Open a granted patent at the claims and you will usually find between fifteen and twenty-five numbered paragraphs. Two or three of them stand alone. The rest begin by pointing back at one of the first group. The split between independent and dependent claims is not a formatting convention — it is the entire risk management strategy of the document.
01 / Anatomy
The Claim That Has to Stand on Its Own
An independent claim recites everything necessary to define the invention without borrowing a single word from anywhere else in the document. Read in isolation, it must still describe a complete, workable thing. Nothing in the drawings rescues it, nothing in the description fills a gap in it, and a reader who has never seen the rest of the specification must be able to identify what falls inside it and what does not.
The practical consequence is length. A well-drafted independent claim commonly runs between eighty and two hundred words as one grammatical sentence, because every element and every relationship between elements has to be stated explicitly. It is also the reason the all-elements principle bites so hard: for a competing product to fall inside the claim, every listed element must be present in it. Omit one and there is no infringement, however similar the rest of the device looks. The same arithmetic runs in reverse during examination, which is why a single earlier document containing all of the listed elements destroys the claim outright.
Most applications carry more than one independent claim because different categories of claim catch different behaviour. An apparatus claim covers a thing that is made, sold or imported. A method claim covers a sequence of steps and reaches a party who performs them — a manufacturer running a process, or an operator using equipment in a particular way. A system claim covers a combination of cooperating components that may never be assembled in one place. A competitor who cleanly avoids one of those categories has often walked straight into another.
02 / Dependency
What a Dependent Claim Inherits, and What It Adds
A dependent claim opens by referring back: "The dispenser of claim 1, wherein the valve seat comprises a moulded elastomer." That opening phrase silently imports every limitation of claim 1 into the new claim, and then adds one more. This is the single fact that makes the structure work, and it is also the one most often misread. A dependent claim is always narrower than the claim it refers to. It cannot widen anything. Nothing written into claim 8 can recover territory that claim 1 failed to reach.
Dependents come in two shapes. The first adds an element that was absent from the parent — a second seal, a return spring, a further processing step. The second narrows an element the parent already recited, by specifying its material, its geometry, a numerical range, or a manufacturing route. Both are legitimate. What is not legitimate is a dependent claim that contradicts or removes a limitation of its parent; that draws a formal objection on its own, before any question of prior art has been reached.
The two directions in which the structure is read run opposite ways, and holding both in mind is the whole skill. On infringement, nobody can fall inside claim 8 without also falling inside claim 1, so the narrow claims add no reach at all against a copyist. On validity, the traffic reverses: claim 1 can be knocked out by an earlier publication while claims 5, 8 and 14 remain perfectly sound, because each of them contains something that publication does not describe. Independent and dependent claims therefore serve two completely different masters — one is written for the day a competitor appears, the other for the day an examiner does.
- Indep.Stands alone, recites every essential element, defines the outer boundary of the patent.
- Dep.Imports its parent in full and adds one restriction. Always narrower, never broader.
- Multi.Refers to several earlier claims at once. Permitted in many territories, surcharged or restricted in others.
A dependent claim can never widen the patent. Its entire purpose is to still be standing when the wider one is taken away.
Why the narrow rungs are drafted at all
03 / Architecture
Designing the Ladder Before Writing the Rungs
The useful way to plan a set of independent and dependent claims is to decide what each rung is for before drafting any of them. Claim 1 states the minimum arrangement that delivers the effect. Every dependent claim beneath it should answer one of two questions: what limitation might an examiner insist on, or what variation might a competitor adopt to step around the wording? A dependent claim that answers neither is decoration, and decoration is charged for at the same rate as substance.
Ordering matters more than it appears to. The limitations most likely to be needed — the ones drawn from the closest known prior art, or from the feature the inventor believes is genuinely new — belong near the front of the sequence, because those are the claims an examiner reads properly and the ones most likely to be folded upward into claim 1 later. Somewhere in the set there should also be a claim reading exactly onto the product actually being sold, in the materials and dimensions it is actually made from. That claim is rarely commercially valuable on its own, but it is the position of last resort, and applications that lack one tend to end in an uncomfortable choice between an unsupported argument and abandonment.
All of this depends on the specification containing the raw material in the first place. Nothing can be written into a dependent claim during examination unless it was disclosed at filing, which is why a description that records alternative materials, tolerance bands and rejected variants is worth so much more than a clean account of one embodiment. Practical guides to the wider journey make the same point in commercial terms; a step-by-step account of how an invention idea is carried through to a finished product reads as a long series of recorded alternatives rather than a single decision. The sequence that turns a rough concept into a first claim is where that raw material is gathered.
04 / Carrying Cost
What Each Extra Rung Costs to Keep
Claims are not free, and the pricing is deliberately shaped to discourage sprawl. A common pattern across granting offices is a base filing fee that covers roughly twenty claims in total and about three independent ones, with a per-claim surcharge above either threshold. The surcharge for extra independent claims is typically several times the surcharge for extra dependents, and in many territories the charge is levied again on amendment, so a set that grows from twenty to thirty-two claims during examination is billed for the increase.
Translation is the cost that surprises people. Where an application is pursued in territories requiring a full translation of the text, the work is usually priced per word. Thirty-five dependent claims, each running forty words, is fourteen hundred words of claim text repeated in every such filing — a figure that multiplies quietly across a family and then recurs at each renewal cycle over the twenty-year term. Against that, a set of fifteen to twenty well-chosen claims does almost all of the defensive work of a set of forty.
The discipline that follows is straightforward: every claim in the set should be nameable. If nobody can say in a sentence what a particular dependent claim is protecting against, it is padding. Products with long development histories illustrate what a purposeful set looks like — the record behind a vibration-based pain relief device brought to market shows protection tracking the features that mattered commercially rather than every variation the drawings could support.
05 / Graceful Failure
How the Structure Is Meant to Break
Here is the ordinary sequence. An examination report arrives and cites a publication that describes everything in claim 1. The report goes on, claim by claim, and concedes that claims 6, 9 and 14 are not described anywhere in the cited material. The response takes the limitation from claim 9, writes it into claim 1, cancels claim 9 as redundant, and the application proceeds. The patent that grants is narrower than the one applied for, but it exists, and the narrowing was chosen from a menu the applicant wrote themselves rather than dictated by an examiner.
Now run the same report against an application whose entire disclosure sits in one long claim with no fallbacks. There are two options: argue that the cited document does not say what it plainly says, or abandon. Neither is a strategy. The difference between those two outcomes was decided at drafting, months or years before the report was written, and cost nothing at the time but attention.
The pattern continues after grant. Validity is assessed claim by claim, so a challenge that succeeds against the broadest claim leaves the narrower ones intact and independently enforceable. A claim set is best understood as a portfolio of positions of decreasing ambition, and the mechanics of trading between them are examined further in this account of how narrowing is negotiated once a report has issued. The inventors' own account of how a household mixing device came together describes the same shape from the inside: a long sequence of small decisions, each of which later had somewhere to retreat to.
Write the fallbacks while nothing is going wrong. They are unavailable the moment you actually need them.
The economics of the claim ladder
One claim defines how far the patent reaches; the rest decide whether it survives being tested. Both jobs have to be done at drafting, because only one of them can be repaired later.
End of report