Patent Knowledge5 min read

The Doctrine of Equivalents Trap: Why Escaping the Literal Claim Can Still Mean Infringement

The most dangerous design-around mistake is treating 'outside the literal scope' as safe. The doctrine of equivalents exists precisely for feature substitutions — understand the function-way-result test to know which design-arounds truly escape and which merely rename the same route.


A logic that surprises engineers

The engineer's intuition: the claim says A, we use B, A ≠ B, therefore no infringement.

Patent law's answer: if B is an insubstantial change from A — performing substantially the same function in substantially the same way with substantially the same result, such that an ordinary artisan could conceive it without inventive effort — then B is A's equivalent, and the product still falls within the protected scope.

That is the doctrine of equivalents. Its rationale is plain: if protection stopped at the literal wording, anyone could escape by trivial changes and claims would be ornamental. The US Supreme Court's phrasing in Graver Tank remains the most quoted — an imitator must not be allowed to steal an invention's substance through "unimportant and insubstantial changes."

For design-around work, equivalents is not an obstacle. It is the quality gate: a "design-around" that cannot survive an equivalents self-check was never a design-around — it was the same route with a new name.

Using the three-prong test: put your weight on "way"

Whether under the US function-way-result test or China's "three basically identical" standard, the prong with real operating room is the way:

  • Function usually cannot change — your product exists to deliver that function;
  • Result must largely hold — a "design-around" with visibly worse results defeats itself;
  • Way is the only dimension both changeable and worth changing. A substantially different technical means delivering the same function is the core proposition of design-around.

Some intuition pairs:

| Original | Substitute | Gut reading | |---|---|---| | Current sensing to detect load | Current sensing with a 10% higher threshold | Substantially the same — a change in degree | | Current sensing to detect load | Hall-sensor RPM differential to detect load | Substantive argument available — different physics | | Bolted fastening | Screwed fastening | Substantially the same — textbook equivalence | | Mechanical reversal to eject debris | Axial vibration to convey debris out | Substantive argument available — different kinematics |

Note the wording: "substantive argument available," not "safe." Equivalence ultimately turns on claim construction, the specification, the prosecution history and case-by-case adjudication — which is exactly why counsel exists.

Three boundary rules often overlooked

Equivalents does not stretch without limits. Its counter-constraints are where design-around opportunities live:

1. Prosecution history estoppel

Narrowing amendments or arguments the patentee made during prosecution to obtain the grant cannot be recaptured through equivalents in litigation. A limitation that was forced narrower during examination usually has a much shorter equivalence reach. (We have a companion article on mining the file wrapper for these openings.)

2. The dedication rule

Subject matter disclosed in the specification but not claimed is dedicated to the public — the patentee cannot fish it back via equivalents. When reading a competitor patent, the specification's "alternatively, … may be used" embodiments deserve their own highlight.

3. The All-Elements floor

Equivalence is judged element by element, not by overall product resemblance. A missing element with no equivalent means no infringement — this is what keeps design-around fundamentally viable.

Make the self-check a process, not a feeling

The most common failure mode in practice: the team shares a collective feeling that "we changed enough," but nobody ever wrote it down element by element. The reusable discipline:

  1. For each design direction, state explicitly which element of which claim it substitutes;
  2. Build a three-row function-way-result comparison for that substitution; the "way" row must articulate a mechanism-level difference — physical, chemical, or algorithmic;
  3. Directions that cannot articulate a mechanism difference are discarded or flagged borderline — do not carry wishful thinking into R&D;
  4. Hand the survivors, tables included, to patent counsel — that record makes the formal FTO dramatically more efficient.

We built this self-check into the product as a mandatory stage: every candidate direction is fed back into the equivalents engine for the three-prong test; candidates judged equivalent are sent back for regeneration; borderline cases are explicitly labeled "counsel review required." Nothing gets packaged as "safe."

Try Smart Patent's Design-Around Studio

Blocked by a competitor patent? Break down its claims and get a different-route product design reference (free related-patent scan first)

Try Free

One honest closing line

Because the doctrine of equivalents exists, there is no such thing as "software-guaranteed non-infringement" — any tool promising that is lying. What an honest tool can do is make the three-prong test unskippable, label the borderline cases truthfully, and send you to counsel with screened directions and a complete self-check record. That beats gut feeling, and it costs far less than going in blind.

Frequently Asked Questions

What is the test for equivalence?

US practice applies the function-way-result test: substantially the same function, in substantially the same way, achieving substantially the same result. Chinese practice uses the closely-related 'three basically identical plus obviousness' standard. The prong with real room for argument is 'way' — a substantially different way is the core goal of any design-around.

Does nudging a claimed numeric range slightly outside count as designing around?

Literally yes, under equivalents extremely risky. Replacing a 0.1–0.5 mm limitation with 0.55 mm keeps function and result nearly identical, and the way differs only in degree — a textbook equivalence scenario. A real design-around changes the implementing mechanism, not the decimals at the boundary.

Are there substitutions naturally harder for equivalents to reach?

Several: elements the patentee narrowed during prosecution to obtain the grant (estoppel); alternatives disclosed in the specification but not claimed (dedication rule); and after-arising technology (treated differently across jurisdictions, case-by-case). All of these require reading the prosecution history — counsel's work.