The doctrine of equivalents lets a patent claim reach a product that does not literally fall within its words. It exists for a straightforward reason: without it, a competitor could avoid patent infringement by substituting one obvious equivalent for a claimed element and taking the invention wholesale.
It is also narrower in practice than patentees expect. Three separate limitations constrain it, and an accused infringer will reach for all of them.
The two formulations
Courts use two overlapping tests, and which one appears depends largely on the technology.
The function-way-result test asks whether the accused element performs substantially the same function, in substantially the same way, to achieve substantially the same result as the claimed element. It comes from Graver Tank and works naturally for mechanical inventions, where function and mechanism are easy to articulate.
The insubstantial differences test asks simply whether the difference between the claimed element and the accused element is insubstantial. Warner-Jenkinson confirmed this as the underlying inquiry, with function-way-result as one useful way of framing it. Chemical and electronic cases tend to use this formulation, because "the same way" is often meaningless where two compounds achieve a result by different mechanisms.
Both are assessed from the perspective of a person of ordinary skill in the art, at the time of infringement rather than at filing. That timing matters — after-arising technology can be an equivalent even though nobody could have claimed it when the application was drafted.
The all-elements rule
The doctrine applies element by element, not to the invention as a whole.
Warner-Jenkinson v. Hilton Davis (1997) settled this, and it is the limitation that defeats most equivalents arguments before the substantive test is ever reached.
| Claim element | Accused product | Result |
|---|---|---|
| A | A present literally | Satisfied |
| B | B present literally | Satisfied |
| C | Substitute performing the same role | Potentially equivalent |
| D | Absent entirely | No infringement |
An accused product missing a claimed element does not infringe, even where it achieves exactly the same overall result by a different route. The patentee cannot argue that the invention as a whole is substantially the same; the comparison is limitation by limitation.
This is why the claim chart is built element by element in both literal and equivalents analysis. The chart makes the missing element visible immediately, which is the point.
Prosecution history estoppel
The principal limit, and the first thing an accused infringer looks for.
The rule. Claim scope surrendered during prosecution to obtain the patent cannot be recaptured through the doctrine of equivalents. An applicant who narrowed a claim to overcome prior art gave up the territory between the original and amended language, and cannot reclaim it in litigation.
The rationale is fairness to the public. Competitors read the file history to determine what the patent covers. If an applicant could narrow a claim to secure allowance and then assert the surrendered scope through equivalents, the public notice function of the file history would be worthless.
Amendment-based estoppel
Any narrowing amendment made for a reason substantially related to patentability creates a presumption of surrender. Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki (2002) established this, replacing an earlier absolute bar with a rebuttable presumption.
The presumption can be rebutted three ways:
| Rebuttal | What must be shown |
|---|---|
| Unforeseeability | The equivalent was not foreseeable at the time of amendment |
| Tangential relation | The rationale for the amendment bore no more than a tangential relation to the equivalent |
| Some other reason | The applicant could not reasonably have been expected to describe the equivalent |
Unforeseeability is the most workable in practice, particularly for after-arising technology. An equivalent that did not exist when the amendment was made could not have been described.
Tangential relation is narrow. It requires showing that the reason for the amendment had nothing much to do with the equivalent now asserted, which is a harder argument than it sounds because the reason is usually documented in the response.
Argument-based estoppel
Statements to the examiner can surrender scope even without an amendment.
An applicant who distinguishes prior art by arguing that their invention does not do X has told the public that X is outside the claim. Asserting equivalence to X later contradicts that.
This catches practitioners who over-argue. Distinguishing a reference more broadly than necessary is a habit that costs nothing during prosecution and can cost the case years later. The discipline is to argue no more broadly than the rejection requires.
Ensnarement: no recapturing the prior art
A claim cannot be extended by equivalents to cover what was already known.
The test is hypothetical. Would a claim drafted broadly enough to literally cover the accused product have been allowable over the prior art? If not, the equivalents theory ensnares the prior art and fails.
This is a question for the court, not the jury, and it operates as a backstop. A patentee whose literal claim is narrow enough to have survived examination cannot use equivalents to reach a scope that would have been rejected.
A worked example
A claim to a fastening mechanism reciting, among other elements, "a threaded bolt". The accused product uses a snap-fit connector.
Literal infringement
No. A snap-fit connector is not a threaded bolt. The analysis stops there for literal infringement.
The equivalents question
| Test element | Threaded bolt | Snap-fit connector | Substantially the same? |
|---|---|---|---|
| Function | Secures two components | Secures two components | Yes |
| Way | Threaded engagement under tension | Interference fit, elastic deformation | Arguably not |
| Result | Components held together | Components held together | Yes |
Function and result match; the way differs. That is exactly the fact pattern the function-way-result test was designed to resolve, and it usually resolves against equivalence — a mechanism that works on an entirely different physical principle is not doing it "in substantially the same way".
Now add the file history
Suppose the original claim recited "a fastener", and the applicant amended to "a threaded bolt" after the examiner cited prior art disclosing a snap-fit connector.
The equivalents argument is dead. The applicant narrowed from "fastener" to "threaded bolt" for a reason directly related to patentability, and the surrendered territory is precisely the snap-fit connector now accused. Festo presumption applies and none of the three rebuttals is available — the equivalent was not merely foreseeable, it was the reference that prompted the amendment.
Change one fact and the outcome changes. If the amendment had been made to overcome a rejection about a different element entirely, the tangential relation rebuttal becomes arguable.
This is why the file history is read before the equivalents analysis, not after. The strongest equivalents theory in the world is worthless against a narrowing amendment that surrendered exactly the accused feature.
Where it fits in an infringement analysis
| Step | Question | If yes | If no |
|---|---|---|---|
| 1 | Every element literally present? | Literal infringement | Continue |
| 2 | Any element missing entirely? | No infringement | Continue |
| 3 | Is the difference insubstantial? | Potentially equivalent | No infringement |
| 4 | Was the scope surrendered in prosecution? | Estoppel — no infringement | Continue |
| 5 | Does the theory ensnare the prior art? | No infringement | Infringement by equivalents |
Equivalents is a fallback, not a primary theory. Most infringement cases are decided on literal infringement, and the doctrine functions as a second line that frequently fails at step four.
Why it succeeds less often than patentees expect
Three independent defences, any one of which is fatal. The all-elements rule, prosecution history estoppel, and ensnarement each defeat the theory independently, and an accused infringer will plead all three.
The file history is public and permanent. Every amendment and every argument made during prosecution is available to anyone assessing the patent, which means the estoppel analysis can be done before the dispute begins.
Which has a practical consequence during prosecution. Every narrowing amendment is a trade — allowance now against equivalents scope later. Practitioners who amend broadly to move an application along are spending something that will not be visible until litigation.
Amend as narrowly as the rejection requires. Argue precisely rather than expansively. Both cost time during prosecution and both preserve options that may matter a decade later.
Prosecution history estoppel in practice
Read the file history first, always. In any infringement analysis — whether you are asserting or defending — the file history tells you what equivalents are actually available before any substantive analysis is worth doing.
What to look for:
- Every claim amendment and the rejection it responded to
- Arguments distinguishing prior art, particularly broad characterisations
- Statements about what the invention is not
- Terminal disclaimers and their effect on family scope
- Interview summaries, which sometimes contain the decisive statement
The response that secured allowance is the most important document in the file history. Whatever the applicant said or did to overcome the final rejection is where the surrender usually lives.
Before relying on an equivalents theory
- Confirm every claim element is present literally or by equivalent — the all-elements rule defeats most theories first.
- Read the entire file history, not just the amendments.
- Identify which limitation you are relying on equivalence for, precisely.
- Check whether that limitation was narrowed during prosecution and why.
- Test the ensnarement question — would a hypothetical broader claim have been allowable?
- Choose your formulation — function-way-result for mechanical subject matter, insubstantial differences elsewhere.
- Treat it as a fallback, and build the literal case first.