The Patent Multiple

The Patent Multiple: a case for the document-object was a talk I gave at 4S in 2011 (Cleveland, OH). This is the transcript of the talk. Unfortunately, the slides have since been lost.

—

The questioning of patents is by no means a new project. Numerous authors who study technology have studied these documents to uncover what they tell us about science, technology, and society. Most have explored the legal uses of patents: how patents are deployed in debates over innovation¹; how the legal status of patents produce an “anticommons”²; how patents legally bind progress, innovation, and R&D.³ Others have historicized the claims within patents against real practice either in tracing the closure of a technology⁴ or in tracing knowledge transfer.⁵ And still there are more.

While the aims, approaches, and arguments are diverse and sometimes contradictory, one feature remains unquestionably consistent: the focus of these discussions involving patents is not patents, but patented objects, whether those objects be mining processes or drill bits. By this token, patents are defined as documents about objects, or more concisely, object-documents.

The patent as object-document denotes many things, a few of which are worth highlighting. First, since patents and patented things are discussed so inextricably tied, patents appear to have the same sociotechnical networks as patented things. In this dependent situatedness, patents exist as embedded waypoints for technologies. To account for patents is nothing more than to account formalizing objects as patentable. Second and derived from the first, an interest in patents is really an interest patented objects, whether that is in the purchase patents or in their debate. The patent officially constitutes an object. Geof Bowker provides a prime example of this, claiming that patents “give internalist and Whig accounts of […] [what] they describe, and as legal instruments they attempt to impose that interpretation on the material world.”⁶. Hence, patents blueprint objects by relying on textual-material slippage and anchorage. Third, in these contexts, patents self-efface. The background discussions, descriptions, claims, and drawings, amongst other things, gain credence in how they point away from the patent as an artifact and toward the patent as a fact. This is not to say patents are the intended deployment of an object regardless of accuracy–Bowker does a fine job of debunking this. Instead, patents are reduced–patents are what they say in spite of their instantiations as documents. Lastly, when entered into debates, patents-as-artifacts appear as immutable matter. Certainly the interpretations can render the document differently, but the patent-as-artifact does not change.

These understandings of patents have their rightful place as dominating the discussion involving patents, at least historically. Patents do describe objects, and, in debates about those objects, patents shed light on what is and is not protected, or innovative, or historically accurate. Likewise, patents necessarily self-efface and appear immutable–what legal status could they claim otherwise? But these understandings beg questioning themselves: Are patents really so tied to their purported innovation? Can patents make claims beyond those listed as such? Are they immutable?

The present paper does not answer these in full, but posits a complementary notion: patents are objects that happen to be documents, that is, patents are document-objects. Accordingly, a study of patents does not need to be brought back to the objects described, but can and should explore the patent itself. If STS is concerned at times with failures and flops, marginalia and minutiae, then concerning ourselves with only the descriptive content of patents denies that patents are not just description, but actual objects in the world.

This paper focuses on four specific aspects of one specific patent. These aspects are litanies, drawings, associated documents, and language, only one of which will be covered here, namely associated documents; the patent is Security system for screening people. The presentation is structured around description: specifically describing the difference between the US patent and the associated WIPO patent in order to unpack the way these patents deploy different objects regardless of what they describe. In exploring patents outside of the traditional legal or historical work, the paper aims to contribute to the ongoing discussion within STS as to how to, as John Law puts it, decenter the object.

To begin, we need to start with the patent: Security system for screening people. Here is an excerpt from the abstract:

“The present invention is directed towards an X-ray people screening system capable of rapidly screening people for detection of metals, low Z materials (plastics, ceramics and illicit drugs) and other contraband which might be concealed beneath the person’s clothing or on the person’s body”.⁷

The device of Security system for screening people is one associated with the Transportation Security Administrations’ program called Advanced Imaging Technology. But just to be preemptively clear, I am not concerned with this object (slide 2, pause) or this object (slide 3, pause) but this object (slide 4) and this object (physical patent in hand). While many are aware of the controversy about the TSA program, the device described in this patent is only a piece of the system: Security system for screening people as described by the patent only produces the raw detection data by way of X-rays and not the hotly contested images of AIT. The digital computer and software processes are elsewhere. While the patent mentions the device as generating and displaying images, along with being categorically aligned with imaging, what this might entail is bracketed as demonstrated by figures 1-3 where the words “digital computer” and “image display monitor” appear in within aptly designed black boxes.

But again, the described device is not the concern here. Instead, the concern is the patent itself. (slide 5)

Filed in December 2008 with the United States Patent and Trademark Office and issued in November 2010, a provisional application is listed as filed in December 2007 under the same name–save different capitalization upon search.

The originally published patent is 14 pages long as a printer-ready PDF. The first page is a standard overview of a patent including one drawing, the abstract, the inventors, the assignee, the filer, the referenced patents, etc.; the second page is a continuation of the referenced patents (there are 162 cited patents in total, spanning 34 years from 1970 to 2004) (slide 6); the next five pages are drawings; the last seven pages are written text divided into sections. The patent contains 6 drawings and 18 claims. A fifteenth page was added in August 2011 detailing an error reported within the body.

Two foreign patents–one with the European Patent Office and the other with the World Intellectual Property Organization–are listed as being foreign patent documents. The former is a patent from 1988 and the latter is a patent from 2007. We will revisit this second document in a moment. (slide 7)

Online and through Google’s patent database, the patent can be viewed primarily as a combination of two-column webpage with sequential scrolling PDF pages (all fifteen) or as converted, and rather buggy, rich-text. The 14-page patent can also be downloaded from Google Patents. (slide 8) The USPTO website offers different options: the patent primarily displays as simply formatted HTML without the drawings or secondarily as scanned 15-page PDF. The USPTO readable database does not offer a downloadable version of the patent at all, though it is accessible elsewhere. In both cases, various document links are all internal links within the US database. (slide 9)

As previously mentioned, Security system for screening people has two associated patent documents. The later document, as it turns out, is an international dzygotic twin titled Improved security system for screening people. This document, despite its titular amendment, was filed in parallel the US patent and subsequently published a year faster.

Improved security system for screening people is 27 pages long. The only internal changes to the content (at least in English) are additions: an international search report added as an appendix and one additional claim inserted between claims 5 and 6 of the US version. The descriptions, the drawings, inventors, and the other 18 claims are identical Security system for screening people save minor changes such as replacing the abbreviated “FIG.” with the full word “figure” when discussing the diagrams. (slide 10)

Online through the WIPO site, this document can be viewed in full text, browsed with a tabbed display without the full set of images. The WIPO site also offers a downloadable PDF of the patent and the various filing documents–status reports, search reports, preliminaries documents–all in multiple digital formats such as XML with TIFF images and scanned PDFs. The WIPO site also includes external links Improved security system for screening people in its various national phases, that is, as external links. (slide 11)

The length increase is more due to formatting than additions: The US patent of Security system for screening people is two-column justified type with tight leading and 9 point font. Line numbers are placed between the columns, and the top and bottom margins are tight at only a 1-inch and a half inch respectively. On the other hand, the WIPO patent is one-column justified type with more open line spacing, a slightly larger font, and a widened left margins to account for line numbering. The top and bottom margins are significantly larger at 1.5 inches and 1 inch respectively. (slide 12)

As for the the added claim–international claim 6–it simply states “The imaging apparatus of claim 1 wherein said processor processes programmable code that causes the first module or the second module to move vertically.”⁸ This is truly the only distinguishing feature of the international patent, maybe accounting for why is improved. Without judging whether this is significant enough, the unchanged descriptions, diagrams, and field of invention make this puzzling. Not only are there are no added features described, nor added processes or enumerated parts of mechanical articulation, the US patent actually includes vertical movement in the optional features, though without specifically highlighting it as a claim. (slide 13)

There is one more thing: proceeding the unchanged body text is a note in both that states “The present invention relies on, for priority, [the] United States Provisional Patent Application […] of the same title, filed on December 25, 2007.”⁹ The referenced provisional title, however, does not actually contain the word Improved. This is a hold over from the US patent, never modified for the WIPO patent.

What we see is, as artifacts, patents are rather mutable both internally and externally. Internally, a patent can be improved upon not by being different, but by separating claims into smaller parts or enumerating claims that are always there. Innovation is a fractal and document-objects are combinatoric even while the material object remains static.

Externally, the patent can change too. While the international version appears to incorrectly list the title of the US provisional patent, it is the international patent that moves most widely, re-filed with Canada, Europe, Japan, Russia, and Australia, protected in most countries around the world, and easily locateable and interlinked on the web. Hence, the international patent assembles a wide network of consenting organizations, secondary-tier documents, patent officials, obliging governments, eager lawyers, search results, and downloadable PDFs. From this perspective, it is the US patent that gets the title wrong. If facts are merely those things that gain enough observational consensus, then the international patent has constructed a fact about the provisional patent regardless of the US or provisional patents’ disagreement. It is not the security system that improves, but the patent for that security system in how it amasses a globally recognized network.

The patent as an artifact is not simply amended, but recombined and reconstituted as other objects by other objects, some of which are documents. Where Bowker recognizes the patent is a text that can be deployed with interpretation, there is an assumption that the patent itself does not change. Bowker stresses how the non-neutral patent-text can be looked at in terms of its falsehood, i.e. how it is out-of-sync with the world at large, and in this space of discord formulates a rhetorical device on this at the level of content. The claim here goes further: the patent composes a rhetorical device in how it composes networks, such as how one document appears by way of another or how the text itself can be reused. The object described is the same in both, but the patent is not. (slide 14)

What does this say about patents? While patents are often primarily understood as descriptors, patents themselves are partially, and by no means primarily, themselves descriptors. Instead, as objects, patents are ontologies–they compose being by exemplifying categorical intersections, litanizing heterogeneous things, recontextualizing text, and assembling networks. As shown in the association of documents, by approaching patents outside their claims, we see patents configure networks which overlap, but also conflict. When centered, the patented device of Security system for screening people obscures this realization, amongst others here unmentioned. Hence, the patent is multiple: it is a series of objects, a set of practices, a deployment of arguments, a collection of topologies. To reiterate, if STS is concerned at times with failures and flops, marginalia and minutiae, then concerning ourselves with only the descriptive content of patents denies that patents are not just description, but actual objects in the world. To recognize the object-ness of patents is to recognize patents are more than claims, and to admit them to STS finally.