Spread the Medium
The same inventive idea, hiding in very different technologies.
Suppose you need to create an intricate pattern on a silicon wafer. Some regions must change; the neighboring regions must remain untouched.
The natural question is: how can we put the right material in exactly the right places? Perhaps we need a smaller tool, a more accurate positioning system, or a finer method of deposition.
But consider a different answer: start by covering everything.
In photolithography, the wafer receives a coating of light-sensitive material called photoresist. Patterned light changes the coating’s chemistry in selected regions. Development then turns that chemical difference into a pattern that can be used in subsequent manufacturing steps.
The coating does not have to be deposited along every contour of the intended pattern. It is already there. The light supplies the selection.
This is the first clue: we have separated where the material is present from where it changes.
Now consider a quite different problem.
We want a therapeutic substance to reach tissue, but we do not want it to produce its effect everywhere it travels. Broad access is useful for transport. Broad action may be precisely what we want to avoid.
Imagine a carrier that retains its payload during transport and releases it only after receiving an external signal. This is more than a thought experiment: University of Florida researchers have developed magnetic nanoparticle carriers designed to release therapeutic agents in response to an alternating magnetic field.
That mechanism does not automatically solve access to tissue or precise field localization. But it makes an important separation possible: transporting the substance and releasing it no longer have to be the same operation.
Again, presence does not have to mean action.
The third example makes the idea even clearer.
Suppose we want to see the path of a charged particle. We cannot position a detector along its trajectory beforehand, because discovering that trajectory is the purpose of the experiment.
A cloud chamber takes a different approach. It establishes a sensitive region of cooled vapor. As a charged particle passes through, tiny droplets form along its route, revealing a visible track.
There is no need to predict the route. The medium is already available throughout the sensitive region.
In the wafer, we know the pattern and want to create it. In the chamber, we do not know the pattern and want to discover it. Yet the underlying arrangement looks remarkably similar: prepare the whole domain to respond, and let a local signal determine where the response occurs.
The part I nearly overlooked
At first, it is tempting to focus on the signal. Light selects one place; a magnetic stimulus controls a release mechanism; a passing particle reveals its own route.
But the signal alone is not the invention pattern.
What caught my attention was the entity receiving it: the medium that makes the signal useful.
It has to possess a rather unusual combination of properties.
It must be able to spread throughout the relevant domain — or at least become available wherever a response may be needed. A perfectly responsive material is not useful at a location it cannot reach.
It must also react to the designated signal, converting that encounter into the required change.
Those two properties initially seemed sufficient. Then a third became impossible to ignore.
The medium must remain stable when the activating signal is absent.
Imagine a detector medium producing marks spontaneously. Or a carrier releasing its contents on the way to its destination. The same broad distribution that was supposed to solve the problem would now spread unwanted effects.
What we need is not simply a sensitive medium. We need a selectively sensitive one: responsive to the intended interaction, but resistant to ordinary conditions and irrelevant disturbances.
“As stable as possible” therefore means as stable as the task requires while still allowing reliable intended activation. It is an engineering requirement, not a demand for a substance that can never change.
Whether the signal is weak or strong is secondary. So is the size of the resulting action. Amplification may help a particular implementation, but it is not what joins these examples together.
Their common structure is coverage, controlled response, and stability against unintended response.
The contradiction hiding underneath
Once I looked at the examples this way, the contradiction became easier to state:
The capability must be available everywhere it may be needed — but it must not act everywhere it is available.
When a design ties presence directly to activity, these requirements pull against each other. Increase coverage, and unwanted action increases too. Restrict the action by restricting distribution, and some possible targets become unreachable.
The solution is not necessarily to find a better compromise.
It is to break the connection between presence and activity.
We distribute something that is ready to act, rather than something that is already acting. The medium solves the availability problem. The signal solves the selection problem.
That seemed like more than a resemblance between a few technologies. It looked like a reusable instruction for resolving a class of contradictions.
But was it already one of the familiar TRIZ principles?
“Isn’t this just an intermediary?”
That was the first obvious candidate.
The Intermediary, or Mediator, principle recommends introducing an intermediate object that carries or transfers an action. Our medium often does exactly that.
But an intermediary need not spread throughout a domain. It need not remain inactive until addressed. It need not distinguish the intended signal from other influences.
Think of an adapter between two incompatible tools. It mediates their interaction, but it does not create a field of locally available, selectively activatable capability.
“Introduce an intermediary” identifies a possible role. The pattern we are examining specifies a particular kind of intermediary — and a particular arrangement in which to use it.
Perhaps Local Quality is a better match?
That principle calls for differentiated properties, functions, or operating conditions. It also includes making an external influence nonuniform, so patterned illumination genuinely overlaps with it.
Get Pinkhas Nisanov’s stories in your inbox
Join Medium for free to get updates from this writer.
Still, consider a tool with an insulated handle and a conductive working end. Different parts perform different jobs, but nothing is waiting throughout a domain to be selectively activated.
Our pattern offers a particular route to local differentiation: make the capability broadly available, then let a signal determine its local behavior. The selected locations need not even be known when the medium is distributed.
Local Quality can describe the differences we obtain. It does not, by itself, specify this method of obtaining them.
Then there is Preliminary Action: do work beforehand, or position objects so they are ready to act without delivery delays.
This is a substantial overlap. Coating a wafer before exposure certainly involves preparation.
But putting a spare tool beside a machine is also preliminary action. It says nothing about a medium covering the possible target domain, remaining inactive there, and responding selectively.
The distinction is not simply early versus late. It is available versus active.
I could describe the examples as a combination of these principles: add an intermediary, put it in place beforehand, and obtain different behavior in different regions.
Yet this left me with an important question:
If I did not already know the solution, would those instructions direct my attention to the medium with the three properties we have identified?
They might. But naming the complete pattern gives me a much more specific thing to search for.
There are closer relatives elsewhere in TRIZ, too. Its substance–field standards discuss structured fields and field-responsive particles introduced into the environment. So I would not claim that this idea has no antecedents. The proposal is to recognize the complete arrangement as a separately named principle — not to claim ownership of the physical mechanisms behind it.
A name for the pattern
I would call it Spread the Medium.
Not “spread the active substance everywhere.” That would recreate the problem.
Instead:
Spread a selectively responsive medium throughout the domain where an effect may be required. Keep it functionally inactive under ordinary conditions, and let the designated signal produce the required response only where and when needed.
Here, “everywhere” means everywhere relevant to the task: a surface, a sensitive volume, or an accessible set of possible targets. The distribution can be established in advance or maintained continuously.
The domain does not even have to be physical space.
Imagine software functionality already available for every account, but disabled unless a particular authorization condition is satisfied. We do not deliver the function anew each time an account becomes eligible. We establish the capability broadly and control its realization selectively.
In that design, the account set is the domain, the deployed functionality and its activation logic play the role of the medium, and authorization supplies the selecting condition.
The materials have disappeared. The arrangement remains.
Why give it its own name?
For me, the strongest reason is not that it lets us place old inventions into another category.
It is that it changes the question we ask about a new problem.
Instead of asking only:
How can I take the active agent precisely to the target?
we can ask:
What medium could I make available throughout the possible target domain — and what signal could make it act only where needed?
That question immediately suggests what we must investigate: a distribution mechanism, a useful signal-sensitive response, and enough stability to prevent unwanted activation.
It does not guarantee a solution. It gives the search a direction.
That is why I think Spread the Medium deserves consideration as a distinct contradiction-resolution principle. Its value is the complete move it asks us to make: separate broad availability from selective action by finding the right responsive medium.
We started with the problem of putting something in exactly the right places.
The unexpected answer was to put a different kind of thing throughout the domain — and postpone the decision about where it should act.
Distribute the capability. Let the signal select the action.
A small family of “Spread the Medium” solutions
Once the principle has a name, inventions that initially look unrelated begin to form a family. They are not mechanically identical, and that is important. In some, an external controller sends the activating signal. In others, the event itself supplies the signal. What repeats is the separation between broad availability and selective realization.
Photolithography — write a pattern that is already known. The photoresist covers the wafer, remains usable before exposure, and changes only where the light pattern reaches it. The signal carries the geometry; the medium turns that geometry into a material difference.
A triggered therapeutic carrier — transport broadly, act conditionally. A carrier can travel through an accessible biological domain while retaining its payload. A selected field or local condition then triggers release. The same entity that solves the transport problem does not have to produce the therapeutic effect everywhere along its route.
A cloud chamber — reveal a pattern that is not known in advance. The sensitive vapor fills the chamber before the particle arrives. The particle creates ionization along its own unknown trajectory, and the medium responds there. Here the signal is not a command; it is the phenomenon being detected.
A dormant computer virus — a harmful use of the same principle. A malicious program may spread across many computers while its destructive payload remains inactive. “Inactive” here applies to the payload, not necessarily to every part of the program: some activity is still required for propagation. The payload is realized only when a narrowly defined condition or external signal appears. To fit the principle strictly, it must remain stable against accidental activation and respond only to its designated trigger. The digital network is the domain, copies of the dormant payload form the distributed medium, and the trigger selects when — and potentially where — the harmful action occurs. This example also shows that the principle itself is ethically neutral; its consequences depend on the purpose for which it is used.
Synthetic-diamond growth — spread the feedstock, localize its conversion. In chemical-vapor-deposition growth, a carbon-containing gas is supplied through the chamber and activated by plasma, while incorporation into diamond occurs at a suitable seed or growth surface. This is a neighboring variant of the principle. The distributed entity is a precursor rather than the finished function, and localization is provided partly by the receptive surface, not only by a projected signal. The deeper move is unchanged: make the resource broadly available, but convert it into the required structure only under selected local conditions.
Self-healing composites — let damage provide its own address. In a classic design, microscopic capsules containing a healing agent are distributed throughout a polymer composite. When a crack reaches them, it ruptures nearby capsules and releases material into the damaged region, where a repair reaction can occur. The designer does not need to predict where the crack will form. Damage itself supplies both the target and the activating signal.
Vat photopolymerization — a three-dimensional extension. A vat contains liquid photopolymer throughout the available build region, while controlled light cures only selected cross-sections. Repetition layer by layer turns distributed liquid possibility into a solid three-dimensional object. This is close to photolithography, but the result is not merely a surface pattern; the selective activation constructs volume.
These examples expose several branches of the same mechanism. Activation may be commanded or event-generated. The spread entity may be the final responsive material, a carrier, or an uncommitted precursor. The domain may be a surface, a volume, a biological transport space, a digital network, or even an abstract set of software contexts.
What remains invariant is the architecture:
Make the capability or resource available throughout the possible target domain; keep it inactive or uncommitted under ordinary conditions; and let a designated local interaction decide where it becomes real.
That is the reason to treat Spread the Medium as more than a collection of examples. It is a repeatable way to resolve the contradiction between being available everywhere and acting only somewhere.
Spread the possibility. Select the realization.
https://nisanovblog.blogspot.com/2023/11/innovation-vs-creativity-or-tactics-vs.html
https://nisanovblog.blogspot.com/2020/06/big-software-contradiction.html
