Raven's Lost Libraries surveys the forces that have destroyed library collections across the ages.² From the sacking of ancient Assyrian libraries by Babylonian conquerors to the confiscation of Jewish libraries by the Nazis, Raven chronicles the destruction of collections by such perils as fires, wars, ideological intolerance, mismanagement, and simple abandonment. What unites these losses is their occurrence in environments that, to varying degrees, were characterized by a scarcity of information. In such environments, the barriers to information distribution and reproduction were high and, if bibliographic resources were not protected, access to the information they carried may be lost forever. Libraries emerged in part to prevent such losses and, for each of the catastrophes recounted in Lost Libraries, in innumerable instances libraries have preserved information that would have otherwise perished.

For most of library history, the primary means of preservation was the effective custodianship of scarce holdings. According to separate studies by Higginbotham and McDonald, today's more sophisticated and proactive preservation practices originate in the decades between the founding of the American Library Association (ALA) in 1876 and the outbreak of World War I in 1914.³ Both studies show that, during this span, libraries gained a better understanding of the vulnerability of their collections and developed a growing sense of a preservation imperative. As the twentieth century progressed, this imperative was boosted by a variety of factors—most importantly, the perceived “brittle books crisis” and the bibliographic destruction resulting from the 1966 flood in Florence, Italy. By the 1970s, preservation became the focus of an ALA unit, and larger libraries started creating preservation departments and participating in broad preservation initiatives.⁴

In subsequent years, the commitment to preservation has become ingrained as an essential responsibility of academic research libraries. Maxwell, for example, declares that preservation is the “sacred duty” of librarians and the reason why libraries exist.⁵ In more measured tones, this commitment to preservation also is reflected in a number of statements published by the ARL over the past few decades. To take a recent example, Hahn and Blixrud preface a 2009 ARL special report on preservation by asserting, “preservation is a core function of the research library and a key element of both the stewardship and access missions of research organizations.”⁶

What distinguishes recent preservation commitments from precursors is the nature of the resources to which libraries provide access. With the ongoing transition from print to electronic formats, preservation practices are becoming more complex. This is particularly true in the case of journals, which have been at the vanguard of the shift to electronic formats. When a library subscribes to a journal in electronic format, it acquires access but usually does not take possession of a copy that it holds locally. While the license agreement associated with the acquisition may contain perpetual access provisions, the means of carrying out these provisions are oftentimes insufficiently developed.

These challenges have sparked a flurry of activity to develop clear, easily adoptable measures that will allow libraries to retain their traditional commitment to preservation. A milestone in this effort was a 2005 meeting hosted by the Andrew W. Mellon Foundation at which library leaders discussed journal preservation. After the meeting, the congregation issued a statement that succinctly captures prevailing views about the severity and urgency of the issues at stake.⁷ This statement, which the ARL later endorsed, ranks journal preservation among the foremost challenges in academia and proceeds to describe the perils:

If a publisher fails to maintain its archive, goes out of business or, for other reasons, stops making available the journal on which scholarship in a particular field depends, no practical means are in place for libraries to exercise their permanent usage rights and the scholarly record represented by that journal would likely be lost. For electronic journals, the academy as of yet has no functional equivalent in long-term maintenance and control over the scholarly record that “owning a copy” provided for printed journals.⁸

To address the crisis, much attention has been devoted to developing solutions for the effective retention of perpetual access provisions. For example, a 2006 report issued by the Council on Library and Information Resources (CLIR) examines twelve emerging archiving initiatives.⁹ Two of these initiatives, LOCKSS (Lots of Copies Keep Stuff Safe) and Portico, have been particularly successful in North America and have achieved participation from a broad spectrum of libraries and publishers. However, even with such initiatives in place, libraries’ efforts to reclaim their mantle as stewards of their journal collections are far from complete, a fact reflected in Atkinson's 2005 statement that the inability to acquire journal access that includes ownership represents “the greatest single failure of research libraries in the past decade.”¹⁰

The transition from print to electronic formats is only one manifestation of broader changes in how information is created, distributed, located, and accessed. Conceptualizing these changes, Darnton writes that they mark civilization's entry into a new stage of information technology.¹¹ In Darnton's framework, the world was previously transformed by the development of written communication around 4000 BCE, the codex around 300 CE, and printing with changeable type in the 1450s. The latest stage in this progression, Darnton explains, commenced with the emergence of electronic communication networks in the 1970s and the World Wide Web in the 1990s. Among this last stage's greatest impacts has been a transformation from past ages of information scarcity to an age in which the rapid proliferation of information can “seem both unstoppable and incomprehensible.”¹² Of course, this superabundance of information resides not in the physical world of paper, bound volumes, and stacks, but rather in the virtual world of large-scale digitization, data sets, and cloud computing. What is emerging is a landscape of what Morville has termed “ambient findability,” a place from which a person “can find anyone or anything from anywhere at anytime.”¹³

Journals are already part of this ambience. Indeed, researchers have come to expect expansive journal access from the time and place of their choosing. While much of this access is due to subscriptions to “big deal” publisher packages, the ambient findability of journals is also the result of vendors making this content available through pay-per-view transactions and the rental service DeepDyve (www.deepdyve.com). The availability of content on these bases is rooted in the “Long Tail,” a marketing concept popularized by Anderson.¹⁴ The Long Tail takes shape in an environment where reductions in providers’ inventory and distribution costs combine with technological advancements to empower large and diverse demographics of consumers to easily find and purchase a wide array of niche products. The Long Tail's inclusion of journals and the momentum toward its further expansion into the depths of scholarly literature suggests that, in the future, market forces are likely to make journal content accessible—either directly to researchers or to libraries acting on their behalf—at an affordable cost. This scenario has the potential to decrease the benefits that future patrons will enjoy because of libraries’ determination to secure perpetual access for newly acquired journals. Traditional commitments to preserve are rooted in an environment of information scarcity in which technological, geographical, and financial barriers often rendered bibliographic resources as irreplaceable. The emerging environment replaces absolute, physical scarcity with what Hazen terms “the artificial scarcity created through restrictive manipulations of the digital marketplace.”¹⁵ If trends progress on their current trajectory, this artificial scarcity, a key basis for Long Tail economics, seems likely to supply a profusion of information that can be easily distributed and affordably acquired.

The full impact of the emerging information age remains unclear, and, as Plutchak comments, the implications on scholarly communication will not be understood for decades.¹⁶ One result of these uncertainties is that today's justifications for an abiding commitment to perpetual journal access are becoming more speculative. In the print-centered era of information technology, threats to scarce bibliographic resources provided a concrete imperative for libraries to preserve their journals. In contrast, grounds for securing perpetual access usually rely more on past precedents, a mistrust of commercial interests, anxieties about the future of the information marketplace, and the sense that the tremendous cost of acquiring journals is itself an imperative for making them a permanent part of the collection.

Although the speculative nature of these justifications has not been used as grounds for abdicating preservation commitments, they do offer compelling premises in support of arguments that perpetual access can be sought with less stringency. Previous arguments involving the tensions between the access and ownership of e-resources have generally been dispelled by repeating a variation on the truism that it is not a question of either/or but both/and.¹⁷ Today, this method of resolution is becoming less tenable. Indeed, funding scarcities and fiercely competing priorities confront libraries with painful either/or decisions. For example, the ability to meet patrons’ expectations for seamless access is more difficult because of reduced or stagnant budgets and inflating subscription costs. Opportunities exist for libraries to decrease journal expenditures without an immediate decrease in access, but many of the most effective solutions involve the loss of perpetual access, including

• converting subscribed access to access through pay-per-view transactions;
• cancelling subscriptions to journals with current issues available through full-text aggregators; and
• discontinuing membership in archiving initiatives such as the LOCKSS Alliance and Portico.

Administrators can resist such measures on the grounds that their lack of perpetual access provisions are at odds with a research library's preservation mission, but the cost of fidelity to this mission may be the inability to acquire certain resources and services to meet patrons’ current needs.

Accompanying the scarcity of funds to acquire journals is a scarcity of personnel to manage them. As Stachokas argues, the few personnel that most libraries assign with e-resource management responsibilities are dwarfed by the libraries’ large and growing e-resource collections.¹⁸ In support of this claim, the author cites Barnes, who conducted research showing that while libraries typically spend about 60 percent of material expenditures on e-resources, only 25 percent of staff have responsibilities for managing these resources.¹⁹ This finding brings to light a dilemma when considered in the context of Stemper and Barribeau's analysis of e-resource licenses, which shows that most publishers are willing to grant perpetual access during license negotiations.²⁰ Although the authors consider these findings heartening, they become considerably less so given the lack of personnel devoted to e-resource management and the significant expertise and time required to carry out effective license negotiations. Administrators can remedy the situation through the allocation of additional personnel to the responsibilities of e-resource management, but such allocations come at the cost of not meeting competing priorities.

The few recent studies of perpetual access provisions suggest that the conflict between libraries’ funding and personnel scarcities and the traditional commitment to long-term access has resulted in a disconnect between libraries’ stated values and what they accept. From November 2006 through January 2007, Rogers surveyed New Zealand university and polytechnic libraries regarding the portion of the libraries’ journals with archival rights or perpetual access provisions.²¹ He found that, despite the libraries’ stewardship commitments, only 30 percent of their licenses explicitly ensured long-term access. Stemper and Barribeau reported similar findings in 2006.²² Their literature review and informal survey suggests that more than 80 percent of research libraries in North America will accept an agreement even if it lacks perpetual access provisions.

At the heart of these contradictions are diverging visions of the role of research libraries in the twenty-first century. In one vision, there is an expanded focus on patrons’ immediate needs and a moderation of stringent perpetual access commitments. Hazen expresses this outlook when advocating as a guiding principle for the future that libraries “frame their information goals in terms of providing access to content that they do not possess, as well as on-site holdings.”²³ Teper charts a different course.²⁴ Discussing the importance of maintaining preservation commitments despite funding scarcities, he argues that libraries must distinguish themselves from commercial interests by “recasting themselves as memory institutions in an environment increasingly concerned with on-the-fly access.”²⁵ While time will reveal the outcomes of the tensions between long-term preservation commitments and immediate priorities, important insights about the nature of these outcomes are available through a clearer understanding of how libraries currently describe their commitments to perpetual access. The results of this paper's survey help provide such insights.

The 136 survey invitations resulted in 48 respondents completing the survey, a response rate of 35 percent. Included within the respondents is the researcher, who completed the survey on behalf of his institution, East Carolina University. An analysis of the replies to survey question 1 shows that 37 of the respondents work at ARL libraries, 10 work at ASERL-only libraries, and 1 works at a GWLA-only library.

Replies to question 2 indicate that 25 respondents work at libraries with a full-time enrollment (FTE) more than 25,000. Of the other respondents, 14 work at libraries serving FTEs ranging from 15,000 to 25,000, and the remaining 9 work at libraries serving less than 15,000 FTE (5 in libraries serving less than 10,000 FTE and 4 in libraries serving between 10,000 and 15,000). As a point of comparison between the respondents’ libraries and research libraries as a whole, 2007–8 statistics show an average FTE of approximately 24,000 among ARL libraries and 20,000 among ASERL libraries.²⁶

Question 3 assessed the funding that the respondents’ libraries are providing in the current fiscal cycle for acquiring journal content (including individual subscriptions in all formats, publisher packages, full-text aggregators, and pay-per-view transactions, but excluding one-time backfile purchases and associated fees). Respondents’ replies to the question indicate that, while there are some outliers, most work at libraries with funding between $3 million and $6 million (see figure 1). In comparison, 2007–8 statistics show that ARL libraries spent an average of about $7.5 million for current serials, and ASERL libraries spent an average of about $6 million for these resources.²⁷

Replies to question 4 show that, in comparison with the previous fiscal cycle, the largest number of libraries (21) have experienced no change in their funds for journal acquisitions (see figure 2). Another 18 of the libraries experienced decreases in funding, and the remaining 9 libraries experienced increases in funding.

Questions 5 and 6 examined the importance that the respondents’ libraries attribute to perpetual access. In response to question 5, which asked how important the respondents’ libraries currently consider perpetual access for journal acquisitions (excluding full-text aggregator subscriptions, nonsubscribed journals that become available through publisher packages, and pay-per-view transactions), most respondents (31) indicated that perpetual access is essential or very important, some (12) ranked perpetual access as somewhat important, and a small number (4) ranked perpetual access as having little or no importance (see figure 3).

The results show a small correlation between the funding and FTE of the respondents’ libraries and the importance that these libraries attribute to perpetual access. Of the 31 respondents ranking perpetual access as either essential or very important, 58 percent (18) of their libraries have an FTE of more than 25,000 versus 52 percent overall, while just 10 percent (3) have an FTE of less than 15,000 versus 19 percent overall. Likewise, 32 percent (10 of 31) of the respondents ranking perpetual access as either essential or very important (versus 25 percent overall) are spending more than $6 million for journal acquisitions in the current fiscal cycle.

The results show no clear correlation between changes in journal expenditures of the respondents’ libraries and the libraries’ differing commitments to perpetual access. Of the 31 respondents who ranked perpetual access as either essential or very important, 45 percent (14) have flat expenditures (versus 43 percent overall) while 39 percent (12) have decreasing expenditures (versus 37.5 percent overall), and 16 percent (5) have increasing budgets (versus 19 percent overall). Of the 12 respondents who ranked perpetual access as somewhat important, 50 percent (6) have flat expenditures (versus 43 percent overall), 25 percent (3) have decreasing expenditures (versus 37.5 percent overall), and another 25 percent (3) having increasing budgets (versus 19 percent overall). Finally, among the 4 respondents ranking perpetual access as either not very important or unimportant, 2 have flat budgets, 1 has an increasing budget, and 1 has a decreasing budget.

In reply to question 6, which asked whether the commitment to perpetual access is increasing, decreasing, or remaining the same, 24 indicated that this commitment is unwavering, 23 indicated that the commitment is increasing, and one indicated that the commitment is decreasing. A comparison between all responses to question 5 and the 23 responses to question 6 from respondents indicating that perpetual access is an increasing priority shows that the differences in importance that these 23 libraries currently ascribe to perpetual access generally mirror the rankings across all responses. A notable discrepancy is the respondents in libraries that already consider perpetual access to be essential. Whereas just 1 of these 23 libraries already rank perpetual access as essential, 12.5 percent (6) of all respondents ranked perpetual access in this way. Inverting this disproportion are the 24 respondents who indicated that, at their libraries, perpetual access remains at the same level of priority; among these libraries, 22 percent (5) ranked perpetual access as essential. However, also of note regarding these 24 respondents at libraries where perpetual access remains at the same level of priority is that 3 of these ranked perpetual access to be either not very important or unimportant.

Question 7 assessed what provisions the respondents’ libraries deem acceptable for securing perpetual access. To do so, it listed five perpetual access provisions that vendors often offer and asked respondents to indicate which of these provisions are acceptable at their libraries (see figure 4). The provision that was most often acceptable (selected by 42 respondents) was ongoing access via a web platform free of charge. The acceptability of this provision is followed closely by participation in an archiving initiative such as LOCKSS or Portico (41) and ongoing access via a web platform subject to a reasonable annual access fee (37). The results show a significant decrease in the acceptability of the other two options listed on the survey. Only 17 respondents affirmed the acceptability of ongoing access via the transmission of files that must be locally hosted, and 15 affirmed the acceptability of ownership of the content in print or microform.

Questions 8 and 9 assessed current and anticipated impacts of budgetary trends on commitments to perpetual access. Question 8 asked if budgetary factors have forced their libraries to undertake any of a list of five actions that often weaken the ability to retain perpetual access:

• Discontinuing participation in one or more archiving initiative (e.g., LOCKSS and Portico)
• Canceling subscribed journal access because of aggregator access to the cancelled content
• Converting subscribed journal access to access through pay-per-view transactions
• Converting print subscriptions to online subscriptions lacking perpetual access provisions
• Downgrading subscribed journal access to subscription levels with decreased perpetual access provisions

Most respondents indicated that their libraries have taken one or more of these measures (see figure 5). The most widely taken measure (in 25 libraries) was journal cancellations because of access through a full-text aggregator. Predictably, many of these libraries have experienced reductions in their journal budgets. Indeed, 52 percent (13) have decreased budgets from the previous fiscal cycle compared to 37.5 overall. The second most frequently taken measure (in 19 libraries) was the conversion of print subscriptions to online subscriptions lacking perpetual access provisions. Here, however, no significant correlation exists between these libraries and budget reductions. Far fewer of the respondents’ libraries have taken any of the three other measures listed. Indeed, only 5 work at libraries that have been forced to downgrade subscribed access to levels with decreased perpetual access provisions, 4 work at libraries that have been forced to convert subscribed access to access through pay-per-view transactions, and 3 work at libraries that have been forced to discontinue participation in one or more journal archiving initiatives.

Correlations are evident between the numbers of measures taken because of budgetary factors and the budgets and commitments to perpetual access among the respondents’ libraries. Of the 15 libraries where at least two measures have been instituted, 47 percent (7) have experienced reductions in their journal budgets from the previous fiscal cycle versus 37.5 percent (18) overall. Likewise, 67 percent (10) valued perpetual access as being either somewhat important or less, versus 33 percent (16) overall. Among the 33 libraries that took any one of these measures, a smaller but still notable correlation with their commitments to perpetual access is present. Forty-two percent (14) valued perpetual access as being either somewhat important or less, versus 33 percent (16) overall. Fourteen respondents indicated that their libraries have not needed to take any of the measures. As would be expected, the budgets and commitments to perpetual access among these libraries invert the correlations observed among the libraries that had to take measures resulting in the loss of perpetual access. Ninety-three percent (13) of these libraries value perpetual access as being very important or essential, and only 21 percent (3) have experienced decreases in journal budgets from the previous fiscal cycle.

Question 9 asked if respondents think that, in the next few years, budgetary factors are likely to force their libraries to take any of the measures listed in question 8. Most respondents expected that their libraries will be forced to take one or more of these measures (see figure 6). The most widely anticipated measure is the cancellation of subscriptions because of access through a full-text aggregator. Twenty-five respondents expected that their libraries will take this measure, the same number that indicated in question 8 that they have already taken the measure.

The second most anticipated measure is converting subscribed access to access through pay-per-view transactions, which 18 of the respondents expected. This number constitutes a drastic increase (450 percent) from the 4 libraries that have already converted subscriptions to pay-per-view transactions because of budget cuts. What is most notable about the libraries of these 18 respondents is that they generally placed a high value on perpetual access. Seventy-eight percent (14) of these libraries considered perpetual access either very important or essential (versus 65 percent overall) and 56 percent (10) indicated that the value being placed on perpetual access is increasing (versus 48 percent overall).

Third on the list of measures that the respondents predicted their libraries will need to take because of budgetary factors is converting print subscriptions to online-only subscriptions that lack perpetual access provisions. Sixteen respondents predicted this measure, a slight decrease from the 19 that have already undertaken it. In comparison to the respondents that predicted pay-per-view to be likely in their libraries’ futures, the respondents that predicted their libraries will convert to online-only subscriptions without perpetual access provisions generally ranked perpetual access as a lower priority. Fifty-six percent (9) of these libraries valued perpetual access as being only somewhat important or not very important, versus 33 percent (16) overall.

Fourteen respondents predicted that budgetary factors will force their libraries to downgrade subscribed journal access to subscription levels with decreased perpetual access provisions. As with respondents predicting their libraries’ uptake of pay-per-view, the 14 respondents predicting downgraded subscribed journal access is far more than the 5 who reported this measure has already been taken. Finally, only two respondents predicted their libraries’ discontinuation of membership in an archiving initiative. These 2 are also distinguished by the fact that they are among the few in the survey who had responded that their libraries consider perpetual access to be either of little or no importance.

Nine respondents did not believe that their libraries will need to take any of the measures. Seven of these 9 respondents also had been among the 14 respondents that answered the previous question by indicating that their libraries have not needed to take any of the measures. As was the case with those 14 respondents, the 9 currently in question have disproportionately healthy journal acquisition budgets. Whereas the percent of respondent libraries with budget cuts was 37.5, none of these 9 respondents work at libraries that have experienced budget cuts.

Question 10 asked what impact any loss of journal content because of lack of provisions for perpetual access has had on their libraries’ abilities to support the research, teaching, and learning of their user communities. No respondents replied that the impact on access has been major, but 25 percent indicated that they had experienced minor losses in access. Seventeen respondents indicated no such losses, and 6 indicated that they are not sure if losses have occurred. The factor that differentiates the respondents indicating impacts—albeit minor ones—because of a lack of perpetual access provisions is the value they ascribe to perpetual access. Compared with the 65 percent (31) of respondents indicating that these provisions are either very important or essential, 52 percent (13) of those respondents with minor losses in access consider these provisions to be either very important or essential.