jsheff

New Draft: Law and Philosophy in IP

I’ve just posted a draft of a new paper to SSRN on law and philosophy scholarship in intellectual property. It is my contribution to a forthcoming handbook from Oxford University Press, edited by Irene Calboli and Lillà Montagnani, on methodologies in IP research. Here’s the abstract:

Intellectual property (IP) law and philosophy is an interdisciplinary approach to scholarship that applies insights and methods from philosophy to the legal, normative, theoretical, political, and empirical questions presented by the project of organizing and regulating the creation and dissemination of knowledge, technology, and culture. In this chapter, I outline four types of IP-law-and-philosophy scholarship, focusing specifically on the discipline of analytic philosophy (with appropriate caveats about the coherence of that discipline). These modes of scholarship can be categorized as (1) the jurisprudence of the IP system, (2) philosophical analysis of IP law, (3) applied philosophy in IP, and (4) normative theory of IP. Category (4) is obviously a special case of category (3), focusing specifically on applications of moral philosophy. Within each category, I provide illustrative examples of past scholarship and suggestions for further research.

As always, comments are welcome.

Article V, the Senate, and the House of Lords Option

It is not news that the United States Senate is a deeply anti-democratic institution. With population shifts sorting whiter, older Americans into low-population states and younger, more diverse Americans into high-population states, the anti-democratic bent of the Senate is becoming a systematic bias that will soon have the effect of delivering a veto over federal legislation and executive and judicial appointments to an increasingly insular and cohesive racial and demographic minority. But people who should know better also say there’s nothing to be done about it, because Article V of the Constitution provides that the Amendment process is qualified to the extent that “no state, without its consent, shall be deprived of its equal suffrage in the Senate.” We’re stuck with the anti-democratic Senate, the naysayers cluck, as if to think otherwise is hopelessly naive. Democracy in America was doomed from the start; the only way to fix it is to ditch the Constitution and start from scratch (and that’ll never happen anyway, so we may as well give up).

This defeatism does not befit a self-governing people. More importantly, it is hogwash. The Senate is not amendment-proof, and Americans who value democratic governance ought to be planning how best to do away with it. Any creative lawyer should be able to recognize that the supposed roadblock of Article V can be easily worked around. Article V says that no constitutional amendment can deprive any state of its equal vote in the Senate. It says nothing about how constitutional amendments may affect the powers of the Senate. If we can’t change who votes in the Senate, we can change what they get to vote on. And though such changes must clear the supermajority thresholds of the Article V amendment process, there are tools available to simple popular majorities to create the pressure necessary to clear such those thresholds. There is precedent for it, in fact: the United Kindgom followed this path a century ago.

First, we can dispense with the supposedly implacable Constitutional text. Article V poses no obstacle to amending the bicameralism requirement of Article I so as to make passage of bills by the House of Representatives sufficient to send a bill to the President for signature. It poses no obstacle to giving the House of Representatives (or some other body) the “advice and consent” responsibilities of the Senate over executive and judicial appointments, or the responsibility for ratifying treaties. It poses no obstacle to modifying the impeachment process so as to vest the power to remove federal officers in the House of Representatives alone, or to assign the trial of impeachments to some other body. There is literally nothing that the Senate does that cannot be assigned to some other institution by constitutional amendment. The fact that we would be left with 100 sinecures on the federal payroll is the only cost of such a strategy—and one that could be remedied by reducing the pay and staff of Senators in a manner consistent with the 27th Amendment’s limitations on varying the pay of legislators between elections.

But now, we come to the practical obstacles to amendment. Article V makes constitutional amendment very difficult. Supermajorities of both houses of Congress and super-duper-majorities of the state legislatures have to concur for such an amendment to pass. And it is difficult to imagine the states that benefit from having greater share of votes in the Senate than they would have in a proportionally apportioned legislature—which by definition constitutes a majority of the states—voluntarily giving those powers away. If the Constitution is to be amended to reallocate power away from the Senate, the states that stand to lose from such a reallocation need to have a reason to go along.

However, the Constitution also provides a vehicle to assert leverage over such states that does not require supermajority support: admitting new states. By an Act of Congress, new states can be admitted to the Union—with two new senators apiece and a vote on any proposed constitutional amendment. If an existing state consents by vote of its legislature, it can also be divided into multiple states by an Act of Congress; each new subdivided state would similarly be entitled to two senators and a vote on any constitutional amendment (application of this latter tool to sever Maine from Massachusetts was the key to the infamous Missouri Compromise). The Constitution imposes no population constraint on the creation of new states; my apartment can be a state if Congress certifies that it has “a republican form of government.” Many people have pointed out that a top priority for the next period of unified Democratic control over both houses of Congress and the White House should be statehood for Puerto Rico and Washington, DC; a few have gone so far as to toy with the “make more states” option. Serious action to ensure a democratically elected federal legislature will require a willingness to go much further: it will require a transparent agenda to “pack the Union.” This will require, at a minimum:

  • a publicly-stated willingness to create enough new states that will predictably commit to reducing the Senate’s powers to command the 3/4 supermajority required to pass a constitutional amendment, and
  • a credible and transparent plan to create these new states.

To implement such a plan would be clumsy, norm-breaking, logistically challenging. But it could be done, and it would not be extra-constitutional to do it. Most importantly, the willingness and ability to do it may make actually doing it unnecessary. That is, a credible threat to pack the Senate may be enough to fix—or effectively end—the Senate. We have good reason to think this, because such threats have been effective before—twice—in accomplishing very similar reforms in the United Kingdom.

The Liberal Party won a landslide in the UK parliamentary elections of 1906, and was (more narrowly) returned to government twice in two successive elections in 1910 on a platform of a progressive budget, home rule for Ireland, and substantial reforms to the House of Lords. The Lords, dominated by Conservatives, had consistently obstructed progressive budgets and other reform measures after the passage of such measures by the House of Commons–as well they might, given that the Liberal government proposed to curtail their power and tax their lands. But by threatening to pack the Lords with enough new Liberal-leaning peers to command a majority, Prime Minister H.H. Asquith ultimately succeeded in strong-arming the Lords into passing the Parliament Act 1911. The Act did to the House of Lords what a moderate pro-democracy reformer in America today might like to see done to the U.S. Senate: it effectively removed the Lords’ power to refuse passage of money bills, and effectively converted their power to refuse passage of other bills to a power to delay such bills for a maximum of two years. (The Parliament Act 1949 reduced the Lords’ power of delay over non-money bills from two years to one—and was enacted using the power of the Commons under the Parliament Act 1911 to pass bills without the Lords’ consent.) The Lords and their allies in the Commons complained bitterly, but ultimately, Conservative peers were put to the choice between giving some of the powers of their House to the Commons forever or losing all of the powers of their House to the Liberals for the foreseeable future, and chose the former.

Asquith’s threat to pack the peerage had precedent too: it was exactly the strategy used by the Whig Prime Minister Charles Grey to expand the franchise and meliorate the shameful malapportionment of the House of Commons with the passage of the Great Reform Act of 1832. Again, the threat to pack the Lords was was enough: the Tories who followed the Duke of Wellington, having successfully blocked Reform Bills for years, capitulated once they learned that William IV was (reluctantly) willing to support Lord Grey’s request to appoint a slew of Whig peers.

What I propose is not “One Simple Trick to Save American Democracy.” I do not mean to suggest that this type of reform was easy in the UK, or that it will be easy in the United States. 18th-19th Century Whigs and 19th-20th Century Liberals had to build a popular movement committed to their reforms, then consistently win elections on the strength of that movement for several years, and ultimately see their democratically validated demands repeatedly rebuffed by anti-democratic institutions before the continued pressure to accede to those demands became irrepressible. In the Whigs’ case, that pressure included the mass violence of the Swing Riots and the Days of May: the closest Britain ever came to a violent popular revolution. The Liberals pushed through the Parliament Act 1911 without such violent uprisings, but with the crucial support of two factions that were structurally underrepresented in the electorate and deeply associated with the threat of such violence in the minds of the English Conservative establishment: Irish nationalists and organized labor.

I do not mean to issue a call to political violence, but I do mean to be clear that the mere availability of constitutional tools for democratic reform will never be sufficient to overcome constitutionally empowered anti-democratic institutions opposed to such reforms. Nor will asking politely and waiting patiently. There must be a credible threat of constitutional norm-breaking in order to use such tools effectively; and such a threat is only credible (and warranted) when supported by persistent, well-organized, and genuinely popular movements that in turn back up the threat of constitutional norm-breaking with the prospect of deep social dissolution if anti-democratic institutions do not accede to democratic demands. Ominously, anti-democratic elites must also be sufficiently sensitive to the risks of social dissolution that they will recognize when their resistance to democratic reform has become untenable. Deep constitutional traditions of royal deference to electoral results (informed by the violent origins and tenuous foundations of constitutional monarchy) in eras of global social transformation did the trick for William IV, Edward VII, and George V; it isn’t at all clear that our anti-democratic elites are sensitive to such considerations.

But the fact that democratic reform is hard is no excuse for not putting in the work to make it happen. Precedent for de-fanging the reactionary and anti-democratic “upper” house of a national legislature is readily available—and it has been proven to be both durable and, ultimately, extremely popular. There is no reason why the United States should be a century behind the United Kingdom in reforming its national legislature to align with democratic principles. Americans who take democracy seriously should be mobilizing to support a credible threat of packing the Senate, as a second-best alternative to amending the Constitution to remove its powers. They should be making plans to admit new states and divide existing states so as to assure passage of constitutional amendments reallocating the Senate’s powers to other institutions. And they should be explicitly including these plans in their agendas and platforms, and be prepared to organize and mobilize voters to support these plans consistently and strategically over a long period of time. And they should, ultimately, be prepared to forego the implementation of such plans only if those who benefit from the anti-democratic structure of the Senate act to relinquish those benefits.

I Choose, You Decide: Structural Tools for Supreme Court Legitimation

Executive Summary: The Supreme Court’s legitimacy is under threat. Efforts to either respond to the crisis of legitimacy or to salvage what legitimacy remains are focusing on reforms to the selection, appointment, and tenure of Justices. I propose an additional, complementary change, which does not require constitutional amendment. The selection of cases for the Court’s discretionary docket should be performed by a different group of Justices than those who hear and decide the cases on that docket. The proposal leverages the insight of the “I cut, you choose” procedure for ensuring fair division of heterogeneous resources–only here, it manifests as “I choose, you decide.” Further refinements of the proposal, for example to include reforms to the length of active tenure of Supreme Court Justices, are also considered.

Continue reading…

Jefferson’s Taper at IPSC 2018 (Berkeley)

In researching my in-progress monograph on value pluralism in knowledge governance, I made a fascinating discovery about the history of ideas of American intellectual property law. That discovery is now the basis of an article-length project, which I am presenting today at the annual Intellectual Property Scholars Conference, hosted this year at UC Berkeley. The long title is “Jefferson’s Taper and Cicero’s Lumen: A Genealogy of Intellectual Property’s Distributive Ethos,” but I’ve taken to referring to it by the shorthand “Jefferson’s Taper.” Here’s the abstract:

This Article reports a new discovery concerning the intellectual genealogy of one of American intellectual property law’s most important texts. The text is Thomas Jefferson’s 1813 letter to Isaac McPherson regarding the absence of a natural right to property in inventions, metaphorically illustrated by a “taper” that spreads light from one person to another without diminishing the light at its source. I demonstrate that Thomas Jefferson directly copied this Parable of the Taper from a nearly identical parable in Cicero’s De Officiis, and I show how this borrowing situates Jefferson’s thoughts on intellectual property firmly within a natural law tradition that others have cited as inconsistent with Jefferson’s views. I further demonstrate how that natural law tradition rests on a classical, pre-Enlightenment notion of distributive justice in which distribution of resources is a matter of private beneficence guided by a principle of proportionality to the merit of the recipient. I then review the ways that notion differs from the modern, post-Enlightenment notion of distributive justice as a collective social obligation that proceeds from an initial assumption of human equality. Jefferson’s lifetime correlates with a historical pivot in the intellectual history of the West from the classical notion to the modern notion, and I argue that his invocation and interpretation of the Parable of the Taper reflect this mixing of traditions. Finally, I discuss the implications of both theories of distributive justice for the law and policy of knowledge governance—including but not limited to intellectual property law—and propose that the debate between classical and modern distributivists is more central to policy design than the familiar debate between utilitarians and Lockeans.

Slides for the presentation are available here.

New Draft: Post-Sale Confusion in Comparative Perspective (Cambridge Handbook on Comparative and International Trademark Law)

It’s the summer of short papers, and here’s another one: Post-Sale Confusion in Comparative Perspective, now available on SSRN. This is a chapter for an edited volume with a fantastic international roster of contributors, under the editorial guidance of Jane Ginsburg and Irene Calboli. My contribution is a condensed adaptation of my previous work on the ways trademark law facilitates conspicuous luxury consumption, with a new comparative angle, comparing post-sale-confusion doctrine to the EU’s misappropriation-based theory of trademark liability. Comments, as always, are welcome.

New Draft: Finding Dilution (An Application of Trademark as Promise)

I’ve just posted to SSRN a draft of a book chapter for a forthcoming volume on trademark law theory and reform edited by Graeme Dinwoodie and Mark Janis. My contribution, entitled “Finding Dilution,” reviews the history and theory of the quixotic theory of liability that everybody loves to hate. As Rebecca Tushnet has noted, in a post-Tam world dilution may not have much of a future, and my analysis in this draft may therefore be moot by the time this volume gets published. But if not, the exercise has given me an opportunity to extend the theoretical framework I established and defended in the Stanford Law Review a few years ago: Trademark as Promise.

In Marks, Morals, and Markets, I argued that a contractualist understanding of trademarks as a tool to facilitate the making and keeping of promises from producers to consumers offered a better descriptive–and more attractive normative–account of producer-consumer relations than the two theoretical frameworks most often applied to trademark law (welfarism and Lockean labor-desert theory). But I “intentionally avoided examining contractualist theory’s implications for trademark law’s regulation of producer-producer relationships” (p. 813), mostly for lack of space, though I conjectured that these implications might well differ from those of a Lockean account. In my new draft, I take on this previously avoided topic and argue that my conjecture was correct, and that the contractualist account of Trademark as Promise offers a justification for the seeming collapse of trademark dilution law into trademark infringement law (draft at 18):

This justification, in turn, seems to depend on a particular kind of consumer reliance—reliance not on stable meaning, which nobody in a free society is in a position to provide, but on performance of promises to deliver goods and services. It is interference with that promise—a promise that does not require the promisor to constrain the action of any third party against their will—that trademark law protects from outside interference. A contractualist trademark right, then, would be considerably narrower than even the infringement-based rights of today. To recast dilution law to conform to such a right would be to do away with dilution as a concept. A promise-based theory of dilution would enforce only those promises the promisor could reasonably perform without constraining the freedom of others to act, while constraining that freedom only to the extent necessary to allow individuals—and particularly consumers—to be able to determine whether a promise has in fact been performed.

As they say, read the whole thing. Comments welcome.

 

New Draft: Brand Renegades Redux

Charlottesville Riot

I have posted to SSRN a draft of the essay I contributed to Ann Bartow’s IP Scholarship Redux conference at the University of New Hampshire (slides from my presentation at the conference are available here.) These are dark times, and the darkness leaves nothing untouched–certainly not the consumer culture in which we all live our daily lives. As I say in the essay, Nazis buy sneakers too, and often with a purpose. We all–brand owners, consumers, lawyers, and judges–should think about how we can best respond to them.

Right and Wrong vs. Right and No-Right

This is a point that is probably too big for a blog post. But as the end of the Supreme Court term rolls around, and we start getting decisions in some of the more divisive cases of our times, something about the political undercurrents of the Court’s annual ritual has me thinking about the way it tends to legalize morality, and how much of our political narrative has to do with our disagreement about the way morals and law interact. I’m not speaking here about the positivist vs. anti-positivist debate in jurisprudence, which I view as being primarily about the ontology of law: what makes law “law” instead of something else. That question holds fairly little interest for me. Instead, I’m interested in the debate underlying that question: about the relationship between law on the one hand and justice or morality on the other. Most of the political energy released over these late decision days is not, I think, about the law, nor even about the morality, of the disputes themselves. Instead, it seems to me to be about the extent to which moral obligations ought or ought not to be legal ones–particularly in a democratic country whose citizens hold to diverse moral systems.

There is a long philosophical tradition that holds there is a difference between the kinds of conduct that can be enforced by legal coercion and the kinds that may attract moral praise or blame but which the state has no role in enforcing. This distinction–between the strict or “perfect” duties of Justice (or Right) and the softer or “imperfect” duties of Virtue (or Ethics)–is most familiar from Kant’s moral philosophy, but it has precursors stretching from Cicero to Grotius. Whether or not the distinction is philosophically sound or useful, I think it is a helpful tool for examining the interaction of morals and laws–but not in the sense in which philosophers have traditionally examined them. For most moral philosophers, justice and virtue are complementary parts of a cohesive whole: a moral system in which some duties are absolute and others contingent; the latter must often be weighed (sometimes against one another) in particular circumstances, but all duties are part of a single overarching normative system. But I think the end of the Supreme Court term generates so much heat precisely because it exposes the friction between two distinct and sometimes incompatible normative systems: the system of legal obligation and the system of moral obligation.

The simplest world would be one in which the law required us to do everything that was morally obligatory on us, forbade us to do everything that was morally wrong, and permitted us to do everything morally neutral–where law and morality perfectly overlap. But that has never been the world we live in–and not only because different people might have different views about morality. Conflicts between law and morality are familiar, and have been identified and examined at length in the scholarly literature, sometimes in exploring the moral legitimacy of legal authority, other times in evaluating the duty (or lack thereof) of obeying (or violating) unjust laws. Such conflicts are at least as old as Socrates’ cup of hemlock; one can trace a line from Finnis back to Aquinas, and from Hart back to Hobbes. Moreover, because in our society moral obligations are often derived from religious convictions, and our Constitution and statutes give religious practices a privileged status under the law, these conflicts are quite familiar to us in the form of claims for religious exemption from generally applicable laws–historically in the context of conscientious objection to military service, and more recently as an expanding web of recurring issues under the Religious Freedom Restoration Act.

But religious convictions are not the only moral convictions that might conflict with legal obligations. And the question whether one ought to obey an unjust law represents only one type of intersection between the two normative systems of law and morals–the most dramatic one, certainly, and the one that has garnered the most attention–but not the only one. Some of those intersections will present a conflict between law and morality, but many will not. Still, I think each such intersection carries a recognizable political valence in American society, precisely because our political allegiances tend to be informed by our moral commitments. I’ve outlined a (very preliminary) attempt to categorize those political valences in the chart below, though your views on the categories may differ (in which case I’d love to hear about it):

 Political Valence
Moral Categories
Wrong Suberogatory Morally Neutral Right Supererogatory
Legal Categories Forbidden Law and Order Nanny State Victimless Crime Civil Disobedience
Permitted Failure of Justice The Price of Liberty The Right to be Let Alone Good Deeds Saints and Heroes
Required Just Following Orders Red Tape Civic Duty Overdemanding Laws

Now of course, people of different political stripes will put different legal and moral situations into different boxes in the above chart, and frame the intersection of law and morals from the points of view of different agents. In our current political debate over enforcement of the immigration laws, for example, an American conservative might frame the issue from the point of view of the immigrant, and put enforcement in the “Law and Order” box; while an American progressive might frame the issue from the point of view of federal agents, and put enforcement in the “Just Following Orders” box. Conversely, to the progressive, the “Law and Order” category might call to mind the current controversy over whether a sitting president can be indicted; to a conservative, the “Just Following Orders” category might call to mind strict environmental regulations. Coordination of health insurance markets through federal law might be seen as an example of the Nanny State (conservative) or of Civic Duty (progressive–though 25 years ago this was a conservative position); permissive firearms laws as either a Failure of Justice (progressive) or as part of the Right to be Let Alone (conservative).

The complexity of these interactions of law and morality strikes me as extremely important to the functioning of a democratic, pluralistic society committed to the rule of law. On the one hand, the coexistence of diverse and mutually incompatible moral systems with a single (federalism aside) legal system means that inevitably some people subject to the legal system will identify some aspects of the law with a negative political valence while other people subject to the same legal system will identify those same aspects of the law with a positive political valence. For a society like ours to function, most of these people must in most circumstances be prepared to translate their moral dispute into a political-legal dispute: to recognize the legitimacy of law’s requirements and channel their moral disagreement into the democratic political process of changing the law (rather than making their moral commitment a law unto itself). The types of political valences I’ve identified may be a vehicle for doing precisely that: they provide a narrative and a framework that focuses moral commitments on political processes with legal outcomes. Our political process thus becomes the intermediating institution of our moral conflicts.

These types of moral disagreements are likely to account for the vast majority of disagreements about which political valence is implicated in a particular legal dispute. But interestingly, I think it is also likely that some disagreements over the political valence at the intersection of a legal and a moral category could arise even where people are in moral agreement. In such cases, the right/wrong axis of moral deliberation is replaced with the right/no-right axis of Hohfeldian legal architecture: we agree on what the parties morally ought to do, and the only question is how the law ought to be structured to reflect that moral agreement. Yesterday’s opinion in the Masterpiece Cakeshop case strikes me as evidence of this possibility, and I’m unsure whether that makes it comforting or concerning.

To be sure, there are deep moral disagreements fueling this litigation. The appellants and their supporters believe as a matter of sincere religious conviction that celebrating any same-sex marriages is morally wrong, and that it is at least supererogatory and perhaps morally required to refrain from contributing their services to the celebration of such a marriage. The respondents and their supporters believe that celebrating loving same-sex marriages is at least morally neutral and more likely morally right, and that refusing to do so is at least suberogatory and more likely morally wrong. (Full disclosure: I’m soundly on the side of the respondents on this moral issue.) But that moral dispute is not addressed in the opinion that issued yesterday. The opinion resolves a legal question–the question whether Colorado’s anti-discrimination laws had been applied in a way that was inconsistent with the appellants’ First Amendment rights. And the Court’s ruling turned on peculiarities of how the Colorado agency charged with enforcing the state’s antidiscrimination law went about its business–particularly, statements that Justice Kennedy believed evince “hostility” to religious claims.

It is quite likely that both political progressives and political conservatives would agree that “hostility” to religious beliefs on the part of state law enforcement officials is morally wrong, or at least suberogatory. And if that–rather than the morality of celebrating same-sex marriages–is the real moral issue in the case, the parties’ deep moral disagreement moves to one side, and instead we must simply ask how the law ought to be fashioned to avoid the moral wrong of anti-religious hostility. Here, interestingly, the typical moral progressive/conservative battle lines are either unclear or absent. If you agree with Justice Kennedy’s characterization of the facts (which you might not, and which I do not), and if you believe the respondents lack the power to compel the appellants to provide their services in connection with same-sex wedding celebrations (or that they ought to lack this power), you likely believe that the case so framed is a vindication of law and order. But even if you think the respondents do (or should) have the power to compel the appellants to provide their services for same-sex weddings, and you agree with Justice Kennedy’s characterization of the facts, invalidating this otherwise permissible state action on grounds that it was motivated by morally suspect “hostility” might be acceptable so framed as a vindication of Law and Order, or at worst seen as an example of the Nanny State. And this reveals an important point that I think likely motivated the Justices in Masterpiece Cakeshop: with this resolution of the case there is a way for moral adversaries to agree on the political valence of the outcome.

There is obviously no guarantee that the litigants or their supporters will come to such agreement. To the contrary, it seems more likely that both parties’ supporters will see the “hostility” reasoning as a distraction from the outcome, which still touches on the deeper moral issues involved: the appellants’ supporters will see the outcome as a vindication of Law and Order and an invitation to Good Deeds, while respondents’ supporters will see it as a Failure of Justice and an instruction to state law enforcement agents to Just Follow Orders. But at the very least, political agreement is possible in a way that it would not be if the Court had aligned the law with one of the conflicting moral frameworks of the litigants and against the other.

It is both the virtue and the weakness of this type of solution is that it solves a legal controversy without taking sides in the moral disputes that ultimately generated the controversy in the first place. In so doing, it insulates the law from contests of morality and, possibly, of politics–but those contests haven’t gone away. Conversely, such solutions might have the effect of insulating politics from law: if the courts will only decide disputes on grounds orthogonal to the moral commitments underlying political movements, such movements may cease to see the law as a useful instrument, and may start casting about for others. I’m not sure that’s a healthy result for a society trying to hold on to democracy, pluralism, and the rule of law all at the same time. On the other hand, I’m not sure there’s a better option. The belief that we can, and somehow will, forge a common legal and political culture despite our deep moral disagreements is not one I think a republic can safely abandon.

Missing the Data for the Algorithm

A short new paper by Steven T. Piantadosi of the University of Rochester has some interesting implications for the types of behavioral modeling that currently drives so much of the technology in the news these days, from targeted advertising to social media manipulations to artificial intelligence. The paper, entitled “One parameter is always enough,” shows that a single, alarmingly simple function–involving only sin and exponentiation on a single parameter–can be used to fit any data scatterplot. The function, if you’re interested, is:

Screen Shot 2018-05-30 at 3.27.27 PM

Dr. Piantadosi illustrates the point with humor by finding the value of the parameter (θ) in his equation for scatterplots of an elephant silhouette and the signature of Joan Miró:

Screen Shot 2018-05-30 at 3.20.23 PM

The implications of this result for our big-data-churning, AI-hunting society are complex. For those involved in creating algorithmic models based on statistical analysis of complex datasets, the paper counsels humility and even skepticism in the hunt for the most parsimonious and elegant solutions: “There can be no guarantees about the performance of [the identified function] in extrapolation, despite its good fit. Thus, … even a single parameter can overfit the data, and therefore it is not always preferable to use a model with fewer parameters.” (p. 4) Or, as Alex Tabarrok puts it: “Occam’s Razor is wrong.” That is, simplicity is not necessarily a virtue of algorithmic models of complex systems: “models with fewer parameters are not necessarily preferable even if they fit the data as well or better than models with more parameters.”

For non-human modelers–that is, for “machine learning” projects that hope to make machines smarter by feeding them more and more data–the paper offers us good reason to think that human intervention is likely to remain extremely important, both in interpreting the data and in constructing the learning exercise itself. As Professor Piantadosi puts it: “great care must be taken in machine learning efforts to discover equations from data since some simple models can fit any data set arbitrarily well.” (p. 1), and AI designers must continue to supply “constraints on scientific theories that are enforced independently from the measured data set, with a focus on careful a priori consideration of the class of models that should be compared.” (p. 5) Or, as Kevin Drum puts it: “A human mathematician is unlikely to be fooled by this, but a machine-learning algorithm could easily decide that the best fit for a bunch of data is an equation like the one above. After all, it works, doesn’t it?”

For my part, I’ll only add that this paper is just a small additional point in support of those lawyers, policymakers, and scholars–such as Brett Frischmann, Frank Pasquale, and my soon-to-be colleague Kate Klonick–who warn that the increased automation and digitization of our lives deserves some pushback from human beings and our democratic institutions.  These scholars have argued forcefully for greater transparency and accountability of the model-makers, both to the individuals who are at once the data inputs and, increasingly the behavioral outputs of those models, and to the institutions by which we construct the meaning of our lives and build plans to put those meanings into practice. If those meanings are–as I’ve argued–socially constructed, social processes–human beings consciously and intentionally forming, maintaining, and renewing connections–will remain an essential part of making sense of an increasingly complex and quantified world.

 

Trademark Clutter at Northwestern Law REMIP

I’m in Chicago at Northwestern Law today to present an early-stage empirical project at the Roundtable on Empirical Methods in Intellectual Property (#REMIP). My project will use Canada’s pending change to its trademark registration system as a natural experiment to investigate the role national IP offices play in reducing “clutter”–registrations for marks that go unused, raising clearance costs and depriving competitors and the public of potentially valuable source identifiers.

Slides for the presentation are available here.

Thanks to Dave Schwartz of Northwestern, Chris Buccafusco of Cardozo, and Andrew Toole of the US Patent and Trademark Office for organizing this conference.