transformations",
e.g. from deep structure to apparent structure
Antonymy. Less clear. One could say that the class of antonyms is
the set constituted by all the synomyms preceded by a
negation .... ???
The "minimal" reprentation in a given system (00000001,0000000 = 1)
Defintions. Dictionary. Generally, you cannot
replace a word by its definition (in natural languages). In maths ? In progrmmaing, somehow you copy the function and classes
A set theory approach. To the set of (global) perception
corresponds a set of meanings.
That creates an equivalence relation into the set of perceptions: are
equivalent all chose which have the same meaning.
Antonums, contrasts or opposition call for
complementary sets.
There can be an inclusion relation, if the meaning of some perception is
included in another meaning.
anguage can be mainly the trigger of a complex
behavior. On-off switch Programming. Recursion. Infinity.
Languages, and any systems of signs cannot be fully described.
Dictionaries and grammars are always partial and approximative. See grammaire synthétique.
4. Externalization . Writing, machines
For permanency and transmission, systems of signs are transferrd from oral to written.
Pour des raisons de permanence et de
transmission, les systèmes de signes passent, si l’on peut dire, de l’oral
à l’écrit.
Link : Meaning, a material autonomy in
respect to their authors
From genetics towoards
culture. Matter still matters. Objectivation of epistemology, then written grammer, "metaXX".
Machines. Sensors. Including headsets. Open CV. Signal processing. The cloud
perceives.
Drawing, gears
Sensorial prostheses Expression prostheses
4. The digital
fully
1947. The bit as such can be only a symbol. But the presentation,
or physical value of a bit, may be an icon or indices. e.g. a light
"on" evokes more naturally a 1, or "true" that a 0 or
"false".
The bit has no meaning by itself. Its meaning is given from
outside: the bits around, hardware and software devices, laws of biology and
society.
Bit vS. Meaning
- position in a logic gates network
- position in a string (e.g. number, string, any variable)
- meaning as what does an automaton with this output (near to the Hodel
definition)
- meaning can be "infinite" while T
(infinite : recursive wit no Stop conditon),
but always countable
First there is the cognition move. From reality, or more exactly
from the bits we detect in reality, through which we see reality, and from
those bits the building of high level objects and
ultimately of the world.
At the conscious level, we generally do'nt
perceive bits.
No specific form, no specific matter.
Measurement in bits
What is the intension of one bit ?
One major point with the bit is that it is not only a parting of
representations and represented, but more deeply the pure stem of separation
between significant and signifier, the basic loss of immediacy.
Linguistics. Phonemes, grahemes, grammar.
The moment when perception becomes binary, emerging from strict
analog physical play of forces.
Nothing is purely digital, ultimately. But the bit is from now on,
the keypoint in meaning.
Examples: DNA and mitochondires, text and prosody.
And "analog" parts in anytjhing digital,
typically an image.
A layering.
At the most surfacic level, communication is
analogous.
Bits can be very near (digital telecommunications) or very far (DNA).
Our five domains of digital
- Genetics. Cross-over and random.
- Animals. Pre-language, recognition.
- Humans and language.
- Machines. Note that only humans make machines. Finite automata. Turing and
Von Neumann machines.
Grammar, a fortiori formal, do not exhaust the reality of language. See Catach, statistics, my book on French syntax, etc.
Id. for dictionary. More open.
Language. Typograhpy, bibliophily,
prosody, oral expression.y.
Ongology. Dictionar
Definitions.
Context
Classic linguistics and semiology are centered on natural
languages. These languages are less and less dominant in human communication,
which itself is more and more replace by machine communication.
Ngrams.
Graphic semiàlogy.
Data visualizatin.
5. Automata, feedback
recursion
cyber loops, electronics. defined by differential equations.
relate that to recursion
by construction, a recursive function is not contiuous.
It's an application of N into some space.
The analog loop of cybernetics lets open an immediate adaptation.
3. Behaviors and community, interaction and mirroring
Its a case of the general law of systems, L. It is more profitable to
interact with the world through signals than "really", for material
reasons and also "logical".
5. Autonomous
subject, as receiver (interpretation of the perception) and an emitter with
feedback (meaning of my life).
Find meaning in the present overflow of messages is the main
challenge of present times. But the books about semantics and semiology don't
help much. Let's try to progress.
Secondly there is the construction of signs, which implies that somehow we know, or that something knows for us, that these
signs will be understood. The agents who create bits may be passive or active.
Full societies, and finally the whole world.
Rectitude des appétit
We find partial expressions of that in linguistic theory. From
sentence to words throught syntactic and semantic
rules. But this view is partial, since the words are coded into graphemes and
phonemes, and from there into bits.
Larsen . Simple "internal loop", with
little of the external wols (but the echo sometimes)
textual emission,
- the coherence of this world, its transmedia and crossmedia
nature, are fostered by the versatile possibilities of the digital, the bit,
- the bit brings also a convergence of human communication, genetics and
machines; human traditonal nature is squeezed between
an omnipresence of DNA and our growing ability to master hitherto deeply hidden
genetic operations; and machines (devices and algorithms) are omnipresent; they
outgrow humans in number and in communicative activities; in economy, genes become
patentable and information capitalism is more important that goods; in
politics, privacy protection and cyberwar become come in foreground; in
entertainment and art, we are immersed in virtual reality and transmedia.
- we live in a multimedia world, where images and interaction are
at least more present than
The test of "me"
5. The will, the
purpose of the emitter
The automaton.
Equivalence of an automaton and a laguage.
The duty. An aspect of the S, requires
from it to know what it has to do. Form me, both a duty, someting
to do for the World, for my childre, grandchildren.
Maximize the yield.
The tetralogy : belonging, purpose,
storytelling, tarnscendance
A main point is the recognition of other humans (Gn 2). And the uncanny effect when the recognition is
doubtful, or creates a doubt about its human nature
Deliberate deception.
Expression of meaning. Face moves. Even though expression on a
face can be considered as meaning the emotion of the character.
6. Art
Art could be considered at the way to create a particular
category of meaning, with ncessariyl
- a pleasure component
- a non-functional (useful) component.
Semantic enhancement.
Semantics in pracoce. "Make" meaning.
Style into a message (or work of art, of general expressive
behavior) is a set of features/rules which meaning relates to the emitter, not
to the things they talk about.
Compression and glitch.
The keypoint is to go from error or
random to meaningful transgression.
Any new bit, the (n+1)th
bit, is the basic transgression.
7. Transcendance
Answer to the stimuli. Depends on the state (finite automata)
The stimulus is yet a particular kind or relation to
the external world.
Send a stimulus.
But there is also the "strenght",
the importance of a message for the receiver. Depents
on their L.
meaning Vs. process
the process is invisible, transparent
the process is the work, shown as such
the code is never shown clearly
pure data.
When descending the scale, processing depth, here also there is a
stop condition. I (the system, being) answers before reaching the bottom.
Otherwise it would be infinite.
Analogy with the resonance model.
Some straightforward cases : green light
-> Go.
Mirroring and recursion.
Clock.
"Factorization" of meaning. Language.
More poweful machines affords deeper
level of behavior, farther and farther from reflexes. Long term memoy, evaluaion in a richer
environment. (the automata parameters are higher and higher).
L theory
get more H with less P expense
In a population. Zipf, EDI.
Norms, standards
Economy, expansion
Spare athe use of hardware, energu,
amter, time
Sapre use of bits? The best codes, compression
From things to representations. But also becase you can do operations on signs that are very costly,
or radically immpossible on real things.
Generalisation, deduction, automatisation.
From representations to bits.
Exception: art, ornament
It is not necessay to have an emitter.
Flowers,
Even less conscious. languages.
Growth. The selfish gene
Finite automata
Reflex
Mirroring. Co-evolution. Recursion. Abyme.
The cloud.
4. Levels, beings
1. Null Signal, 0 signal
If I = O. No complexity of a message = no signal at all. That does
not exist.
If R is complex enough, the absence of a signal, or the null signal can always
be interpreted as an input (no news = good news). More exactly, in a S with
memory, what is the input is several successive (possible infinite, or limite by memory) e.g. in an exponential smoothing
, a permanent signal will incrrease each time,
convergence. Asymtotic.
if the number of bits is finite, after a moment, the asympote
is actally reached, no more dimintuion
of distance
But, at a higher level, the S may count the number of cycles wher the asymptote distnce is not
changed
If the value of the input is constantly equal to the asymptote, afotre some number of steps, all that we can count is the
number of steps.
2. 1 bit signals
may be on/off, forwards/backwards
it it's on, may trigger the full complexity of the receiver
When there is only one bit, the meaning is purely external.
When 2 or more bits, progressive internalization of the meaning OOP
and self contained (drone)
A bit, isolated, has no meaning but,
- if you have many bits, you will generally be able to find some regularities,
to decode (do reverse engineering to get the code, the language, the core
meaning)
- a large set of bits can by itself be a high level automaton, as such as its
structure allows this and there is a clock working as a motor.
Low levels : characters, phonemes, words.
Construction laws. just concat.
Then appeear special combinations of bits.
In DNA, some pairs
limits of matter : cohezionforces,
Human: Alexander, the Greek schism,
Limit violence. Rituals of submission
Limit risk : detect risks Meteo
Growth demands a structural sophistication.
Names, identifiers
When a set, a population, becomes too large, "natural language" is no
longer sufficient. You have to build upon more formal,
artificiel naming systems :
NIR, IP, URL, EAN, "nomenclatures", addresses
rules for database identifiers
internet of objects
The keypoint is the separation of
significant and signifier into the sign.
An analog sign is only a partial rupture.
A digital sign of one bit is a total rupture.
When several bits, the rupture may be more or less
total. This depends on the kind of coding. Generally, il
will remain some analogy. Would it be only because it is more efficient for our
senses and mind
A bit is not even "something" by itself.
It must have some physical/material localization, with an "absolute"
address
Only one basic structure: string, streaming. Concat.
other structures are given from outside,
may be (and more ande more) combined with contrl its. CNA
the string. or parallel channel (message "width" and "lenght"
Quantitative approach. Number of bits.
In perception
In emission
Internalization of meaning with size growth.
See telecom protocols.
Headers,e xtensions CSS (fatctorizing)
Recursion, (Smithj
Small groupings. Characters, words, DB item, classes, files. Codes
history.
Elementary components are structures (or values of
)
Genetics. Anthropometry as ADN. ADN as identifier. - Genetics.
Genes. Global genome. Population
Small levels: pairs, genes.
High level : globl genome,
species population, the whole life community.
DNA is a counter-example : its
a source code going to analog interface. The code-to-code communication is
hidden deep into the mothers uterus.
The cell, basic element of life.
Animals.- Animals. Neurons, brain. Society
Cells. Functions.
Tribe.
Humans.- Conscience. Senses, then explicitaion.
The human. Identity.
Machines. Agents. Classes. The cloud.
The only final meaning: build the global mind. The God's plan
Ephésiens 3:1-13
See Bouge
mots
--Notes about signs
and semiology
Sign requires two separations :
emitter/receiver and signified/significant,
sign is always a kind of dematerialization. otherwise would not be useful,
economically profiable
emitters : nature, men, machine
receiver : nature men machine
some signs are generated by machines (sensors, thermoeters,
magnetic compass...)
some signs are direct to machines (button, crank, keyboard, mouse)
Semiology or semiotics, generally, deal only with language
Let's try do look a wider landscape.
Types of signs /
visual, graphic
letter, alphabet
road sign
color
flag/blaon
insign (sheriff star), clothes, tattoos
gesture, facial expression
Yin-Yang.
Maps, legends. scale rules on typonymy pringing
any meaning atributed to any signifier
by convention. icon/indexical/symbol
audio
sound, pitch, phoneme, prosody (language, accent)
animals (birds, etc.)
corpus prepatration
cutting, digitizing
atoms/integration
sememe
bits/characters/books..;
laws of combination. grammar, dictionary
style
emplotment, storyboard
datavisualisation
integration is in space (1D 2D 3D 4D) or only time
integration is "analog" . e.g.
sampled music, photography, film without moungint
integration is mouted, emplotted
translation from one system into another
evaluation of systems, and languages