The University of Chicago Law School Faculty Blog

Sonia Katyal's Comments [Mobblog: The New Servitudes]

Hi everyone,

I just want to begin my commentary by echoing how great Molly’s paper is.  I have long been a fan of Molly’s work (her Distributive Values in Copyright piece is one of my favorite recent papers in IP, and presciently applies to the areas of IP that often raise vexing questions about grey areas of legality and fair use).  Seen from this context, in addition, Van Houweling’s newest paper is useful precisely because it asks us, albeit implicitly, to think about distributive complexities, but from a vantage point that focuses on the overlay between contract and property principles in the servitude context.  When you add the additional complexities generated by the different contexts she explores—real property, personal property, and intellectual property, the result is a really insightful framework for understanding both the limitations and the possibilities of servitudes in this context.

Moreover, part of the reason why I was so happy to see this group is because it’s a great mix of commentators, some of whom are known for their lauding of freedom of contract, and others who are known for their measured allegiance to property principles.  And others, of course, who are perhaps slightly more agnostic between these areas, but who have focused on the interactivity between these two in their own work.  For me, Molly’s piece, the New Servitudes, is useful, not only because it forces us to rethink the property/contract interface in a new way, but because it also leads us to contemplate how application of this interface raises an interesting host of problems in the clickwrap and shrinkwrap worlds we live in today.

As a primary matter, I want to echo Lee Ann Fennell’s useful point about looking deeper into the different forms of servitudes.  Fennell usefully analogizes between the negative easements in gross, conservation servitudes and the Microsoft EULA.  She also points out that the GPL and Creative Commons licenses look more like covenants in common interest communities.  I think we can also go deeper into the analogy, too, and explore some of the distributive ramifications of these different types of servitudes, as well, and who gravitates to these arrangements in particular.  In returning to Fennell’s assertion regarding the different types of servitudes that might be in play—it also might be helpful to explore a potential divergence among the servitudes regarding whether they seek to consolidate (and therefore strengthen) an existing bundle of rights (as in Microsoft’s EULA) or whether they seek to disaggregate (and potentially weaken) the underlying IP rights involved (arguably, both the CC and GPL tend to fragment particular entitlements from copyright, and add other things like attribution to the bundle).

Moreover, real servitudes, it seems to me, are directed towards a mutually beneficial relationship between the parties. whereas servitudes in gross seem to be designed more for the benefit of one party in mind.

These considerations may require us to imagine a reblending/remixing of other types of servitudes, ie maybe an affirmative easement analogy would help here in specific contexts.  But in the end, I completely agree that these considerations are quite distinct from the “running” problems due to their inherent public policy implications.

For that matter (on the public policy angle), I found myself wanting to hear more regarding the observation that “[t]he information cost problem would be more serious if these licenses imposed conditions outside the scope of copyright law’s exclusive rights….” (p. 52).  In some ways, this is at the heart of the property v. contract interface—if we see things from a slightly more restrictive standpoint (something tcould easily imagine, given the myriad of end user license agreements that are far more restrictive than the ones we might see here), servitudes empower owners to utilize contract-based agreements that foreclose things like reverse engineering, as Molly points out, and things like the first sale exception.  The question, however, about how these restrictions affect information costs, ex ante, is something that I found myself wanting to hear more from Molly about how more restrictive licenses alter the information cost problems that she describes—is the change in information costs a difference of degree, of kind, or both?  What new information costs do we see, and how can these agreements shift to respond to them up front?  These are some things that I wondered whether it might be helpful to elucidate at a later point, perhaps.

On her really insightful points regarding the ‘problem of the future,’ in licensing here, I enjoyed her identification of two kinds of problems: (a) the problem of licensing incompatibilities, and (b) the problem of protecting downstream innovation from potential thickets.

Here, however, I found myself wondering about the potential costs and benefits that we could reasonably operationalize from the kind of ‘changed circumstances’ scenarios that we see in servitude cases in the real property context, and if so, what that kind of a solution might look like, given the remoteness, durability, and ubiquity issues she raises in the electronic context.  Is the doctrine survivable in the IP/information age context, given Henry’s points about attention scarcity?   

Furthermore, it seems to me that one additional complexity that we have to grapple with is the special status of copyright, as opposed to land: that certain licenses, even of the Microsoft variety, involve restrictions that are meant to “run with the work,” as Molly points out.

But sometimes, at least part of the time, the underlying “work” that is being protected might arguably comprise a mixture of both copyrighted and non-copyrighted elements, i.e elements that are taken from the public domain or elsewhere, and hence might produce a more complex mixture of commonly and privately held information (a semi-commons, perhaps?).  Unlike the tangible property scenario, where these elements are almost easily always characterized as part of the domain of the owner, the copyright context suggests otherwise—it suggests that some elements might not be owned (or ownable) and others might be privately held.  For this reason, it seems to me that we have to grapple with the different types of digital content we are dealing with—and the implications that these forms might take for the parties concerned.

This leads me to a comment I wanted to make on Henry’s post, since his work on semi-commons has been deeply influential to my own thinking on IP.  I had a slightly different reaction to the three licenses in applying the exclusion/governance strategy to this context.  In my view, precisely because of the consolidating/enlarging implications of EULA (in that it goes beyond copyright restrictions in foreclosing reverse engineering, etc)—one might conceivably argue that it might resemble, at least in some ways, an attempt at a private exclusion strategy, although I definitely agreed in viewing both the CC and GPL scenarios as more governance-related because they seemed to be much more focused on delineating particular rights and expectations to the end user.  Of course, one could easily argue in the opposite direction, as Henry points out, but I wondered if it might be interesting to think about the exclusionary approach here as well.  Given our ongoing discussion about enforcement—injunctive relief vs. damages – the licenses themselves seemed to suggest the need to view these strategies along a slightly looser formulation of the exclusion/governance continuum of private property agreements—and Molly’s presentation of the licensing variations helps us to apply Smith’s framework in an interesting new angle.

Like many of the other commentators, I found myself particularly fascinated by questions of enforcement.  (After all, many of us property professors find ourselves, year after year, trying to explain Shelley v.

Kraemer’s constellation of state action in enforcement of covenants…and we could only imagine the effects a similar principle might have here.

But I digress.).  As we know, however, there are some pockets of state/judicial enforcement that could conceivably raise interesting parallels to explore here:  private rules and bylaws in homeowner’s contexts have reasonableness restrictions, and courts sometimes won’t enforce covenants if they violate public policy or trample on constitutionally protected rights.  Further, in the real property context, however, we DO have helpful tests that we can use to guide state intervention or regulation: nuisance offers us the ‘unreasonable interference with use and enjoyment,’ test in the real property context.

Is there a nexus—or even a justification—for state oversight here, at all, given the public policy implications Molly identifies?

Finally, as for Stuart’s great point about looking for the reason behind enforcement of servitudes in the IP context—his question sets up a basic scenario about the interactivity between intellectual property and contract.  In his scenario, if the servitude accomplishes all of the same goals as copyright does, then why opt for copyright, he asks?  This is a great question.  We see, after all, why servitudes are attractive, in that they operate as helpful overlays onto copyright law (from the licensor’s perspective) is because the nature of the private agreement allows for the foreclosure of options like fair use, options that the licensee might enjoy with only copyright protection.  But here—we might wonder why the drafting of the servitudes themselves have not shifted to comport with the externalities that they produce.  One answer could be that the externalities that Molly identifies come into play at a much later point—and it is hard to imagine how we could craft agreements that might anticipate these costs at the outset (ie loss of fair use, etc.).

However, one potential, free market defense could be to argue that in such cases, the license agreements would ideally change to reflect a shift in market demands along these lines, and I confess that the optimist in me wants to agree.  After all, some license agreements force the end user to disclose whether the use is for commercial vs. personal purposes, thereby facilitating price discrimination along those lines.

Maybe it is possible for licenses to shift to respond to the externalities Molly identifies, then, and perhaps it might be interesting to compare some of those cases to the licenses at issue here.

The last area that I want to touch on is remedy and one of my favorite areas to think about: the role of property rules, liability rules, and “pliability” rules (with a tip o’ the hat to the imaginative team of Avi Bell and Gideon Parchomovsky in crafting this term).  Like Stuart, I also believe that the employment of money damages might shift the paradigm in inventive ways—I am fascinated by the proposition that money damages might reduce the incentives for folks to engage in “due diligence’ in determining whether works are protected, and therefore might be less incentivized to seek Microsoft’s permission.  To me, I think these licenses are perfect opportunities for us to think through the formation of ‘pliability’ solutions here (here, maybe copyright misuse might be an interesting avenue to think through).  Though, in the end, however the point about remedies is also a distributive point—we might incentivize a different type of innovation by utilizing a different type of remedy.

Van Houweling Responds Part II: First Sale and Patent Exhaustion [Mobblog: The New Servitudes]

As Henry and Jim both note in their comments, there are doctrines within intellectual property law that are entangled with the servitude issues that I discuss in The New Servitudes.  Specifically, both the first sale doctrine embodied in section 109 of the Copyright Act and the software-specific provisions of Section 117 that Jim mentions allow owners of copies of copyrighted works to do specified things with those copies notwithstanding the exclusive rights of copyright holders.  In the patent context the analogous judge-made doctrine is patent “exhaustion”—the meaning of which the Supreme Court is currently considering in the LG v. Quanta case to which Henry refers.  I’m anxious to discuss Quanta.  But first let me give a little background (familiar to those who have read the paper) about the connections I see between first sale, exhaustion, and more general concerns with servitudes.

I argue in the paper that the early twentieth century cases in which these IP doctrines developed reflect an anxiety about enforcing servitudes attached to items of personal property (an anxiety that is in turn explained by the concerns with notice, “the problem of the future,” and externalities that I described in my initial post).

For example, in Bobbs-Merrill Co. v. Straus, the case that established the (later codified) first sale doctrine, a book publisher attempted to enforce a restriction printed inside books that purported to limit the price at they could be resold.  The plaintiff copyright owner argued that by distributing books subject to restrictive terms it was granting purchasers only a conditional license to exercise the vending right that the Copyright Act otherwise granted exclusively to it; vending outside the terms of that license (that is, selling books for more than the minimum price) therefore amounted to copyright infringement.  The Copyright Act, according to this logic, provides a mechanism for imposing running restrictions on chattels that the common law lacks.  The Second Circuit rejected that argument, and the Supreme Court affirmed.  210 U.S. 339 (1908).

The Court’s opinion purports merely to interpret the language of the Copyright Act, concluding that the right to “vend” granted in the Act is exhausted as to a given copy of a copyrighted work once that copy is sold.  But this interpretation seems to have been motivated in part by the more general notion that chattels should not be burdened with running restrictions.  Now, to be sure, the statutory rights granted to copyright holders supersede that hostility to some extent: there are some things that the owner of a copyright-embodying chattel is not permitted to do with it (for example, reproduce each of its pages) on account of the non-possessory intellectual property rights created by copyright (which have, as I noted in my earlier post, a servitude-like character).  But Bobbs-Merrill articulates limits on even a copyright holder’s power to impose running restrictions on personal property.  In the patent context, a similar logic seems to motivate cases like Bauer v. O’Donnell, 229 U.S. 1 (1913), in which the Court refused to enforce a retail price restriction printed on packages of patented medicine because “a patentee who has parted with a patented machine by passing title to a purchaser has placed the article beyond the limits of the monopoly secured by the patent act.”

These cases (and other similar cases including Straus v. Victor Talking Machine and Motion Picture Patents Co. v. Universal Film Manufacturing Co.) are complicated by the fact that the substance of the restrictions was arguably anti-competitive--a fact that has led several commentators to suggest that the general hostility toward chattel servitudes (and the specific IP doctrines of first sale and exhaustion) could be replaced with particularized antitrust scrutiny--especially now that antitrust law is more doctrinally mature than it was when these cases were decided.  Richard’s post suggests that he would endorse this view.  And it is also reflected in the Federal Circuit’s 1992 decision in Mallinckrodt, Inc. v. Medipart, 976 F.2d 700 (1991). 

In Mallinckrodt the Federal Circuit characterized the early patent exhaustion cases as “establish[ing] that price-fixing and tying restrictions accompanying the sale of patented goods were per se illegal.  These cases did not hold, and it did not follow, that all restrictions accompanying the sale of patented goods were deemed illegal.”  The court, therefore, refused to interpret those cases as necessarily invalidating a use restriction that neither fixed prices nor tied multiple products but rather limited a patented medical device to “single use only.”  The court suggested that so long as the condition did not violate the antitrust laws, it was enforceable against a purchaser of the medical device through a suit for patent infringement.  The Federal Circuit reaffirmed this logic in LG v. Quanta, in which the Supreme Court recently heard oral argument. 

To my mind, the Federal Circuit has adopted an overly narrow view of the purpose of the exhaustion doctrine.  Safeguarding competition through antitrust might deal with some types of negative externalities, but it does not eliminate every problem associated with chattel servitudes.  Problems with notice and information costs, and with unjustified but inescapable burdens on future activity (the problem of the future) can still loom large. 
      
There are some interesting allusions to these problems in the LG v. Quanta oral argument.  Justice Breyer suggested that the “doctrine that you cannot impose equitable servitudes upon chattel” might help to decide the case.  And he returned several times to a hypothetical in which a patent holder tries to enforce restrictions on a consumer’s use of patented bicycle pedals: 

“Imagine that I want to buy some bicycle pedals, so I go to the bicycle shop….The inventor has licensed somebody to make them, and he sold them to the shop.  I go and buy the pedals.  I put it in my bicycle.  I start pedaling down the road.  Now, we don’t want 19 patent inspectors chasing me….” 

Breyer apparently finds this a troubling scenario even if the pedals come with some sort express notice of the use restrictions.  He asked whether

“if … I go in the bicycle shop and I buy the pedals and then they give me … one of these pieces of paper that has all of the 42,000 words on it and there are in these 42,000 words it says, and now you are put on notice that once you put it in your bicycle and you pedal away, they’re going to get you and you’re going to be hauled into Patent Court, then--then that’s ok?”

I don’t want to read too much into the tea leaves of this oral argument transcript.  But Breyer seems to be suggesting (in part relying on the traditional rule against chattel servitudes) that consumers would be justifiably surprised to find themselves “hauled into Patent Court” for merely using things they had lawfully acquired, even where the restriction was imposed by a patent holder by means of a written notice attached to the good.  As Henry's post suggests, the problem here is not exactly lack of notice--the information is available in the “42,000 words” in Breyer’s hypothetical.  But it may not be reasonable or even desirable to expect people (at least not retail pedal purchasers) to absorb that information.  The Court expressed this same anxiety back in Straus v. Victor Talking Machine Co., in which it refused to enforce a restriction printed on plates attached to record players after noting that “not one purchaser in many would read such a notice, and that not one in a much greater number, if he did read it, could understand its involved and intricate phraseology.” 

This gets us back to Lee’s question:  don’t the problems that I associate with running servitudes also plague restrictions that don’t run--e.g. plain vanilla contracts (which are also often unread, misunderstood, and not fully contemplated)?  What is especially problematic about servitudes?  That is a critical question to which I will return in my next post.

Henry Smith's Comments [Mobblog: The New Servitudes]

First of all, I’d like to echo previous posts in praise of Molly’s article. It is systematic and illuminating on the question of servitudes in three related but very distinct contexts (real property, personal property, and IP), despite the theoretical and empirical difficulties servitudes have always presented. Think of poor Bigelow heroically tackling the touch and concern requirement. 

In both the article and especially in her post, Molly, correctly in my view, makes the desirability of enforcing servitudes a comparative exercise. This is truly Coasean in spirit, even if the answer does not always point in the direction of contract enforcement. As a number of posts have reflected upon, one main issue here is notice, and in particular the most cost-effective method of furnishing it. While Richard is right that systematic and centralized land records do often provide effective notice, it is an empirical question how they stack up against other methods in any given type of situation. Other methods include standardization, equitable doctrines of notice (which incidentally apply in personam and not in rem), and doctrines absolving those encountering rights from liability. Where a legal device fall between in personam contract and in rem property, we should expect intermediate strategies to deal with the potentially large but still limited set of dutyholders. See Thomas W. Merrill & Henry E. Smith, The Property/Contract Interface, 101 Colum. L. Rev. 773 (2001). 

When notice is the issue it is often important to keep in mind here that it is not information that is scarce but rather attention as Herbert Simon pointed out a long time ago. See Herbert A. Simon, Designing Organizations for an Information-Rich World, in Computers, Communication, and the Public Interest 37, 40-41 (Martin Greenberger ed., 1971). Thus, even where land records or notices printed on a product may give notice in some sense, there might still be reasons to force a standardized format (cf. nutrition information, or the terms of consumer loans). Even the land records are not a “data dump” but limit the types and form of documents that are permitted to be recorded. Format can matter. For example, a rule that rent is incompatible with fee ownership means that once one knows that an interest is a fee simple, one can stop looking for information along this dimension. Similar problems arise in contract, and are solved with a different mix of private and public solutions (making contracts shorter, enforcing reasonable consumer expectations etc.) 

One reason servitudes present a problem of informational detail is that they implement a governance strategy. Basic exclusion (keep off, boundaries) is a platform upon which we can build governance regimes, i.e. rules of proper ruse. Governance rules refine and supplement the basic exclusionary regime when particular use conflicts are important enough. Governance rules can be contractual, common law, or some combination of statute and regulation. (For an application to IP, see Henry E. Smith, Intellectual Property as Property: Delineating Entitlements in Information, 116 Yale Law Journal 1742, 1784-98 (2007).)Servitudes are a largely private governance regime. One possibility here is that courts have little problem with servitudes as long as they can be said to refine and supplement the basic exclusionary regime. Servitudes that are not refinements but unrelated (e.g. the haircut) or more than a mere refinement (e.g. going outside the copyright baseline) present information problems that normal governance regimes do not. 

A particularly interesting aspect of Molly’s paper is its treatment of new servitudes and how they implicate some of the traditional concerns with servitudes both more and less than real and personal property servitudes do. In particular, the possibility that licenses can conflict downstream is a very important point. It is a little reminiscent of water law in which property rights definition is difficult because it is desirable or unavoidable that water rights interlock tightly (the return flow issue in first appropriation is a dramatic example). One feature of basic exclusion rights is their modularity: behind the boundary a lot of “internal” information about uses and the users is irrelevant to an outside dutyholder: the interface between the owner’s property and the rest of the world (in rem) is a simple keep off. I do not need to know much about use or owners to keep off Balckacre or not to steal a car in a parking lot.

For a variety of purposes we need refinements (governance) which complicate this interface where uses interact (e.g. nuisance and servitudes). One difference between land, chattels, and intangibles is that the exclusion strategy is easier to carry out for tangible property. The baseline is clearer: in the case of land there is a physical bubble that corresponds to the module that the exclusion strategy provides. By contrast, in IP this is necessarily artificial. Thus it is easier for servitudes in IP to fail to have reference to an exclusion baseline. Some do have reference to a clear ex ante baseline, as with the creative commons licenses favoring use but within the scope of the copyright. But some of Molly’s other examples do not (as where rights to criticize are being contracted away). The conflicting license issue too would not arise if IP were more naturally modularized: servitudes here can in principle be about anything and interact with each other in any way. The modularity of land rights through spatially defined exclusion limits the extent to which servitudes will come into conflict. Owners will be aware what a servitude will “cover” (almost literally) in the case of land. (If however, we followed the Legal Realists, and assert that there is no core to the bundle of sticks of rights in land, the situation would be a lot more like the one Molly points out for software). Software as a resource does not ensure this. 

Complex interfaces can reduce transferability, as in the case of water. In some kinds of property, those setting up property desire liquidity and this more than enough incentive for standardization (financial instruments are an example). In other cases, idiosyncratic rights (fancies) may “pollute” the general informational atmosphere, making information costs for others go up. The resulting need on the part of others generally to be on the lookout for more types of information in no set format can present an externality that exceeds the benefits of the idiosyncrasy to the transacting parties. What private incentives there are for liquidity and how large the externality is partly determine the need for standardization. Also, as long as the state is involved in enforcing property rights there can be economies of scope in the state taking on the standardization function as well. 

Many of the issues in this article are of relevance to the case of LG v. Quanta now at the U.S. Supreme Court. In this regard the distinction Molly draws between commercial producing entities and individual consumers (who may have more of an everyday expectation of being able to use a physical article) is potentially a good rule of thumb. Maybe this is a subject for a future post, but for one thing, those manufacturing under a license are a more expert audience with more at stake, than in the case of consumers. There is less reason for the law to worry about the processing costs of closer, more expert dutyholders. There are many issues raised by the LG v. Quanta case, and I’d be interested in how Molly and others think this fits into her overall scheme of concerns.

Jim Gibson’s Comments [Mobblog: The New Servitudes]

   Molly’s insights and observations in The New Servitudes seem self-evident—almost obvious—until one considers that no one has really made them before. For me, that’s the definition of a great article, and reading it was a rare treat. I’m delighted to be included in the mob that has a chance to blog on it here.

   I’ll share two thoughts here:

   1. The article got me thinking about the preferred structure for regulating transactions in information goods: the top-down, public law approach of copyright and other intellectual property regimes versus the bottom-up, private law approach of contracts and servitudes. This is obviously not an either/or proposition; indeed, as Molly points out, copyright law’s default entitlements form the point of departure for the private-law restrictions in the licenses that she examines (the Microsoft Vista EULA, the GPL, and the Creative Commons licenses) and thus influence the degree to which such licensing schemes impose additional notice costs, generate negative externalities, and create difficulties for downstream users.

   But it struck me that some of the problems that Molly identifies might be mitigated if public law provided a more flexible range of entitlement options. Consider the “Problem of the Future” that has arisen in the context of copyleft licenses. One is not likely to find a more cooperative and collaborative community than open-source programmers and their Creative Commons cousins. Yet their focus on long-term benefit and their repeat transactions have created and then locked in an increasingly obsolete set of norms; as Molly explains, the sheer number of contributors to these projects makes altering the existing license terms nearly impossible, even if everyone recognizes the advantages of doing so. (I can see the article’s title now: The Creative (Anti)Commons.)

   If Congress were to offer a smorgasbord of opt-in copyright protections, however, this anticommons issue might not arise. So why shouldn’t the government develop intermediate sets of entitlements, like the Creative Commons licenses? Such a regime would provide only default rules, so individualized bargaining could still take place, but it would mitigate the lock-in problem of large-scale, collaborative licensing regimes. More generally, an increase in public law options would reduce the need for individualized tailoring at the contract/servitude level, which could help address some of the other problems that Molly identifies.

   Among other things, such an approach would require resurrecting copyright’s defunct opt-in provisions (particularly notice), but as I’ve argued elsewhere that may be a good idea regardless of the servitudes issue. Also, it would generate some new costs, in the form of the need for the consumer to become familiar with these various regimes and identify which one governs a given transaction; I can only hypothesize that the corresponding benefits that I have identified would be greater.

   2. Although I took Property from Glen Robinson, my knowledge of the intricacies of servitudes is a more than a bit lacking (my fault, not Glen’s), and like many others I have tended to view shrink-wrap and click-wrap licenses through a contract lens. And in the contract context, the aspect of such licenses that bothers me the most is that notice of the licensing terms often arrives quite late in the transaction, if it arrives at all. The end user may have a sense that there are additional terms awaiting him or her in the box or during software installation, but the cost of acquiring these terms is considerable and usually involves a fair amount of investment on the purchaser’s part (at least relative to the overall value of the transaction). This insulates such terms from the market pressures that we might normally expect to do the heavy regulatory lifting here.

   In contract law, my understanding is that we handle late-arriving terms through the infamous “battle of the forms” (and its counterpart in UCC § 2-207). This approach has not gained much purchase in the world of shrink-wrap and click-wrap licensing (somewhat inexplicably, in my view—see ProCD, Inc. v. Zeidenberg, 86 F.3d 1447 (7th Cir. 1996); Hill v. Gateway 2000, Inc., 105 F.3d 1147 (7th Cir. 1997)), but it at least provides a theoretical framework within which we can think about late-arriving contract terms. I cannot think of a similar concept in the law of servitudes, but I thought I would throw this out there for those more knowledgeable about such things.

   Parsing this issue raises other more tangential issues. I will touch on three, all from the software world:

a. First, where do we stand if the seller loses the battle of the forms (or its servitudes equivalent)? In other words, what happens if the shrink-wrap or click-wrap license does not become part of the bargain or encumber the chattel? On this note, let me just mention the underappreciated section 117 of the Copyright Act, which creates a use privilege for the owner of a copy of software, thus obviating the need to get any license from the copyright holder.

b. Second, section 117 in turn raises the “license versus sale” issue, because software companies insist that the section does not apply; apparently no one ever “owns” any “copies” of their programs. I have never understood the “license versus sale” dichotomy; for me, asking, “Is this transaction is a license or a sale?” makes about as much sense as asking, “Which is faster, New York or by car?”

c. Finally, even if the user has a legal privilege to install and run a program, technological restrictions may prevent the exercise of this privilege. And the Digital Millennium Copyright Act puts legal weight behind those technological protections, so that even a skilled hacker may decide to forgo a privileged use.

   Many thanks to Molly for a great read and to the Chicago folks for including me here!

Van Houweling Responds: The Affirmative Case for Enforcing Servitudes [Mobblog: The New Servitudes]

A number of themes have emerged from the comments so far.  I’ll jump back into the fray by addressing--for starters--the question of the affirmative case for enforcing servitudes. 

Both Stew and Randy are interested in exploring the affirmative case for enforcing servitude-like arrangements in the intellectual property context.  Stew reminds us of one element of the affirmative case in the land context:  “the servitude was the only effective mechanism for a landowner to deal with the monopoly power of his neighbor.”  And he asks whether enforcement of IP servitudes would also yield efficiency gains. 

Randy suggests that the answer is yes--pointing to price discrimination (which Lee also mentions) and systems competition as two examples.  Richard refers more generally to the “superior internalization of costs and benefits” that enforcement might yield.  The literature suggests additional possibilities.  Consider, for example, Henry Hansmann and Reineir Kraakman’s observation that coordinating downstream treatment of artwork can protect and promote an artist’s valuable reputation:  “The value of one work by an artist generally depends heavily on the quantity and quality of the artist’s other works. . . . The destruction or alteration of individual works can therefore damage the artist’s overall reputation and thus decrease the value of the artist’s other works as well.”  Hansmann & Kraakman, Property, Contract, and Verification:  The Numerus Clausus Problem and the Divisibility of Rights, 31 J. Legal Stud. 373, 409 (2002) (pre-print version).

First, let me acknowledge the importance of addressing (and assessing) the affirmative case.  In The New Servitudes I focus primarily on the concerns raised by servitude enforcement across the contexts of real, personal, and intellectual property.  But in order to figure out what (if any) doctrinal and policy implications should be triggered by those concerns (as I hope to do in ongoing work), I of course need to weigh both the costs and benefits. 

Let’s focus for now on two potential benefits:  facilitating price discrimination and protecting creators' reputations.  Elsewhere I have briefly discussed the potential benefits of price discrimination for poorly-financed players in the copyright marketplace.  (See Distributive Values in Copyright, 83 Texas L. Rev. 1535,1569-71 (2005).)  As for protecting reputations--I see this as a minor theme in copyright, one that looms increasingly large with the rise of technologically-empowered creators who appear to be more motivated by reputational rewards than by the monetary rewards more traditionally associated with the exclusive rights granted by copyright law.  Indeed, reputational benefits are an important explanation for the growth of the free- and open-source software movements, and they seem critical to many Creative Commons licensors as well.

(I should say something here about how the servitude mechanism may be superior to other mechanisms--e.g. negotiated bilateral contracts--for facilitating price discrimination and protecting reputations.  For now let me concede the point that seems to be motivating some of Richard’s enthusiasm for the servitude mechanism:  those alternatives would be a pain in the neck.)

So should the benefits of enforcing servitude-like mechanisms that facilitate price discrimination and protect creators’ reputations trump the concerns I raise in The New Servitudes?  I’m not ready yet to offer a definitive answer, but I’ll introduce one way of thinking about the problem. 

We might say that intellectual property law itself acknowledges the need to coordinate uses of intangible works in ways that are more ubiquitous and durable than mere contractual agreements would be.  (Stew alludes to this when he points out that “[c]opyright itself is largely premised on the fear that creators of intellectual works cannot contract with all potential users, and that without statutory protection an author would be at the mercy of any user who could reproduce, display, produce derivative works, etc., thus destroying the market for the author’s work.”)  As I point out in the article and in a companion piece, the mechanism that intellectual property law deploys to achieve this coordination is itself servitude-like (creating non-possessory property rights that limit what end-users may do with a thing in their possession).  And, indeed, intellectual property itself triggers many of the concerns I associate with servitudes. 

So one way to think about the cost/benefit question is this:  Congress has already weighed the costs and benefits of servitude-like protection for intangible works of creativity and invention.  The results of that cost-benefit analysis are embodied in copyright and patent law--which create running restrictions but also impose limits (first sale, exhaustion, fair use, etc.) that address at least some of the concerns raised by servitude enforcement.  On this view, efforts to impose servitude-like restrictions above and beyond what is provided for by copyright and patent are (at least presumptively) invalid, because they upset the balance established by this Congressional scheme. 

(An aside:  I look forward to seeing what Henry has to say on this topic.  It seems to me that this approach resonates with some of his writing about the numerus clausus principle--in particular the points that he and Thomas Merrill make about the relative institutional competence of legislators versus common law courts when it comes to creating new types of property interests.  See Thomas W. Merrill & Henry E. Smith, Optimal Standardization in the Law of Property:  The Numerus Clausus Principle, 110 Yale L.J. 1, 58-68 (2000)).

To me, this approach has the virtue of taking seriously everything that Congress and the courts have said about the importance of the various balances built into intellectual property law.  On the other hand, the balance that Congress has struck as a general matter may not be the ideal way to promote creativity and innovation in individual cases.  The assumptions that Congress has made about what typically motivates creators and innovators may justify the general contours of intellectual property law without negating the possibility that some creators have idiosyncratic preferences that could be most effectively enforced via servitudes--and that serving those preferences would benefit society by inducing the production of new works, or better works, or works by a wider variety of creators, etc.  (Not to mention the possibility that the Congressional balance has been rendered obsolete by changes in technology and transaction costs--a possibility that Randy raises in From Edison to the Broadcast Flag.)

But this brings us back to the costs of servitudes.  It’s one thing (albeit not an unproblematic thing--an issue to which we might return) to expect consumers to understand the generally-applicable background law of copyright and patent.  It’s something else to expect them to parse every unilaterally imposed idiosyncratic variation.  Furthermore, even putting notice problems aside, enforcement of some of these idiosyncratic preferences will surely not be efficiency-enhancing.  It’s particularly easy to imagine reputation-protection run-amok.  (“By opening this book you agree never to say anything bad about it or its author.”)  Lee and Richard have both suggested that we have other sources of law (antitrust, civil rights, Lee’s Fair Housing Act analogy) that might take care of some of these concerns (which would be raised to some extent even by traditional contracts with similar terms).  I hope to return to that issue in a separate post. 

Stewart Sterk's Remarks on The New Servitudes [Mobblog: The New Servitudes]

Molly’s article ties together doctrinal strands in such a compelling way that it leaves me asking myself why I hadn’t thought of the analogies in the same way.  For me, that’s the mark of a successful piece.  On top of that, her article raises some interesting questions for another day.  I’ll raise three of them.
    (1) Molly’s piece focuses on the reasons for judicial skepticism about enforcement of servitudes.  With respect to land, she identifies three: notice and information costs, the problem of the future, and externalities, and then skillfully examines how those same reasons for skepticism apply with respect to intellectual property servitudes.  But one could ask the question from another perspective.  In light of the problems she identifies, why did courts enforce servitudes, despite the problems she identifies, all of which cast doubt on the assumption that a consensual servitude is value enhancing?  One answer is that the servitude was the only effective mechanism for a landowner to deal with the monopoly power of his neighbor.  Only a next-door neighbor can provide me with the assurance that I won’t live next door to a factory or apartment house; only my neighbor can assure me that my view won’t be blocked by a wall, etc.  If easements and restrictive covenants were not enforceable, every time neighboring land were sold, the purchaser would be able to exert that monopoly power by demanding payment for agreeing not to build the factory or the wall, and I wouldn’t pay, because my neighbor could then turn around and sell, leaving me at the mercy of the new purchaser.  This explanation dovetails with the common law’s reluctance to enforce affirmative covenants against successors.  With affirmative covenants, neighbors typically enjoyed no monopoly power.  If my neighbor didn’t agree to maintain a driveway over my land, I could always hire someone on the market to do the same thing.  On this view, courts had a positive reason to enforce restrictive (but not affirmative) covenants, and that positive reason contributed to judicial willingness to overcome the problems Molly identifies.  Is there a comparable positive reason for enforcing IP servitudes?  What efficiency gains does enforcement of these servitudes promote?  It seems to me that an answer to those questions would provide a nice bookend to Molly’s focus on the reasons for not enforcing servitudes.
    (2) The increasing use of IP servitudes raises another interesting question: should licensors who extract servitudes on what we call intellectual property have to “opt out” of copyright protection?  Copyright itself is largely premised on the fear that creators of intellectual works cannot contract with all potential users, and that without statutory protection, an author would be at the mercy of any user who could reproduce, display, produce derivative works, etc., thus destroying the market for the author’s work.  But if that premise is wrong – if the author can impose a servitude that binds all potential users to license terms – then why afford copyright protection at all?  Of course, a regime that required licensors to choose between servitudes and copyright would provide little benefit to potential users; the typical licensor would choose more onerous license terms that incorporated all copyright protections, and then some.  But there might also be some leakage: users who, for some reason would not be charged with notice (because, for instance, someone had engineered the “click-wrap” out of a piece of software).   
    (3) Molly’s article does not focus at all on remedy, and I wonder whether the problems she identifies with enforcing IP servitudes might be significantly ameliorated by limiting licensors to money damages rather than injunctive relief.   Many of the costs Molly identifies of enforcing servitudes (particularly underutilization of resources) would become less significant if a potential user knew that injunctive relief were not available.  Consider, for instance, Molly’s example of the researchers who reverse-engineer Windows 95.  If Microsoft could only obtain money damages, the researchers might not be terribly concerned about the difficulties of obtaining Microsoft’s permission.  Limiting IP owners to money damages generally reduces the incentive for potential users to engage in expensive searches to learn whether works are protected, and to learn who owns the associated IP rights ( a point I make in my forthcoming article, Property Rules, Liability Rules, and Uncertainty About Property Rights, 106 Mich L Rev ___ (2008).)  Moreover, there is certainly precedent for limiting servitude owners to money damages as a remedy in a variety of cases.  (see, for instance, NY RPAPL sec. 1951 (prohibiting award of injunctive relief when servitude is of “no actual and substantial benefit” to the person holding the servitude).
    Thanks to Molly for providing such a provocative article!

Stew Sterk

The New Servitudes

I’m delighted to be a guest here at the Faculty Blog.  And I’m looking forward to this week’s mobblog about my article, “The New Servitudes” (forthcoming in the Georgetown Law Journal).  The article lies at the intersection of the two fields in which I teach—intellectual property and tangible property.  So it’s a real honor to be discussing it online with luminaries from both of those fields.

Let me get things started with an overview of the paper, which is ultimately about those pesky “licenses” that purport to define how consumers may use computer programs, digital music, and other intangible works of the information age.  For example, you have undoubtedly downloaded a computer program from the Internet after viewing a screen of text limiting how the program may be used; installation began only after you clicked “I agree.” 

Courts in the United States have increasingly enforced such restrictions—labeling them “click-wrap licenses” and applying to them the same contractual concepts that govern face-to-face exchanges of promises.  But I argue in the paper that the law of tangible property offers a different and more powerful lens through which to view them. 

As others have noted (including Glen Robinson in the University of Chicago Law Review), these licenses can usefully be likened to “servitudes”—non-possessory property interests that attach to land and impose their restrictions and obligations on generations of landowners (e.g. a covenant that prohibits a homeowner from painting her house pink).  Like the licenses that characterize the digital age, use restrictions imposed by servitudes bind remote purchasers with whom the beneficiaries of the restrictions may have no direct relationship.  They do not arise from any human communication, but instead “run with” the burdened assets and automatically bind current owners.

Although servitudes are a familiar feature of contemporary real property law, they have long encountered judicial skepticism that has generated a host of doctrinal complications.  This skepticism has been even more pronounced in the context of servitudes applied to items of tangible personal property (imagine a covenant prohibiting the owner of a horse from riding on a pink saddle). 

In the article I develop an account of the evolving jurisprudence of servitudes as applied to both land and personal property, identifying the sources of traditional servitude skepticism in order to evaluate contemporary licensing practices.  I apply the lessons I draw from the old servitudes to three paradigmatic “new servitudes”—the Microsoft Vista End-User License Agreement, the Free Software Foundation’s General Public License, and Creative Commons licenses.  (NB:  I am a board member and former staff member of Creative Commons.  But of course the paper expresses only my personal views.)

My examination of the tangible servitude jurisprudence finds that servitude skepticism arises from three types of concerns: (1) concerns related to notice and information costs; (2) concerns related to dead hand control and other aspects of what Julia Mahoney has usefully referred to as the problem of the future, and (3) concerns related to harmful externalities. 

Turning to the new servitudes, I find that some of the concerns that have animated skepticism about servitudes on land and personal property may in fact be more relevant to contemporary licensing practices than they are in the contexts in which they originally arose.  But the new servitudes differ from each other in respects that are critical to the applicability of the servitude analysis: each of the paradigmatic licenses that I examine exhibits a different mix of problematically servitude-like features. 

For example, the Microsoft EULA purports to impose restrictions beyond the background restrictions of copyright law (e.g. by limiting uses that courts have declared to be non-infringing “fair uses” under copyright law) on multiple generations of remote consumers by attaching itself ubiquitously to works of authorship for which there is limited effective competition.  These features raise nearly all of the concerns traditionally associated with servitudes—potentially confusing consumers for whom the restrictions are not initially salient (notice and information costs) and impinging on future creativity and innovation (the problem of the future) by imposing restrictions that threaten the positive externalities preserved by copyright’s public-regarding limitations (harmful externalities).

The GPL and Creative Commons licenses are fundamentally different in that they arguably impose conditions that are merely a subset of those restrictions already imposed by the background law of copyright—a characteristic that mitigates both notice and externality problems.  Indeed, by releasing creative works from unnecessary copyright restrictions, these licenses promise to generate positive spillovers from collaborative creativity.  But, ironically, this environment of collaborative creativity can result in complex webs of overlapping and potentially incompatible conditions, causing unanticipated future problems that may even transcend the problems that necessarily arise as a consequence of copyright law’s own running restrictions and long duration.  I have a few examples in mind that I am happy to describe once our discussion gets rolling. 

So with those basics on the table, let the mobblogging begin!

Next Week: Mobblog on "The New Servitudes"

On Monday, February 4, the Faculty Blog will host a mobblog (online workshop) centered around "The New Servitudes" by Molly Van Houweling of Boalt Hall. The paper attempts to apply ideas from the law of servitudes to modern licensing practices involving intangible products. Molly will begin with a post summarizing her paper, with participants then chiming in with their comments and questions. Confirmed participants include: Jim Gibson (Richmond), Sonia Katyal (Fordham), Henry Smith (Yale), and Stewart Sterk (Cardozo) in addition to Chicago's Richard Epstein, Lee Fennell, Randy Picker, and Shyam Balganesh. We hope you'll join us for what promises to be a fascinating discussion.

H2H: Jack Balkin's Last Word

(cross-posted at Balkinization)

The discussion between David Strauss and myself about whether originalism is a good idea for liberals has by now gone well beyond the narrow question of originalism itself and moved on to far deeper topics-- rhetoric, faith, and legitimacy. I cannot address all the valuable things David has to say in his last post, but here are a few comments on what I thought was most worthy of note.

I. "Rhetoric" versus "Legal Argument"

In his second post, David agrees with me that my text and principle approach "provides resources that [his] does not, particularly to groups that are trying to change the status quo," but that "the resources are rhetorical." He argues that conservative uses of common law and precedent and appeals to "deeply rooted practice[s]" are also powerful weapons for progressives, too.

Assuming that David is right about the latter claim, nothing in my text and principle approach rules out the use of appeals to precedent or tradition, or to consequences for that matter. It treats them as fully appropriate modalities of legal argument for fleshing out and implementing text and principle, and for making structural arguments where the Constitution is silent.

Indeed, I would argue that not only is common law development of doctrine by courts not hostile to originalism, it is implicit in the grant of judicial power in Article III (other aspects of common law adjudication are also assumed by the Seventh Amendment). Thus, there is a pretty good originalist argument for judicial construction of doctrine over time using common law methods.

Thus, under my approach, at least, one can have the best of both worlds, and employ all the resources of constitutional argument.

David's approach, by contrast, regards appeals to the constitutional text and its underlying principles as merely "rhetorical," rather than genuine legal arguments. As he says, when we make originalist arguments, "we should understand that we are making a rhetorical move. That's not disreputable; it's just not a legal argument."

I found this claim puzzling. What is the nature of David's distinction between "legal argument" on the one hand, and "rhetoric" on the the other? I had thought that all of the traditional modalities of legal argument were simultaneously rhetorical forms and valid modes of legal argument. David's claim seems to be that we should write off two of the modalities of legal argument-- arguments from text and appeals to principles derived from history-- as not really legal arguments at all. They are not real reasons; they are just window dressing. This makes arguments from precedent, tradition, and consequences (and possibly structure) the only appropriate forms of legal argument; the rest is mere rhetoric. You may notice the irony: in the past progressives have accused originalists of writing off (or downgrading) every form of argument except appeals to text, history and structure because they are "just politics" in disguise.

I am hoping that David does not really mean to say that arguments from text and principle aren't legal arguments. What is it that would make them mere "rhetoric?" Surely appeals to precedent and tradition are rhetorical: so too is the act of characterizing traditions and generalizing from and distinguishing cases. Perhaps it is David's concerns about the manipulability and uncertainty of choosing appropriate levels of generality. But all of David's previously expressed concerns about levels of generality also apply to arguments from precedent and tradition. Usually more than one principle can be derived from a case or a traditional practice, and you can read a case or a tradition either broadly or narrowly. Both principles and traditions, moreover, can be articulated at higher and lower levels of generality.

Nothing David has said really convinces me that appeals to text and principle aren't genuine legal arguments (like arguments from precedent and tradition) that offer reasons to interpret the Constitution one way rather than another. And if they are, why would one adopt an approach to constitutional interpretation that gives progressives fewer resources for legal argument rather than more?

In any case, I'm not sure that David is right that appeals to deeply rooted practices are the best way to promote a progressive constitutional agenda, especially in our current age. They might be a good way to defend against further changes by conservative social movements, but not necessarily to seek significant reform or to reject traditional practices that are unjust. My sense, rather, has been that appeals to deeply rooted practices and precedents tend to put liberals in a defensive posture. And my understanding of American history is that energized social movements for change usually call on the Constitution and its text and principles to reject what people previously thought were settled questions. They succeed by unsettling those questions and making them controversial. And when social movements succeed in changing people's minds about what the Constitution means, they tend to win in courts as well.

History seems to show that constitutional change often comes from social movements and political parties changing constitutional culture, which is then ratified by courts. If that is so, then arguments from text and principle are probably just as important as arguments from precedent in promoting either a progressive or a conservative agenda, and probably more so. Indeed, I don't think it is an accident that conservative social movements in the past generation have regularly made claims about the text of the Constitution and the purposes of the founders. Appeals to what I call text and principle have been a common feature in the arguments of most successful social movements in the United States.

Liberals have been fighting a defensive battle against insurgent social movements from the right for a generation. If I am correct about this, then for liberals to stop playing defense and push their claims forcefully once again, they cannot afford to throw away the modalities of text and principle.

II. Precedent and the Hardwired Constitution

One puzzle with David's common law approach, as I noted in my first post, is that it is hard to explain why the hardwired features of the Constitution are not subject to common law modification in the same way that judicial precedents are. If you look at David's account of why the text is biding in his Yale Law Journal article, you can see that he argues (1) that it provides a focal point for discussion, and (2) that it solves problems of coordination. As he explains, "it is more important that things be settled than that they be settled right. A legal provision can settle things, and sometimes the importance of settlement alone is enough to make the provision binding." However, "[t]he binding force of the provision rests on its functional ability to settle disputes, and not at all on whether the entity that enacted the provision is entitled to obedience or "fidelity.""

The puzzle for me is that the same things could be said of precedents. Precedents can also serve as focal points for discussion. (Think about McCulloch v. Maryland or Brown v. Board of Education.). Precedents also can settle contested questions, because, after all, they also have the force of law, and one might equally say that it is better that they be decided than that they be decided correctly. But if that is so, then why shouldn't our views about the bindingness of certain precedents be precisely the same as our views about the binding character of certain "hardwired" features of the Constitution, for example, that the President serves only four years per term, that there are only two houses of Congress, and so on?

That equivalence goes in two directions. If we decide that the length of the President's term must remain four years because the text says so, whether or not that is a good idea, why isn't the result in an old precedent like Plessy v. Ferguson equally binding on us, because it also settled an important question of law, whether or not the decision was a good one? Conversely, if we think that a long standing precedent like Plessy can be overturned or modified on a sufficient showing to a court that it has outlived its usefulness as a settled decision and now imposes very serious injustices, why couldn't we say the same thing of the limitation on Presidential terms, the Presidential veto, the malapportionment of the Senate, the Electoral College, and so on? If there is nothing special about the text that differentiates it from precedent, why couldn't courts change these rules through common law adjudication? Why couldn't the Supreme Court decide, for example, that a President can serve a longer term in office than four years because the country needs him?

The answer that an originalist like me would give is that the Constitutional text stands on a different footing than common law implementations of the text. Common law adjudication fills out contested issues surrounding text and principle, but it cannot contradict the text. But because David wants to blur that distinction, I'm not sure that the same answer is available to him.

III. Legitimacy and the Role of Commitment

David argues that a common law constitutional system poses no special problems of legitimacy, because the British system is also a common law system. I'm not sure this is a complete answer.

First, the British system has a different theory of sovereignty than the American model. The British system is based on Parliamentary sovereignty-- Parliament can change pretty much any aspect of the constitutional system at will, and Parliament can overturn decisions of courts. (The extent to which Parliamentary sovereignty has been affected by greater European integration and by the passage of the Human Rights Act I leave for others to determine). By contrast, the American system specifically broke from the English model. It was based on a system of popular sovereignty with delegated powers to the states and the Federal government, separation of powers, and demarcations of rights as limitations on government power. These delegations of power and restrictions on power were established by a written constitution, which was demarcated as supreme law. Therefore it seems to me that common law adjudication of constitutional doctrine has a different place in the American system than common law decisionmaking does in the British system. The American system seems to place the written constitution above ordinary common law and even above judicial constructions of the written constitution using common law methods. And the choice of a written constitution was designed to establish a particular model of popular sovereignty.

The theory of popular sovereignty creates a problem for David's common law model, under which the text is binding only because it serves as a focal point or solves coordination problems. What connects the work of courts to the theory of popular sovereignty in his model? The answer to that question is particularly important if common law courts can disregard or modify the commands of the constitutional text, even the hardwired features. David's is not the only theory that faces this question, but because he regards the text as on the same level as precedent, it arises even more urgently for him.

Finally, David argues that faith is not really necessary to constitutional legitimacy. He also denies that legitimacy requires that we see the Constitution as the project of many generations. (This is a puzzling claim to make for somebody who believes in a traditional constitution, but put that question aside for the moment). David reasons that even if judgments of legitimacy depend on what may happen in the future, making that judgment "is just a matter of collecting and weighing the evidence in order to figure out what the system is like today and will be in the future." That is to say, legitimacy does not depend on our attachment to the Constitution or our attitude toward it; it depends rather on a "cold-blooded" weighing of future probabilities.

I think this overlooks the attitudes of commitment and attachment that are necessary for a successful constitutional system to operate over time. For the Constitution to be legitimate, people cannot be in the position of perpetually calculating whether to defect from it or not. Rather, Constitutional government requires commitment to make the constitutional system work over time. That requires that you decide to put yourself on the side of the Constitution, wanting it to succeed, and feeling that you have a stake in it, rather than straddling the fence of obedience, so to speak. This is not a question of the presence or absence of rationality: Commitment or the lack thereof can create self-fulfilling prophecies of success or failure for constitutional systems.

David offers the example of a military coup where the junta promises that it will restore democracy soon and asks people to be patient. But in that case the Constitution has been suspended. The junta is not asking for faith in the constitution; it is asking for belief in their good faith as individuals. A far better example of the need for constitutional faith concerns whether people should have joined the Constitution in 1787 and worked for its success, when it was by no means certain that this new form of government would last. Benjamin Franklin famously said at the end of the Philadelphia Convention that he would support the new Constitution not because it was free of defects-- for in his view it was not-- but because he thought it was the best that could be done under the circumstances, and that he believed and hoped that it would succeed, and that it would eventually live up to its promises. Therefore he argued that it was worth working for its success. I call this attitude of attachment and commitment faith; it was crucial to the new Constitution's success, and is so even to this day. David may not like the use of the word "faith," perhaps because he is suspicious of religious language. But I see no reason to shy away from it as a way of understanding the sort of attitudes of commitment and attachment that are necessary to preserve the legitimacy of constitutional government in a system like ours.

H2H: Balkin's Response to David Strauss on Originalism -- First Round

(cross-posted at Balkinization)

Like David Strauss, I’m delighted to be able to participate in this exchange on liberalism and originalism.

In his first post, David makes two criticisms of originalism. The first is that it is empty; it has no real content. The second is that it is illiberal; it has bad content. (Note that these two claims are at least potentially in tension with each other.)

David’s first point is that originalism doesn’t determine hard cases by itself. We do not disagree. All the other modalities are usually necessary to resolve these cases. Much of constitutional law involves what I would call (following Keith Whittington) constitutional construction-- fleshing out the constitutional text and principles through implementing rules and institutions. That construction is not limited to courts; the political branches also implement constitutional values all the time. And their implementations have path dependent effects on how we construe the Constitution later on. There is not one single way the Constitution-in-practice had to end up. Much depends on what each succeeding generation does with what they are bequeathed, what institutions they create, what precedents they set, and so on.

I think David is running together the question of original meaning with the question of how best to implement the Constitution. At least, that is what I think is going on with his discussion of levels of generality. From the standpoint of the meaning of the equal protection clause, the proper level of generality is the level that we find in the text itself. The text says “equal protection of the laws.” That is the proper level of generality. David asks why the proper level of generality for understanding the clause is not “racial equality” or “equality with respect to common law rights.” The reason is that the text does not say “racial equality” or “equality with respect to common law rights.” There are important historical reasons why it did not do so.

David’s concern about the proper level of generality makes sense only if you think that the clause must somehow be confined or limited to original expected applications. Then you have to engage in debates about how to characterize those expectations, broadly or narrowly. But I reject that view. Asking about levels of generality also makes sense if you believe that determining the original meaning of the clause will do most of the work of implementing it. But I also reject that view.

David wonders why I say that the principles underlying the Fourteenth Amendment ban class and caste legislation. I say this because these are principles at the same level of generality as the text and the history strongly suggests that these were key principles underlying the Fourteenth Amendment. These ideas come straight out of Jacksonian and abolitionist ideology. When Senator Jacob Howard, representing the Committee of Fifteen, introduces the Fourteenth Amendment in the Senate in May of 1866 he explains each clause and its function. When he gets to the Equal Protection and Due Process Clauses he says that they “abolish[] all class legislation in the States and do[] away with the injustice of subjecting one caste of persons to a code not applicable to another.”

How would we implement the equal protection clause? The answer is pretty much what you would expect. We would come up with doctrines that tried to articulate when laws were likely to unfairly single out groups for special burdens or benefits, or produce or maintain unjust social subordination. Our modern scrutiny doctrines attempt to do this. (So did the old police power jurisprudence of the Lochner era.). These doctrines don’t always do it well, but that is their point.

David and I agree that the class and caste legislation principles, by themselves, do not determine the scope of their own extension. But, not to put too fine a point on it, neither does the language of the Equal Protection Clause itself. Everything he says about these two principles could also be said about the words “equal protection of the laws.” Rather, we have to implement text and principles in doctrines, constructions and institutions. We have to build methods of applying these principles in practice. To do this, we use the various modalities of legal argument that are familiar to lawyers. When lawyers develop doctrine, they try to develop conceptual tools to implement text and principle and concretize them. Think of doctrines as heuristics. Our contemporary scrutiny rules are implementations of the Fourteenth Amendment’s text and principles, which sometimes work fairly well to promote these principles and sometimes do not work very well at all.

When I say that laws banning abortion violate the class legislation and caste legislation principle, that means that I think this is the best implementation of those principles in today’s world. In my article Abortion and Original Meaning I give fairly elaborate reasons for why that is so. My arguments don’t always track existing doctrine, but in implementing these principles in concrete settings I am trying to do what existing doctrine does. The key point, once again, is that doctrine isn’t the same thing as constitutional meaning. It implements meaning.

Thus, David’s claim that originalism by itself doesn’t resolve lots of controversies is not really much of an objection from my perspective. In fact, if David and I both wrote accounts of various constitutional questions involving the Constitution’s vague and abstract clauses, we would probably rely on similar materials, and the results might often look quite similar. The major difference perhaps, is that I would start by talking a bit about text and underlying principles before launching into the other modalities and explaining how they helped implement text and principle.

In fact, I think that starting with text and principle can be very enlightening in some cases. One recent example is Romer v. Evans. If you start with our standard doctrinal models, you have to twist yourself in knots to explain the result in Romer. But if you start with the assumption that these doctrines are designed to implement the Fourteenth Amendment’s ban on class and caste legislation, Romer makes much more sense. In fact, I would argue that the law at issue in Romer is a paradigmatic example (in our day, at least) of what the Reconstruction Republicans were getting at when they spoke of class legislation. Romer is a case that shows the limits of our current models of doctrine. But doctrine isn’t written in the sky. It’s a heuristic device for implementing and applying constitutional values. Doctrinal structures that implement the Fourteenth Amendment have been different before, and they will be different again.

So far, it does not look as if David and I disagree very much. What is the nature of our disagreement, then? I think it is this: David sees text and principles as not particularly helpful or important to constitutional law; for him all of the work (or most, anyway) is done through the common law process of reasoned argument and doctrinal elaboration. By contrast, I regard the purpose of common law elaboration as in the service of fleshing out and implementing text and principles. Why does this difference matter? I can think of three reasons.

First, if doctrine and common law development serve text and principle, and not the other way around, then text and principle are always available to critique present-day doctrines-- and encrustations of doctrine-- that are no longer reasonable implementations of text and principle and no longer serve their purpose. This is especially important for social and political movements, which are the key drivers of changes in constitutional thought. These movements press for change by calling on the Constitution’s text and principles as they understand them. Sometimes their arguments are lawyerly and doctrinal, but often times they are not. Thus, the centrality of text and principle to constitutional interpretation is important if you think that much constitutional interpretation goes on (and even should go on) outside of the courts that create and depend on the elaboration of systems of doctrine. In the long run, changes in doctrine respond to changes in societal understandings of constitutional values-- and not the other way around.

Second, David thinks that talk of text and principles obscures constitutional argument. Quite the contrary, I think it immensely clarifies it. It shows that we are always doing is trying to implement and build on a framework, and it is a framework we are always permitted to return to, restore and redeem if we think that years of doctrinal glosses have become unwieldy, overly formalistic, or have otherwise taken us in the wrong direction. This presents a very different picture from a system of common law development that David favors, because sometimes it is difficult to explain why we can reject some precedents we don’t like (Bowers) if we want to insist that others have to stay in place (Roe).

This is a particular problem for liberals today: The present generation of living constitutionalists have found themselves on the defensive against conservative social movement energies. Like most social movements before them, these conservative mobilizations have called for a return to the Constitution’s text and to the principles of the founding generation, even if their notions of what that entails are disputable. Faced with incessant demands for constitutional revolution, living constitutionalists have become today’s “conservatives”: they have resorted to arguing for preserving the status quo, and for respecting older precedents created in politically more liberal times. But earlier social and political movements helped produce the doctrinal changes they now defend; those movements would not have succeeded if courts had applied the same concept of precedent that liberals insist on today. Arguments for respecting precedent make the most sense when they are directed at persons who do not share your constitutional views, but in that case they are a modus vivendi, not an independent criterion of constitutional fidelity. The best argument for decisions living constitutionalists admire is not that they are settled precedents; it is that they are faithful implementations of the Constitution's textual commitments and underlying principles.

Third, I think that David’s dismissal of original meaning originalism as mere rhetoric proves too much. I don’t think he can give a good account of why the "hardwired" rules in the Constitution are binding on us as law today. The common law method by itself can’t explain the binding nature of the "hardwired" features of the Constitution– the fact that there are two Houses of Congress, and so on. My version of originalism can: it argues that the Constitution contains rules, standards, and principles. All of them are binding on us in the present; however, the latter two, because they are standards and principles, require fleshing out and implementation through doctrine, institutions, and statecraft.

In fact, my sense is that most progressives already are original meaning originalists with respect to the "hardwired" Constitution. They don’t assume that doctrinal elaboration or common law reasoning can increase the number of houses of Congress or the length of the President’s term. For most progressives, living constitutionalism is primarily focused on the vague and abstract clauses of the Constitution, like those in the Bill of Rights and the Fourteenth Amendment; they want to make sure that these clauses are not limited to the original expected application. I think that my version of originalism– rather than a common law theory like David's– better explains why this attitude makes sense.

David’s second major concern is that originalism leads to inappropriate hero worship of founding generations, and it imposes the values of 18th and 19th century dead white men on the very diverse and vibrant society in which we live today. This is no small matter. The Constitution’s legitimacy comes from our present ability to see it as responsive to our needs for governance, for justice, and for the protection of our rights. Why should “people who do not feel any affinity to American traditions” have the constitutional values of an alien past forced on them? There is no special reason, David insists, why “we owe it to the earlier generations to maintain some kind of continuity with them.”

For my part, I don’t understand why these concerns don’t also arise in a common law system of constitutional development of the sort that David favors, one that also relies on reasoned elaboration of existing traditions and precedents. Won’t that sort of model be equally alienating to persons “who do not feel affinity to American traditions?” And don’t common law conceptions of reasoned development also involve continuity with previous generations? I think that David may be worried that originalism too readily binds us to the concrete expectations of people living in “the late-18th or mid-19th century”– like their views about “gender roles.” But certainly precedents of the past (and even the recent past) are also inflected with the attitudes and values of the past. Moreover, this is a strange criticism to level at my version of originalism, because, as David himself notes, I don’t believe that originalism commits us to the original expected application of the adopters.

There is a deeper question at stake, however, that I think David is aiming at. The question is whether legitimacy depends merely on our present day satisfaction with the Constitution or whether it requires that we imagine ourselves as having some continuity with earlier generations; whether legitimacy requires Americans to imagine ourselves as part of the American people, existing over time, and engaged in a common constitutional project. David’s remarks (at least here) suggest no. I say yes.

In my second article, Original Meaning and Constitutional Redemption, I argue that the legitimacy of the Constitution depends on the public’s attachment to it. Attachment is not the same thing as consent. You consent to something you have a choice in; you become attached to something that you live in and feel you are a part of. The Constitution is legitimate if people from their different perspectives can reasonably believe that, understood in its best light, it is worthy of their respect. (Frank Michelman has developed this point in his work). That means either that they either believe that it currently adequately protects their rights and respects their values or that they have faith that, in time, it will come to do so.

That means, among other things, that the legitimacy of the Constitution depends on our belief in its redeem-ability, and this requires us to imagine the Constitution as a project that extends over time, which in turn requires that we see it as a joint project of many generations, a project that extends backward into the past and forward into the future. In addition, I argue that for the Constitution to be legitimate, it must not only function as what I call “basic law” and “higher law,” it must also be “our law.” The Constitution works as “our law” when we view it as our achievement and the product of our collective efforts as a people, which simultaneously involves a collective identification with those who came before us and with those who will come after us. I don’t think this is the quite the same thing as hero worship. It does, however, require some degree of attachment to and pride in the constitutional project as a whole and a faith in its future redemption.

As you may have guessed by now, this is not your grandfather’s originalism. For one thing, it is not bound by original expected applications; for another, it carries with it a different conception of legitimacy. I’ll be interested in what David has to say about the question of legitimacy in his preferred method of constitutional interpretation in our next go-round.