A provisional filing is examined by nobody. It receives a serial number and a date, and then it sits. No official reads it, no search is run against it, no objection is ever raised to its contents. For twelve months the inventor hears nothing at all, and the silence is easily mistaken for approval. It is not approval. It is the absence of any process capable of telling them the document does not work.
By the time the deficiency surfaces — during drafting of the substantive filing, or years later when an examiner or an opponent applies prior art dated between the two filings — the window in which it could have been fixed has closed. That is the whole failure mode, and it is worth understanding in mechanical detail.
§ 01 Function
What the Placeholder Actually Does
A provisional application establishes a filing date and nothing else. It never publishes, never issues, and expires automatically twelve months after filing. It does not require claims, an oath or a declaration, or an information disclosure statement. It does not begin the twenty-year term, which runs instead from the later non-provisional filing — one of the few genuine advantages of the route, since it effectively adds a year to the commercial life of the eventual patent.
What it buys is time on the market side: twelve months in which to approach manufacturers, test whether anyone wants the thing, refine the design, and use the phrase "patent pending" accurately. Those are real benefits, and the low cost of entry is real too. The difficulty is that the low cost is frequently read as a low standard, and the legal requirements attaching to the document are not reduced at all. A provisional application must satisfy the same written description and enablement requirements as any other filing for the priority claim to hold. It simply does not get checked.
§ 02 Mechanism
Benefit Is Granted Claim by Claim, Not Document by Document
The most consequential misunderstanding is that priority attaches to the invention. It does not. It attaches to each claim individually, and a claim receives the earlier date only if the earlier document described that specific subject matter well enough that a person skilled in the field would recognise the inventor had possession of it, and could make and use it without undue experimentation.
Consider a provisional that describes a filtration device generally, notes that "a suitable membrane is used", and lists no materials, pore ranges or assembly method. Twelve months later the non-provisional claims a device with a membrane of a defined pore range in a defined housing. The general claim may hold its early date. The specific claims — which are the ones with any chance of surviving examination — take the later date, because the earlier document did not describe them. A competitor's publication appearing in between now reads against exactly the claims the inventor needs.
Nothing in a provisional is preserved by intention. The date protects the sentences that were written, in the order they were written, and no more.
On what priority actually attaches to
The same principle blocks the obvious workaround. New matter cannot be added to an application after filing. Whatever the provisional omitted cannot be back-filled into it; it can only be introduced into the later filing, where it carries the later date. The document is therefore not a draft. It is a fixed record of what was understood on the day it was submitted.
§ 03 The Year
Twelve Months That Do Not Extend
The deadline is a hard one. Twelve months from the provisional filing, the non-provisional must be on file, and the same twelve months governs foreign filings claiming that priority. There is a narrow restoration route for genuinely unintentional delay in the international system, but it is a petition supported by a statement of reasons, decided by an office, and not accepted by every national authority — a remedy, not an extension.
A second date runs alongside it. If a public demonstration or sale occurred in reliance on the provisional, the domestic grace period counts twelve months from that disclosure, and it will often expire before or shortly after the provisional does. Inventors who file, then demonstrate at a show three months later, then take eleven months to raise the money for the substantive filing, arrive at a point where both clocks have run. The remaining option is a fresh filing with a fresh date, standing behind the inventor's own demonstration in the prior art.
Serial provisionals do not solve this either. Filing a second, better provisional at month eleven produces a second date; it does not extend the first, and the twelve months for anything claiming the original date still expires on the original anniversary.
§ 04 Sufficiency
What a Document Written to Hold Up Contains
A provisional that does its job reads like the specification it will become. It sets out the technical problem and the shortcomings of existing approaches. It describes the structure element by element, with the relationships between the elements stated rather than implied. It gives materials, dimensions and operating ranges, and it gives them as ranges with stated endpoints rather than as a single working example, because a claim can only be supported to the breadth the description reaches. It names alternatives — other materials, other geometries, other ways of achieving the same function — since each unlisted alternative is a design-around handed to a competitor.
It includes figures, labelled and referenced in the text. Drawings are disclosure, and dimensioned drawings are dense disclosure. It describes the method of manufacture and of use, step by step. It records the variants that were tried and rejected, and why, because that record supports both breadth and the later argument that the result was not obvious. Ten to thirty pages is unremarkable for a mechanical device. Three pages of prose about the benefits of the concept is not a short version of that document; it is a different document that happens to share a name.
§ 05 Economics
The Arithmetic of the Cheap Filing
The appeal of the thin filing is obvious. Official fees for a provisional are modest, the drafting can be done in an evening, and the inventor emerges with a date and a phrase they can use in conversation. Set against that, the substantive filing costs an order of magnitude more once drafting, drawings and prosecution are counted, and the temptation to defer that cost by treating the first filing as sufficient is entirely understandable.
What the arithmetic omits is that the cheap document does not defer the cost; it relocates it, usually to the point where the money has already been spent. This is one of the recurring failures set out in the account of how applications are lost before examination, and it is the least visible of them. Public reporting on independent inventors tends to focus on the outcome — the coverage of how the MixAid device reached production, or the longer account of the inventors behind it — and rarely on the unphotographable months of specification writing that preceded it. That asymmetry is itself part of the problem: the visible parts of invention are the prototype and the product, and the part that decides ownership is a document nobody sees. Even the cultural appetite for stories about innovation reliably skips it.
The working rule is simple enough to state. Write the provisional as though no second chance exists, because for everything it fails to describe, none does.
End of sheet 05b