Skip to content

Patent Examination Types, Part II

Patent Examination Types, Part II published on 1 Comment on Patent Examination Types, Part II

The first post about patent examination types focused on examination tactics that amount to a wholesale dismissal or avoidance of the disclosed invention. In those cases, the examiner isn’t even considering the contents of the claim. Nothing that you put in the claim will cause the examiner to alter the outcome or rationale.

At the other end of the spectrum, some examination tactics focus on the contents of the claim in excruciating detail. Through a variety of tactics, these examiners endeavor to shift the conversation away from technology, logic, and the ordinary rules of language, and into the realm of linguistic gamesmanship.

(Disclaimer: These observations are drawn generally from extensive interaction with the examining corps at large, and do not describe any particular examiner or application. If you’re an examiner and you suspect a description is about you, it’s definitely not – it describes some other, lesser examiner.)

Examination Type #4: Word Search

Your client has invented a new and totally different type of encryption, which critically involves (and the patent claims recite) “a hashtable of prime numbers indexed by a solution of Euler’s formula for the plaintext message.” The specification presents a groundbreaking mathematical proof for why this encryption technique works.

Your claims are rejected under 35 USC 103, in view of the combination of:

  1. A computer science data structures textbook that recites the definition of a hashtable;

  2. A grade-school math textbook that describes prime numbers; and

  3. The Wikipedia entry for Euler’s formula.
In patent law, the patent claim is regarded as an expression of the disclosed technology. The elements of the claim are intended to be read and understood “as a whole,” and in the context of the disclosure. This type of examiner rejects that principle, and regards the patent claim as a string of disconnected words or phrases.

This tactic reduces the task of examination to a rote word search: dissect the claim into n individual phrases, and find a sentence somewhere in the prior art that kinda-sorta resembles phrase n. The specification doesn’t matter; the field of the art doesn’t matter; the context of each reference doesn’t matter; the resemblance between the disclosure and the references can be purely superficial.

Semantically, the office action reads like a ransom note – cobbled together with words snipped out of textbook pages, all different shapes and sizes, and pasted together with Elmer’s glue.

Notably, this examination methodology could be performed by a simple algorithm: a text-comparison library applied to the corpus of prior art. An undergraduate computer science major could knock together such an algorithm with about 30 lines of Python. Even better: the algorithm could just map every word of every claim to a page of the Oxford English Dictionary presenting the definition of the word. The entire examination backlog of the USPTO could be solved with a week of computational time – with a 100% rejection rate! Surely, the task of patent examination requires something more, but this point is occasionally not appreciated.

Variations of this tactic:

  • All arguments presented about what the words of the claim actually mean, in the context of the specification, are rejected en masse: “The applicant’s arguments involve subject matter in the disclosure, but the applicant is not allowed to import limitations from the specification into the claims.”

  • All arguments presented about what the cited portions of the reference actually mean, in the context of the reference at large, are also rejected en masse: “The applicant’s arguments about the teaching of the reference are irrelevant. The claim uses the terms (XYZ); the cited portion of the reference uses the terms (XYZ). Q.E.D.”

  • On the off-chance that the examiner can be compelled not to use a particular reference – i.e., a clear teaching-away in the reference, or a disqualification of the reference as prior art (or, more likely, instructions from the examiner’s supervisor that the rejection is unfounded) – the examiner will find another reference presenting another superficially similar passage.

    This parade of superficial comparisons can be performed until the applicant runs out of budget – or files an appeal, which causes the case to vanish for 4+ years. As an added bonus, the examiner racks up high production counts with minimal effort.

Examination Type #5: The Lexicographer

Your client has invented a text parsing technique based on identifying pairs of words that appear together in the document. The patent claim recites: “a first word of the document within a two-word proximity of a second word.” The specification and figures show an example where the words “alpaca” and “farm” appear together in the document.

The examiner rejects the claim by citing a reference that involves pairs of words that have the same frequency, and the example of a 100-word document featuring the words “alpaca” and “llama” each appearing six times… somewhere in the document.

You argue that “frequency” and “proximity” are unrelated concepts. You present examples showing why techniques involving “frequency” lead to totally different results than techniques involving “proximity.” These arguments are dismissed: “The examiner is entitled to the broadest reasonable interpretation, under which ‘proximity’ can mean ‘frequency.'”

The Manual of Patent Examining Procedure indicates that when interpreting a claim, “the words are presumed to take on the ordinary and customary meanings attributed to them by those of ordinary skill in the art.” However, the MPEP also specifies that each applicant is “free to be his or her own lexicographer,” and cautions examiners not to construe claim terms “in a lexicographic vacuum, but in the context of the specification and drawings.”

However, some examiners understand this concept backwards: that the examiner may act as lexicographer, and is entitled to assign any desired meaning to claim terms. Under this understanding of the “broadest reasonable interpretation” principle, “volume” can mean “mass”; “temperature” can mean “pressure”; “height” can mean “width”; “prediction” can mean “detection.”

This examination style can be motivated by a variety of causes. A technical deficiency, or even a language barrier, may cause the examiner not to understand the claim term – e.g., that “temperature” has a specific definition in the field of physics. Alternatively, the examiner may understand the terminology difference perfectly well… but may not care, believing that the invention (or, perhaps, the applicant) simply does not deserve patent protection. In either case, when presented with arguments about the meaning of claim terms, the lexicographer examiner will double down and assert that his or her selected meaning is valid.

Variations of this tactic:

  • When pressed that those of ordinary skill (or, indeed, even passing familiarity!) in a particular field all understand that term (X) means (Y) and not (Z) – e.g., that physicists generally agree that “volume” and “mass” are not synonymous or interchangeable – the examiner says: “I am one of ordinary skill in the art, and I think that (X) can mean (Z).”

  • This type of examiner may advise the applicant: “If you want a specific definition of a term like ‘temperature,’ put the definition in the claim.” While amending the claim to define “density” explicitly as “mass per unit volume” should not have any meaningful impact on patentability, such steps may, inexplicably, prompt an allowance.

  • In more extreme versions of this tactic, even adding the definition to the claim cannot suffice. The examiner may assert that the new definition raises further questions: “Your claim now defines ‘data size’ as a total number of words, but how do you define a ‘word’?” Or the examiner may raise a 112 rejection over the new claim: “You added the dictionary definition of the term ‘temperature’ to your claim, but that dictionary definition does not appear in your specification.” (Your failure to include an explicit definition of the term “temperature” in the specification is, apparently, your lack of foresight.)

Examination Type #6: The Kitchen Sink

Your technology includes a combination of parts (ABC). You present independent claims reciting (ABC). The examiner rejects it using a combination of references (Q) and (R), which together show (ABC). Fair enough…

The specification also discusses optional element (D), which can be used to extend the base invention with a specific feature. You amend the independent claims to recite (ABCD). The examiner adds reference (S), which shows (D), to (QR), and maintains the rejection.

The specification also discusses a further variation where (D) is more specifically (E). You amend the independent claims to recite (ABCDE). The examiner adds reference (T), which shows (E), to (QRS), and maintains the rejection.

After many iterations, your claims are two pages long, reciting elements (ABCDEFGHIJ) – a picture claim that’s so specific that the only entity on the planet infringing it is your client – and yet, the examiner maintains the rejection using the combination (QRSTUVWXYZ).

Modern patent examination practice places no objective limits on the combination of references. Indeed, the Manual of Patent Examining Procedure explicitly endorses this tactic:


Reliance on a large number of references in a rejection does not, without more, weigh against the obviousness of the claimed invention. In re Gorman, 933 F.2d 982, 18 USPQ2d 1885 (Fed. Cir. 1991) (Court affirmed a rejection of a detailed claim to a candy sucker shaped like a thumb on a stick based on thirteen prior art references.)

In some cases, this examination tactic is perfectly legitimate. If an application recites a type of automobile engine, and the prior art shows that engine, adding unrelated features to the claim – tires, brakes, turn signals, and windshield wipers – should not confer patentability.

However, in other cases, this tactic amounts to a dismissal of the nuance of an invention. A new machine can may exhibit a new principle of engineering, but may be built using a combination of off-the-shelf components: nuts, bolts, washers, pistons, gaskets, flanges. A kitchen-sink rejection can cite a massive combination of references teaching each of these components, despite the absence of any teaching actually relating to the engineering principle.

In other words, this examiner rejects the notion that an invention can be more than the sum of its parts.

Variations of this tactic:

  • A specific technique can be rejected using a generic teaching. The invention of the Stirling engine could be dismissed citing an antique text describing the basic concept of heat transfer.

  • Significant connecting language may be silently ignored. “A flange connected at a 45-degree angle to a strut” may be rejected in view of a first reference teaching a flange, and a second reference teaching a strut… with no mention of the connecting angle. Arguments about the failure of the art to teach the connecting angle are dismissed as a “design choice” or an “inherent” feature of the combination.

Examination Type #7: Form Over Function

Your client has invented a machine that calculates the mean weight of a collection of particles in a chamber when a user presses a button. The claim recites: “upon detecting a particle entering a chamber, increment a particle count; and upon detecting a button press, divide a mass of the particles in the chamber by the particle count.”

The claim is rejected under 35 USC 112 for one or more of the following reasons:

  • The term “the particles” has no antecedent basis. The preceding term “a particle” fails to provide antecedent basis, because of the difference in number between “a particle” and “the particles.” The examiner asserts that the rules of antecedent basis require absolute, rigid consistency: if a claim introduces a term as “a big, brown dog,” the rest of the claim must refer to it every single time as “the big, brown” – not simply “the dog,” or “the big dog,” or “the brown dog,” all of which lacks antecedent basis.

  • The invention has no utility. Because the first statement is only specified once, it is only performed once. Accordingly, when the user presses the button, the chamber will always contain exactly one particle, and the particle count will always be exactly one.

  • The claim does not REQUIRE that even one particle enters the chamber, and does not REQUIRE the user to press the button. If either step doesn’t occur, then nothing happens, so the claim is meaningless.

  • If the steps are performed in the opposite order, the particle count is zero, resulting in a division by zero, causing the method to… explode, or something.
It is a fact of patent law that claim language is a highly stylized form of legal writing, filled with recognized conventions of interpretation – e.g., terms like “comprising,” “wherein,” and “respective” have a specific meaning that’s not typical of the English language. (This sylistic language is also a chronic hindrance to public understanding of the patent system: some who encounter patent language for the first time have tremendous difficulty parsing it, and perceive it as a form of obfuscation or trickery.)

However, some examiners supplement these accepted conventions with their own rules of language and interpretation, or enjoy distorting the plain logic of the claim to include a bizarre scenario where the invention becomes pointless or inoperative. Satisfying their peculiar rationale requires loading the plain claim language with disclaimers (“wherein the machine divides by the particle count only if at least one particle has entered the chamber”) and convoluted language constructs (“the respective at least one particle”).

Variations of this tactic:

  • Depending on the examiner’s temperament, a parade of objections and 112 rejections may be asserted to bury the application in a blizzard of perceived technical issues. Favorites include:
    • Objecting to the title as not “descriptive.”
    • Objecting to the abstract as not sufficiently detailed.
    • Objecting to the figures as “not showing every feature of the claims” – even elements that are trivial or not amenable to illustration.
    • Demanding that portions of the figures showing alternatives to the disclosed invention be identified as “PRIOR ART,” even if the applicant does not admit that they are “prior.”
    • Objecting to trivial graphical elements of the figures:
      • “In Figure 8, an arrow points from element 100 to element 102, but the examiner would prefer that the arrow point from element 102 to element 100. A substitute Figure 8 is required.”
      • “In Figure 11, box 100 is illustrated with a dashed line, but the examiner does not understand why the line is dashed. Figure 11 must be resubmitted with box 100 shown with a solid line.”
      • “In Figure 13, element 100 is a picture of a mobile phone, and is referred to in the specification as a mobile phone, but is lacking a label. Figure 13 must be resubmitted with a ‘MOBILE PHONE’ label affixed to element 100.”

  • These rejections rarely cite a specific section of the CFR or MPEP in support of their peculiar rules, but just assert that these rules exist. If pressed to provide support, the examiner may adjust the rejection to an “objection” based on “informalities” in the claims, again without citing any type of support.

1 Comment

Leave a Reply

Your email address will not be published. Required fields are marked *