What is the difference between an independent and a dependent patent claim? An independent claim is self-contained: it recites a full set of limitations and does not refer to any other claim, so its scope is read entirely from its own words. A dependent claim, by contrast, refers back to another claim in the same patent and adds at least one further limitation. Because the dependent claim incorporates everything in the claim it refers to and then adds more, it is necessarily narrower than its parent. That asymmetry is the key to reading a patent quickly: the independent claims tell you the broadest thing the patent reaches, and each dependent claim is a narrower fallback that carves out a more specific embodiment.

The structure is set by 37 CFR 1.75, the rule on the claim or claims in a patent application. Its text on dependent claims reads:

One or more claims may be presented in dependent form, referring back to and further limiting another claim or claims in the same application. Any dependent claim which refers to more than one other claim (“multiple dependent claim”) shall refer to such other claims in the alternative only. A multiple dependent claim shall not serve as a basis for any other multiple dependent claim.— 37 CFR § 1.75(c), source

Three rules live in that passage. First, a dependent claim must be "referring back to and further limiting another claim" — referring back without adding a limitation does not make a proper dependent claim; the further limitation is required. Second, a claim that refers to more than one other claim is a "multiple dependent claim" and must refer to those other claims "in the alternative only" (claim 5 as in claim 1 or claim 2, not claim 1 and claim 2 simultaneously). Third, a multiple dependent claim cannot itself be the basis for another multiple dependent claim — the rule forbids chaining multiple dependencies. These constraints shape how a claim set is drafted and counted, and they are why the regulation, alongside 35 U.S.C. 112, is the authority on claim form.

Why the distinction drives scope and validity

The independent/dependent split has two practical consequences a reader cares about: scope and validity. On scope, the broadest reach of a patent is always in its independent claims, because every dependent claim adds a limitation and therefore narrows. If a product avoids an independent claim, it avoids all of that claim's dependents automatically, since the dependents include every limitation of the parent plus more. That is why a freedom-to-operate read starts at the independent claims and only descends to the dependents to understand specific embodiments. On validity, the relationship runs the other way: a dependent claim can survive even if its parent independent claim is held invalid, because the added limitation may distinguish prior art that the broader claim did not. The statute on the presumption of validity, 35 U.S.C. 282, expressly provides that dependent claims are presumed valid even though dependent upon an invalid claim — so a narrower dependent is a hedge against the broader claim falling.

How claim counting and fees follow the form

The independent/dependent distinction also drives the mechanics of how claims are counted and charged, which is part of why the form is regulated. The USPTO's fee structure counts independent claims and total claims separately, with additional fees once an application exceeds set numbers of each, and multiple dependent claims are counted as the several claims they reference. This is not a trivia point: it shapes how applicants draft. A claim set is often built as a small number of independent claims, each followed by a ladder of single-dependent claims that progressively add limitations, precisely because that structure manages both the fee count and the validity hedge. The rule that a multiple dependent claim must refer to other claims "in the alternative only" and cannot serve as the basis for another multiple dependent claim — straight from 37 CFR 1.75 — exists to keep claim dependencies from branching combinatorially, which would make both examination and counting unworkable. For a reader auditing a claim set, the shape of the dependency tree (how many independents, how deep the dependent ladders run, whether any multiple dependent claims appear) is itself information about how the applicant structured the application under these constraints.

Reading a real claim set

The structure is visible on any granted patent. United States Patent US10672707B2, "Low aspect ratio interconnect," is a real granted semiconductor patent whose claim set follows this pattern: one or more independent claims reciting the full interconnect structure, followed by dependent claims that refer back and add limitations — a particular material, a dimensional relationship, an additional layer. To gauge what the patent broadly covers, a reader parses the independent claims element by element; to understand the patentee's fallback positions and the specific embodiments emphasized, the reader then walks the dependents. The independent claim is the wall; the dependents are the reinforcements behind it.

What the record shows is the regulatory and statutory rule on claim form. Under 37 CFR 1.75, a dependent claim refers back to and further limits another claim in the same application; a multiple dependent claim must refer to other claims in the alternative only and cannot serve as the basis for another multiple dependent claim. Because dependent claims add limitations, they are narrower than the claims they depend from, so independent claims define the broadest scope and dependents define narrower fallbacks. And under 35 U.S.C. 282, dependent claims are presumed valid even if their parent independent claim is invalid. The rules do not say which claims are infringed or valid in any case — that turns on the specific claim language, the accused product, and the prior art — but they fix the form, and the form is what lets a reader navigate a semiconductor claim set from broadest to narrowest.