01
When does difference become a pattern?
Three marks are only three marks. The same three again, and the eye starts to keep a count. A third time, and the spacing matters. One of the four lines is set apart.
A pattern, at this first step, is a noticed regularity. It is not yet a message, a cause, or a reason to care. The same capacity that groups the marks will later group faces, words, and false alarms. The question is not whether minds seek pattern. It is what sort of claim a pattern can support.
02
Can something carry information without meaning anything?
H = −Σ p log p
Unequal chances. Entropy here is uncertainty over the set, not the importance of any one outcome.
In 1948 Shannon defined the engineering problem as reproducing at one point a message selected at another. Messages often have meaning. He set that aside on purpose. “These semantic aspects of communication are irrelevant to the engineering problem.” What can be counted is how surprised a receiver should be, given the probabilities of the alternatives.
Warren Weaver, writing with him in 1949, was explicit that this use of “information” must not be confused with meaning. A sentence full of significance and a sentence of nonsense can be equivalent as information if they were equally unlikely selections. Entropy does not measure what a person cares about. Redundancy, noise, and channel capacity are about reliable distinction, not about significance.
A spectral line, a pulsar’s period, a codon, and an error-correcting code can all be informative in this stricter sense, or in the neighboring sense that they constrain what is true. None of that, by itself, is a claim about personal meaning. The next chapter asks when a pattern becomes a sign.
03
When does a pattern become a sign?
An icon stands for something by resemblance. A drawn circle can stand for the full moon because of a character the drawing already has. Peirce’s point is stricter than “it looks like it.” The likeness can hold as a quality even if, in a given case, the object is absent. We still use the icon as a sign only when someone takes it that way.
An index is tied to its object by a real connection. Smoke indicates fire because fire affects it. A footprint indicates a step. A weathercock indicates the wind by being moved. The index is not mainly a matter of agreement. Remove the object’s effect and the sign, as an index, fails.
A symbol operates by habit, law, or learned convention. The word “tree” does not resemble a tree, and a tree does not cause that exact sound. A red lamp does not physically contain the instruction to stop. Drivers share a rule. Peirce calls this dependence on a law, “usually an association of general ideas.”
These are not three boxes to drop every sign into forever. Peirce treated them as ways a sign can relate to its object, and he noted that actual signs often mix them. A photograph is iconic and indexical. A diagram can be an icon of a relation. A proper name is mostly a symbol with an indexical job.
The sign is also not a pair of “thing plus dictionary definition.” In the Collected Papers formulation, a sign “stands to somebody for something in some respect or capacity.” It creates a further sign in that somebody, which Peirce calls the interpretant. Object, sign, interpretant. Leave one out and the relation he is describing collapses.
04
Where does meaning happen?
Photons of a certain range reach the eye. Nothing in the wavelength is the English word “stop.” The lamp can burn out. The photons change. The rule, for a while, does not.
“Stop” is held in place by teaching, law, and the expectation that others will also stop. In the United States, the Manual on Uniform Traffic Control Devices specifies red as the stop indication for traffic signals. That is a convention with force. It is not an arbitrary whim restarted by each driver, and it is not a fact about photons.
A driver who has learned a code meets the lamp. So does a child who has not, and a visitor from a place with a different signal. The interpretant is not a sticker on the glass. It is the further understanding the sign produces, which may be a foot moving to the brake, or confusion, or nothing.
Meaning here is relational. It lives in the practice that joins object, sign, and the people who can use it. Different traditions put the weight in different places. A naive view puts meaning in the object. An associationist view puts it in the private mind.
Peirce puts it in a triad that always includes an interpretant. Later, Wittgenstein suggests that for a large class of cases the meaning of a word is its use in the language. None of these is a theory of everything. They disagree about where to look when a sign succeeds or fails.
05
Why is the ox not the same word on the other side of the border?
Ferdinand de Saussure, in lectures at Geneva between 1907 and 1911, described the linguistic sign as a bond between a signifier and a signified. The bond, he argued, is arbitrary. Not arbitrary in the sense that you may rename the ox and still be understood. Arbitrary in the sense that the link is unmotivated. His example travels badly across a border: the signified “ox” has b-ö-f on one side and Ochs on the other.
He also refused to treat a sign as a positive lump of meaning. “In language there are only differences without positive terms.” Sister is what it is partly because it is not brother. The value of a term comes from its place in the system. Signifier and signified, in this account, are two faces of one social fact, not a word stuck onto a thing.
Do not merge this with Peirce. Saussure’s sign, as the 1916 Course presents it, is a pair inside a language. Peirce’s sign is a triad, and his symbol is only one way of relating to an object. Saussure even warns that “symbol” is a poor name for the linguistic sign, because a symbol, such as the scales of justice, still has a rudiment of natural bond. The Course itself was not a book he finished. Charles Bally and Albert Sechehaye assembled it from students’ notes after his death. The arbitrariness claim is his teaching. The exact page is an editorial object.
06
A guess, then a miss
Nervous systems that learn do a few unglamorous things well. They detect regularities, bind them to outcomes, predict what comes next, and generalize. A face seen once is easier the second time. A word heard in a pattern of use becomes available. The same machinery produces false positives. Signal detection theory treats that as a trade between hits and false alarms, not as a character flaw.
How does a system tell a useful pattern from noise? It cannot, with certainty, from one coincidence. It can keep score: how often the pattern appears, what else would predict the same outcome, whether the link survives a new sample, whether acting on it helps. Those are tests. They are not available to a creature that must move before the sample is large. So animals, including us, sometimes act on thin evidence. That is a cost of learning in a hurry.
In 1958 the psychiatrist Klaus Conrad used Apophänie for an “unmotivated seeing of connections” joined to a feeling of abnormal meaningfulness, in the beginning of schizophrenia. The clinical term is not a synonym for ordinary curiosity. Later writers widened it to name the everyday habit of finding a figure in noise. Both uses matter. Neither is a reason to stop looking. The capacity that over-reads a coincidence is close to the capacity that learns a language.
07
In the classical birthday problem, 23 people are enough for the chance of a shared birthday to rise above one half. The surprise is the number of pairs, not a hidden design. Among 23 people there are 253 pairs. Each pair is a chance to match.
Two paths can cross on a page without becoming one event. A bus is late. A song you have not heard in years begins as you wait. Drawn together once, they look joined. The drawing is ours.
A relationship can be statistically real without being causal. A cause can exist without personal meaning. A personally meaningful coincidence can be statistically ordinary. The ladder is there so those sentences do not collapse into one another.
The traps have names. Base rates: a rare wonderful match is easier to remember than the many quiet days. Multiple comparisons: look at enough pairs and some will look sharp. Selection: the pattern was chosen after it appeared. Confirmation: hits are kept, misses thin out. A “significant” result can be tiny. A repeated result is more interesting than a single surprise. None of this debunks a life. It asks what kind of inference the evidence can actually bear.
08
Chance, given enough pairs. Selective memory. A shared cause not yet seen. A story fitted after the fact. These can be true of an experience that still feels exact. A causal account does not have to insult the feeling in order to explain the conjunction.
Not settled
Jung’s name for a different proposal was synchronicity: a coincidence in time that means more, to the observer, than mere chance, linking outer events with a psychic state. He offered it as an acausal connecting principle, not as a finished law of physics.
He had used the word in public by 1930. The long argument came later, beside Wolfgang Pauli, in the 1952 volume Naturerklärung und Psyche. Pauli pressed for a narrow use. Effects that show up in a few striking cases and vanish in large numbers are not the same thing as a physical law. Jung’s astrological trial, which he himself revised, is not a demonstration that the principle holds.
In the foreword to the Wilhelm and Baynes I Ching, Jung wrote that synchronicity “takes the coincidence of events in space and time as meaning something more than mere chance, namely, a peculiar interdependence of objective events among themselves as well as with the subjective (psychic) states of the observer or observers.” That is a proposal about significance. It is not a measurement of mutual information, and it is not established physics. It is also not answered by saying “only apophenia.” Apophenia names one way meaning-feelings can be produced. It does not exhaust the description of what the coincidence did in a life.
09
Ring
The object does not have to change for the field around it to change. A ring is an alloy, a piece of jewelry, a legal and cultural sign, sometimes a religious symbol, sometimes a dead person’s object still on a living hand. Memory, grief, and promise are different claims about the same thing, with different kinds of evidence.
Flags, rituals, dreams, and numerals raise the same split. A flag can coordinate strangers who will never meet. A ritual can make a passage feel real by repeating it. A dream can organize a private history. A mathematical notation can be invented and still describe structures no one invented. The numeral “17” is ours. The fact that 17 is prime, in ordinary arithmetic, is not a mood. We did not invent the stars. We did invent the constellations. Once named, those figures became tools: navigation, calendars, myths, and later a record of how science itself once looked.
What begins as interpretation can change what people are able to discover. A named pattern is easier to test than an unnamed scatter. That does not erase the difference between the constraint and the construction. Some meanings are mostly found, because the world pushes back. Some are mostly made, because only a community holds them up. Many are mixed.
10
Detect a regularity. Estimate a probability. Classify an image. Move symbols by a rule. Surface a correlation a person missed. Produce a sentence that has the shape of an interpretation.
The test is ordinary. Did the pattern predict anything new? Was the sample fair? Would another method find it? None of those tests is a test for meaning.
Pattern recognition without a mind is already real: a filter can separate signal from noise.
Has an interpretant occurred? John Searle’s Chinese Room, set out in 1980, is one pressure on that question. A person who follows rules for Chinese symbols can return fluent answers and still, Searle argues, not understand Chinese. His conclusion: syntax is not enough for semantics. Programs, on this view, are not enough for minds.
The replies belong to the argument. The systems reply says the person in the room is the wrong place to look for understanding. The robot reply asks for a body in the world. Sorting pictures is not the same question.
Meaning, if Peirce is right that a sign must determine an interpretant, asks for more than a successful sort. If meaning is only use inside a practice, then a machine joins the question only where it actually joins a practice. Both conditions can be argued. Neither is settled by a demonstration that the machine found a cluster.
11
John Archibald Wheeler asked a narrower question than the one often borrowed from him. In “Information, Physics, Quantum: The Search for Links” he wrote: “It from bit symbolizes the idea that every item of the physical world has at bottom — at a very deep bottom, in most instances — an immaterial source and explanation; that which we call reality arises in the last analysis from the posing of yes-no questions and the registering of equipment-evoked responses; in short, that all things physical are information-theoretic in origin and this is a participatory universe.”
Read the apparatus in that sentence. The questions are yes-no questions put by equipment. A photon click is not a private hunch. Wheeler’s “participator” is not a license to say that a mind creates whatever it notices. It is a claim about what can be said, measured, and known at the bottom of a physical description. Information in his slogan is closer to Shannon’s distinctions than to a felt significance. He uses the word “meaning” inside a physicist’s problem. He does not settle whether meaning needs a mind.
A regularity can be in the world before anyone counts it. A sign cannot do its work with no interpreter.
Does the universe contain meaning before there is something capable of noticing it?
What happened between seeing the pattern and asking what it meant?
12
Only now. 42 is still the integer after 41. 69 is still the integer before 70. Nothing in the arithmetic changes because a person has started to recognize them. What changes is the interpretive history that can gather around a recurrence: repetition, biography, literature, astronomy, ciphers, memory, attention.
Keep the kinds apart. Quantity is the number. Pattern is the recurrence. Association is what the recurrence calls up. Symbol is the habit of treating it as standing for something. Personal meaning is what it does in one life. Statistical evidence is a claim about how surprising the recurrence is under a stated model. A causal claim says the number is produced or linked by a mechanism. Those are not interchangeable. A life can hold a symbol that is not evidence. A dataset can hold evidence that is not a symbol.
The question is not what 42 and 69 “objectively mean,” as if a numeral came with a message inside it. The question is how a number becomes capable of carrying meaning for a consciousness. When does recurrence become evidence, and when does it remain a symbol?
No probability is stated here for any personal series of 42s or 69s. A fair test would need a predeclared pattern, a base rate, and a count of the misses.
13
An instrument, not a verdict.
This sits near an index, and near causation.
Smoke is not a convention for fire.
Fire produces it.
Someone can fail to notice and the smoke is still there.
What the smoke means to a frightened person is a further question.
The physics does not settle the fear.
Not as a repeated symbol. As a regular effect of combustion.
Yes. Combustion sends particles into the air. That is a mechanism, not a resemblance.
The alarm you feel may be learned. The link between smoke and fire is not a dictionary entry.
A personal history can change the fear. It does not create the smoke.
Yes. Seeing smoke raises the probability of fire, and the reverse is often true.
The physical relation would. A private reading of omen would not.
The readings are distinctions argued in the chapters above: index and cause, convention, personal history, prediction, and the unsettled status of a machine-found pattern. They are not measurements of the visitor.