Journal of the Society for American Music (2015), Volume 9, Number 4, pp. 445–454.
C© The Society for American Music 2015 doi:10.1017/S175219631500036X

AmeriGrove II: Perspectives and Assessments
The Grove Dictionary of American Music. 2nd ed. Edited by Charles Hiroshi
Garrett. Oxford and New York: Oxford University Press, 2013.

Introduction

CHRISTINA BAADE and EMILY GALLOMAZZEI

Cue: Lalo Schifrin’s “Theme from Mission: Impossible.” If the return of Tom Cruise,
explosions, and fast-paced chases to the screen this summer is any indication,
Americans (and the global audience for Hollywood action film) love an impossible
mission. Certainly, U.S. music scholars do. Reading the eight reviews of the second
edition of The Grove Dictionary of American Music (hereafter, AmeriGrove II), edited
by Charles Garrett (with a large and distinguished editorial team and nearly fifteen
hundred contributors), as well as dipping frequently into its entries, we were deeply
impressed both by the quality and ambition of the eight-volume, 5.4 million-word
encyclopedia and by the “gargantuan” task (as Leta Miller puts it) with which the
reviewers had been charged.

We will focus our introduction upon the reviews themselves, asking (to rephrase
from Richard Crawford), “How does one read AmeriGrove II?”1 John Koegel,
Christina Baade’s predecessor as book reviews editor for JSAM, envisioned this
multipart review of AmeriGrove II and recruited eight intrepid scholars to each
consider a distinct topic, to which they all brought their own perspectives, strategies,
and voices. The reviewers and topics are as follows: Berndt Ostendorf and Wolfgang
Rathert, overview, European perspective; John Graziano, overview, U.S. perspective;
Glenda Goodman, music before 1800; Douglas Shadle, nineteenth-century music;
Kip Lornell, folk and traditional music; Leta Miller, twentieth-century art music;
Sherrie Tucker, popular music and jazz, 1900–50; and Theo Cateforis, popular
music and jazz, 1950–present.2

Reading their reviews was an exhilarating experience that helped us better appre-
ciate the achievement of AmeriGrove II in encompassing the growth and changes
in “American” music studies, allowing us to reflect on the current state of the
field, particularly in relation to big questions of belonging: Who are the “we” who
study American music? Who and what is included? How do we decide? These are
challenging questions indeed, given how many of us (readers and reviewers) have

1 Richard Crawford, “Amerigrove’s Pedigree: On The New Grove Dictionary of American Music,”
College Music Symposium 27 (1987): 174–75, quoted by Berndt Ostendorf and Wolfgang Rathert in
their review of AmeriGrove II in this issue.

2 Many thanks to John Koegel for envisioning this multipart review and engaging in the persuasive
tour de force involved in recruiting this outstanding team of reviewers. Thanks also to Karen Ahlquist
for arranging to publish all of the reviews in a single issue. Finally, our deep appreciation to all the
reviewers for taking on this tremendous task and carrying it out so well!

445

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

http://dx.doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


446 Baade and Gallomazzei

contributed to the volume and our temporal and personal closeness to the scholarly
conversations that shaped it.

Few reviewers could resist praising AmeriGrove II’s scope or remarking upon its
size; all of them engaged critically with the dictionary as a document showing the
development of the field of U.S. music studies and its central concerns. Glenda
Goodman, noting the ideological nature of such projects, linked AmeriGrove to
Diderot and d’Alembert’s Encyclopédie (1770–77), a project that “attempted to
map the world of knowledge,” and Noah Webster’s Dictionary of the American
Language, a lexicographical and nation-building endeavor. Indeed, music studies
is one of the few disciplines to have sustained this Enlightenment commitment to
cataloguing its field. With a family of Grove resources available at the stroke of a
key (at least for those with access to good academic libraries), exhaustiveness was
not the goal for Amerigrove II. It is explicitly a selective and nationally bounded
resource: “a repository of historically significant information” related to American
music (i.e., “musical life and cultures of the region now covered by the fifty states,
the District of Columbia, and US territories,” viii, vii). For reviewers, this selectivity
was both a strength—Theo Cateforis praised the ways in which scholars’ voices and
interpretations made their way into the entries—and an invitation to critique and
engagement. AmeriGrove II, like the first edition (hereafter, AmeriGrove I), edited
by H. Wiley Hitchcock and Stanley Sadie and published in 1986, is less a map than
a self-portrait of our field, inviting us all to greater reflexivity about what it is we
study and why.

To arrive at their analyses, each reviewer had to read AmeriGrove II quite dif-
ferently from the ways most “users” usually read the encyclopedia. Most users will
likely encounter AmeriGrove II online, entering search terms into the Grove Music
Online interface to find specific entries to answer a quick factual question or get an
overview of research in a given area. By contrast, the reviewers all focused on the
standalone version of Amerigrove II, in hard copy or in PDF, although some also
referred to the online version. Of course, it was unrealistic to read and comment
upon the entire encyclopedia. We were fascinated by the multiple ways in which
the reviewers engaged with this “impossible” text, drawing out important themes
while also negotiating its sheer size: they sampled, browsed, followed thematic
threads, and drew comparisons between the new edition and the old (as well as
with other references in the Grove family). They attended not only to what was
covered but also to how many words were assigned to a given topic, suggesting that
assigned word count was an important indicator of perceived significance. They
also read a great deal, closely and critically. Many focused their engagement upon
the wealth of essay-length entries surveying theoretical concepts, major genres,
social categories, important historical events, and cities—offering insights both
compelling and expertly informed.

In the rest of this introduction, we would like to bring one more reading strategy
to AmeriGrove II: that of text analysis using a digital humanities tool, Voyant,
developed by Geoffrey Rockwell and Stéfan Sinclair.3 If our expert reviewers could

3 Stéfan Sinclair, Geoffrey Rockwell, and the Voyant Tools Team, Voyant Tools (2012) (web
application), http://docs.voyant-tools.org. Our ability to carry out this analysis was enabled by

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

http://docs.voyant-tools.org
https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


Introduction 447

Figure 1. Screenshot of Voyant Tools interface with AmeriGrove II corpus. (Stéfan Sinclair, Geoffrey Rock-
well, and the Voyant Tools Team, Voyant Tools [2012], http://docs.voyant-tools.org).

not read the entire AmeriGrove, the Voyant tool, designed specifically to assist in
analyzing large corpuses, could do so with ease. Impossible Mission solved? Not
so fast. As Stephen Ramsay asserts in Reading Machines, the point of text analysis
is less to bring empirical, scientific fact to humanities inquiry and more to open
up new interpretive possibilities: “Text analysis, because it allows navigation of the
unread . . . [is] capable of presenting the bare, trivial truths of textuality in a way
that allows connection with other narratives—in particular, those narratives that
seek to install the text into a network of critical activity.”4 Thus inspired, we set
about putting our quantitative findings with Voyant into dialogue with the critical
observations of the reviewers.

Before we discuss our findings, a few more words about Voyant, the methods
we used, and the strengths and limitations of text analysis are in order. Voyant,
like other popular digital humanities tools, produces word frequency tabulations,
tracks word trends, allows users to investigate how keywords are used in context, and
provides a range of visualization options, among other possibilities (Figure 1). The
key to sharing results from these projects is visualization. One type of visualization
Voyant generates is the word cloud. Word clouds offer a visually striking way into
texts, but they are also imprecise and methodologically opaque (Figures 2 and 3).
For example, the two word clouds pictured here are the same size, although in a
more accurate representation, Amerigrove II would be twice as big as Amerigrove
I; further, they do not reveal how the text was manipulated to produce the results.
More useful, we found, were the word frequency lists generated for each edition,

Anna-Lise Santella of Oxford University Press, who facilitated our access to text files for both editions
of AmeriGrove. We also received significant support in facilities and expertise from the Lewis & Ruth
Sherman Centre for Digital Scholarship at McMaster University and its administrative director, Dale
Askey. A special thank you to Dr. Paige Morgan, a postdoctoral fellow at the centre, whose expertise
and helpfulness were invaluable throughout our research process.

4 Stephen Ramsay, Reading Machines: Toward an Algorithmic Criticism (Champaign: University
of Illinois Press, 2011), 16–17, 78–79. Many thanks to Dr. Jennifer Askey for recommending this book.

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

http://docs.voyant-tools.org
https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


448 Baade and Gallomazzei

Figure 2. Word cloud: most frequent terms in Amerigrove I. Generated by Voyant Tools.

Figure 3. Word cloud: most frequent terms in Amerigrove II. Generated by Voyant Tools.

which we were able to analyze—not only for a word’s frequency (i.e., its number of
occurrences) relative to that of other words within each edition but also for how a
word’s frequency differed between the editions.

To determine whether the trends (i.e., the word frequency differences between
editions) we discovered were significant, we employed a keyword significance test
often used by digital humanities scholars and corpus linguists: the log likelihood test
(LL). Although generating the word frequency lists was a useful starting point, their
values need to be normalized before they can add to our discussion: in this case,
the LL test compares the frequency of a word in two corpuses (Amerigrove I and II)
against their total word count to give a relative representation of the word in both.

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


Introduction 449

Figure 4. Screenshot of log likelihood calculator (from Tony McEnery and Andrew Hardie, “Statistics in Cor-
pus Linguistics,” Support Website for Corpus Linguistics: Method, Theory and Practice [Cambridge: Cambridge
University Press, 2012]. http://corpora.lancs.ac.uk/clmtp/2-stat.php).

To form a basis of comparison, the text from both editions of Amerigrove was
analyzed using Voyant. Before entering each corpus into the Voyant system, the
text had to be reworked to suit our purposes. For example, the bibliographies
accompanying entries were removed. This information would not tell us much
about how this new edition talks about U.S. music and could even skew the final
results gathered, particularly with regard to publication details like cities, presses,
and dates. Once the text was entered, we applied a standard English “stop words”
list in order to exclude common words like “the,” “and,” and “in” from our results.
Applying this stop words list ensured that the “top terms” explored in this intro-
duction were representative of musical genres, practices, and populations. The LL
calculator (Figure 4) then compared the frequency of the top fifty words within the
corpus of Amerigrove I vs. II to generate the log likelihood, which shows whether
the difference in frequency can be considered significant (Table 1).

We found performing word frequency analysis a useful point of entry into these
two large corpuses, especially when considered in relation to the reviewers’ in-depth
explorations. Given the importance the reviewers ascribed to the size of entries, we
found the ability to tabulate the number of words throughout both editions a
particularly compelling application of the tool. However, the context of terms—
especially terms used in multiple ways—was harder to decipher. Although many
of our findings correlated with themes identified by the reviewers, we also found
many intriguing results we were unable to fully explain given our constraints of
time and expertise.

One of our most striking findings is also obvious from a glance at the word
clouds: that both editions had a similar “most used” vocabulary after the stop
words were removed. They included obvious choices like “music” (and “musical”),
“american” (and “usa” in Amerigrove 1, “united” and “states” in Amerigrove II),
“song” (and “songs”); the intriguingly prominent “university”; and words linked to
chronological accounts of genres and lives: “became,” “early,” and “later.” Given that
we retained works and recordings lists (even though we stripped the bibliographies),
the prominence of “works” is unsurprising, while the high ranking of “pf” (and
“piano”) leaves no doubt as to the centrality of the piano in a range of musical
genres. Investigating the ranks of “new” and “york” shows “New York” to be second
only to “music” on both word frequency lists, confirming the city’s preeminence
in U.S. musical life. Finally, the almost geometric increase in the references (by

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

http://corpora.lancs.ac.uk/clmtp/2-stat.php
https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


450 Baade and Gallomazzei

Table 1. Top fifty words in AmeriGrove I and II (after English stop words list applied) with log likelihood.
The log likelihood number expresses the difference in word frequency between the two editions. The
higher the number, the more statistically significant the difference. Generated with Voyant Tools. Log
likelihoods calculated with McEnery and Hardie’s tool.

AmeriGrove I AmeriGrove II
More or Less Frequent Log

Term Frequency Term Frequency in Amerigrove I∗ Likelihood∗

music 22,490 music 49,302 Less 110.36
new 12,298 new 23,387 More 27.00
york 7,786 york 14,007 More 63.82
pf 6,965 american 13,812 Less 12.54
american 6,499 pf 12,244 More 81.83
works 6,165 musical 11,213 Less 26.14
songs 5,521 works 9,446 More 274.15
musical 5,106 songs 9,202 More 122.37
opera 4,421 jazz 7,701 More 2.44∗∗

university 4,348 university 7,479 More 67.98
jazz 3,940 early 6,784 More 0.75∗∗

orch 3,932 band 6,721 Less 66.47
became 3,586 became 6,479 More 27.26
piano 3,578 opera 6,404 More 278.38
early 3,428 orch 6,291 More 126.4
usa 3,109 including 6,182 Less 122.57
orchestra 2,986 piano 6,098 More 62.39
composer 2,850 united 5,359 Less 3105.09
band 2,781 states 5,341 Less 2997.72
recordings 2,723 composer 5,215 More 17.02
later 2,690 dance 5,133 Less 9.69
song 2,690 later 5,063 More 8.16
style 2,647 orchestra 5,004 More 20.26
years 2,628 song 4,999 More 11.41
boston 2,602 work 4,952 Less 14.51
studied 2,581 began 4,890 Less 3.24∗∗

chamber 2,474 popular 4,883 More 0.16∗∗

popular 2,447 school 4,807 Less 0.41∗∗

dance 2,358 years 4,789 More 17.03
including 2,358 group 4,347 Less 0.78∗∗

school 2,347 fl 4,277 Less 52.31
began 2,319 recordings 4,229 More 110.09
work 2,231 time 4,128 More 9.93
time 2,226 style 4,014 More 126.18
str 2,132 boston 3,952 More 122.75
group 2,107 john 3,915 Less 4.56
college 2,085 musicians 3,810 Less 23.20
vn 2,007 vn 3,794 More 5.34
composers 2,004 str 3,744 More 25.42
pieces 1,916 studied 3,733 More 163.91
chorus 1,893 world 3,726 Less 151.26
concert 1,875 performed 3,702 More 0.07∗∗

performed 1,851 city 3,683 Less 50.08
instruments 1,848 recorded 3,548 Less 3.55∗∗

composition 1,841 composers 3,464 More 29.49
john 1,829 college 3,441 More 50.92
chicago 1,818 century 3,359 Less 6.99
conductor 1,810 rock 3,341 Less 139.89
played 1,795 sound 3,337 Less 141.23
1970 1,743 played 3,256 More 12.7

∗Log likelihoods are given for words in the AmeriGrove II column.
∗∗This log likelihood suggests that the difference in frequency between the two editions is not significant.

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


Introduction 451

Table 2. Occurrences of century references in AmeriGrove I and II. Generated with Voyant Tools.
Century references include dates and ordinals (e.g., 1886 and 19th).

Aggregated Word Frequency Aggregated Word Frequency
Century Amerigrove I Amerigrove II

16th (1500s) 295 312
17th (1600s) 462 636
18th (1700s) 2690 3559
19th (1800s) 17862 25698
20th (1900s) 97071 164020

date and title) to each ensuing century is remarkable within both editions, as is the
striking growth for all periods between editions. Goodman’s and Douglas Shadle’s
observations that their centuries have been “repopulated” seems to be borne out
by these numbers—as are their observations that coverage of the eighteenth and
nineteenth centuries is dwarfed by coverage of the twentieth. In AmeriGrove II, there
are 30,205 appearances of pre-twentieth century dates and titles, whereas twentieth
century dates and titles increased from the first edition by more than twice that
number (Table 2).

As the reviewers observe, this explosion of twentieth-century coverage responds
to significant changes in the field, including the growth of research in non-classical
genres, especially popular music; the musics of African Americans and other racial-
ized people in the Americas; and broader questions of cultural hybridity, transna-
tional movement, and power. The word frequency lists register how seriously the
AmeriGrove II editorial team responded to these developments. A scan of the
word frequency charts confirms that with words like “rock,” “country,” “guitar,”
“broadway,” “gospel,” “traditional,” and even “band,” “album,” and “studio” mak-
ing striking leaps in the rankings with large, statistically significant increases in
representation.

Above all, however, “jazz” stands out as the most frequently occurring genre term
in AmeriGrove II (with 7701 appearances), reflecting both the increased coverage
of jazz in the new edition and the generous word counts for many entries, which,
as Cateforis point out, often align with word counts assigned to canonical classical
topics and people. A quick look also shows that “jazz” has surpassed “opera” as the
top-ranked genre. However, this “switch” in position requires a closer investigation.
The occurrences of both words has grown, but although the frequency with which
“opera” appears in the second edition is significantly lower than in the first edition,
the differences for “jazz” are not statistically significant. As Sherrie Tucker notes,
the coverage of jazz is greater in the new edition, but it was also well represented in
the first edition. Meanwhile, occurrences of “opera” have not increased at the same
rate as have the popular terms listed above.

Where goes opera, so goes art music? Finding the answer using text analysis is
more challenging than in the case of jazz or rock, for which single-word genre
designations are widely used. One reason, as Leta Miller observes, is that it has been
difficult to arrive at a satisfactory overarching term to describe “art” or “classical”
music (both of these words are quite low in the frequency lists for both editions). A

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


452 Baade and Gallomazzei

Table 3. Occurrences of references to racial and ethnic groups in AmeriGrove I and II (preliminary).
Generated with Voyant Tools.

Frequency Frequency More or Less Frequent Log
Categorization Amerigrove I Amerigrove II in Amerigrove I Likelihood

African Americana 2120 4984 Less 35.99
Asianb 235 825 Less 61.99
Native Americanc 1109 1725 More 44.34
Latinod 636 2873 Less 397.63
Whitee 1828 3062 More 38.50

a Represents primarily words incorporating “black,” “African,” and “Afro-.”
b Represents words incorporating “Asian” and “Oriental.”
c Represents words incorporating “Indian,” “native,” “First Nations,” and “aboriginal.” This count is incomplete:
it does not include names of nations and peoples.
d Represents primarily words incorporating “Latin,” “Latino/a,” “Spanish-,” “Mexican,” and “Puerto Rican.”
This count is incomplete: it does not include a full list of nationalities.
e Represents words incorporating “white,” “European,” and “Caucasian.”

second reason is that art music operates as an unmarked category in musicology:
Van Cliburn is a “pianist,” not a classical pianist, whereas Teddy Wilson is a “jazz
pianist.” Our workaround was to turn our attention to specifically classical genres
and terminologies, such as “opera,” “chamber,” and “symphony.” Two trends are
observable when we compare AmeriGrove I and II. First, with the exception of
“chamber,” the count for each of these terms grew.5 Second, all of these terms have
a lower overall rate of representation in AmeriGrove II. At least at the quantitative
level, AmeriGrove II has not stinted art music, but it has increased its representation
of a number of other non-classical genres to represent a much wider swath of U.S.
musical life.

What about other sorts of diversity, such as ethnicity, sexuality, and gender?
The reviewers all praised the quality, scope, and theoretical sophistication of the
entries on race, gender, sexuality, and ethnic groups and their musics, as well as the
expansion of biographical coverage of women, queer, and racialized people. Our
analysis confirmed that the occurrence of words used to describe African Americans,
Asian Americans, Latina/os, and Native Americans increased in the second edition;
the growth in the rate of occurrence was statistically significant for all of these
groups (with the probable exception of Native Americans; Table 3). Because we had
to identify and search for each individual term, using text analysis to investigate
representation for racialized groups in publications dating from 1986 and 2013
demonstrated to us the degree to which preferred language has changed (“the bare,
trivial truths of textuality,” indeed). It also presented significant methodological
challenges, particularly for Latina/os and Native Americans, groups often discussed
in AmeriGrove with reference to their nations (e.g., “Mexican” and “Choctaw”).
The politics of naming and representation are important, and we acknowledge that
our preliminary analysis only scratched the surface. Several of the reviews attend to
these questions with far more nuance.

5 “Chamber” fell from 2474 occurrences in AmeriGrove I to 2195 in AmeriGrove II, but
“symphony” grew from 1289 to 2395.

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


Introduction 453

If tracking references to racialized groups was difficult, using text analysis
to investigate the representation of white people in AmeriGrove was still more
challenging because whiteness tends to be an unmarked category. Critical schol-
ars, especially feminist, critical race, and queer theorists, have long argued that
unmarked categories (such as maleness, whiteness, or heterosexuality) reinforce
dominance by remaining “relatively invisible as an unstated, privileged norm,” as
Loren Kajikawa writes in his “Race and ethnicity” entry. Text analysis is blind
to what is not named, thus references to African American and black identi-
ties outnumber references to European American and white identities within
both editions. However, there was a numerical increase in the rate of occur-
rence for terms relating to white and European identities in AmeriGrove II—
reflecting, perhaps, the impact on our field of whiteness studies, which seeks to
make the dynamics of whiteness (and white privilege) visible and open to critical
investigation.

When considering questions of visibility, the changes in our scholarly field and
broader culture are perhaps most striking when comparing the representation of
sexual and gender minorities documented between the two editions. “Homosex-
ual,” “homosexuals,” “lesbianism,” “transsexual,” and “gay” in reference to gay men
(in the Bette Midler entry) each appeared once in the 1986 AmeriGrove I. In Ameri-
Grove II, there are nearly five hundred occurrences of words relating to transgender,
lesbian, gay, bisexual, and other queer identities. The musical and scholarly work
that transformed these identities from that which was almost literally unspeakable
into significant areas of research is charted in a number of AmeriGrove II subject
entries, including “Lesbian, gay, bisexual, transgender and queer music,” Nadine
Hubbs’s updating of Philip Brett and Elizabeth Wood’s classic entry; “Transgender,”
by Stephan Pennington; and “Sex, sexuality” by Fred Maus.

Turning to a final set of marked and unmarked categories, “women” made
significant gains, from 425 to 1494, although occurrences of the word still fell behind
references to “john” (3915), “george” (2326), “william” (2052), “charles” (1886),
“james” (1766), “paul” (1726), “david” (1700), “robert” (1538), and “thomas”
(1501). A fairer comparison would perhaps be occurrences of “he” and “she.” In
AmeriGrove I, there were roughly 5.7 occurrences of “he” for every “she” (28,135
to 4912); in AmeriGrove II, there were roughly 4.5 (45,606 to 10,229). This change
in ratio is due to a decreasing rate of occurrence for “he”; there was no statistically
significant change in the rate of occurrence for “she,” even though there was
significant numerical growth (the number doubled). Of course, these raw numbers
lack the nuanced analyses to be found in the reviews. It is also worth remembering
that the numbers reflect our field and its histories, contexts, and concerns as much
as they reflect AmeriGrove II, whose editorial team made attending to difference and
power, as well as the inclusion of underrepresented groups, a priority—at which
they succeeded in many ways.

If our rather basic text analysis offers insights into the shifting representations
of genres, centuries, race, sexuality, and gender, it also complements our reviewers’
sense that AmeriGrove II has become more attentive to border crossing, hybridity,
non-compositional musicking, and sound. Consider the surprising entrance of
“sound” to the second edition’s top fifty words, the decreasing shares of “composer”

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core


454 Baade and Gallomazzei

and “composition,” and the increased rate of representation for “dance,” “culture,”
“international,” and even “radio.”

Ultimately, our digital humanities adventure with Voyant gave us greater insight
into these two monumental corpuses, the tool’s colorful screen inviting us to explore
these texts in new ways, to dig below the relatively crude analysis we have presented
here, and to further interpret the lists it has produced. As our reviewers show,
AmeriGrove II (and I) is a capacious, engaging, impossible text that rewards many
sorts of reading—whether by human or machine, online or on paper. We hope you
are engaged and challenged by the ways into AmeriGrove charted by these reviews,
and hope they inspire you in your own impossible missions in American music.

at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S175219631500036X
Downloaded from https://www.cambridge.org/core. Carnegie Mellon University, on 06 Apr 2021 at 00:40:05, subject to the Cambridge Core terms of use, available

https://www.cambridge.org/core/terms
https://doi.org/10.1017/S175219631500036X
https://www.cambridge.org/core