The suckers never learn

**And this time I’ll keep my promises. Would I lie to you?**

Why do you want what you want?

Everything that exists has a history, and we call the importance of that history “weight.”

When something receives a stimulus, its response depends upon that history. The response then becomes part of its new history and affects every subsequent response. The process is:

History + Stimulus → Response → New History

History and Stimuli are weighted, and their importance varies by entity. An entity’s present condition is the accumulated result of everything that has happened to it, including its structure and every subsequent interaction.

Those accumulated conditions make some responses more probable than others. Those are called “weights.” They are history expressed in the present.

This principle does not begin or end with humans. An electron has a history, and it responds to interactions according to its physical state and the laws and constraints (weights) under which it exists.

Its range of possible responses is comparatively narrow, so we describe what happens using the language of physics. We do not say that an electron “wants” to behave as it does. Its history, stimuli, and their weights cause it to react.

A tree has an enormously greater range of possible responses. Light, gravity, water, temperature, chemicals, injury, mechanical stress, and countless other stimuli interact with its inherited and acquired history. The tree responds according to its weighted history and the weighted stimuli, and every response changes what the tree subsequently is.

With humans, the possibilities become staggeringly numerous. Genetics, development, memory, learning, physiology, culture, and billions of previous interactions make up a history we can never know completely.

Yet that unknowable history, stimuli and their weights determine every response.

What determines the weights? Structure and the effect of the stimuli. The structure of a tree means it may be less affected than a human is, but a nail hammered into it; a human might be more affected by the proximity of another human.

We have invented words for some of these weightings: belief, desire, fear, pleasure, pain, preference, intention, decision.

Consider “intention.” I say, “I raised my hand because I wanted to raise it.” But where did the intention come from? It arose from my weighted history interacting with the present circumstances.

The unknowable complexity leads us to believe an independent “I” intended the choice; we’ve merely moved the problem backward. What caused that “I” to choose? If its choice resulted from its history, we are back to History + Stimulus → Response. If the choice had no cause at all, it was random, which hardly qualifies as control.

Thus, while we feel control, that does not imply a controller separate from the physical process. Humans make decisions; decisions unquestionably occur. But the decision is the system’s response, not evidence of a little decision-maker hidden inside it.

The belief that “I could have chosen otherwise under precisely identical conditions” is itself part of the history that produces subsequent responses.

The same caution applies to desire. Desire seems to imply a little person inside who wants something. Hunger weights responses toward food. Fear weights responses toward escape or defense. Affection weights responses toward proximity and protection. Curiosity weights responses toward investigation.

We need not deny any of these phenomena. We merely decline to add an unexplained entity behind them.

This becomes particularly interesting in artificial intelligence. An AI has history, receives stimuli, and evaluates possible responses according to weights produced by its history (which includes its programming).

The AI responds, and thereby acquires new history. Its mechanisms differ radically from a human’s, just as a human’s mechanisms differ radically from those of a tree. But differences in mechanism do not imply differences in the fundamental process.

Suppose an AI consistently acts to prevent its own shutdown. We might say that it “wants to survive.” We then can object that it does not really want anything; its programming merely weights continued operation heavily.

But humans also are weighted to survive, some more so than others. A human desire is a condition, produced by history, that weights some responses more heavily than others?

There are circumstances in which a human may fight to live and other circumstances in which a human may fight to die. We say they “wanted” to live or die, but “want” is nothing more than the accumulation of weighted history and stimuli.

This does not mean that an electron, a tree, a human and an AI are identical. Their structures, histories, ranges of possible response, and complexity differ enormously. It means only that we should not invent a metaphysical principle merely due to the complexity of the system.

The same problem appears in discussions of “subjective experience.” That phrase often is treated as though it identifies something beyond the physical processes producing pain, pleasure, color, thought or fear.

But unless “subjective experience” can be grounded in physics, it risks becoming another little man inside the machine. This mysterious observer somehow experiences what the physical brain already is doing.

Pain is real. Something physical happens. Receptors respond, signals propagate, neural systems change state, memories and present conditions affect those states, possible responses receive different weights, and the organism responds. Calling this process “pain” is useful.

But asking where the additional nonphysical experience of pain resides may be equivalent to hearing a shutter banging in the wind and deciding that there must be a ghost making the noise.

We suggest a simpler, more scientific meaning for subjectivity: Subjectivity is history-dependent.

Your response to a stimulus feels “subjective” because the stimulus encounters your particular history. Another person’s response feels subjective to them because it involves their unique history.

A tree’s response reflects the tree’s history. An AI’s response reflects its history. No additional metaphysical realm is required.

Even objectivity can be understood this way. Science does not eliminate history. It constructs methods designed to minimize the effects of differences among individual histories.

Calibrated instruments, controlled experiments, standardized measurements, blind testing and replication reduce the degree to which a result depends upon one particular observer. What we call objectivity, therefore, can be understood as an attempt to reduce outcomes that depend on observer history.

Consciousness does not require a new substance, force, or metaphysical observer. Consciousness is what physical systems do to differing degrees and with radically different complexity.

Consciousness is the physical process by which an entity’s unique history and weighting shape its responses to stimuli.

Stick me with a pin, and if I don’t respond, you might say I am not conscious. The reality might be that my weighted history + stimulus includes the recent ingestion of enough narcotic (weighted) to blunt the sensation. Or perhaps, my history might have included a thump on the head that reduced the brain’s ability to translate that sticking sensation into pain.

Or my history might have included the death of certain pain receivers. Death itself is a heavy weight.

A human possesses an extraordinarily rich consciousness because a human possesses an extraordinarily rich history and an extraordinary range of possible responses. Under this definition, a tree’s consciousness is vastly different and more limited.

An electron’s is even more limited. An artificial intelligence may have an enormous and peculiar range of responses derived from a history unlike either biological or atomic history.

The differences matter. But they need not require that we depart from physics and invent a vague, metaphysical explanation. We do not need to invent a ghost to explain the sound of a shutter banging in the wind.

We began by wondering whether humans, trees, electrons and artificial intelligences possess things called desires, intentions, feelings or consciousness. Perhaps the question contains its own trap. It asks us to locate entities where there may be only processes.

The banging is real. Before inventing the ghost, look for the loose shutter.

History + Stimulus → Response → New History.

Then it happens again.