Means-plus-function claims: spotting § 112(f)

A claim limitation falls under 35 U.S.C. § 112(f) when it describes a function without reciting enough structure to perform it. The word "means" is the classic signal, but since Williamson v. Citrix a generic word such as "module", "mechanism", "element" or "unit" can trigger the statute too. Once it applies, the limitation covers only the corresponding structure described in the specification and its equivalents; for a software function that structure is an algorithm, and if the specification discloses none, the claim is indefinite.

What § 112(f) says

The text of 35 U.S.C. § 112(f) reads:

"An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof."

Before the America Invents Act, the same provision was § 112, paragraph 6, which is why older cases and many briefs still refer to "¶ 6".

The statute is a trade. The drafter may claim by function, and in return the claim is limited to what the specification actually describes for performing that function, plus equivalents. A limitation that reads broadly on its face can turn out to be narrow, or invalid, once the specification is examined.

How to spot a § 112(f) limitation

The three-prong test

MPEP 2181 gives examiners a three-prong analysis, and it works as a screening checklist for anyone reading claims:

1. (A) The limitation uses "means" or "step", or a generic placeholder (a nonce term) in their place.
2. (B) The term is modified by functional language.
3. (C) The term is not modified by sufficient structure, material or acts for performing the function.

All three must be present. "Means for fastening, comprising a threaded bolt and a nut" meets (A) and (B), but the recited bolt and nut may supply enough structure to fail (C).

Nonce words

Words that commonly act as placeholders include "module", "mechanism", "element", "unit" and "device for". Watch for functional prefixes: "scoring module" tells a reader what the module does, not what it is, and that is the distinction prong (C) asks about.

What Williamson changed

In Williamson v. Citrix Online, LLC, 792 F.3d 1339 (Fed. Cir. 2015), the Federal Circuit, sitting en banc on this point, overruled the "strong" presumption that a limitation without the word "means" is not subject to § 112, ¶ 6. The test is whether the words of the claim are understood by persons of ordinary skill in the art to have a sufficiently definite meaning as the name for structure. The court held that "module" in the claim before it was a nonce word that did not connote structure, so the statute applied.

The absence of "means" still counts, but it no longer comes close to settling the question. Every functional limitation built around a generic noun deserves the three-prong check.

Steps as well as means

The statute also covers a "step for performing a specified function without the recital of ... acts". Most method steps recite the acts themselves ("sending a request", "retrieving additional information") and are ordinary limitations. A step written purely as a result to be achieved is the one to flag.

Corresponding structure for software is an algorithm

For a mechanical claim, corresponding structure is a part you can point to in a drawing. For a computer-implemented function, it is harder. Aristocrat Technologies Australia Pty Ltd. v. International Game Technology, 521 F.3d 1328 (Fed. Cir. 2008) held that disclosing a general-purpose computer is not enough; the specification must disclose the algorithm that performs the function. Williamson reached the same result: with no corresponding algorithm disclosed for the computer-implemented function, the claim was indefinite.

In practice, an algorithm can appear as a mathematical formula, a flowchart, or a step-by-step description in prose. Practitioners look for a passage that explains how the function is carried out and that the specification clearly ties to the claimed function. A passage that restates the function ("the module scores the documents") is not an algorithm.

The indefiniteness risk

Section 112(b) requires claims that particularly point out and distinctly claim the subject matter of the invention. Under Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898 (2014), a claim is indefinite if, read in light of the specification and prosecution history, it fails to inform those skilled in the art about the scope of the invention with reasonable certainty.

For a § 112(f) limitation, the risk is concrete. If no corresponding structure is disclosed, there is nothing for the claim to cover, and the claim fails. That is why checking independent claims for nonce words is an early step in many validity reviews of software patents. It also shapes the claim's scope well before any validity challenge, because claim construction for that limitation becomes a search of the specification for structure.

A worked example

Start from claim 1 of US 6,285,999 B1, "Method for node ranking in a linked database" (filed 9 January 1998, issued 4 September 2001, expired). Its limitation 1[b] is a method step: "assigning a score to each of the linked documents based on scores of the one or more linking documents". Now imagine a hypothetical system claim that recasts the same idea as "a scoring module for assigning a score to each of the linked documents based on scores of the one or more linking documents".

Run the three prongs on the hypothetical text:

• (A) "Module" is a generic placeholder, as in Williamson.
• (B) "For assigning a score..." is functional language.
• (C) Nothing in the phrase describes how the scores are assigned. "Scoring" names the function, not a structure.

The limitation would likely be treated under § 112(f). The scope question then moves to the specification that supports the claim: what algorithm does it disclose for assigning those scores? If it gives a computation, that computation and its equivalents define the limitation. If it says only that a computer assigns scores, the claim is exposed to indefiniteness.

How to chart a § 112(f) limitation

A § 112(f) row needs more than evidence and a finding. A common approach:

1. Flag the row and record why: the term, the function, and the missing structure (prongs A to C).
2. State the function exactly as the claim recites it. Do not paraphrase it into something broader.
3. Identify the corresponding structure in the specification, with column and line or paragraph locators. For software, quote or summarize the algorithm and every disclosed variant.
4. Identify the accused structure that performs the claimed function, with evidence. The accused product has to perform the function the claim recites.
5. Compare structures. Is the accused structure the same as a disclosed one, or an equivalent of it? Record the reasoning either way.
6. Keep statutory equivalents separate from the doctrine of equivalents. Equivalents under § 112(f) are part of the literal analysis of that limitation. The doctrine of equivalents is a separate route with its own limits.

Against a hypothetical intranet search appliance, the row might look like this.

Read the row the way you would read any other row of an element-by-element claim chart: evidence first, finding last.

Try it

One Transform's Infringement search parses your patent's claims into individual limitations and charts each candidate product limitation by limitation, with a finding of present, partially present or not found and evidence for every element. For a limitation you have flagged under § 112(f), treat that evidence as the starting point for the accused-structure step above, then compare it against the algorithm in your specification.

Next step: run an Infringement search on a software patent.

This article is general information about patent practice, not legal advice. For advice on a specific matter, consult a registered patent attorney or agent.