What is the difference between a §102 rejection and a §103 rejection?

Anticipation under §102 requires a single prior-art reference that discloses every limitation of the claim, arranged as in the claim, expressly or inherently. Obviousness under §103 lets the examiner combine references and fill gaps with what a person of ordinary skill would have known, provided there is an articulated reason to combine. A §102 rejection has no motivation-to-combine element to attack — but it has an all-elements requirement that a §103 rejection does not, and that is where it breaks.

Can I argue that the reference does not qualify as prior art?

Often, and it is checked less than it should be. Under the AIA a reference is prior art if it was publicly available before your effective filing date (§102(a)(1)), or if it is a US patent or published application by another that was effectively filed before yours (§102(a)(2)). Two exceptions matter: your own disclosure, or one derived from you, within the year before filing does not count (§102(b)(1)), and an (a)(2) reference is removed if you had publicly disclosed the subject matter first or it was commonly owned (§102(b)(2)). If the reference is a published application, confirm the examiner is relying on a date it is actually entitled to.

What is inherent anticipation, and how do I respond to it?

A reference anticipates inherently when a limitation it does not state is nonetheless necessarily present in what it describes — not probably, not usually, necessarily. Examiners reach for inherency when a reference shows the structure but not the claimed property or result. The response is to show that the limitation could be absent from the reference's disclosure as written: if the reference's device can be built as described without the claimed feature, it is not inherent, and the rejection fails.

The reference really does anticipate. Now what?

Amend, but not blindly. The examiner will search again for whatever you add, in the same field, and will often keep the same reference as a §103 primary and add a secondary reference for the new limitation. A scope search of the examiner's classification before you choose the amendment tells you which candidate limitation the field does not already show. One search is cheaper than the extra round of prosecution it prevents.

Does pre-AIA §102 still matter?

Only for applications with an effective filing date before March 16, 2013, and for any application in a family that ever contained a claim with such a date. Those still have the old §102(a), (b) and (e) categories, the one-year statutory bar, and the ability to swear behind a reference with evidence of earlier invention — which the AIA removed. Check which regime the application is under before arguing a date.

Guides

Responding to a §102 rejection: the three questions that decide it

Michael W. Smith, former USPTO patent examiner · September 13, 2026

A §102 rejection stands or falls on three questions, in this order. First, does the one cited reference disclose every limitation of the claim, arranged as the claim arranges them — a single missing element defeats anticipation outright, which makes this the easiest rejection to overturn when the examiner has over-read a passage. Second, is the reference actually prior art: its effective date against your effective filing date, the AIA §102(b) grace-period exceptions, and whether an inherency argument is doing work it cannot do. Third, if it genuinely anticipates, amend — but search the examiner's field first, because the same reference plus one more is the §103 rejection waiting in the next action.

Question one: every limitation, arranged as claimed?

Anticipation is all-or-nothing. The rejection has to point to a disclosure of each and every limitation in a single reference, and the elements have to be arranged or combined in the same way the claim arranges them — a reference that discloses all the parts in separate embodiments, or in a list of alternatives from which the claimed combination could be picked, does not anticipate. So the first act of the response is a table: each limitation of the rejected claim down the left, the examiner's citation on the right, and a verdict per row after reading the reference at that pinpoint.

In my experience on both sides of the desk, this is where a meaningful share of §102 rejections fail. An examiner reading fast finds the elements and cites the passages; whether the passages show the elements together, in the claimed relationship, is the question that gets less attention under a production count. A single limitation that the reference does not show, or shows only in a different embodiment, is a complete answer — no amendment, no claim scope surrendered.

Question two: is it prior art at all?

Before arguing what the reference says, confirm it counts. Under the AIA, a reference qualifies if it was patented, described in a printed publication, in public use, on sale, or otherwise available to the public before your effective filing date (§102(a)(1)); or if it is a US patent or published application naming another inventor that was effectively filed before your effective filing date, even if it published after (§102(a)(2)). Then check the exceptions:

Your own disclosure within the year before filing — a paper, a talk, a product — is not prior art against you (§102(b)(1)(A)), and neither is a third-party disclosure of the same subject matter made after yours (§102(b)(1)(B)).
An (a)(2) reference is removed if the subject matter was obtained from you, if you publicly disclosed it first, or if it was commonly owned or under a joint research agreement by the effective filing date (§102(b)(2)).
The reference's own date. A published application is prior art as of the earliest application it is entitled to claim for the relied-upon subject matter — not automatically its earliest priority claim. If the disclosure the examiner relies on appears only in the later filing, the earlier date does not carry it.
Your own date. If your claims are supported by a provisional, the effective filing date is the provisional's, and a reference between the two dates drops out.

Inherency deserves its own check. An examiner who cannot find a limitation stated in the reference may assert it is inherent. Inherency requires that the missing limitation be necessarily present in what the reference describes — not that it probably is, or would be in the usual case. If the reference's device could be made as described without the claimed feature, the feature is not inherent and the rejection is incomplete.

Question three: if it anticipates, what do you amend to?

When the reference genuinely discloses every limitation, the claim as written is not new and the response is an amendment. The instinct is to add whatever the reference lacks. The examiner then searches for that limitation in the same classification, and the next action is frequently a §103 over the same reference in view of a second one that shows the added feature — because the feature you chose was the first thing that distinguished the reference, not the thing that is actually novel.

The fix is the same as for a §103 rejection: read the examiner's search record in the file wrapper, run a scope search of that field against the two or three amendments you are considering, and amend to the one the field does not show. That turns a §102-then-§103-then-RCE sequence into a single amendment and an allowance.

Every amendment is prosecution history. Narrowing around a specific anticipatory reference can surrender equivalents you did not intend to give up. A search-guided amendment narrows once, to the limitation that is actually new, rather than in successive rounds to whatever the examiner cited last.

The interview

On a §102 rejection an interview is unusually productive, because the disagreement is concrete: does this passage show this element, arranged this way, or not. Go in with the limitation table and the reference open to the cited column. An examiner who agrees a limitation is missing will often withdraw the rejection on the spot, or tell you exactly what amendment would be allowed.

What this costs here

Published on the office-action page, flat:

Verification — $750$400. Every limitation checked against the reference at the pinpoint cited, and the reference's prior-art status checked against your dates. About a week.
Scope check — $750$400. A search of the examiner's field against your candidate amendments, before you commit to one.
Drafted response — $1,400$800. The response written for your signature. Amendments come only in the package, with the scope check.
All three — $1,750$1,000. Amendments included.