Of the failures that quietly disable an application, this is the one inventors treat as social rather than legal. A colleague helped; a friend built the first prototype over a weekend; an investor pushed the project forward when it had stalled. Adding their name feels generous and costs nothing. It costs a great deal, and so does leaving off somebody who belongs there.
Inventorship is a determination made against a defined legal standard, and the granting office takes the applicant's word for it at filing. Nobody verifies the list. The verification happens later, usually when an adversary has a reason to look — during a licence negotiation, an ownership dispute, or an attempt to enforce the patent against a competitor. At that point an error is not merely embarrassing.
§ 01 Standard
Conception Is the Test, and Nothing Else Is
An inventor is a person who contributed to conception: the formation in the mind of a definite and permanent idea of the complete and operative invention, sufficiently specific that a person of ordinary skill could reduce it to practice without further inventive input. The word doing the work is complete. A vague wish for a device that solves a problem is not conception. A specific mechanism for solving it is.
Everything downstream of conception falls outside the test, however much labour it involved. A person who builds a prototype to the inventor's instructions is exercising ordinary skill, not conceiving. A person who runs the test programme, refines tolerances, sources materials, funds the work, identifies a market, suggests a name, or manages the project is not, on those facts, an inventor. The distinction is unintuitive precisely because it ignores effort. Two people can spend a year each on a project and only one of them be named.
The reverse case is equally live. Where a machinist, given a functional requirement, solves it by devising a linkage the inventor had not conceived, and that linkage ends up recited in a claim, the machinist contributed to conception of that claim. The invoice they issued for the work is irrelevant. So is the fact that nobody thought of them as an inventor at the time.
§ 02 Granularity
The Determination Runs Claim by Claim
The determination is not made for the invention as a whole. It is assessed for each claim, and a person who contributed to the conception of a single claim — including a dependent claim buried at position fourteen — is a joint inventor of the patent. Joint inventors need not have worked at the same time, need not have contributed equally, and need not each have contributed to every claim.
This granularity has a consequence that catches applicants during prosecution. Claims change. Examiners require amendments, claims are cancelled, and new claims are added within the bounds of the original disclosure. If the only claim a particular person contributed to is cancelled in response to a rejection, that person is no longer an inventor and the list must be corrected. If a claim added during prosecution captures a contribution made by someone unnamed, they must be added. The list is therefore not a filing-day formality but a live record that has to be reviewed each time the claim set moves — which, across a two to four year prosecution, is several times.
Cancel one dependent claim and a named inventor may cease to be one. The list on the front page describes the claims as they stand, not the project as it happened.
On why the inventor list is reviewed at every amendment
§ 03 Consequence
Every Name Added Is a Co-Owner Created
Here is the mechanism that makes generosity expensive. Absent a written agreement to the contrary, each joint inventor holds an undivided interest in the entire patent — not a fractional share of it. Each may make, use, sell and grant non-exclusive licences to the whole invention, without the consent of the others and without accounting to them for a penny of the proceeds, in many territories.
Read that twice, because it is the whole risk. A person added to the list as a courtesy can lawfully license the invention to a direct competitor on any terms they choose, and keep the money. They can also refuse to join an enforcement action, and in a number of systems an action cannot proceed without all co-owners. Meaning an unwilling or absent co-owner can render the patent effectively unenforceable while remaining perfectly free to exploit it themselves.
The remedy is not to falsify the list; it is to fix ownership separately from inventorship. The two are distinct concepts. Inventorship is a factual determination that cannot be assigned or negotiated. Ownership can be, and normally is, by written assignment — executed at the outset, recorded with the granting office, and covering every named inventor including the sole one. Where employees or contractors are involved, obligations to assign should sit in their engagement terms and be signed before the work starts rather than reconstructed afterwards, when the contributor has both leverage and a reason to use it.
§ 04 Repair
Correction Is Available, and the Timing Is Never Yours
Errors in inventorship made without deceptive intent can generally be corrected, both during prosecution and after grant, by petition supported by statements from the affected parties and payment of a fee. Where all parties agree, the process is administrative rather than adversarial and the patent survives intact. That is the good news, and it is genuine: these are among the few defects in an application that remain fixable after the fact.
The difficulty is who initiates the correction and when. In practice the error surfaces because somebody with an interest in the patent's weakness has gone looking for it — a defendant's counsel reading the file history, a prospective licensee's diligence team interviewing former colleagues, or the omitted contributor themselves, who has now noticed a product on sale and taken advice. Correction then happens in the middle of a dispute, at the moment of minimum leverage, and it typically requires the cooperation of the very person who has been left off. Where deceptive intent is alleged, the exposure is worse than a correction fee, because a finding of inequitable conduct can render the whole patent unenforceable.
Certain patterns generate the problem reliably. Informal collaborations with no written record. Work done by students, interns or short-term contractors. Improvements suggested in a supplier's email and adopted without acknowledgement. Two friends developing a device together, one of whom files alone because they paid the fees. And the mirror image: a founder who names a business partner on the application as a gesture of good faith, then discovers years later that the partner's undivided interest travelled with them when they left. Understanding what a patent actually grants and to whom makes both errors much harder to make.
§ 05 Practice
The Record That Settles It in Advance
The discipline is unglamorous and takes minutes a week. Keep a dated log of technical decisions that records who proposed each feature, in their own words where possible, and keep the drafts rather than overwriting them. When a drawing changes, note who asked for the change and why. Where a supplier or contractor contributes a solution, resolve it immediately: either their contribution is ordinary skill applied to your specification, or it is conception, and a written assignment removes the ambiguity either way.
Then run the review at three points, not one. Before filing, walk the claim set element by element and ask, for each limitation, who first proposed it. Before each substantive amendment during prosecution, repeat the exercise against the amended claims. And before signing any licence or assignment, confirm the list still matches the claims as granted, because that is the document a counterparty's diligence will test. Handled this way the question costs an hour or two across the life of a file; handled after the fact it is the most expensive hour in the process, and often the one that ends it.
The consequences also differ by application type, since a filing that begins as a provisional and continues through subsequent related filings can accumulate different contributors at each stage — a reason to be clear on how the various application routes and proceedings relate to one another before the family grows. The pattern that produces a clean file is the one described in most practical accounts of taking an invention from idea to product, and it is visible in the paper trail behind commercialised devices such as a vibration-based pain-relief product that reached retail: contributions logged as they happened, assignments signed early, and no argument about names once money appeared. This is the last of the failure modes set out in the account of applications lost before examination, and the only one whose damage is done by other people.
End of sheet 05c