Trevor Deck, Jan Guise, James Neufeldt, James Mason, and Rebecca Shaw, University of Toronto Music

Library

1. Introduction

Hello everyone. My name is Jan Guise, and I’m the Director of the Music Library at the University of Toronto. I have the great privilege of welcoming you here to this virtual space to have a series of discussions about music score collection assessment.

In this pre-recorded session, each of our team members will describe a portion of the collection assessment project we’re working on. I’ll start us off with a little background on the Music Library and why we are tackling music score collection assessment.

If you’ve had a chance to view our virtual tour on the Summit website, you’ll know The University of Toronto Music Library is the largest academic music collection in Canada, with over 300,000 books, music scores, and periodicals, and over 200,000 sound recordings (about half of which are CDs). The majority of the collections are housed on-site in the Music Library, with about one-third located in U of T’s off-site storage facility, UTL @ Downsview. Downsview, as we call it, is similar in scope and construction to those storage facilities at Harvard and Yale.

There are four reasons we are undertaking a collection assessment. The first is Space Planning:

The current music library space opened in 1992, and thanks to a healthy acquisitions budget, we are now exceeding our capacity. We are in the midst of a space audit alongside the Faculty of Music, with the potential for a major renovation on the horizon. We need to figure out: What does the music library of the future look like? What do users expect to find on-site, what do they expect to see when they walk into our space? What are they willing to wait for, since part of the collection will be at Downsview?

We also want to Clarify Policy:

Currently about a third of our total collection is already housed at Downsview. We know that items at Downsview should be low-use and need to be able to circulate. Not only will they circulate to U of T users, but to all five “Keep @ Downsview” partner institutions in the region. We want to devise a clear set of criteria for what stays on-site, and what goes to Downsview, and then communicate these criteria clearly to our users and our staff.

We need to Address Inequities:

We are living with a legacy decision that sees all new music score acquisitions sent straight to

Downsview. This decision was made to overcome a backlog in the binding and preservation department: new scores get catalogued and sent straight to Downsview prior to being bound. Then, once a patron requests the item, it circulates unbound to that person, then gets routed to the bindery and transferred to the Music Library. The unfortunate consequence of this decision is that there are very few music scores on-site with publication dates after 2000. Naturally this causes confusion among browsing patrons who want to play contemporary music.

Finally, we seek to Enable Discovery:

We know that crowded stacks inhibit discovery. Our stacks are so narrow they are not compliant with modern accessibility standards. The narrow stacks are so dark it’s impossible to read call numbers in many areas. As our faculty makes strides toward building diversity, equity, and inclusion into the curricula, users are seeking more music written by composers who are Black, Indigenous, or People of Colour. In the library we are working to identify gaps in our collections, but also working to illuminate the diversity that already exists therein. We don’t want to inadvertently send works by marginalized composers to Downsview simply because they’re “low use.”

Now I’ll pass the mic to Trevor to talk about our literature review.

2.1 Literature Review: Overview

A literature review on the subject of collections assessment for special collections presented a number of considerations and strategies. It also confirmed our suspicion that, while there is a fair amount of literature discussing collections assessment in the context of special collections, there is very little that deals with music scores, specifically.

We found several common themes in regard to motivations for collections assessment projects in the context of special collections. One common motivation was financial challenges, particularly around the economic downturn of 2008. A 2011 article by Colorado State University’s Michelle Wilde and Allison Level notes that during the financial crash of 2008, “as universities contemplated how they were going to make up for the losses in their endowments and state funding, libraries became an obvious target for budget cuts” (Wilde, 2011). Despite the financial challenges that resulted from the global financial crisis, special collections in research libraries were experiencing exponential growth. An OCLC Research Survey of Special Collections and Archives from 2010 noted that despite the fact that 75% of respondents had indicated their general library budgets had been reduced in recent years, special collections were nonetheless growing rapidly. The study found that the “mean number of both printed volumes and archival and manuscript collections [had] risen by an average of 50% in the aggregate since 1998” (Dooley, 2011). Moreover, audiovisual materials had increased at a rate of 240%-300%. It was, therefore, not a surprise to discover that two-thirds of libraries had reported having some special collections materials in remote storage facilities (Dooley, 2011). This ongoing increase in collection size has spurred many institutions to explore collaborative collections through the use of off-site storage facilities, as evidenced in the 2016 case study of the University of Texas, which saw University of Texas Library reduce their print book collection from approximately 132,000 titles down to 66,000 over the course of a year (Acadia, 2016). This was accomplished through a major deselection process in conjunction with the development of a ‘joint library facility’ that utilized a ‘resource in common’ model, which mandated that “only material not currently in JLF’s [Joint Library Facility’s] collection is accepted” (Acadia, 2016). A single copy of a book accepted by joint library facility then became collectively owned by the University of Texas and its partnering institutions.

In addition to considerations around financial constraints and collection size, competing demand for space was a common theme regarding motivations for collections assessment projects. For instance, a 2020 paper from library researchers at the Texas Tech University Health Sciences Center details how collections that have traditionally intentionally avoided any significant weeding are now experiencing a need to reduce collection size, as print usage continues to decline and demand for student space continues to grow (Bowers et al., 2020). Similarly, a 2019 article in The Serials Librarian by library researchers at Villanova University, documents a mass weeding project that was precipitated by a campus-wide report, which articulated a need for more public study space for students on campus, as well as a planned building renovation that would affect amount of shelf space available in the library (Burke and Kilb, 2019). Likewise, a paper from Collection Management in 2019 explores a mass deselection project library staff at Pollak Library at California State University took on. This project was in response to a ‘Library of the Future Taskforce’, whose final report mandated CSU ought to prioritize ‘student study space’ and "provide more space for collaborative projects and spaces for student-faculty interaction’. Library staff resorted to a novel approach in weeding, developing their own software inhouse, which would allow “faculty to provide input on a title-by-title basis through a webform”, which eventually resulted in over 95,000 unused or underused titles being removed from the shelves, freeing up over 7900 linear feet of shelving (DeMars et al., 2019).

These case studies mentioned contain several common themes, especially when it comes to discussion around the changes in publishing and collection models. Libraries are continuing to see a shift towards electronic publication, especially in relation to journals, and to a growing extent, in monographs as well. Burke and Kilb’s case study of Villanova University, for example, notes that its collection building trends had evolved over the past decade to reflect an intentional transition from print to electronic materials. Of the 600,000 records the library sent to OCLC for analysis through Greenglass (OCLC’s web based collection analytics tool), about one third of the records were for electronic content, which allowed Villanova to analyze overlap between their perpetual electronic holdings and their print collection. While this data proved essential in aiding in decisions around deselection of print materials, the authors note that subject librarians did encounter some faculty who were committed to retaining content in print, even if the library had perpetual online access (Burke and Kilb, 2019). While considerations around maintaining print collections tend to focus especially on monographs, challenges in format preference are increasingly relevant when it comes to music scores, as the landscape of music publishing is slowly but surely catching up to other areas. We are seeing more opportunities to acquire digital scores through new products link the digital score platform, nKoda, however print scores are likely to remain the preferred and predominant format for musicians in the near term, at least.

OCLC’s Greenglass was just one of several assessment tools mentioned in literature we reviewed. Others included custom-designed software (as noted in the California State University case study), as well custom-made surveys. Despite a number of options in the market for collection library decision support systems (or DSSs), Paynter (2009, 33) notes that “currently no single commercial DSS provides all the functionality to analyze collections in ways commonly used or desired by collections management librarians (i.e. comparatively, by usage, by package, or by resource sharing).” Wilde and Level point out a common refrain amongst literature on collection assessment: “librarians must work to provide a balanced approach to assessment”, one that utilizes the tools available, while avoiding overly complicated reporters that go beyond the level of usefulness (Schwartz, 2007). The need for a multifaceted approach is indeed essential to the assessment of music score collections, as traditional assessment criteria and methodologies, such as analysis of circulation stats, often bear little relevance to decisions around what materials are most important to include in our Music Library stacks on site, as opposed to being shipped to off-site storage.

2.2 Literature Review: Methodologies for Assessment

In the literature there was a fair bit of discussion around various tools and methods used to conduct assessments. Much of what we read showed a desire to broaden the scope of assessment from simpler, more limited point of view, methods, like circulation statistics and item condition, to something more holistic, something considering multiple metric points.

This resonated with us as we have a very expansive and deep collection intended to support the research needs of our Faculty. Clearly condition is an important factor, as well as circulation, but the novel items are what, in some ways, distinguish our library. Examples include scores published from the United Soviet Socialist Republic, Renaissance lute tablature, rare books and scores, and of course, our archival material. We want to, in a way, highlight these items, to make sure they are discovered by those whose research can be enhanced by them.

Though circulation statistics and evaluating condition remain important tools, the discussion around assessment highlights a shift towards a more qualitative approach.

Some authors make the differentiation between analysis, assessment, and evaluation fairly clear. With collection analysis we see a focus on quantitative metrics. Things like:

circulation rates,
the number of titles in a particular subject era, or  financial support of a particular subject area.

With collection assessment focused on more qualitative tools, attempting to measure things like

how well a library’s collection supports a particular subject, or
how well the collection compares to that of another library, or
if the collection even represents what the users’ needs and wants are.

When looking at more holistic tools, evaluation takes both qualitative and quantitative measurements into consideration. Multiple data points which can be compared and contrasted with other collections, or other components of the various collections being assessed takes on an important role.

From our Lit. Review, some of the methods for assessment, and tools used, included:

Subject analysis: This came up in diverse ways. Often, the ideas here, involved dividing the collection up by subject to determine the collections totals in these various subject areas. Then, this data would be compared with other data points, such as circulation, which then could be analyzed at a subject level.

We saw some comparing this data with holdings at peer institutions—some even used inter-library loan statistics; others considered cataloguing data from the item metadata. Other suggestions included comparisons with faculty programs to determine which are supported by specific subject areas, and how well they are supported comparatively with other subject areas.

We saw collection mapping and the conspectus method, which similarly use subject areas to analyze data on larger criteria than individual items. With to conspectus method, for example, we see “scores” attached to areas based on subjective criteria to allow for strengths of collection areas to be determined.

Consulting the user base to understand needs and desires was also a common tool discussed. Surveys, focus groups, observation techniques were all used.

Citation analysis was used to measure outcomes as well as activity. This was one tool that came up a few times. With this type of analysis, for example, you saw sample citation data from subject specific journals compared to holdings.

Systematic and programmatic analysis of data was discussed as well. There was definitely interest in finding ways to automate analysis based on data. There was a desire to find an efficient way to assess collections, especially in regard to budget cuts, limited resources, and short time frames. Though, this idea seems to be most fully developed regarding e-resources and working with things like vendor data. We did see some ideas involving, for example, programmatic analysis of circulation data compared to subject areas.

It’s easy to imagine the relevance of these concepts and tools for analyzing a score collection. However, the literature was fairly sparse in that regard. Books, journals, and electronic resources were by far the main focus of much of what we read in our Literature Review. So now, I think our next task would be to see if we could come up with ways of applying what we learned from this Literature Review to our score collection.

3. Data collection

So, reflecting on what we’d learnt from reading through the literature, we aimed to find ways to apply ideas to our score collection. Looking at our catalogue records is where we decided to start. This was data that was rich in detail, it was on hand, and was ready to be worked with. We saw this as a way to move forward with a data driven method allowing us to explore multiple data points.To start, the goal was to gather the information and organize it in a way that allowed us to ask questions and draw conclusions about our collection.

We extracted all the MARC data from our score collection as a large xml file. From there, we used a Python script to divide that xml file into folders and subfolders based on subject. The way we chose the subject was by consulting the Library of Congress classification system.

We decided to use the LC classification system as it was developed, and is maintained, by subject experts. Also, the classification system is what we use in our library. So, it seemed like a reasonable and useful authority to derive “subjects” from. We used the M class schedule for music to isolate 18 divisions, which each in turn were further subdivided.

In each of the folders the Python script created an “overview” document, charts, and spreadsheets with data represented in different ways. Our hope is that this will allow us to evaluate aspects of the collection against the whole collection, against particular areas of the collection and against external sources. This gives us a way of analyzing our holdings with multiple data points and helps us create a sustainable future vision for what our library should look like.

Our 18 large categories included broad areas like Solo Voice, Duets, Orchestral Music and Sacred Vocal Music. The large categories were also subdivided into more finite categories, which, again, were suggested by the LC class numbers.

Like the larger categories, these smaller subject areas also tended to form around instrumentation. We see music for Organ, Keyboard, Strings, Winds, Plucked strings, and Other (which would include percussion). In the chamber music sections this pattern was augmented by the addition of other instrumental elements. So, Organ became Organ and string(s); or keyboard was added to making Keyboard with 1 string and 1 wind. This pattern followed, with variation suggested by consulting the LC Class schedule through orchestral music and large ensemble music.

So, if we look at duets, for example, we see that it was further subdivided into 13 subject areas. These categories demonstrate that pattern well. We see things like Duets for Piano, which would include things like 2 pianos, piano 4 hands, piano 3 hands, etcetera. Or, we see Duets for Two stringed instruments. We see Duets for Strings and winds; we see Duets for Strings and keyboard and plucked strings; various combinations of these instrumentation patterns. And, again, this pattern continued onto things like Trios, Quartets, Septets, Quintets, etcetera.

So when we get to orchestral music, the pattern was altered a bit, however, the instrumentation pattern suggested by the LC Class system was followed.

We saw the biggest departure from instrumentation when dividing up vocal music. Here we saw subjects emerging around form: Operas, for example, or Solo cantatas, and we also saw divisions emerging for geographic area: Folk and ethnic music being the chief example here.

Breaking the collection down into these “subjects” (or perhaps genres would be a more appropriate term) allowed us to compare the parts to the whole, parts to other parts, as well as external elements (which will hopefully help us to understand issues of diversity, for example).

The data gathered at the collection level gave us a snapshot of our whole collection of scores. We see the breakdown of what we have in various subject areas, and their locations. So, for example, for the whole collection, we have 163,361 scores in the music library, and 71,999 at our Downsview facility.

We can also see the distribution of our collection at a subject level. Here is an example of 10 areas, with solo/ chamber instrumental subjects combined into one category.

Here you will notice that Solo and chamber music takes up a substantial part of our score collection. If we take that section and look at it by itself, we can see how it breaks down.

Here we see Solo music is the largest part of that section, dipping down in numbers as the ensemble size increases.

Furthermore, we can look at our subject areas to see how they compare to the total. Here is an example using Duets. This allows us to contrast that with similar charts for the whole collection, where we see the numbers are fairly similar, but there are subtle variations.

We can also look at charts that are created for the subdivision level. So, here, for example, we can add in piano and wind instruments, where we start to see a noticeable variation or differentiations from the norm established by the total collection. For piano and wind, there are more scores at the Music Library than there was for the collection as a whole.

So, the overview document was one of the other documents created by the Python script and put into the folders. This tells a bit more information about the category, or the subject area. It gives us information about item counts, item makeup, and circulation statistics. It organizes these as bullet points at the start of the document. So, we see, for example looking at Piano and winds that there are:

Item counts:

Items at Music: 7442

Items at Downsview: 2319 Total items: 9761 Score make up:

Average publication date: 1989

Number including facsimiles: 26

Number of miniature scores: 1

Number of vocal scores: 0 Number of items with parts: 8627 Circulation statistics:

Circulations at Music: 9140

Circulations at Downsview 35

Total circulations: 9175

Other charts that are created by the Python script allow us to look at data in other ways. So, for example, we have charts created around publication dates. Here we see a chart for publication dates for all scores and we notice that the highest number of scores are coming from the 1980s and into the 1990s. If we look at similar publication data for items stored in the music library, we again see items being published in the 80s and 90s being the highest representation but a fair dip when it comes to the 2000s and on. Then if we look at publication dates for scores at Downsview, we see why. Most of the scores from the 2000s on—or at least a high percentage of the scores from the 2000s on—are finding their way out to Downsview.

The charts also give us views of publication data. So, here, for example, we can see the top the top 20 publishers of that category of music that are out at Downsview, including Editions Bim, Billaudot, Schott, Emerson Editions, and so on.

Looking at the spreadsheets that the program creates, we see that we get a sheet called “overview” with columns for call number, name, analytical entry, title, uniform title, note, circulations, last use, library, collation, subjects. It also gives us other “sheets” that allow us to see circulation statistics, for example. This sheet is very similar to the overview sheet, except it organizes the data by item, not title. So, if there are two copies, we see both copies represented, or if it is an issue with multiple volumes, we would see all of the multiple volumes as entries with the associated data.

It also gives us sheets allowing us to see information on pagination, facsimiles, mini scores, vocal scores and scores with parts. This data was all derived from the catalogue record.

Looking at the spreadsheets we can see patterns start to emerge. Work duplications for example: same edition, different editions, different publishers, different format, selections from complete works, collections of pieces. All of these can represent duplication of data that is on our shelves—musical data.

We can see circulation information and publication data presented in a tabular format, which allows us to skim through with our eyes and facilitate analysis.

Using subdivisions for piano and wind as an example, with a quick analysis of the spreadsheets and the overview document we see:

9416 titles
0.04% of circulation from Downsview
23.8 % at Downsview
Items at Downsview are almost entirely published after 2000
Publication dates at the Music Library are highest in the 80s and 90s. Significant amount from the 50s on.
Publication dates for this entire section show a dip in the 2000s, yet still significant numbers.  USA is most common nationality, followed by France  Circulation:

o 3070 items at music have 0 circs; 4377 have 1 or more circs o 2319 items at Downsview have 0 circs

7.8% of composers are female
items with parts are the norm
Selections and collections of pieces are common
Multiple editions are common
We can notice publication information for types as well: Urtext, Facsimiles, Performance editions, Critical editions, etcetera

We also attempted to get a better understanding of the makeup of who created the items in our collection and get an idea of the diversity represented. To accomplish this, the program extracted the names of the composers from the divisions and subdivisions and searched against Wikidata.

So, for the total collection, as an example we see:

Male	61359
Female	4861
Transgender female	1
Unknown	1696

We can also get a bit of an idea of the creation age of the material in our collection. The program did this by looking at the date of death of the composers and organizing them into half century categories. With this data we can see that composers living in and around the twentieth century are well represented. The zero category is for composers who are still alive or did not have that data element represented in their Wikidata page.

It helps us to look at the composer’s country of origin for the material in our collection. Here we see Germany is well represented, with the US not too far behind.

We can also do this on a division level. Looking at Duets, for example:

Male	4783
Female	412
Unknown	198
Transgender Female	1

Here we see women are represented as composers in our duets category at a similar rate as the total collection.

Composer origins are a little different, but a Eurocentric picture still emerges. More “Unknown” here, than in the total collection. This suggests that more composers in this category are missing that data point on their Wikidata pages.

And for death dates, we see similar picture for duets as we did for the whole collection.

If we look at a subdivision level now, piano and winds again, we see women are represented at almost a percentage point greater rate than the entire collection: we see 7.8% female compared to 92.2% male.

Composers from the United States are now clearly in the majority, with Germany a little ways down the list.

Not all composers were able to be found in Wikidata however. Approximately one-third spread fairly evenly across the collection retrieved no results, and some brought back incomplete results. Due to the size of the data set, we feel that fairly accurate results were retrieved that allow us to work with confidence. However, perhaps more targeted surveys or research techniques could help enhance this data.

We also searched IMSLP to get an idea of representation from that “subject area” online. This technique worked best with large subject areas as organizational nature of IMSLP differs substantially from that of a library catalogue. With IMSLP you see an emphasis on the Work. So, if we look at the Canon and Gigue in D major, P.37 by Johann Pachelbel, we see this piece is represented at a work level, not an expression. So, with one URL we see a piece represented in over 72 different ways. Therefore, this work would be found for virtually every subdivision in our folders. Looking at it by broader categories, however, we can get an idea of who is represented and how often, though details on specifics would be difficult to parse out.

Other ideas for external sources to consult and contrast: we looked at the International Composer Diversity Project. The data that was retrieved there though did not add much to our knowledge as it was not marked up in a very meaningful way for our purposes. There was no differentiation made between distinct types of diversity: Women, Aboriginal, Black were all notated the same.

Other possibilities we are considering include comparing the data against our CD collection, or against our Naxos holdings, comparing the data against program files, our program files (for concerts and events, what was performed). We would need datasets to work on here and we would need some way of normalizing the names and titles of pieces, and this seems like a difficult obstacle for me to overcome. But it is one of the thoughts that we have. We also though about comparing this data against enrollment at the Faculty, or programs that the Faculty offers, perhaps noticing trends and how they have evolved over time and comparing these with publications dates in our collection.

So, with this data we hope to be able to get a better understanding of our collection, be able to articulate that understanding, and have data that will help us question the makeup of our collection.

4. Surveys, Focus Groups, and Ethics Approval

Once James had pulled and organized the data, the team met and reviewed the Orchestral Music folder together. We discussed what data was present, what we felt was missing, how to interpret the charts. We began a list of questions raised by the data.

Once we were on the same page with how to read and question the data, we divided up the remaining folders to review individually, then came together two weeks later to share our findings. We collected our notes, thoughts, and questions in a messy, shared document. Here are several examples from this document.

It appears that most of the collection was composed by German men. Is this accurate? How can we test this?
When a piece is duplicated in an anthology or volume of “selections,” should we treat it differently? (e.g., send all anthologies/selections to Downsview?)  How do our students/faculty use Miniature Scores?
How important are Facsimile Editions to users?
Some collections have an older average publication date: is this an indicator of condition? Are there implications for preservation?
How many hymn books do we need on-site?

We recognized that many of our questions should be put to our users to corroborate or explain the data.

In order to turn our messy document into a set of survey questions for our users, I consulted the University of Toronto’s Assessment Librarian. She suggested we first do a short survey of about ten minutes of our primary users: music students, staff, and faculty. In order to design survey questions, we should think about what statements we want to be able to make after the results are analyzed:

X percent of the music faculty do Y.
In a typical year, undergraduate students consulted the music score collection X number of times per week.
X proportion of graduate students requested a score from Downsview in the past year.
X proportion of faculty said that if a music score is at Downsview, they won’t bother requesting it.

From the survey data, we will identify “personas” that share the same traits and invite these personas to focus groups. For example, if undergraduate performance majors report that they have never requested a score from Downsview, we might want to consult a focus group made up of undergraduate performance majors to have a conversation with them as to WHY this might be.

Given the lack of literature in this area, we know we will want to publish and present the results of this assessment project, and that means going through our Ethics Review board. We’re currently developing the survey questions; our Assessment Librarian and our student workers will help test the instrument. Then we will seek approval from the Ethics Board. We don’t know how long this will take during the pandemic, but we hope to be able to administer the surveys and focus groups in the Fall.

5.1 Collection Development Policies

A perhaps obvious, but nonetheless central, question in determinations around our collection development policies is: “what do we want our library to look like”. As we have outlined in the discussion of our literature review and collections assessment methodologies, there are a number of different, and sometimes competing, elements to consider when attempting to devise sustainable and forward-thinking collections development policies.

While the task of developing our policies is incomplete and ongoing, we do, at this point have a strong idea of the questions we want our policies to inform. These questions include:

How do we continue to grow our collection in a focused manner, and one that will continue to serve our users’ needs both today and five to ten years down the road?
How can we best inform decisions around what materials are visible to our users? This includes decisions around what to keep on site versus what to send to off-site storage. These decisions will, in effect, decide what materials shall be visible in our catalogues, but invisible to patrons browsing our stacks.
How should we prioritize different types of scores? For instance, should collected works be prioritized for on-site reference? What about newly published materials, or performance scores, or piano reductions, etcetera?
Another important consideration is how much should traditional assessment criteria, such as usage stats and the condition of materials, play into our policies and decisions?

In regard to how to best prioritize what materials should live on-site versus be directed to off-site storage, there are many different aspects to consider. Currently, as Jan mentioned in her introduction, about 30% of our collection is housed at Downsview, and this number is certainly only going to increase. Our previous criteria for routing material to Downsview involved held that any new score that did not fall into the following criteria would be automatically directed to Downsview:

‘Serious works’ by Canadian composers
Solo piano music
Songs (one voice + piano)
Vocal scores (except choral, unless by major composers)
Major composers and award winners
Ensembles with unusual but playable instrumentation

Additionally, any score that had not circulated in at least 25 years, or are oversize and have not circulated recently, will be considered as candidates for deselection to Downsview.

These policies have resulted in the bulk of music score collection at Downsview being made up of scores published in the 2000s. They only make their way to the Music Library if someone requests them, at which point they are bound and added to our Music Library on-site stacks.

This begs the question, ‘is it a problem that the majority of our on-site Music Library collection is made up of older, and perhaps often less aesthetically pleasing materials? Similarly, is it a problem that many of our new compositions are invisible to users that rely on shelf-browsing to discover new materials? These are just a couple of the many questions at hand.

In order to remedy these questions and potential problems, we have come up with a number of preliminary ideas. One possibility lies in taking advantage of digitization possibilities, with respect to scores that have fallen out of Copyright (i.e. older than 1924). This will by no means solve our physical space-crunch, but could at least serve to increase access to our music scores, while maximizing space on-site for newer or otherwise most desirable materials.

As we continue to endeavour on this collections assessment project, it is our expectation that our planned student and faculty surveys and focus groups, in conjunction with a further analysis of the collections data we gathered, will help clarify and support our conclusions and determinations that will form the basis of our music score collections development policies. While we intend this thorough process to address needs at the Music Library both short term and long term, it is, at the same time, important to note that these will not be considered ‘final’ or ‘permanent’ collection development policies, but rather living documents that we will require regular revisiting in order to update the policies and reflect on, as well as anticipate changes in the Faculty of Music.

To this end, it is of utmost importance that we consider strategies and methodologies to address the increasing importance being placed on diversity within our collections. The past year has brought the long overdue topic of diversification of library collections to the forefront, not only in the library world in general, but specifically through calls to action from many prominent U of T Music Faculty stakeholders. This includes both current faculty and students, as well as alumni. It is our hope that the outcomes resulting from our collections assessment will directly support these calls to action, through ongoing conscious deliberation of our past shortcomings and future directions when it comes to the broad and comprehensive collecting of music materials.

This means ensuring that decisions around what lives at the Music Library versus Downsview reflect the rich diversity of not only our many different programs within the Faculty of Music, but also the incredible diversity of people and cultures within our department and the mosaic that is Toronto and Canada, broadly speaking.

Moreover, while a large part of our collection is our music score collection, of which this talk and our analysis today is the focus, we cannot forget about the other parts of our collection, for instance, our MLs, performances collections, archives, and special collections. Nor can we forget the importance of space needs for our users, such as study space and listening rooms.

Successful strategies for collections development policies will consider all these elements in relation to each other, and to our future envisioning of our library. We will endeavour to create a space that has ample room for exhibitions that can showcase the rich history and diversity within our collections, and one that provides a place to learn, a place to study, and a place to discover.

5.2 Side-bar: Archives Mandate and Policies

The space conundrum and necessity for a clear and informed collection development policy is not limited to our music score collection, although that is the main focus of our presentation and discussion at the Summit. We are simply running out of space in all areas of the library and the necessity of continued growth means this issue will only be exacerbated by time.

Let’s indulge for a moment in considering another of our collections—our archives—which recently underwent a comparable—although much smaller-scale and less formal—assessment project to determine “what stays” and “what goes”. The project consisted of two parts: the development of an archival collections mandate; and determining criteria for keeping materials onsite, and sending materials to Downsview.

Our archival collections mandate was formalized in the fall of 2019. It defines the scope of the collection and helps inform decisions made by the Librarians and Archivist at the Music Library on whether to accept or refuse donations. “The Music Library Archives acquires, makes accessible, and preserves records created or collected by individuals associated with the Faculty of Music at the University of Toronto, including materials created by faculty and staff, and materials that document their professional contributions to local, national, and international music communities.”

The mandate was written in part to define the relationship between the Music Library archival collections and other archival repositories at the University of Toronto and in Canada. Essentially, what is our role in preserving musical heritage and how do ensure that collection development efforts—in this case, donations—meet our goals?

The second part of the project was dealing with the space issue.

Until early 2020, all of our archival collections were stored onsite at the Music Library. However, they were quickly outgrowing the shelves available to them and were stored in a number of different locations within the library. To start the process of consolidating storage locations, and to make space for incoming donations, we decided to send some of our collections offsite to the climate-controlled, storage facility, Downsview.

The decision of which collections to store offsite and which to keep at the Music Library was based on four main criteria:

frequency of access over the past 20 years (based on handwritten sign-in books kept for our rare book and archival collections). Most frequently accessed collections were kept on site.
the size of the collection - sending a collection that consists of a single box, or a few folders would have little to no impact on our space problem.
whether further accruals were expected
and, the completion of a comprehensive, searchable finding aid in our online archival database to ensure the accessibility and discoverability of the collection.

13 collections were sent initially in February 2020, and this number is expected to grow as more collections meet the criteria.

It is important to note, that while these collections are less frequently accessed than those stored onsite, they are still accessible to our users, and materials can, and are, returned temporarily to the Music Library Reading Room for research access when required.

So far, this project has been successful. I have mediated a few requests over the past year for material stored at Downsview and we have acquired several new collections that fall under our collection mandate. As we move forward we will need to continually assess new and current collections to determine where they should be stored, much like the rest of our collections at the Music Library.

6. Conclusion (with a short Coda)

To sum up what my colleagues have presented, the University of Toronto Music Library has a number of collections-related challenges that we are addressing. These generally fall into three categories.

The first is physical space. One hundred years of collecting music resources has made for a collection and archive that is comprehensive in a traditional western classical music sense, and is also equally comprehensive in its physical footprint. There is not enough space in the present building to hold the collection, and you’ve heard my colleagues talk about the resulting workflows and legacy decisions that are shifting a now-sizable percentage of our older and brand-new materials to an offsite storage facility called Downsview. We’re working on comprehensive plans to address the challenges arising from these workflows through analytics, and forthcoming user surveys and focus groups, with a goal of creating some combined quantitative and qualitative data that will allow us to take concrete steps towards what we would like the future music library to be, and in doing so, chart a path towards a revised use of space that balances our in-house collection with our off-site storage.

Our second challenge is: how do we create sustainable collection growth to match the re-envisioned space? As you’ve heard from my colleagues, re-envisioned collection development policies will play a substantial role, as will digitization, collection analytics, and revised offsite storage policies. These traditional western methods will be significantly augmented by stakeholder focus groups that reach out to community members whose methods of learning and research needs fall outside of traditional western academic patterns, and in so doing, will help us shape the future library as a place that supports western and non-western research and learning methodologies.

Addressing this second challenge also requires addressing our third goal: How do we meaningfully diversify our holdings and ways of accessing information to promote the music of a broader spectrum of people? We recognize that our current collection primarily reflects the white male Eurocentric canon that has dominated conventional western classical music history, and we know this needs to change. Addressing this historical bias is one of the most challenging tasks. As mentioned earlier in this video, last summer the University of Toronto Faculty of Music received a Call to Action from its alumni, asking it to address systematic oppression, racism, and coloniality within the faculty’s programming and pedagogical practises. This call not only built on Black Lives Matter protests following George Floyd’s death, but also on a movement towards equity and inclusion following the Truth and Reconciliation Commission’s 2015 investigation into Canada’s Indian Residential Schools, and its subsequent Calls to Action. The Faculty of Music responded to the alumni’s Call to Action with a number of initiatives, including creating an Anti-Racism, Anti-Oppression committee, and is starting a meaningful deeper dive into its curriculum and pedagogical practises. Beyond needing to ensure the library is appropriately positioned to support the changing curriculum, we independently initiated our own self-assessment, determining through analytics what BIPOC-related materials the library already holds. We also initiated conversations on how we can improve equity of representation going forward, and make our existing BIPOC-related materials more visible and accessible.

Meeting the goals set out by these three challenges is key to ensuring that the library of tomorrow not only serves the students of tomorrow, but also the research, performance, space, and other as yet unknown needs of students in the decades to come. As part of working towards these goals, we look forward to hearing the other libraries’ summit presentations and the resulting conversations with you listeners, with the hope that these discussions will inspire transformative changes. Thank you.

What Stays and What Goes?