Technology Extends Cognition Only by Redrawing Where the Mind Ends

A route is seven things long. A left turn, a bus, a bridge, a bakery, a second bus, a building, a door.

Try to hold it in mind, then write it down.

Nothing mystical happens when the ink arrives. No tiny soul marches out of the skull and rents a room in the notebook.

But the task is no longer the same. The paper now keeps the order. The hand can check it. The body can follow it.

Memory has acquired a colleague with terrible handwriting.

This is the scene in which philosophy of technology and cognitive science keep meeting, each pretending to have arrived for a different reason.

Heidegger begins with a refusal.

Technology is indeed an instrument for ends and a human activity. It is also, rather inconveniently, more than that. Modern technology is a mode of revealing: a way the world comes to show up before anyone has selected an app or misplaced a password.

The older craft image is poiesis, a bringing-forth. Modern technology is enframing, a challenging-forth.

The Rhine becomes hydraulic pressure. A forest becomes timber inventory. A person becomes human resources.

Under Gestell, things appear as standing-reserve, a stock arranged to be called upon. Neutrality has acquired a warehouse and a procurement department.

Jacques Ellul published his own version the same year, from a sociologist’s desk rather than Heidegger’s ontological attic.

Ellul’s technique is broader than machinery. It is any rationalized, maximally efficient method.

He gives it a ferocious thesis in autonomy: efficiency ceases to be one human value among others and becomes the organizer that rearranges the others.

Both accounts make technology sound like a large, competent animal that has learned to fill in its own forms.

This is where the corrective arrives carrying a bicycle.

Pinch and Bijker took the empirical program of relativism from sociology of scientific knowledge into artifacts. The high-wheeler and the safety bicycle were live rivals for decades.

Young male riders could prize speed and the glamour of risking a header. Women, older riders, and safety-conscious riders could see a machine designed for a very short appointment with the pavement.

The safety bicycle did not simply reveal itself as engineering’s one true child.

Relevant social groups negotiated what counted as a problem. Rhetorical closure made safety persuasive. Problem redefinition made the pneumatic tire a comfort solution rather than a failed speed machine.

An artifact has a task structure before it has a destiny.

That is why writing down the route is not a small event.

The marks do not merely aid an already finished act of remembering. They change who or what bears which part of it: sequence on paper, recognition by eye, movement by body, correction by hand.

The distributed task is ordinary enough to escape notice, which is how machinery does its best philosophical work.

There is a sober version of this insight.

Value-sensitive design asks its practitioners to run conceptual, empirical, and technical investigations. Direct and indirect stakeholders enter the first. Actual use enters the second. Particular system properties enter the third.

Its four criteria are stakeholders, time, values, and pervasiveness.

Values do not drift into a design after lunch. They arrive with the requirements.

Langdon Winner’s question gives the discomfort a sharper edge. Do artifacts have politics?

He distinguishes arrangements that function as order from technologies inherently compatible with a political form. His bridge illustration must remain an illustration, whose historical particulars need primary verification.

The thought it stages is still clear. Concrete can allocate access without waiting for a rulebook to become cruel.

Cognitive science supplies the other half of the room.

George Miller dates the field’s conception to one 1956 symposium, where Newell and Simon presented the Logic Theorist, an engineer described a machine with two thousand words of core memory, and Chomsky presented transformational generative grammar.

A mind became, among other things, something that could be modeled as information processing.

The model soon became a family quarrel.

Logic Theorist proved a substantial run of the early theorems in Principia Mathematica. Rumelhart and McClelland’s networks gave simple units, weighted connections, and past-tense overregularizations such as “broked.”

Fodor and Pylyshyn’s systematicity objection then asked why a system that can think “John loves Mary” must be able to think “Mary loves John.”

It is hard to keep a philosophy of mind tidy when the diagrams keep answering back.

Embodied cognition makes the route look less peculiar.

Reading a word such as “kick” can recruit motor areas relevant to the leg. Perceptual symbol systems treat concepts as partial reactivations of sensory and motor states.

Action-sentence compatibility adds the same provocation. Comprehension is not necessarily a small clerk processing abstract slips in a sealed office.

But an experiment can show sensorimotor recruitment without issuing property deeds to notebooks. Some celebrated embodied findings have replication concerns, and the caution belongs with the specific claim rather than the whole programme.

Predictive processing would happily redraw this entire scene as a hierarchy of predictions and error signals.

The free-energy principle became a philosophical programme in which perception is controlled hallucination constrained by what arrives.

It is a candidate unifying framework, not the imperial seal of cognitive science. A good theory has not yet earned the right to annex the stationery cupboard.

Which is the point at which the border stops looking defensible.

If a notebook stores the route, a phone completes the address, and a body acts through each, then perhaps the notebook and phone literally become parts of the mind. The brain has outsourced a department, and the guards should go home.

That conclusion runs ahead of its evidence.

Functional coupling shows causal help and altered task structure. It does not settle whether the artifact shares an agent’s beliefs, whether responsibility travels across the coupling, or whether ontology should follow convenience.

There may be a device so deeply integrated that calling it merely external becomes misleading. That is speculation, not a finding obtained by watching someone consult a map.

What holds is neither technological determinism, which makes artifacts too sovereign, nor social construction, which can make them too obedient.

It is that thinking has never been only a private episode under the skull.

It takes shape through bodies, symbols, rooms, habits, instruments, and institutions. Arrangements that carry information and values together.

The route is no longer only remembered. It has been arranged.