MPEP 2181: Identifying 35 U.S.C. 112(f) Claim Limitations

MPEP 2181 is the section of the Manual of Patent Examining Procedure that tells patent examiners how to identify a means-plus-function or step-plus-function claim limitation under 35 U.S.C. 112(f), and how to interpret it once identified. The short version: a limitation written as a function without reciting the structure that performs it gets construed to cover only the structure, material, or acts disclosed in the specification, plus equivalents, rather than everything that could perform the function.¹ That tradeoff, drafting convenience in exchange for narrower scope, runs through everything the section covers.

The Three-Prong Test

An examiner uses a three-prong test to decide whether a claim limitation triggers 112(f). All three must be satisfied:¹

Prong A: the limitation uses “means,” “step,” or a generic placeholder term that lacks specific structural meaning.
Prong B: that term is modified by functional language, usually connected by “for,” “configured to,” or “so that.”
Prong C: the term is not modified by enough structure, material, or acts to perform the claimed function on its own.

A limitation reciting “means for amplifying a signal” with no further structural detail satisfies all three prongs. It gets read under 112(f), and its scope is confined to whatever structure the specification identifies as performing the amplification, plus equivalents. Change the language to “a transistor amplifier for amplifying a signal” and Prong C fails, because the claim itself names a known structure. That version is treated as ordinary claim language with its full ordinary meaning.

When the Claim Doesn’t Use the Word “Means”

Whether the word “means” appears sets up a rebuttable presumption. Using “means” presumes 112(f) applies. Omitting it presumes 112(f) does not apply.¹ Neither presumption is the end of the analysis.

Before 2015, the presumption against 112(f) when “means” was absent was treated as “strong,” which made it easy for applicants to sidestep 112(f) by replacing “means” with a vague substitute. The Federal Circuit eliminated that heightened standard in Williamson v. Citrix Online, LLC.² The presumption against 112(f) is now overcome whenever the claim term fails to recite sufficiently definite structure, or recites function without reciting sufficient structure to perform it. The test is whether a person of ordinary skill in the relevant field would understand the words as the name for a definite structure.¹

The practical result is that examiners scrutinize substitute terms more closely and are willing to apply 112(f) even when “means” is nowhere in the claim. Common placeholder words that often trigger the analysis include “mechanism for,” “module for,” “unit for,” “device for,” “member for,” “apparatus for,” and “element for.”¹ Context controls. “Filter” can be structural in one technical setting and vague in another. “Circuit for” tends to name a structure; “system for” often does not. What matters is whether the term, read with the rest of the limitation, conveys a definite structure to someone skilled in that field.

Step-Plus-Function in Method Claims

The analysis is not limited to apparatus claims. A method claim that uses “step for” performing a function without reciting the specific acts that accomplish it is a step-plus-function limitation, and the specification must then disclose the corresponding acts.³

Ordinary method steps do not become 112(f) limitations just because they use a gerund. “Heating,” “filtering,” or “transmitting” recite actual acts. The trigger is “step for” used to describe a result without saying how to reach it. Language like “steps of” or “step” alone tends to show 112(f) does not apply.¹ The underlying function corresponds to what the element accomplishes; the acts correspond to how it is accomplished. A step-plus-function limitation states the “what” and leaves the “how” to the specification.

What the Specification Must Disclose

Once a limitation invokes 112(f), the specification has to hold up its end. MPEP 2181 sets three requirements for adequate disclosure:¹

The corresponding structure, material, or acts must be disclosed so that someone skilled in the field will understand what performs the recited function.
That disclosed structure must be capable of performing the entire claimed function, not just part of it.
The specification must clearly link or associate the disclosed structure to the claimed function.

The MPEP describes clear linkage as the quid pro quo for the convenience of using functional claim language. Broad-sounding claim language is allowed, but the inventor has to pin it to specific structures in the specification. Linkage can be established through explicit statements, consistent terminology, or cross-references between text and drawings. A statement that “known techniques can be used” is not enough.¹ The linkage can be implicit if a skilled reader would understand which structure performs the function, but vague hand-waving does not qualify.

When a single means-plus-function term is tied to multiple functions, the specification must disclose structure capable of performing all of them. If separate structures handle separate functions, each pairing needs its own clear linkage.

Computer-Implemented Claims and the Algorithm Requirement

This is where most applications get into trouble. For a computer-implemented means-plus-function limitation, saying “a processor configured to perform X” is almost never enough on its own. The corresponding structure is not the general-purpose computer, it is the special-purpose computer as programmed to carry out the disclosed algorithm.¹

The specification has to disclose the algorithm in some understandable form. A flowchart, a mathematical formula, prose steps, or any other description that shows the sequence of operations will do. Reciting “software” without saying what the software actually does step by step is inadequate. Only in the rare case where any general-purpose computer without special programming can perform the function does the algorithm requirement drop out.¹

Missing algorithmic disclosure triggers a rejection under 35 U.S.C. 112(b) for indefiniteness. The same defect usually carries a written-description rejection under 112(a), because a claim that lacks corresponding structure also lacks adequate written description. Digging out of that double rejection during prosecution is hard, and often impossible, if the application as originally filed does not already contain the algorithmic detail needed to support an amendment.

Indefiniteness When Structure Is Missing

A claim that invokes 112(f) with no corresponding structure in the specification is indefinite under 35 U.S.C. 112(b).⁴ The definiteness standard is whether a person of ordinary skill in the art can determine the scope of the claim with reasonable certainty.⁵ Without structure to anchor the functional language, the claim has no boundaries to evaluate.

Indefiniteness also arises when the specification discloses some structure but never clearly ties it to the claimed function. A specification that describes several components without identifying which one performs the recited function leaves both the examiner and the public guessing. The link does not need to appear in a specific format, but it has to be discernible to a skilled reader.

Scope and How to Respond

A claim construed under 112(f) is generally narrower than a structurally drafted claim. It covers the structure disclosed in the specification and equivalents of that structure, not every possible structure that could perform the function.³ The USPTO notes that 112(f) limitations will in some cases receive a more narrow interpretation than limitations not written in means-plus-function format.¹ Equivalence for 112(f) purposes is a structural inquiry built into claim construction, asking whether an alternative structure performs the same function in the same way to achieve the same result as the disclosed structure. It is distinct from the doctrine of equivalents used in infringement analysis.⁶

When an examiner determines that a limitation invokes 112(f), an applicant who disagrees has two options:¹

Argue that the claim term already recites enough structure to perform the claimed function, so 112(f) should not apply.
Amend the claim to add structural language that avoids triggering 112(f).

An applicant who is comfortable with 112(f) treatment can instead focus on confirming that the specification adequately discloses corresponding structure with clear linkage. That path accepts the narrower scope but avoids the indefiniteness rejection. For computer-implemented claims it means confirming the specification contains a sufficiently detailed algorithm. For mechanical or electrical claims it means verifying that a specific component or configuration is tied to the claimed function. The choice is strategic: structural claims are usually broader, but functional claims sometimes describe an invention more naturally by what it achieves. The risk with 112(f) is that a competitor who uses a fundamentally different structure to reach the same result may fall outside the narrower scope.

Pre-AIA Terminology

Older patents, case law, and MPEP references cite “35 U.S.C. 112, sixth paragraph.” The America Invents Act reorganized the statute from numbered paragraphs to lettered subsections, and the sixth paragraph became subsection (f). The substance did not change. The MPEP 2181 guidance applies equally to the pre-AIA sixth paragraph, and examiners use the same three-prong test and disclosure requirements regardless of which version governs a given application.¹

1
United States Patent and Trademark Office. Identifying and Interpreting a 35 U.S.C. 112(f) or Pre-AIA 35 U.S.C. 112, Sixth Paragraph Limitation
2
UC Berkeley Law. Williamson v. Citrix Online, LLC, 792 F.3d 1339 (Fed. Cir. 2015)
3
Office of the Law Revision Counsel. 35 U.S. Code 112 – Specification
4
United States Patent and Trademark Office. Related Issues Under 35 U.S.C. 112(a) or (b) and Pre-AIA 35 U.S.C. 112
5
United States Patent and Trademark Office. Manual of Patent Examining Procedure Section 2173
6
United States Patent and Trademark Office. Relationship to the Doctrine of Equivalents