Crowd Sourcing Gun Control

Background checks deter only those with prior offenses. For more comprehensive gun-violence prevention, we need systems that help law enforcement focus on high-risk individuals before their first offense. Such prevention systems must operate within important legal and political constraints: they must be compatible with the second amendment, they must be relatively effortless for the vast majority of low-risk gun buyers/owners, and they must respect the privacy of all citizens.


Graph theory makes it possible to extract strong information from a large set of weak relationships. It is the foundation of companies like Google and Facebook. It is the basis for alternative metrics that assess the quality of scholarly papers. Graph theory also applies to gun control. As individuals, we absorb all kinds of information about others and decide to like or dislike, befriend or avoid, trust or distrust them. These individual assessments are rather unreliable, but they become a powerful resource when connected with one another in a graph. For gun control, imagine a system in which gun owners recommend one another. This creates a graph on which to unleash the power of graph theory.

Consider a concrete example (no graph-theory knowledge required):

1. Every gun buyer/owner registers three Qualified Recommenders (QRs). Without assuming any legal liability, QRs state that the gun buyer/owner is not presently a danger to himself/herself or to society and that they will withdraw as QR if their opinion changes or if they should lose contact with the gun owner.
2. Most legal gun owners may act as a QR. However, law enforcement may revoke anyone's QR status at any time for any reason. Such action prevents this person from submitting new recommendations, and it voids his/her existing recommendations. Revoking the QR status carries no other consequences. One's own QR status does not affect one's right to own guns.
3. A gun buyer who is unable to obtain three QRs is subject to a background investigation that may include interviews with family, friends, and neighbors.
4. A gun owner may lose recommendations because (1) QRs may withdraw their recommendation, (2) QRs may die, or (3) QRs may lose their QR status. A gun owner who does not maintain three QRs is subject to increased scrutiny.

Maintaining three QRs is relatively effortless for low-risk gun owners, who tend to have strong ties to their communities. If they should become high risk (through mental illness, for example), their QRs are likely to withdraw their recommendations, thereby creating an early warning.

The impact on high-risk gun owners is more substantial. Very few law-abiding gun owners would be willing to sponsor someone they consider dangerous because of mental health, behavioral issues, or criminal connections. Convicted criminals cannot be QRs, because they cannot be legal gun owners. This reduces the pool of available QRs for their non-convicted associates. Law enforcement can reduce that pool even further: when they arrest someone, they could revoke the QR status of the arrestee and of his/her QRs (under the theory that these QRs are unreliable).

Consider a few scenarios.

Scenario 1. Joe is a gun owner. His QRs are his wife (Mary) and two friends (John and Pete).

Scenario 1a. Joe develops a mental-health problem and is in therapy. Mary convinces Joe to dispose of the guns. Mary withdraws her recommendation and, perhaps, approaches John and Pete to withdraw theirs. There are no legal proceedings of any kind.

Scenario 1b. Joe develops a mental-health problem, refuses therapy, and develops violent tendencies. The QRs and the police develop a plan to withdraw their recommendations and, simultaneously, to impound Joe's weapons. After the emergency intervention, Joe's fitness to own guns is evaluated.

Scenario 1c. Joe is arrested under suspicion of committing a violent crime.

Upon his arrest, Joe's QR status is revoked immediately. Joe's gun rights depend on the resolution of the criminal case against him. Mary, John, and Pete are not legally liable for failing to intervene in time, but their QR status is revoked.

Mary, John, and Pete remain legal gun owners. If they acted as each other's QRs, they have to find replacement QRs. For most gun owners, these circumstances are an extremely rare occurrence, and replacing QRs is a low hurdle. For criminal gangs, every arrest eliminates QRs from their environment and increases their difficulty of replacing QRs.

Scenario 2. Bill sells his recommendations. As soon as he is identified as a shady operator, the police subjects his network to heightened scrutiny. With the first arrest in Bill's network, he loses his QR status, and his recommendations are voided. As Bill's clients find replacement QRs, they leave useful network trails.

Scenario 3. Mary, a law-abiding citizen living in a gang-infested neighborhood, wants a gun for protection. Some of her relatives and friends have arrest records, and she cannot find three QRs. She is a high-risk gun buyer because of the circumstances of her life: her gun is at much higher risk to be stolen and/or misused. She is able to buy a gun after passing an in-depth background investigation.

Scenario 4. Bob is a criminal and does not care about gun laws. Because his guns are illegal, he faces a much harsher sentence if arrested and convicted.

Bob's guns are most likely stolen or obtained from sources that claim their guns were stolen or lost. If one of his guns is traced back to Mary, her QR status is revoked immediately. If Mary is an innocent bystander, losing her QR status is no big deal. If she is a member of a network that traffics guns illegally, this incident reduces the pool of available QRs.

Other Scenarios. The QR system implements a form of social control that occurs organically in small communities where everyone knows their neighbors (a rarity nowadays). The process of finding QRs provides opportunities for a community to intervene in a non-confrontational manner and to help prospective gun buyers make a realistic assessment of their own risk profile.

The QR system shows that gun-violence prevention does not require storing detailed personal information. It shows that an alert and involved community does not need to become a society of snoops that report every eccentricity to the police. Yet, the QR system is just one example. Graph theory provides a lot of flexibility to respond to legal and political concerns.

I call on the graph-theory community to contribute to the gun-control effort. It is up to you. Criticize. Amend. Propose. Do something. Anything. Save lives.

The Fourth Branch Library

Today’s library is the result of twenty years of incremental changes: an institution buying access to information wholesale and restricting that access retail. As discussed here and here, the wholesale market for information is distorted and creates artificially high site license prices. Another expense is the inordinate amount of time staff spend on usage studies and community outreach to gauge collective information needs, negotiations with consortia that pool resources to obtain imaginary discounts from inflated list prices, negotiations with publishers and their agents, and internal library discussions. After they are acquired, site licenses remain expensive. As protectors of publishers’ digital rights, libraries spend significant resources restricting access at the retail level.

The time for incremental change is over. We must rebuild the library from scratch on a foundation of traditional library values. Here is my attempt.

The mission of the library is to serve the members of its community by:
1. Helping them create high-quality information,
2. Collecting, organizing, and archiving that information, and
3. Making that information widely available, subject to legal and ethical constraints.

This mission is steeped in tradition. Libraries of the antiquity were more about secrecy than openness, but their primary purpose was to archive locally produced information. The purest modern implementation of the vision I am proposing is the American presidential library, which collects, manages, and makes available the information from one administration. Public libraries routinely accept manuscripts and personal correspondence of authors and other luminaries. University archives preserve scholarly history. Many academic libraries have implemented various open-access initiatives and have set up databases containing publications of faculty and students (scholarly articles, books, theses, and dissertations).

This mission allows for specialization. Libraries are ideally positioned to add value to information produced by the communities they serve. A public library that serves a particular location may help its constituents with educational programs in information literacy. Other libraries may specialize in particular disciplines and serve communities that are dispersed worldwide. This is particularly the case for data archives, which require deep specialization.

This mission includes nonprofit and for-profit organizations. In this view, publishers are for-profit libraries. As such, they shoulder all the responsibilities of a library, including archiving the information under its purview.

This mission exploits the network effect. Through collaboration, libraries can create a worldwide network of high-quality information that is more than the sum of its parts.

This mission is critically important. We produce an exponentially rising amount of information that is poorly managed and in danger of being irretrievably lost.

For concrete examples, I could point to existing open access initiatives. Peter Suber’s The Open Access Overview is a good place to start. Most of these initiatives share the trait of being focused on disseminating information from a community to the world and letting the web take care of bringing the world to individuals. These are great initiatives, but I want to push the limits. I do not want to be boxed in by what is feasible today.

The largest producer of public-domain information is the government. Legal information, legislative records, and official government reports are readily available through established channels. Other government records, however, are more problematic. As a matter of expediency, officials tend to have a bias towards opaqueness. Impenetrable government records are managed by a hodgepodge of government agencies. The system hides problems ranging from bad judgment to corruption and complicates good governance.

What if we had an independent agency to manage the government’s records? This agency would create the systems to gather this information. It would decide the appropriate level of public access. By imposing standards, it would ensure that government records were machine-readable and discoverable. The infrastructure for such an independent agency is already in place: the public library system at the local, county, state, and federal levels. In its most extreme form, this independent agency could evolve into a fourth branch of government, one dedicated to transparency of the other branches.

As a practical matter, this may be an overreach, and more modest initiatives are more realistic starting points. However, considering the profound impact of digital information on our lives and considering that the information age is here to stay, we are forced to think big.

On the other hand, thinking small comes naturally. The latest innovation of the Los Angeles Public Library sets free the all-important Sony Music catalog, saving Los Angeles residents from the unspeakable burden of $1 song downloads. The Librarian in Black has a detailed critique.

<Note: edited title 8/26/2011>

The Law: What's logic got to do with it?

One of my favorite daily reads is the Volokh conspiracy, a blog by a group of law professors. As one would expect, their arguments are exquisitely reasoned, honed by years of classroom teaching, academic writing, and litigation. Although they have a definite philosophical point of view (libertarian), these scholars apply the law with intellectual integrity, developing their arguments logically from existing law and precedent. However, I cannot help but wonder about the use of logic in law.

In formal logic, the starting point is an axiomatic system, a set of statements assumed as true (axioms) that defines a particular field. The rules of logic are then used to derive theorems within that field. The most famous example is the field of Euclidean geometry, which is built upon Euclid's axioms. When developing an axiomatic system, mathematicians want a system that is consistent, independent, and complete. For our purposes here, we can ignore independence and completeness. The crucial property is consistency: axioms within an axiomatic system should not contradict one another.

Legal analysis applies the rules of logic (for example, A implies B, B implies C, hence A implies C), but it does so outside of anything resembling an axiomatic system. Laws, statutes, and precedents vaguely resemble axioms in the sense that they are assumed to be true. But the legal system is rife with contradictions. When an axiomatic system is not consistent, it contains at least one statement that is simultaneously true and false, and from that statement one can derive an unlimited number of other statements that are simultaneously true and false. In law, one can reduce the inconsistencies by only considering those laws that are deemed relevant to a particular case. No matter which subset of laws one gets to work with, it is virtually impossible for it to be anywhere near a consistent axiomatic system. Formal logic seems doomed in legal reasoning.

In spite of their mutual resemblance, legal logic lives in a universe quite different from formal logic. Legal logic is about convincing others of the merits of a case. A legal argument is successful only if accepted by some authority, and this acceptance lasts only until a higher authority overturns it. With every decision, legal authorities help shape the nature of successful legal argument. This creates a legal logic that evolves over time and reflects the nature of the power of the state. The law is about power, not logic... Who knew?