The Ring Cycle

Journal of Tolkien Research
Volume 18 Issue 1 Article 4
2023
The Ring Cycle: Journeying Through the Language of Tolkien’s

Third Age with Corpus Linguistics
Michael Livesey
University of Sheffield, m.livesey@sheffield.ac.uk
Follow this and additional works at: https://scholar.valpo.edu/journaloftolkienresearch
Part of the Computational Linguistics Commons, and the Literature in English, British Isles Commons
Recommended Citation
Livesey, Michael (2023) "The Ring Cycle: Journeying Through the Language of Tolkien’s Third Age with
Corpus Linguistics," Journal of Tolkien Research: Vol. 18: Iss. 1, Article 4.
Available at: https://scholar.valpo.edu/journaloftolkienresearch/vol18/iss1/4
This Peer-Reviewed Article is brought to you for free and open access by ValpoScholar. It has been accepted for
inclusion in Journal of Tolkien Research by an authorized administrator of ValpoScholar. For more information,
please contact a ValpoScholar staff member at scholar@valpo.edu.
Livesey: The Ring Cycle
The Ring Cycle: Journeying Through the Language of Tolkien’s Third

Age with Corpus Linguistics
Journeying is a master theme in J.R.R. Tolkien’s published works.1 Indeed,

Middle-earth could be considered a world of journeys. Its stories are oriented
around a series of journeys: including those taken by Thorin Oakenshield’s
company, in their efforts to reclaim Erebor; by the Nine Walkers, in seeking to
destroy Bilbo’s ring; and, not least, by the Ring itself – as its centrality and
imagined nature unfolded across Tolkien’s writings. This article explores the
latter journey in greater depth: employing computational methods to assess
trends in Tolkien’s language use, across The Hobbit and The Lord of the Rings
(LotR).
The article feeds into a growing movement towards digital humanities
readings of Tolkien’s (and other authors’) published works. It proposes
statistical techniques enabled by recent advances in text programming can help
inform our interpretations of significant texts.2 Indeed, it argues analysts can
employ these techniques to pop the bonnet on language use within such texts:
reading materials through alternative (non-human) eyes, and mapping their
contents in such a way as to illuminate elusive linguistic and stylistic choices
authors make in crafting their worlds, characters, and narratives.3
In the below, I illustrate this contention by considering linguistic/stylistic
choices from Tolkien’s Third Age writings. I use tools from software-driven
corpus linguistics (a method of quantitative text analysis which has made
inroads into linguistic,4 literary,5 pedagogical,6 and social scientific7 research in
1
In Brian Rosebury’s words, in Tolkien “it is the journey, rather than the quest, which serves
as the unifying image”… Brian Rosebury, Tolkien: A Cultural Phenomenon (Basingstoke:
Palgrave Macmillan, 2003), p. 29.
2
Franco MoreN, Distant Reading (London: Verso, 2013); Franco MoreN, Graphs, Maps, Trees
(London: Verso, 2005); MarTn Paul Eve, Close Reading with Computers: Textual Scholarship,
Computa@onal Formalism, and David Mitchell’s Cloud Atlas (Redwood City: Stanford
University Press, 2019).
3
Ted Underwood, Distant Horizons: Digital Evidence and Literary Change (Chicago: University
of Chicago Press, 2019).
4
Paul Baker, ‘Times May Change, But We Will Always Have Money: Diachronic VariaTon in
Recent BriTsh English’, Journal of English Linguis@cs, 39.1 (2011), 65–88
<hbps://doi.org/10.1177/0075424210368368>.
5
Michaela Mahlberg, Corpus Stylis@cs and Dickens’s Fic@on (Abingdon: Routledge, 2013).
6
Dana Gablasova, Vaclav Brezina, and Tony McEnery, ‘CollocaTons in Corpus-Based Language
Learning Research: IdenTfying, Comparing, and InterpreTng the Evidence’, Language
Learning, 67.S1 (2017), 155–79 <hbps://doi.org/10.1111/lang.12225>.
7
Amir Salama, ‘Ideological CollocaTon and the RecontexualizaTon of Wahhabi-Saudi Islam
Post-9/11: A Synergy of Corpus LinguisTcs and CriTcal Discourse Analysis’, Discourse &
Society, 22.3 (2011), 315–42 <hbps://doi.org/10.1177/0957926510395445>; Tony McEnery,
Helen Baker, and Vaclav Brezina, ‘Slavery and Britain in the 19th Century’, in Time in
Languages, Languages in Time, ed. by Anna Čermáková and others (Amsterdam: John
Benjamins, 2021), pp. 9–38.
Published by ValpoScholar, 2023 1

Journal of Tolkien Research, Vol. 18 [2023], Iss. 1, Art. 4
recent years) to expose the hidden wiring of language use in The Hobbit, The
Fellowship of the Ring, The Two Towers, and The Return of the King. I consider
how Tolkien lexicalised his One Ring (“the central force of his symbolic
conception… the most powerful and most dangerous artefact in his imaginary
world”8) across these texts. Specifically, I employ corpus linguistic methods of
keyness and collocation analysis to: a) assess the One Ring’s centrality to
Tolkien’s unfolding Third Age narrative; and b) establish how the Ring’s nature
might have changed – as the semantic network within which Tolkien
operationalised it evolved.
Using these methods, I find the Ring did indeed undergo a significant
journey across Tolkien’s Third Age texts. Firstly, the Ring shifted from the
centre of the Third Age’s narrative arc to its periphery (as revealed by falling
keyness scores between texts). Secondly, the Ring transitioned from being
operationalised as an innocent, magical device – to one of ominous power or
heavy burden (as illustrated by changing collocation returns, whenever the word
ring reappeared). Collectively, I suggest these findings illuminate questions of
worldly disenchantment in Tolkien’s writings. They point to Middle-earth’s fall,
from a state of child-like mystery, adventure, and faerie – to one of darkness,
hierarchy, and amnesia. In doing so, they also shed light on possibilities for
interpreting Tolkien’s story-telling as a reflection of his (contested)
disenchantment with Britain’s industrialisation/modernisation.
Such is the argument I make vis-à-vis Tolkien’s writings. This argument
manifests a useful contribution to Tolkien studies: feeding into a digital turn in
Middle-earth scholarship, in its use of computational methods to advance the
interpretation of Tolkien’s legendarium. This argument also contributes to
literary studies more broadly: elaborating how readers might combine statistical
techniques with human-led interpretations, in evolving our sensitivity to
patterns of language use and authorial style.
The article began as an experiment with a friend. One lunchtime,
following a conversation in which I sought to convince that friend of corpus
linguistic softwares’ ease-of-use and analytical potential, I decided to run an
analysis of texts we both love (namely, The Hobbit and The Lord of the Rings).
This experimental study’s outcomes were surprisingly illuminating. They both
confirmed and challenged, deepened and reoriented, interpretations we had
intuited from our own private human readings. Hence, my decision to share
these outcomes more widely. During the experiment, I sought to answer two
research questions:
1. How does the One Ring’s centrality to the Middle-earth story evolve,
over its lifetime?
2. Do the Ring’s imagined qualities change, as Tolkien’s Third Age
writings unfold?
8
Adam Roberts, ‘The One Ring’, in Reading The Lord of the Rings: New Wri@ngs on Tolkien’s
Classic, ed. by Robert Eaglestone (London: ConTnuum, 2005), pp. 59–72 (p. 59).
https://scholar.valpo.edu/journaloftolkienresearch/vol18/iss1/4 2
These questions remain the basis for the present article. The article begins by
reviewing scholarly literatures. I consider recent shifts in Tolkien studies – away
from abstracted contests concerning Tolkien’s status within the literary canon,
and towards more detailed assessments of style and lexis. I introduce the scope
for corpus linguistic/digital humanities approaches to contribute to these shifts,
and reflect on insights derived from the handful of works undertaking such
contribution. I outline the methodology I employed to answer my two research
questions: noting my study’s conceptual and logistical parameters, and flagging
important weaknesses in computer-led readings (emphasising the necessity of
combining computational readings with, rather than treating them as a substitute
for, qualitative readings). Having set out my research design, I divulge my
findings: considering the One Ring’s growing peripheralization across Third
Age texts (as measured by keyness returns), as well as the transformation of its
meaning profile (per collocation returns). I conclude by noting these findings’
import for assessing Tolkien’s changing language use (the disenchantment of
Tolkien’s legendarium, as it evolved), as well as what such changes imply for
our efforts to interpret Middle-earth.
Tolkien studies: in the beginning was the word
I begin by situating this article in relation to literatures on Tolkien’s

legendarium, and on the use of corpus linguistics in literary studies. I argue
conventional critical reception of Tolkien’s works has become lost in debates
around their literary value (or lack thereof). These debates often become
cyclical, and shed little light on the work itself – the production of rich, affective
stories through specific techniques of style and lexis. More recently, and in
response to this circularity, Tolkien scholars have proposed a descent to the level
of word and language: foregoing attacks on/defence of Tolkienian texts’ status
as “literature” – in favour of a more specific assessment of their technical
parameters. I suggest corpus linguistics can add value in this descent to word. I
introduce corpus linguistics as a particular method for reading texts – which,
when combined with human readings, can illuminate questions of style/lexis in
interesting and unexpected ways. I illustrate this value-add by referencing the
handful of works which have undertaken corpus linguistic Tolkien studies, and
considering insights their authors have delivered.
Historically, contests over his works’ literary value have been a major
feature of Tolkien’s critical reception. Indeed, in reviewing past scholarship at
the turn of the century, Michael D. C. Drout and Hilary Wynne identified
“defence of Tolkien” as one of four major strands within Tolkien studies.9
According to this tendency, critical work on Tolkien divides between two
9
Michael D. C. Drout and Hilary Wynne, ‘Tom Shippey’s J.R.R. Tolkien: Author of the Century
and a Look Back at Tolkien CriTcism since 1982’, Envoi, 9.2 (2000), 101–67 (pp. 113–17).

antagonistic schools. On the one hand, Tolkien’s “detractors”10 amongst

“contemporary literary studies, modernist and post-modernist [schools] base
their dismissal of Tolkien’s work as unworthy of study based on his supposedly
‘poor writing’”.11 Burton Raffel’s early claim that LotR, for all its “magnificent
performance, full of charm, excitement and affection”, was not “literature”12 set
the tone in this regard. Operationalising an exclusive (yet also non-measurable,
and therefore simultaneously elusive) standard for literature, Raffel found LotR
fell short of standing alongside texts like “The Iliad or The Odyssey… Paradise
Lost, or The Great Gatsby”13 – on grounds that Tolkien’s writing was
“embarrassingly bad”, or enjoyed “almost no independent literary merit”.14 For
scholars of Raffel’s school, Tolkien’s sustained popularity had less to do with
“correct and proper [literary] taste” – than with “a kind of literary disease, whose
sufferers… should be scorned, pitied, or rehabilitated”15 (as interpreted by Tom
Shippey).
Unsurprisingly, and on the other side of the coin Drout and Wynne
identify, Tolkien’s apologists (Shippey foremost amongst them) react to such
attacks on his works’ literary merit through a defensive posture. This second
critical school responds to “intense critical hostility… the refusal to allow
[Tolkien] to be even a part of ‘English literature’”16 through a rear-guard action:
seeking to re-establish Tolkien as “a significant figure of the twentieth century,
someone of comparable stature to, say, Poe or Peacock among nineteenth-
century writers”, rather than merely “a best-seller”.17 This defensive line has
become a foundational one for Tolkien Studies. As Drout and Wynne put it,
“nearly every Tolkien critic has worked to some degree or another on the
problem of defending Tolkien against his detractors”.18 The defence takes
various forms – including especially through abstracted political/sociological
claims-making regarding the “class hostility” of literary criticism’s “haute
bourgeoisie”;19 the intrinsic value of morality tales;20 or an innate yearning for
pre-modern ecologies.21
10
Drout and Wynne, pp. 113–17.
11
Robin Anne Reid, ‘Mythology and History: A StylisTc Analysis of The Lord of the Rings’,
Style, 43.4 (2009), 517–38 (p. 517).
12
Burton Raffel, ‘The Lord of the Rings as Literature’, in Tolkien and the Cri@cs, ed. by Neil
Isaacs and Rose Zimbardo (London: University of Notre Dame Press, 1968), pp. 218–46 (p.
218).
13
Raffel, p. 220.
14
Raffel, pp. 229–331.
15
Tom Shippey, J.R.R. Tolkien: Author of the Century (London: HarperCollins, 2000), p. 4.
16
Shippey, p. 128.
17
Rosebury, p. 1.
18
Drout and Wynne, p. 114.
19
Shippey, pp. 129–30.
20
Hal Colebatch, Return of the Heroes: The Lord of the Rings, Star Wars, Harry PoOer and
Social Conflict (Perth: Australian InsTtute, 1990).
21
Patrick Curry, Defending Middle-Earth: Tolkien, Myth, and Modernity (New York: Mariner
Books, 2004).
For all that Tolkien’s defence is a laudable endeavour, and for all the
truth in accusations of establishment snobbery vis-à-vis Tolkien’s “‘anorak-
clad’ followers”,22 the energy expended in this cycle of attack and defence is
counter-productive. Put simply, the debate between critics and apologists
generates more heat than light when it comes to our understanding of, and
appreciation for, Tolkien’s literature. For this debate tends to occur at a level of
remove from the texts themselves. As Shippey discovered of “one of [Tolkien’s]
most vehement” critics during a private conversation following a radio debate,
for example: “he had never actually read The Lord of the Rings which he had
just been attacking”.23 Similar inattention to textuality applies to defensive
scholarship, moreover. Again, Drout and Wynne’s review concluded Tolkien
studies’ “biggest failing” was the field’s “lack of discussion of Tolkien’s style,
his sentence-level writing, his word choice and syntax”.24 Four years later,
Drout found similarly that, though “J.R.R. Tolkien’s prose style… has been both
attacked and defended, its details have seldom been analysed in terms of specific
aesthetic effect”;25 while Robin Anne Reid proposed “stylistic or applied
linguistic scholarship” on Tolkien’s legendarium remained “limited”,26 by the
end of the decade.
Operating a primary concern with Tolkien’s status as an author, rather
than on his use of language, defensive debates constrain deeper understandings
of Tolkien’s craft as a story-teller. This includes understandings concerning
narrative techniques, aesthetic styles, or lexical choices Tolkien made, in forging
his legendarium. Such is the logic behind more recent scholarship, proposing a
descent in analysis – away from abstracted debates around status/meaning, and
towards the assessment of language use. Scholars like Brian Rosebury argue for
“understand[ing] and evaluat[ing] Tolkien’s works as compositions, that is, as
products of literary art which are for readers aesthetic experiences”27 – rather
than engaging in contests over their connection to/placement alongside literary
canon. This means developing micro-level tools and mechanisms that tell us
how the composition took shape, at word- or sentence-level – as much as why,
or to what end.
It’s worth remembering in this vein that, for Tolkien himself, it was “the
words [which] create[d] the story”.28 Tolkien considered himself a “philologist
before he was a mythologist”:29 words and “linguistic aesthetic”30 were at the
22
Shippey, pp. 129–30.
23
Shippey, p. 14.
24
25
Michael D. C. Drout, ‘Tolkien’s Prose Style and Its Literary and Rhetorical Effects’, Tolkien
Studies, 1 (2004), 137–63 (p. 137).
26
Reid, p. 517.
27
Rosebury, p. 5.
28
29
Shippey, p. 7.
30
Elizabeth Kirk, ‘“I Would Rather Have Wriben in Elvish”: Language, FicTon and “The Lord of
the Rings”’, Novel, 5.1 (1971), 5–18 (p. 7).

forefront of his thinking, in producing works that were (in Tolkien’s own
reflections) “fundamentally linguistic in inspiration”.31 In contrast to all the heat
generated by debates around their literary status, following Tolkien’s self-
professed interest in the specific craft of word use advances scholarship by
shedding light on ways the legendarium took shape… Namely: word by word.
Per Drout, “critics who have focused solely on source or theme should note that
the analysis of style may unearth new sources and shed light on traditional
themes as well”…32 As, for instance, in parallels Drout draws between Return
of the King and King Lear in the former’s “grammatical, syntactic, lexical, and
even aural effects”;33 or, in “the impression of textual depth” which Drout,
Namiko Hitotsubashi, and Rachel Scavera note of consistent “references to
other, absent texts”34 across Third Age writings. By addressing “the issue of
style, of sentence-level writing”35 in Tolkien, these works alter the tenor of
scholarship. Not only, developing our knowledge of the detail of his creation –
but, further (and in so doing), contributing to that creation’s very defence:
demonstrating artful ways it took shape through linguistic practice. Such
accounts of technical, lexical, stylistic craft present a useful response to
“mainstream”36 critiques that the texts themselves were not of sufficient
aesthetic standard to merit consideration as literature: by demonstrating how
“elegant and powerful”37 Tolkien’s specific word choices really were –
concluding, forcefully, “that LotR would be a lesser work if it were written any
other way”.38
What can the present article contribute to this shift from cyclical debates
on status, to evolving awareness of language? This article makes a case for the
inclusion of digital humanities methods within this descent to word: as one path
towards delivering its objectives of illuminating Tolkien’s linguistic craft. For,
though studies paying attention to word use have indeed advanced existing
scholarship, many of them draw upon similar methods of qualitative analysis in
generating their findings – and few have drawn upon innovations in
computational analysis that open up new ways of thinking about word use. To
dwell on one example: Nils Ivar Agøy’s 2013 assessment of the use of adjectives
in LotR came to valuable conclusions concerning Tolkien’s “invitational
style”.39 This was a style that adopted thin description – not for lack of literary
31
Tolkien in Shippey, p. 6.
32
Drout, p. 155.
33
Drout, p. 137.
34
Michael D. C. Drout, Namiko Hitotsubashi, and Rachel Scavera, ‘Tolkien’s CreaTon of the
Impression of Depth’, Tolkien Studies, 11 (2014), 167–211 (p. 179).
35
36
Sharon Bolding, ‘Review: Tolkien as a Literary ArTst by Thomas Kullmann and Dirk
Siepmann’, Mythlore, 41.1 (2022), 277–84 (p. 277).
37
38
39
Nils Ivar Agøy, ‘Vague or Vivid? DescripTons in The Lord of the Rings’, Tolkien Studies, 10
(2013), 49–67 (p. 63).
ability (as Tolkien’s sceptics claim), but with the “deliberate” goal of leaving
the depiction of characters/landscapes so “open that they are mentally filled in
by their readers”.40 Ivar Agøy’s findings advance the field: providing a lens for
thinking about Tolkien’s specific literary style (one running contrary to his
historical critical dismissal). Yet, Ivar Agøy arrived at these findings according
to a self-confessedly unsophisticated method. Being “methodically hampered
by the fact that I am a historian, not a literary scholar”, Ivar Agøy’s “approach
[was] extremely simple”:41 hand-counting the number of descriptions in each
LotR chapter, to understand (in raw, absolute terms) the sheer volume of depth
Tolkien engaged when envisaging Middle-earth’s people/places.42 More
sophisticated methods for calculating language use can add complexity to this
picture – deepening Agøy’s significant argument, by enhancing tools by which
it was attained. In critiquing Tolkien, Raffel felt the former’s language must be
“both more deeply felt and more deeply worked”43 if it was to be considered
alongside the Fitzgeralds and Lawrences of twentieth-century literature.
Ironically, it is exactly this deep working which a digital humanities approach
to Tolkien permits: illuminating patterns of style and lexis through the
deployment of tools beyond the capacities of conventional human readings.
Corpus linguistics: reading rhizomatically
As a method of digital humanities scholarship, corpus linguistics can advance

Tolkien studies’ descent to the level of word: affording procedures and measures
that raise our sensitivity to Tolkien’s authorial craft, by identifying patterns of
style/lexis across large-n corpora. But what is corpus linguistics? Corpus
linguistics is a field of quantitative text analysis: in which computer softwares
are employed to read a text sample (the corpus), and identify patterns of word
use within it. This field emerged from advances in language programming at
Brown University during the 1960s;44 and has since gained a footprint in
disciplines from medicine45 to mathematics.46
When it comes to analysing texts like those considered in this article,
there are two major arguments in favour of corpus linguistic methods. The first
is as regards the scope and substance of claims made about linguistic samples.
40
Ivar Agøy, p. 63.
41
Ivar Agøy, p. 50.
42
Ivar Agøy, p. 59.
43
Raffel, p. 221.
44
Henry Kučera and W. Nelson Francis, Computa@onal Analysis of Present-Day American
English (Providence: Brown University Press, 1967).
45
Jane Demmen and others, ‘Language Mabers: RepresentaTons of “Heart Failure” in English
Discourse - a Large-Scale LinguisTc Study’, Open Heart, 9.1 (2022), e001988
<hbps://doi.org/10.1136/openhrt-2022-001988>.
46
Juan Mejia-Ramos and others, ‘Using Corpus LinguisTcs to InvesTgate MathemaTcal
ExplanaTon’, in Methodological Advances in Experimental Philosophy, ed. by Eugen Fischer
and Mark CurTs (London: Bloomsbury, 2019), pp. 239–64.

Corpus linguistic tools can enhance these claims’ validity: by setting them
against a systematic assessment of word use within the entirety of the relevant
population. As Gerlinde Mautner has put it, with the support of statistical
findings from corpus linguistic techniques, analysts can advance “less
speculative”47 claims about the quantity and quality of language use. Being
drawn from replicable counts of all lexical items within a body of text, these
claims are not inhibited by “researcher bias”,48 including “cherry-picking”49 of
unrepresentative data. It’s hard to take issue, for example, with the claim a theme
is losing relevance across a corpus, if the volume of use for words proxying that
theme can be shown to be in systematic (and statistically significant) decline.50
This is the most basic argument in favour of quantitative language
analyses: that they advance a dispassionate, “scientific” characterisation of all
materials within a large text population. Interestingly, other proponents of
digital humanities theorise these strengths of systematic computational readings
vis-à-vis the “universal, but often unspoken, bounding of mortality” amongst
human readers – with the major reason for reading with computers being that,
for human subjects, “death cuts short every totalising attempt to read
everything”.51 Computers can help critics overcome this “finitude of
humanity”,52 in relation to the scale of data pertaining to our analysis: by simply
reading that data much faster than we ever could. Such circumvention of
mortality seems curiously appropriate to the analysis of texts like Tolkien’s.
After all, Tolkien’s literature pivots, similarly, on questions of morality –
including, a device conferring immortality! (Not a computer, but a ring.) As
Rosebury notes, “Tolkien himself spoke of the wish to escape from death as ‘the
oldest and deepest desire’”.53 Computational methods parallel this desire in
curious ways – overcoming human constraints, by making accessible/assessable
volumes of text beyond our temporal bounds.
The second argument in favour of corpus linguistic tools for text analysis
is the often-unexpected findings such totalising readings draw from their text
sample – findings which can challenge, deepen, or reorient intuitive human
cognition. When using corpus linguistic techniques in my work, I’ve often
found they open doors to new lines of research – lines I hadn’t anticipated, or
which I’d overlooked, when undertaking qualitative readings. By nature,
computers read textual materials differently to humans. This reading is not
47
Gerlinde Mautner, ‘Checks and Balances: How Corpus LinguisTcs Can Contribute to CDA’, in
Methods of Cri@cal Discourse Studies, ed. by Ruth Wodak and Michael Meyer, 3rd edn
(London: SAGE, 2016), pp. 154–79 (p. 42).
48
Mautner, ‘Checks and Balances: How Corpus LinguisTcs Can Contribute to CDA’, p. 156.
49
Markus Rheindorf, Revisi@ng the Toolbox of Discourse Studies: New Trajectories in
Methodology, Open Data, and Visualisa@on (London: Palgrave Macmillan, 2019), p. 2.
50
Baker, ‘Times May Change, But We Will Always Have Money: Diachronic VariaTon in Recent
BriTsh English’.
51
Eve, p. 3.
52
Eve, p. 12.
53
Rosebury, p. 57.
sequential (word by word), but global (swallowing an entire corpus in one gulp).
Moreover, computational readings are not intuitive/interpretive (allowing a
text’s meaning to take shape without needing to enumerate that meaning’s exact
parameters), but literal/direct (drawing an understanding of textual substance
from a precise counting of contents).
This alternative form of reading produces different results to human-led
approaches. Often, such results can appear meaningless, abstruse, or overly
exact. If one was to ask a computer what Shakespeare’s sonnets say about love,
for example, one would be given the literal list of words Shakespeare uses to
say love (perhaps ordered by frequency), rather than any insight into how these
words connect to each other, what texts/meanings they reference from outside
the corpus, or how they build particular visions of relationship. However, results
of computerised reading can also be thought-provoking. By dint of having
received the text differently, computer programmes can open our eyes to
patterns undetected by human cognition. Just as individuals with different life
experiences, and different knowledges of intertextuality, bring different
interpretations to a single sample of text; so computers that read corpora non-
intuitively can shed “different perspectives on the data”.54 Computational
readings can map textual data in ways that deviate from conventional readings
– “pinpointing [new] areas of interest for a subsequent close analysis”.55 Such
multiplicitous mapping is useful from an analytical perspective, insofar as it
changes angles from which we view our corpus. This is what the father of
critical discourse analytic method, Norman Fairclough, found in learning to
combine corpus linguistic methods with qualitative approaches: suggesting the
former
can be useful in checking out impressionistic conclusions about which

words and co-occurrences of words are most significant, in alerting
analysts to words and co-occurrences which they had not noticed, and in
providing statistical information about certain features of discourse.
Above all it can stimulate new ideas which might lead to new directions
of investigation and analysis…56
Corpus linguistics can bring systematicity and multiplicity to the analysis of

language. Both these arguments have been advanced in applying computational
tools to literary studies. A small but growing number of scholars have chosen to
employ corpus linguistics to aid their readings of works, ranging from Lewis
54
55
Paul Baker and others, ‘A Useful Methodological Synergy? Combining CriTcal Discourse
Analysis and Corpus LinguisTcs to Examine Discourses of Refugees and Asylum Seekers in the
UK Press’, Discourse & Society, 19.3 (2008), 273–306 (p. 284)
<hbps://doi.org/10.1177/0957926508088962>.
56
Norman Fairclough, Language and Power, 3rd edn (Abingdon: Routledge, 2015), p. 21.

Carroll’s Alice57 to Virginia Woolf’s The Waves.58 These scholars have noted
corpus methods’ primary facility in “aid[ing] systematicity and objectivity in
the analysis of literary texts”.59 But they have also noted the “alternative
interpretative engagement”,60 which digital readings sustain. Kieran
O’Halloran, in particular, makes a helpful comparison between computational
methods of text analysis and Gilles Deleuze/Felix Guattari’s notion of
“rhizomatic” investigation. Citing Deleuze and Guattari’s botanical metaphor of
“the rhizome61 as a productive image of creative thought”, O’Halloran suggests
corpus linguistic searches “often” return results “which would have been
difficult to predict. When a corpus search leads to unpredictable results, this
discovery could be construed as rhizomatic”62 – by which O’Halloran means
that, like the rhizome, the corpus linguistic researcher is able to draw unforeseen
connections and discover unanticipated sources of interpretive inspiration.
In my use of corpus linguistic methods to assess literary texts, I came to
a different metaphor: that of corpus linguistics as a means to “pop the bonnet”
on language use within a text. When we watch a vehicle pass by, we perceive
its movement, power, and aesthetic, without sight of the cogs and pistons that
drive these elements. Likewise, when we read a text intuitively, we receive its
meaning, profundity, and style, without recording precise mechanics by which
such principles function. In listing the units that make a text in their naked
volumes and operation, however, corpus linguistics can bring that text’s
mechanics into sharp focus: lifting the lid on “patterns of usage of which [we
might have] had only a vague notion or even no knowledge at all”,63 based on
human readings.
This popping the bonnet is what the present article sets out to achieve:
using corpus linguistic techniques to expose hidden mechanics of language in
Tolkien’s Third Age writings. Doing so situates this study amongst the (even
smaller, but likewise growing) group of works applying corpus linguistics to
Tolkien’s legendarium. Part of the wider turn to digital humanities in Tolkien
57
Paul Rayson, ‘ComputaTonal Tools and Methods for Corpus CompilaTon and Analysis’, in
The Cambridge Handbook of English Corpus Linguis@cs, ed. by Douglas Biber and Randi
Reppen (Cambridge: Cambridge University Press, 2015), pp. 32–49 (p. 45)
<hbps://doi.org/10.1017/CBO9781139764377.003>.
58
Giuseppina Balossi, Corpus Linguis@c Approach to Literary Language and Characteriza@on:
Virginia Woolf’s The Waves (Amsterdam: John Benjamins, 2014).
59
Mahlberg, p. 5.
60
Kieran O’Halloran, ‘Performance StylisTcs: Deleuze and Guabari, Poetry and (Corpus)
LinguisTcs’, Interna@onal Journal of English Studies, 12.2 (2012), 171–99 (p. 172).
61
(A creeping underground plant stem which sends roots out at random from mulTple nodes
– thus delivering new opportuniTes for growth, and producing new fruit in different
locaTons.)
62
O’Halloran, pp. 174–78.
63
Ángela Almela and Irina Keshabyan, ‘A New Approach to Literature: Corpus LinguisTcs’,
Interna@onal Journal of English Studies, 12.2 (2012), i–iv (p. i).
studies,64 a handful of scholars have employed corpus linguistics to gain insight

into Tolkien’s story-telling styles. The most significant work in this group is
Thomas Kullmann and Dirk Siepmann’s 2021 Tolkien as a Literary Artist.65 By
quantising language from LotR, Kullmann and Siepmann shed light on the
highly-precise, deliberate, and artistic ways in which Tolkien crafted his
universe. Of “the impression of archaicity which any reader will experience on
reading Lord of the Rings”,66 for example, Kullmann and Siepmann find this
stylistic quality was no accident. Rather, they suggest readers’ intuitive sense of
archaicity emerges from deliberate forms of linguistic curation: involving the
unusual combination of new and old registers which, when brought together,
engender a jarring lexical rhythm forcing the reader to take account of non-
modern lexicons. Kullmann and Siepmann draw this finding from collocation
analysis – with the systematic consistency of modern-medieval collocations
elaborating the authors’ core contention, that “far from being ‘amateurish’,
Tolkien’s prose effectively employs several centuries’ worth of linguistic
developments, putting old words to new uses”.67
In the only other peer-reviewed work I’m aware of that uses corpus
linguistics to analyse Tolkien’s writings, Vanessa Milom comes to a similar
conclusion. Milom draws on statistical techniques to assess Tolkien’s
orthographic choices: the different spellings he employed when constructing
dialogue for the Gollum character, versus dialogue for Sméagol. Drawing on
statistics for word keyness, Milom finds high levels of “sibilant reduplication”
whenever Gollum speaks (as compared with Sméagol): the repetition of s’s
(fissh, pocketses), which “illustrate the cave-dwelling Gollum as a snake-like
and surreptitious character”.68 She concludes “the linguistic mannerisms of the
Gollum-Sméagol [character] are distinct and change depending on which of the
two are speaking”. Milom’s use of statistical techniques bring this feature of the
text unambiguously to light: illuminating just how “thoughtfully curated”69
Tolkien’s language choices were.
Earlier, I proposed corpus linguistics as a pathway to enhancing Tolkien
studies’ descent to the level of words – including, by illuminating specific
stylistic/lexical choices Tolkien made in crafting his legendarium. These two
studies demonstrate possibilities of this corpus linguistic pathway. Together,
they advance an argument regarding Tolkien’s place amongst other literary
64
See especially James Tauber and others, ‘Digital Tolkien Project’
<hbps://digitaltolkien.com/> [accessed 15 December 2023].
65
Thomas Kullmann and Dirk Siepmann, Tolkien as a Literary Ar@st: Exploring Rhetoric,
Language and Style in The Lord of the Rings (London: Palgrave Macmillan, 2021).
66
Dirk Siepmann, ‘Tolkien as a Stylist: A Corpus-Based InvesTgaTon into “The Lord of the
Rings”’, in 9th Interna@onal Corpus Linguis@c Conference (University of Birmingham, 2017), p.
2.
67
Siepmann, p. 3.
68
Vanessa Milom, ‘Corpus LinguisTc Analysis of the Idiolects of Gollum and Sméagol’, Journal
of Linguis@cs and Literature, 5.1 (2022), 1–5 (p. 5).
69
Milom, p. 5.

greats – not just a powerful story-teller (as is well-known), but also an artful
writer. Importantly, they don’t advance this argument in abstraction from the
writing itself. Rather, they employ the advantages of rhizomatic, computer-led
readings to clarify Tolkien’s precise and deliberate linguistic strategies – a
precision and deliberateness whose oversight elsewhere has kept Tolkien “out
of the mainstream literary tradition and sidelined his writing… from receiving
its due consideration”.70 Using corpus linguistics to re-read literary texts can
unravel such tendencies in past criticism: illuminating unnoticed patterns of
language/style, and taking our interpretations in new directions. The present
article builds on this argument, as it pertains to literary studies broadly or
Tolkien studies specifically: elucidating Tolkien’s lexicalisation of the One Ring
across The Hobbit and The Lord of the Rings. Before I divulge my findings vis-
à-vis mechanics of this lexicalisation, however, let me first outline my study’s
research design. Having set out the uses and value of corpus linguistic methods
for analysing literary texts, I now proceed to précis specific processes by which
I applied them to my Third Age corpus.
Methodology: keyness, collocation, weaknesses
As flagged in my introduction, my analysis of Third Age texts drew on two

research questions:71
1. How does the One Ring’s centrality to the Middle-earth story evolve,
over its lifetime?
2. Do the Ring’s imagined qualities change, as Tolkien’s Third Age
writings unfold?
These questions map sympathetically onto corpus linguistics’ two core

techniques: keyness and collocation analysis.
Keyness analysis: focus corpus, reference corpus, Simple Maths Parameter
In corpus linguistics, keyness analysis represents the assessment of a text

sample’s “aboutness… its topic and the central elements of its content”.72
Corpus linguistic softwares ascertain this aboutness by comparing a focus
corpus (the corpus of interest, about which an analysis is being constructed)
against a reference corpus (a comparable sample of language, excluding the
language of the focus corpus itself). This comparison draws out elements which
are unique to the focus corpus: words which appear in statistically significant
high or low volumes, vis-à-vis the reference. Other scholars have used this
70
Bolding, p. 277.
71
I used LancsBox 6.0 so{ware for my corpus linguisTc analysis. Vaclav Brezina, Pierre Weill-
Tessier, and Tony McEnery, ‘#LancsBox v.6.0.0’ (Lancaster: Lancaster University, 2021).
72
Baker and others, p. 278.
process to clarify language use’s diachronic evolution: by comparing a corpus

of language from one period against that of another, to see whether words/topics
rise or fall in centrality over time.73
I propose a similar approach here: exploring the One Ring’s diachronic
evolution, by comparing each of Tolkien’s four Third Age texts74 against a wider
Third Age reference corpus: to identify whether/how the Ring’s keyness
changed as the Third Age unfolded. If different Third Age texts return different
keyness scores for the word ring, that would suggest the extent of the Ring’s
centrality to the Third Age’s narrative arc varied over time. Moreover, if
variance in ring keyness is in a particular direction (up or down), that would
suggest the Ring is undergoing a sustained increase or decrease in relevance to
Middle-earth story-telling.
I created my focus and reference corpora by using Optical Character
Recognition (OCR) software to scan75 my copies of HarperCollins’ 2020
editions of The Hobbit76 and LotR.77 This created a fully digitised database of
Tolkien’s Third Age texts – which I read through in detail to correct errors and
remove extraneous data like page numbers. I then divided my database by Third
Age volume – giving me individual .txt files for each volume. These .txt files
were my four focus corpora for keyness analysis. I created my reference corpus
by combining these four .txt files into a single Third Age file. This procedure
left me with four Third Age focus corpora (The Hobbit, The Fellowship of the
Ring, The Two Towers, The Return of the King), and one Third Age reference
corpus (against which I could compare my focus corpora, to identify relative
levels of word keyness in each volume).
Whenever using my Third Age reference corpus in keyness analysis,
however, I removed the specific focus corpus under investigation from the wider
reference file. Each keyness analysis thus proceeded by comparing the focus
corpus to a reference corpus of all Third Age texts minus the focus corpus itself
(i.e., comparing The Hobbit to all Third Age texts minus The Hobbit; comparing
73
Helen Baker, Tony McEnery, and Andrew Hardie, ‘A Corpus-Based InvesTgaTon into English
RepresentaTons of Turks and Obomans in the Early Modern Period’, in Lexical Priming:
Applica@ons and Advances, ed. by Michael Pace-Sigge and KaTe Paberson (Amsterdam: John
Benjamins, 2017), pp. 42–66 <hbps://doi.org/10.1075/scl.79.02bak>; Baker, ‘Times May
Change, But We Will Always Have Money: Diachronic VariaTon in Recent BriTsh English’;
Michael Livesey, ‘Introducing the “Conceptual Archive”: A Genealogy of Counterterrorism in
1970s Britain’, European Journal of Interna@onal Security, 8.4 (2023), 471–92
<hbps://doi.org/10.1017/eis.2023.10>.
74
(The Hobbit, The Fellowship of the Ring, The Two Towers, The Return of the King.)
75
The use of OCR to digiTse copyrighted texts for non-commercial research is legal in the UK,
under the Copyright, Designs and Patents Act 1998 secTons 29/29A (provided sufficient
source acknowledgement is made).
76
J.R.R. Tolkien, The Hobbit, 2nd edn (London: HarperCollins, 2020).
77
J.R.R. Tolkien, The Lord of the Rings: The Fellowship of the Ring, 2nd edn (London:
HarperCollins, 2020); J.R.R. Tolkien, The Lord of the Rings: The Two Towers, 2nd edn (London:
HarperCollins, 2020); J.R.R. Tolkien, The Lord of the Rings: The Return of the King, 2nd edn
(London: HarperCollins, 2020).

The Fellowship to all Third Age texts minus The Fellowship). This was an
appropriate step in ensuring my analysis highlighted details which were unique
to each Third Age volume. It’s important to note the important work that other
Tolkien scholars have undertaken, in compiling digitised corpora for use in
research like mine. James Tauber, in particular, has created a remarkable source
of marked-up, XML versions of Tolkien texts under the aegis of the Digital
Tolkien Project.78 I did not use these XML files in my analysis. However, I am
willing to share .txt files I generated for this article with any colleagues seeking
to conduct similar (non-commercial) research in future.
So much for the focus/reference corpora I used to generate keyness
metrics. But what of the metrics themselves? There are various ways to measure
keyness. The version of keyness I use in this analysis is Simple Maths Parameter
(SMP), as proposed by Adam Kilgariff.79 SMP keyness works by dividing the
relativised count80 of each word within the focus corpus, by the relativised count
of the same word in a reference corpus. The result of this formula is a number
quantifying how much more, or less, that word features in the focus corpus vis-
à-vis its reference. The beauty of SMP is in the simplicity of its interpretation
(hence simple maths parameter). If a word returns SMP keyness of 2.00, for
example, it can be read as being twice as key in the focus corpus vis-à-vis the
reference corpus. Conversely, if that word returns an SMP of 0.50, it can be read
as half as key in the focus vis-à-vis the reference.
Collocation analysis: MI2
Keyness metrics illuminate the quantity of word use in a corpus. By contrast,

collocation metrics shed light on word qualities: the meanings by which words
are used within the language sample. I use collocation to answer my second
research question: supplementing analysis of how much the Ring featured in
each of Tolkien’s Third Age texts, with further analysis of how it featured. The
definition of collocation is “the above-chance frequent co-occurrence of two
words within a pre-determined span”.81 Corpus linguistics approaches this co-
occurrence as the essence of meaning-making – on the Firthian principle that
“you shall know a word by the company it keeps”.82 (I.e., that word meanings
emerge from contexts of use, such that when words connoting time consistently
appear in the presence of words connoting money, for instance, this points to
these semantic fields’ co-constitution: the mutual notion of value and resource
78
Tauber and others.
79
Adam Kilgarriff, ‘Simple Maths for Keywords’, in Proceedings of Corpus Linguis@cs
Conference CL2009 (Liverpool: University of Liverpool, 2009), pp. 1–6.
80
RelaTvisaTon is important here – ensuring comparability of keyness counts across corpora
of different sizes.
81
82
John Rupert Firth, Papers in Linguis@cs, 1934-1951 (Oxford: Oxford University Press, 1957),
p. 179.
by which English language conceptualises time and money.83) Collocation

analysis uncovers the network of meanings pertaining to a word within a text by
listing terms which appear alongside that word with every appearance, and
calculating these co-occurrences’ probabilities to indicate the strength of each
co-relation. This procedure is more complex than keyness analysis, and involves
a reading beyond the capacity of intuitive, human cognition.84
As with keyness analysis, there are various ways to operationalise
collocation. In my study, I employ the Mutual Information (MI2) metric to
clarify the quality of the Ring’s usage in Tolkien’s writings. Unlike other
collocation measures quantifying mere frequency of words’ collocations, MI2
captures the “exclusivity” or “tightness”85 of collocational relationships. MI2
draws on a formula comparing observed collocations between a word/its
collocates, with chance collocations – if the corpus under analysis was reordered
at random. This procedure filters out common words like the or and, which
naturally appear in high frequencies alongside any keyword (a by-product of
their ubiquity in general language, rather than an indication of meaningful co-
constitution). Instead of highlighting common terms, MI2 identifies the co-
appearance of meaningful terms like royal and family – terms which might
appear in lower absolute frequencies than common terms, but whose consistent
co-occurrence signifies a non-chance meaning relation.
MI2 scores are harder to interpret than those for SMP keyness. They are
best read comparatively, to ascertain high or low values. As a point of reference,
when applied to the BE06 corpus of modern British English,86 the search
parameters used in this study87 returned an MI2 of 8.66 for the collocation royal-
family. As above, this is a very strong (and common) collocation in British
English. MI2s of 8.66+ can therefore be read as evidence of very strong
relationships between words within a corpus of interest. In searching for ring
collocates within Third Age texts, I set a threshold MI2 of 9.00. This means that
all collocates considered in my study exist in a stronger relation to ring in
Tolkien’s writings, than do royal and family in BE06. I short, I limit my analysis
to co-configurations which meaningfully affect the quality of the Ring’s
lexicalisation. (For clarity, I didn’t use BE06 as a reference corpus in my
83
Vaclav Brezina, Sta@s@cs in Corpus Linguis@cs: A Prac@cal Guide (Cambridge: Cambridge
University Press, 2018), p. 78 <hbps://doi.org/10.1017/9781316410899>; Vaclav Brezina,
Tony McEnery, and Stephen Wabam, ‘CollocaTons in Context: A New PerspecTve on
CollocaTon Networks’, Interna@onal Journal of Corpus Linguis@cs, 20.2 (2015), 139–73 (pp.
153–153) <hbps://doi.org/10.1075/ijcl.20.2.01bre>.
84
The results produced by collocaTon analysis enable a similar analysis to Ivar Agøy's analysis
on Tolkien's descripTve style, but at a much higher level of sophisTcaTon.
85
Gablasova, Brezina, and McEnery, pp. 163–64.
86
Paul Baker, ‘The BE06 Corpus of BriTsh English and Recent Language Change’, Interna@onal
Journal of Corpus Linguis@cs, 14.3 (2009), 312–37 <hbps://doi.org/ijcl.14.3.02bak>.
87
(Returning collocates within a span of 5 words either side of my keyword, and with a
minimum frequency of 3 collocaTons across the focus corpus – search parameters replicaTng
Kullmann and Siepmann’s usage.)

research. I merely used it to acquire a baseline for defining statistical

significance in MI2 scores.)
Weaknesses: decontextualisation, chronology, pagination
Finally, a caveat on weaknesses of my study (which also illuminate generic

weaknesses of “scientific” method in analysing language). For, despite my
claims regarding corpus linguistics’ analytical strengths, “a word of warning
[remains] in order”.88 The apparent neutrality of statistical results, and the
apparent validity of large-n quantitative sampling, can lull scholars/their readers
into a false sense of security vis-à-vis their findings’ significance and surgical
provenance. In fact, it is important for readers not to “overinterpret”89 the
significance of corpus linguistic findings – including findings from the present
article. The primary problem with quantitative analysis of text is the
“necessarily decontextualised”90 character of statistical outputs. Computers read
corpora at distance – swallowing the entirety of a text in one gulp, and returning
insights abstracted from the mess of in-text word use. Corpus linguists need to
be cautious in reading statistical returns, therefore, which might appear
significant at first sight – but which may actually be overblown, or misleading.
One of the most significant collocations in The Hobbit, for example, is
that between bell and ring (a configuration with a very high MI2 of 12.13).
Reading this score alone, one could misinterpret the Ring’s meaning within The
Hobbit – finding Tolkien used ring in the context of other metal artefacts, for
instance. A closer reading of instances where these words co-occur, however,
clears up the confusion. When ring and bell appear in The Hobbit it is not in the
sense of the One Ring, but in the sense of the verb “to ring” – as in the following
concordance lines:
Just before tea-time there came a tremendous ring on the front-door bell,
and then he remembered!91
They had not been at table long… when there came another even louder
ring at the bell. “Excuse me!” said the hobbit, and off he went to the
door…92
Avoiding statistics’ misinterpretation, as in these examples, requires the

researcher to shuttle continuously between quantitative returns and a close
reading of relevant texts. These quantitative returns should never be considered
in isolation from the lines from which they are drawn. On the contrary,
88
89
Mahlberg, p. 22.
90
Mahlberg, p. 7.
91
From Chapter 1: An Unexpected Party.
92
Same chapter.
quantitative and qualitative readings should be considered complementary –

with findings from each being used to “triangulate”93 insights from the other.
Hence my earlier emphasis on combining computational readings with, rather
than treating them as a substitute for, qualitative readings. This is what Martin
Paul Eve meant in describing computational methods as an “environment in
which we can ‘think along’ with machines”94 – using digital methods “where
they are helpful and appropriate”, but “abandon[ing them] when they become
overly forced”.95 I operationalise this restraint throughout my write-up, to guard
against misleading claims. Like Eve, “I aim to avoid [treating] everything like
a technological nail, just because I have a digital hammer”.96
A second weakness concerns my study’s chronological ordering. In
reporting my analysis below, I work through each volume97 of Tolkien’s Third
Age writings sequentially. This design may appear incongruous to Tolkienists.
For Tolkien approached LotR not as “a ‘trilogy’ but [as] a unified work of some
600,000 words”,98 whose presentation in three separate volumes had less to do
with narrative flow than printing constraints. Nonetheless, I chose to structure
my analysis volume-by-volume for three reasons.
The first two reasons for this design are methodological. Descending
from volume-level analysis (as in my study) to book- or chapter-level analysis
(in common with Tolkien’s intentions) would have raised issues of corpus
granularity. Large-n quantitative methods work best when applied to large-n
data samples. And, though some insight can be gained from smaller/bespoke
corpora,99 there remains consensus amongst corpus linguists that corpora need
to be of a minimum size to return meaningful statistical calculations.100 The
Hobbit already tests this minimum size, with a total wordcount of 95,559 tokens.
Some LotR books fall even further below this count – with books four and six,
for instance, including only 66,374 and 64,738 tokens each. Guarding against
misleading claims for collocation in particular (MI2’s calculation is based on
the number of possible collocations, which in turn correlates with total tokens),
required grouping Third Age books into trilogy volumes (with each volume
capturing well over 100,000 words – a good baseline for robust findings). A
further methodological rationale for grouping my analysis trilogically concerns
the Ring’s non-balanced appearance across LotR books. For, as a physical
93
94
Eve, p. 2.
95
Eve, p. 12.
96
Eve, p. 12.
97
(The Hobbit, The Fellowship of the Ring, The Two Towers, The Return of the King.)
98
Rosebury, p. 11.
99
Gerlinde Mautner, ‘Corpora and CriTcal Discourse Analysis’, in Contemporary Corpus
Linguis@cs, ed. by Paul Baker (London: Bloomsbury, 2012), pp. 32–46 (p. 37).
100
Hai Zhao, Yan Song, and Chunyu Kit, ‘How Large a Corpus Do We Need: StaTsTcal Method
Versus Rule-Based Method’, in Proceedings of the Seventh Interna@onal Conference on
Language Resources and Evalua@on (Valleba: European Language Resources AssociaTon,
2010), pp. 1672–77.

character-object, the One Ring is absent from LotR books three and five – which
follow the trajectories of characters other than Frodo and Sam, post-breaking of
the Fellowship. This absence has implications for keyness returns (logically,
ring would appear less frequently in books where it is physically absent – which
is not an especially insightful finding). More importantly, it also has
implications for collocation analysis. MI2 returns risk distortion when applied
to keywords returning low wordcounts, because of the smaller pool of observed
collocations (an incidental outcome of corpus size, rather than a meaningful
reflection of meaning). I chose to avoid such distortion, by structuring my
analysis trilogically.
The third reason for organising my analysis this way had to do with my
argument’s flow and complexity. In writing this article, per its beginnings in an
attempt to persuade my colleague of corpus linguistics’ ease-of-use, I set out to
tell an accessible and fluent story regarding the Ring’s progress through
Tolkien’s Third Age lexis. Structuring this story according to seven books, or 81
chapters (as opposed to four volumes), would have added unnecessary
illegibility to this story. This illegibility would apply for Tolkien specialists: who
might understand a book-level focus, but who would find the additional
complexities of a seven-fold keyness/collocation analysis hard to follow. It
would also apply for corpus linguists: who might have an interest in the
application of established techniques to a new source base, but for whom LotR
is indeed most familiar as a trilogy (whose six-fold division would render the
reading of my findings fragmentary or inaccessible). Besides necessary
considerations of methodological robustness noted above, therefore, I chose to
stick with a volume-based assessment to ensure my account (whether on
computational readings, or on Tolkien) remained legible to multiple audiences.
My last caveat concerns ways I present my findings. As noted earlier, in
entering Tolkien’s texts into my corpus linguistic software,101 I removed all
matter extraneous to the Third Age narrative itself (page numbers, foreword,
prologue, appendices, etc.). This was to avoid confusing software, which cannot
distinguish between analytically-relevant and irrelevant contents. The downside
of taking this step was that I lost page numbers for materials considered in my
study. In my write-up, therefore, I am limited to referencing book and chapter
numbers for all excerpts cited. This limitation is a common and accepted feature
of corpus linguistic studies102 (and, indeed, of Tolkien studies itself103).
Keyness analysis: the disenchantment of Tolkien’s world?
Having established my study’s argument, its relationship to scholarly literatures,

and its methodological design, I now turn to its substantive findings. In
101
Brezina, Weill-Tessier, and McEnery.
102
As, for instance, in Baker, ‘Times May Change, But We Will Always Have Money: Diachronic
VariaTon in Recent BriTsh English’; or in Baker, Brezina, and McEnery.
103
reflecting on LotR, Adam Roberts concludes that “when one starts to look at [the
text] ‘through the Ring’, as it were, it starts to assume a certain ubiquity”.104
According to Roberts, the Ring enjoys a pre-eminent position in Tolkien’s Third
Age, as the centrepiece to the latter’s narrative structure. Certainly, on a
narrative level, Tolkien’s Third Age pivots on the Ring: whether in Bilbo’s
finding of it, Frodo’s quest to destroy it, or Sauron’s attempts to recover it. But
does this narrative centrality manifest in Tolkien’s use of language? Is the Ring
really ubiquitous to Third Age style/lexis?
According to my keyness analysis, the answer is no. This analysis
actually found the Ring’s prevalence to be a passing, and indeed declining,
feature of Third Age language. The word ring appears in varying frequencies in
each Third Age text: 68 times in The Hobbit (0.71 mentions per thousand
words); 306 times in The Fellowship (1.72/thousand words); 70 times in The
Two Towers (0.45/thousand words); and 73 times in The Return of the King
(0.54/thousand words). Per these figures, the Ring enjoys relatively high
frequency of mentions in The Hobbit, followed by a peak of frequency in The
Fellowship. But the frequency of its mentions subsequently declined to low
levels in The Two Towers and The Return of the King. The same pattern applies
to the associated word rings (as in “rings of power”): which appears 0.08 times
per thousand words in The Hobbit; rising to 0.28 per thousand in The
Fellowship; but then falling back to 0.05 and 0.04 mentions per thousand words
in The Two Towers/Return of the King.
To convert these naked frequencies into readable SMP ratios: ring enjoys
SMP keyness of 0.83 in The Hobbit; 2.85 in The Fellowship; 0.43 in The Two
Towers; and 0.56 in Return of the King (for rings, the scores are 0.81; 2.64; 0.55;
0.50). As a reminder, these scores quantify the prevalence of each word in each
text, by comparison to a wider Third Age corpus (all texts, minus the one being
assessed). The scores indicate that the Ring was 83% as key in The Hobbit as in
the wider Third Age (i.e., a ratio of nearly 1:1); 285% as key in The Fellowship
(nearly three times as prevalent as in other texts); and 43%/56% as key in The
Two Towers/Return of the King. In short, the Ring’s keyness journey began with
a moderate plateau, followed by a substantial peak, and then a precipitous
decline (to roughly half its original prevalence).
To put it in even simpler terms, the Ring is not ubiquitous to Third Age
writings. On the contrary, its centrality to the narrative arc is broadly in decline
as these writings evolve. This finding precipitates a follow-up question: which
motifs might have displaced the One Ring’s centrality, as the Ring itself went
into decline? My keyness analysis revealed a series of significant terms enjoying
steady growth in keyness across The Hobbit/The Lord of the Rings. I list ten of
these terms in the table below: giving SMP scores for each term, in each text.
104
Roberts, p. 60.

Table one: keyness scores for ten rising keywords in Third Age texts
The Hobbit The The Two The Return of

Fellowship of Towers the King
the Ring
king 1.19 0.18 0.71 3.45

captain 0.76 0.44 0.91 2.48
lord 0.35 0.40 0.83 3.59
men 0.55 0.37 1.35 2.21
war 0.48 0.48 1.25 2.14
blood 0.58 0.72 1.07 1.74
death 0.63 0.46 1.25 1.98
shadow 0.38 1.10 1.05 1.36
terror 0.55 0.98 0.99 1.42
darkness 0.68 0.86 1.09 1.35
For ease of interpretation, I’ve also visualised these scores in the following
graph.
What the table and graph show is continuous growth in keyness for certain
significant words, across Tolkien’s Third Age. These include words like king,
captain, lord, and men – capturing motifs associated with order, hierarchy, and
masculinity. The keyword lord, for example, enjoys continuously rising keyness
across the Third Age: from 0.35 in The Hobbit; to 0.40 in The Fellowship, 0.83
in The Two Towers, and 2.48 in Return of the King. These scores identify
lordship as only 42% as central as ring in The Hobbit and 14% as central as ring
in The Fellowship – but fully 641% as central by Return of the King. Likewise,
men grows from keyness of 0.55 in The Hobbit, to keyness of 2.21 in Return of
the King; meaning masculinity became exactly four times as central to the
Middle-earth story by the time of its fourth chapter as it had been in its first –
and, indeed, 4.5 times as central as the ring motif. Other rising keywords
connote with war – including war itself, blood, and death. War, for example,
rises from keyness 0.48 in both Hobbit and Fellowship; to keyness 1.35 in The
Two Towers; and 2.14 in Return of the King. Again, by the Third Age’s final
instalment, war had displaced ring at the story’s heart – with higher centrality
at a factor of 3.8. Finally, words like shadow, terror, and darkness also enjoy
steadily-growing centrality across the Third Age – with the three words being,
respectively, 2.43, 2.54, and 2.41 times more central to language than ring in
Return of the King (having been only 0.46/0.66/0.82 times as central in The
Hobbit and only 0.39/0.34/0.30 times as central in The Fellowship).
These statistics reveal that the Ring became less central to Tolkien’s
language as the Third Age progressed, whilst themes of hierarchy, masculinity,
war, and darkness became more central. We can add a layer of meaning to these
findings, if we consider what other words, besides ring, also lost centrality
across the Third Age. The table below details ten further falling keywords:
words whose usage, like ring, was in decline with each Third Age instalment.
Table two: keyness scores for ten falling keywords in Third Age texts
The Hobbit The The Two The Return of

Fellowship of Towers the King
the Ring
adventures 2.90 0.90 0.54 0.68
wizard 2.90 1.16 0.59 0.38
spells 1.39 1.01 0.86 0.88
elves 1.64 2.08 0.64 0.24
music 1.17 1.39 0.80 0.73
history 0.64 1.63 0.84 0.84
ancient 1.09 1.24 1.00 0.70
friendship 1.25 1.02 0.97 0.84
wandering 1.48 1.44 0.68 0.65
forest 2.56 0.90 1.15 0.23

Again, I’ve visualised these scores in graphical form, for ease of interpretation.
Just as my first graph demonstrated the trend towards rising keyness for words
like men or darkness, so my second graph demonstrates falling keyness amongst
words like elves or friendship.
Scores from the table, and its graphic representation, balance my findings on
rising keywords. They show that the Ring was not the only element to lose
centrality as the Third Age narrative evolved. On the contrary, several words fell
out of Tolkien’s usage in the Third Age’s latter instalments. Adventures, for
example, was only 24% as key by Return of the King as it had been in The
Hobbit. Likewise, the keyness of wizard fell by a factor of 7.63 across each text;
music was half as relevant by the end of the Third Age as it was at its start; and
the footprint of forest in Return of the King was only 10% what it had been in
The Hobbit.
Like my rising keywords, we could group these falling keywords into a
series of thematic categories: capturing motifs of magic (wizard, spells, elves),
adventure (adventures, wandering, friendship), memory (history, ancient), and
mystery (music, forest). Of course, these categorisations are imperfect, and there
are many overlaps between them. Nonetheless, that themes of magic, adventure,
memory, and mystery became more peripheral as the Third Age wore on, whilst
themes of hierarchy, masculinity, war, and darkness became more central, tells
us something significant about Tolkien’s evolving Middle-earth imaginings.
Revisiting Tolkien’s unfolding Third Age language use through corpus
linguistic analysis, we see that as the Ring became less ubiquitous to the
legendarium, so too that legendarium lost its enchanting, mysterious, and
adventurous characteristics. At the same time, keyness returns indicate Middle-
earth became increasingly dark, hierarchical, and amnesiac in parallel with the
Ring’s marginalisation (losing sight of its ancient history, even as it became
embroiled in the blood of men’s wars).
Such disenchantment of Middle-earth ties into significant narratives
from the wider story. It reflects, for example, the otherworldly elves’ departure
from Middle-earth – and their displacement by fallen men. (As in Frodo’s
exchanges with Gildor/Galadriel in The Fellowship: “we are Exiles, and most
of our kindred have long ago departed and we too are now only tarrying here a
while”105/“our power is diminished, and Lothlórien will fade, and the tides of
Time will sweep it away. We must depart into the West, or dwindle to a rustic
folk of dell and cave, slowly to forget and to be forgotten”.106 Or, in an exchange
between Legolas and Gimli in Return of the King: “seldom do [men] fail of their
seed… The deeds of Men will outlast us, Gimli”.107) This theme of the elves’
departure is repeated in the quantities of Tolkien’s word choice – with elves
falling in prevalence, and men rising.
This disenchantment also reflects the legend-isation of ents and hobbits
amongst Middle-earth’s peoples. (Per Aragorn’s/Theoden’s surprise at the
former’s appearance in The Two Towers: “The Ents!… Are there still Ents in the
world? I thought they were only a memory of ancient days, if indeed they were
ever more than a legend of Rohan”108/“Ents!… Songs we have that tell of these
things, but we are forgetting them, teaching them only to children, as a careless
custom”.109 Or, comments on hobbits from a man of Rohan/Faramir, in The Two
Towers/Return of the King: “Halflings! But they are only a little people in old
songs and children’s tales”110/“now we come to strange matters… For this is not
the first halfling that I have seen walking out of northern legends into the
Southlands”111 – not to mention Tolkien’s own reflections on hobbits in the
prologue to The Fellowship: “Hobbits are an unobtrusive but very ancient
people, more numerous formerly than they are today”.) Like elves, hobbits also
suffered declining keyness with each Third Age instalment: with the word
105
Book 1, chapter 3: Three is Company.
106
Book 2, chapter 7: The Mirror of Galadriel.
107
Book 5, chapter 9: The Last Debate.
108
Book 3, chapter 5: The White Rider.
109
Book 3, chapter 8: The Road to Isengard.
110
Book 3, chapter 2: The Riders of Rohan.
111
Book 5, chapter 4: The Siege of Gondor.

hobbit falling from SMP 3.68 in The Hobbit; to 0.62 in The Fellowship; 0.74 in
The Two Towers; and as low as 0.53 in Return of the King (14% what it had
been in the first text). Likewise, the word hobbits: which rose in keyness from
0.48 to 2.10 between The Hobbit and Fellowship of the Ring, but subsequently
fell to 1.48 in The Two Towers and as low as 0.57 in Return of the King (again,
only 27% as key as it had been mid-Third Age; and only 25% as key as men by
the Third Age’s end).
Combining my returns from keyness analysis with these excerpts points
to a trend in Tolkien’s writings: Middle-earth’s transition from a legendary land
of music/adventure, elves/ents, to a modern world of men, blood, and steel. This
transition runs parallel to the One Ring’s peripheralization in the Third Age’s
narrative arc. This is not to say that the Ring symbolises adventure or magic,
rather than darkness or terror. After all, the Ring’s semantic profile (as we will
see) crosses both these thematic categories. Nonetheless, these findings give
pause for thought. They reveal patterns regarding Middle-earth’s aboutness: the
specific linguistic choices Tolkien made in lexicalising his universe – vis-à-vis
which motifs Tolkien allowed to acquire centrality, and which he confined
increasingly to its margins. If other corpus linguistic Tolkien scholars are right
in finding his language use to be highly deliberate, we would have to conclude
that this shifting balance between centre and periphery is by design. Namely,
that it sheds light on Tolkien’s vision for Middle-earth’s evolving history. And,
further, that understanding this balance in depth might equally illuminate
Tolkien’s own journey – as regards his disenchantment with the real world (his
dissatisfaction with his country’s “shabby destruction”112 under
industrialisation/modernisation), and how this might have bled into his
conceptions of Arda.
I now undertake this effort at greater understanding. I consider ways
Tolkien disenchanted his One Ring, specifically: turning to different meanings
by which Tolkien operationalised the Ring in each Third Age text (as illustrated
by variation in collocation returns), and exploring what changing qualities of
the Ring’s operationalisation (alongside its declining quantities) might tell us
about Tolkien’s disenchanted Middle-earth visions.
Collocation analysis: the disenchantment of the Ring?
To recap this article’s start-point… My study began with two research questions;
the second of which was: do the Ring’s imagined qualities change, as Tolkien’s
Third Age writings unfold? The most straightforward answer to this question is:
yes. According to my findings from collocation analysis, the word ring appears
with a shifting semantic profile across Third Age texts.
112
Paraphrasing Tolkien’s foreword to the second ediTon of The Lord of the Rings.
Within my search parameters,113 my collocation analysis returned a list

of 197 words that appear alongside ring whenever Tolkien used the latter word.
This is itself a relatively narrow lexical field. Ring appeared 517 times across
my four texts, meaning the word could have returned a total of 1,723 collocates
within my search parameters. In reality, ring returned only 11% of this potential
maximum – suggesting Tolkien exercised a high degree of caution when it came
to the breadth of the One Ring’s lexicalisation (a finding synergising nicely with
Rosebury’s argument on Tolkien’s style as “distinguished by an unobtrusive
economy and precision”,114 or Ivar Agøy’s on Tolkien’s deliberately thin
descriptiveness115 – whilst also adding value to both claims, by substantiating
them through the use of sophisticated computer-led readings).
And yet, even within this restricted network of meanings, only 23 of the
Ring’s 197 total collocates appeared in all four Third Age texts. Again, this
represents 11% of all returned collocates – suggesting that, despite his lexical
restraint whenever using ring, Tolkien tended to target different parts of his
Ring’s vocabulary to different periods of his Third Age story. Moreover, of the
23 words which reappeared alongside ring in every Third Age text, 22 are so-
called stopwords. That is, words like a, and, but, is, at, it, etc. – highly common
words in English language, which don’t meaningfully affect the Ring’s semantic
profile. Only one meaningful word appeared alongside ring in every Third Age
text: great. The notion of the Ring’s greatness was the only part of its meaning-
in-use that remained continuous across the Third Age. Otherwise, each Third
Age text operationalised ring according to different qualities.
I visualise the variation in Tolkien’s vocabulary for ring in the network
graph below.116
113
Returning collocates within a span of 5 words either side of the word ring, and with a
minimum frequency of 3 collocaTons across each text.
114
Rosebury, p. 22.
115
Ivar Agøy.
116
Code for visualising MI2 adapted from Guillaume Desagulier, ‘PloNng CollocaTon
Networks with R and Ggraph’, Around the World, 2020.

What the graph shows is all words appearing alongside ring, excluding most
stopwords, in each Third Age text. The four texts are labelled as the central
nodes from which the collocates emerge. And the collocates are arrayed in a
network graph. Where collocates are connected to only one text node (i.e., by
only one line), this reveals those collocates as unique to that single text. Where
collocates are connected to more than one text node (by more than one line),
this reveals the word as appearing alongside ring in multiple texts. This graph’s
significance is in visualising the extent of semantic variance in the Ring’s every
Third Age usage. Few of the collocates in the graph appeared in more than one
text, as represented by multiple nodal lines. And a minimal number sit in the
middle of the network, with relationships to every text node. Indeed, most
collocates are unique to each text – appearing alongside ring in only one of
Tolkien’s Third Age texts. This initial finding serves as an entry-point to my
collocation analysis. Whenever bringing the One Ring into his language,
Tolkien did so according to different meanings. Thus, the One Ring did indeed
undergo a journey across the Third Age – with its semantic qualities changing
in each text.
Once more, this finding precipitates a follow-up: what was the character
of this unfolding journey? According to what meanings did Tolkien
operationalise ring, in each Third Age text? To answer these questions, I turn to
analysis of collocation relationships in each text. Here, I limit my search to
collocates returning a minimum MI2 of 9.00. This means focussing on strong
collocational relationships – relationships which, like royal and family in British
English, meaningfully affect the quality of each word within the configuration.
The Hobbit: a magic ring
I begin by assessing collocation returns for the word ring in The Hobbit. The
table below provides a list of all ring collocates returning an MI2 of 9.00 or
more in The Hobbit (excluding stopwords and bell, as noted earlier).
Table three: words collocating with ring in The Hobbit (MI2>9.00)
Collocate MI2 Frequency of

collocation
slipped 12.13 8
invisible 12.10 5
magic 11.12 7
pocket 9.72 3
his 9.68 23
golden 9.17 3
These collocation returns exemplify Tolkien’s operationalisation of ring in The

Hobbit according to a playful, almost childish quality. The Ring’s most
significant collocation relationship in The Hobbit is with an inoffensive verb:
slipped. This collocation emerges in a playful sense, as in Bilbo’s light-hearted
and innocent usage of his ring across the story:
Hobbits are clever at quietness, especially in woods, as I have already

told you; also Bilbo had slipped on his ring before he started. That is
why the spiders neither saw nor heard him coming…117
117
Chapter 8: Flies and Spiders.

Or
“Let’s have a light!” he said. “I am here, if you want me!” and he slipped
off his ring, and popped from behind a rock.118
These usages situate the Ring as closer to a toy than a weapon or burden –
something to be slipped on and off light-heartedly, rather than a device to be
feared or a weight to be borne.
The same light-hearted quality applies to the Ring’s second, third, and
sixth most-significant collocation relationships in The Hobbit: those with
invisible, magic, and golden. These collocations identify the Ring as a device of
wonder or child-like pleasure. Take invisible, for instance: Tolkien’s consistent
use of this collocate alongside ring framed the latter as a source of exciting and
wonderful possibilities (especially, in connection with Bilbo’s naughty-
schoolboy “burglar” role, within Thorin’s company). Hence, Bilbo’s thoughts
upon first discovering the Ring’s powers:
His head was in a whirl of hope and wonder. It seemed that the ring he
had was a magic ring: it made you invisible! He had heard of such things,
of course, in old old tales; but it was hard to believe that he really had
found one, by accident…119
Of a similar character is the Ring’s collocation with magic (also manifested in

this excerpt). This usage reflects the high keyness enjoyed by words like wizard
or elves early in the Third Age, as noted above. Here, the Ring appears as an
artefact to make children go oooh, rather than one to make grown-ups go arghhh
– an artefact Bilbo is lucky (“I am Ringwinner and Luckwearer”120) to have
come across, per the following passage:
Knowing the truth about the vanishing did not lessen their opinion of
Bilbo at all; for they saw that he had some wits, as well as luck and a
magic ring – and all three are very useful possessions…121
The magic pertaining to the Ring in The Hobbit is a childish magic of quick
escapes and casual usages – rather than the dark and shadowy sorcery we come
to know later in the Third Age. As, for example, in Tolkien’s description of
Bilbo’s carefree attitude vis-à-vis the Ring, upon his return to the Shire:
His gold and silver was largely spent in presents, both useful and
extravagant – which to a certain extent accounts for the affection of his
118
Chapter 16: A Thief in the Night.
119
Chapter 5: Riddles in the Dark.
120
Chapter 12: Inside InformaTon.
121
nephews and his nieces. His magic ring he kept a great secret, for he
chiefly used it when unpleasant callers came…122
In short, when lexicalising ring in The Hobbit, Tolkien did so according to an

innocent, magical, and undaunting meaning profile. Bilbo’s ring was not
something to be feared, but something to be used in pursuit of Tookish
shenanigans. That being said, there is one further collocation worth mentioning
– which adds a layer of meaning, and which will reappear in subsequent texts.
This is the collocation between ring and his. At various points, Bilbo thinks of
“the secret of his ring”,123 or “blesse[s] the luck of his ring”.124 These usages are
significant. They cast the Ring as a possessive item: a thing to be owned. This
possessive quality foreshadows the darkness that will later envelop the Ring.
The Ring’s possessiveness, even in early Third Age writings, serves as a hook
for later parts of its journey.
For the most part, however, the Ring in The Hobbit is a magical,
wondrous, and innocent device – one exemplifying Bilbo’s good luck within the
narrative, and one enabling him to pursue his adventures without the burden of
fear and care.
The Fellowship of the Ring: a powerful ring
In early drafts, Tolkien planned to title the second instalment of his Third Age
writings “The Magic Ring”.125 Such a title would have carried the Ring’s initial
meaning profile into its subsequent narrative development. However, Tolkien
changed his mind early on in his plans, opting instead for the title “The Lord of
the Rings”. This shift in word use, from magic to something more sinister, was
also reflected in the meaning network by which Tolkien operationalised ring in
post-Hobbit writings.
In The Fellowship, Tolkien operationalised the One Ring as a powerful
device: one its wearer could wield to acquire dominance over others. The table
below lists all collocations for ring returning an MI2 above 9.00 in The
Fellowship.
122
Chapter 19: The Last Stage.
123
124
Chapter 12: Inside InformaTon.
125
See hbps://twiber.com/TolkienWonder/status/1586448691224752128.

Table four: words collocating with ring in The Fellowship (MI2>9.00)

collocation
the 12.51 332
ruling 12.22 7
finger 11.21 9
rule 11.19 8
one 11.02 38
to 10.46 95
chain 10.45 6
isildur 10.24 8
pocket 10.11 6
itself 10.05 7
power 9.93 13
of 9.64 80
his 9.62 45
find 9.59 14
bearer 9.55 3
bring 9.03 5
Given the Ring appeared more in The Fellowship than in other texts, it also
returns more collocates. For brevity, I will focus on a few of these.
The most important collocations for ring in The Fellowship are those
connoting power or rule. Ruling, rule, power, and bring (as in “One Ring to
bring them all”) all appear with high MI2s in this text. These words shift the
Ring’s meaning profile in significant ways: from a lucky device of careless
magic to a more ominous tool for control/conquest. Hence, the co-occurrence
of ring and ruling in Fellowship book 2, chapter 4 (A Journey in the Dark):
The Ringwraiths are deadly enemies, but they are only shadows yet of
the power and terror they would possess if the Ruling Ring was on their
master's hand again…
Or, that between ring and rule in book 1, chapter 2 (The Shadow of the Past):
This is the Master-ring, the One Ring to rule them all. This is the One
Ring that he lost many ages ago, to the great weakening of his power…
And, finally, that between ring and power in the same chapter:
Clearly the ring had an unwholesome power that set to work on its
keeper at once. That was the first real warning I had that all was not
well…
A master-ring to rule them all, forged by deadly enemies and wielding an

unwholesome power… We are a long way from the magic ring of Bilbo’s youth,
here. Tolkien has developed the Ring’s significance: from childish to adult
themology, from an opportunity with Bilbo to a dangerous responsibility for
Frodo.
The collocations noted in my table also point to a complementary shift
in the level of agency Tolkien gave his Ring during The Fellowship. In this
second of Tolkien’s four Third Age texts, the Ring came to collocate with the
definite article, the, with very high MI2 of 12.51. This coinage brought the Ring
into being as a proper noun – not a magic ring, one among many; but the Ring,
the centre-point of all unseen power in Middle-earth. Thus, also, the equally-
high score for the collocation ring-one. In The Fellowship, the Ring became
“The One Ring” – as in Frodo’s offer to Galadriel (“I will give you the One
Ring, if you ask for it”126); or his question to her upon her refusal (“I am
permitted to wear the One Ring: why cannot I see all the others…?”127). The
Ring became a singularity in The Fellowship: with its meaning profile being
clarified through addition of the word one (MI2 11.02). It also developed an
agency which it hadn’t enjoyed in The Hobbit, per ring’s novel collocation with
itself. Gandalf tells Frodo in Bag End that
It was not Gollum… but the Ring itself that decided things. The Ring left
him…128
Equally, he tells others in Rivendell that
in my despair I thought again of a test that might make the finding of

Gollum unneeded. The ring itself might tell if it were the One…129
Finally, in Bree, Frodo
wondered if the Ring itself had not played him a trick; perhaps it had
tried to reveal itself in response to some wish or command that was felt
in the room…130
This novel collocation between ring and itself (a strong one, returning MI2 of
10.05) endowed the Ring with new qualities: the ability to act on its own terms,
rather than merely being used by others. This contrasts with the possessive use
of ring in The Hobbit (collocating with his). Whilst ring continues to collocate
126
Book 2, chapter 7: The Mirror of Galadriel.
127
Same chapter.
128
Book 2, chapter 2: The Shadow of the Past.
129
Book 2, chapter 2: The Council of Elrond.
130
Book 1, chapter 9: At the Sign of the Prancing Pony.

with his in The Fellowship, it is not with any increase in MI2. And, in fact, in
later texts, we’ll see that this possessive character of the Ring falls out of usage
entirely. The Ring came into its own across the Third Age. It developed agency
and singularity (the/one). It also developed a relationship with vocabularies
communicating dominance through exercise of its power. This relationship
displaced child-like connotations of Bilbo’s lucky magic ring, which had
prevailed in earlier language use.
The Two Towers/The Return of the King: a burdensome ring
The word ring returns only two collocates above MI2 9.00 in The Two Towers:
the (MI2 10.99) and of (9.17). These minimal collocations point to the Ring’s
declining relevance within the Third Age narrative, as similarly noted in my
section on keyness. Put simply, the Ring is mentioned less in The Two Towers
than in previous texts – and, as a result, appears alongside fewer words. There
is one significant thing to note about these two collocations, though. And that is
the reduced MI2 score for the collocation the-ring they reveal (meaning the
strength of the configuration “the Ring” fell by 12% between LotR’s first and
second texts). In fact, exploring collocates for the word the in The Two Towers,
we find that the definite article came to collocate more closely with a range of
other words, including tower (MI2 11.29, up from 8.80 in Fellowship); riders
(11.26, up from 10.27); ents (11.24, up from 0); or darkness (10.96, up from
10.87), than it did with ring (10.91). This trend, the weakening of the Ring’s
proper-noun-isation continued in Return of the King, where the came to
collocate more closely with city (12.71, up from 8.71 in Fellowship/9.70 in
Towers); king (12.44, up from 7.64/10.87); west (11.82, up from 11.40/11.34);
lord (11.64, up from 10.35/10.58); or captains (10.73, up from 0/6.45), than
with ring (10.67).
These growing collocations for words denoting hierarchy and
masculinity affirm earlier comments vis-à-vis the displacement of elves by men
in keyness centrality. This time, however, men displace ring in the exclusivity
of their relationship with the definite article. Men became more proper-nouned
than the One Ring, as Tolkien’s writing evolved. Where the Ring enjoyed the
status of singularity in The Fellowship, it came to compete with other proper
nouns in The Two Towers. This is all to deepen my findings from keyness
analysis: confirming that not only did the Ring become more peripheral as the
Third Age wore on, it also lost its gravitas (slipping down the hierarchy of
Middle-earth’s proper-noun beings).
In Tolkien’s final Third Age text, Return of the King, ring enjoys highly-
exclusive relationships131 with seven other words. These are listed in the table
below:
131
I.e., MI2 scores 9.00+.
Table five: words collocating with ring in Return of the King (MI2>9.00)

collocation
chain 12.06 5
wraiths 12.05 3
bearer 11.72 3
the 10.76 89
finger 10.46 3
put 10.32 7
weight 10.14 3
As these collocation returns indicate, the Ring underwent a further significant

shift in Tolkien’s language use as his Third Age narrative drew to a close. Once
more, it came to be associated with a new lexical field: the most significant
unifying feature of which was the Ring’s appearance alongside words connoting
burden. For example, the strength of the collocation between ring and chain
grew from 10.45 in The Fellowship to 12.06 in Return of the King. This
configuration appears in the struggle between Gollum and Frodo in the
Sammath Naur: “Gollum was tearing at his master, trying to get at the chain and
the Ring”132 – a usage which combines chain and ring as two features of the
same phenomenon… Or, similarly, in Sam’s reassurances in Cirith Ungol:
“I’ve kept it safe. It’s round my neck now, and a terrible burden it is,
too.” Sam fumbled for the Ring and its chain. “But I suppose you must
take it back.” Now it had come to it, Sam felt reluctant to give up the
Ring and burden his master with it again…133
Note, here, that not only are “the-Ring-and-its-chain” becoming one

phenomenon; but, also, that this combined artefact is described consistently as
a burden. Hence, the significant collocation between ring and weight in Return
of the King – a collocation returning high MI2 of 10.14, and appearing in
passages like those below:
The Ring is enough. This extra weight is killing me. It must go…134
Sam guessed that among all their pains he bore the worst, the growing
weight of the Ring, a burden on the body and a torment to his mind…135
132
Book 6, chapter 3: Mount Doom.
133
Book 6, chapter 1: The Tower of Cirith Ungol.
134
Book 6, chapter 2: The Land of Shadow.
135

He had feared that he would have barely strength to lift his master alone,
and beyond that he had expected to share in the dreadful dragging
weight of the accursed Ring…136
Evidently, and uniquely to the final Third Age text, Tolkien came to think of the
Ring as a weighty burden – one that must be endured by its bearer (ring and
bearer also collocate in Return of the King with a very high MI2 of 11.72, up
from 9.55 in The Fellowship). Again, this new quality of burdensome weight
adjusts the Ring’s nature in significant ways. Across the Third Age, the Ring
travelled from the status of a wonderful, childish plaything to that of a painful
and heavy load (via a profile of dominant singularity in The Fellowship). This
final lexical shift adds physicality to the Ring’s evolving meaning profile – no
longer an artefact which might be slipped nonchalantly onto one’s finger, but a
cross for its bearer to carry.
In my keyness analysis, we saw the gradual disenchantment of Tolkien’s
legendarium – the fall in usage of words like elves, music, and friendship; versus
a rise for men, shadow, and war. My collocation findings develop this picture:
bringing the Ring itself into Middle-earth’s disenchanting journey. Like Middle-
earth generally, Tolkien’s One Ring underwent a decline from the status of child-
like mystery, adventure, and faerie, to one of darkness, dominance, and strain –
across his Third Age writings. We can see this decline in visual form via a
filtered network graph, shown below.
136
This final visualisation filters collocates from my earlier network graph –

focussing only on those returning meaningful MI2 of 9.00 or higher. Looking
closely at the words in the graph reveals the unique meanings pertaining to the
Ring in each of its textual instantiations. Invisible, magic, slipped, golden are
all unique to The Hobbit – words capturing the Ring’s child-like, innocent
qualities at the outset of the Third Age. Many words collocate uniquely with
ring in The Fellowship (a product of its peaking centrality within this text).
Amongst these are power, rule, ruling, itself, and one – words signifying the
Ring’s unsettling power properties, its agency, and its singularity. Few words
appear alongside ring in The Two Towers (a product of its peripheralization in
Tolkien’s language use). But a handful of Ring collocates are unique to the
Return of the King – including weight, exemplifying the Ring’s growing,
burdensome physicality.
The network graph simplifies the story I’ve told regarding the Ring’s
Third Age journey within this article. This is a story of declining relevance (the
Ring’s peripheralization in language, including falling keyness and a slip down
the pecking order of proper-noun beings); as well as changing meanings (from
innocent magic to burdensome physicality, via shadowy power). These findings
bring the Ring’s Third Age development (its evolving meaning profile as
Tolkien’s writing style unfolded) into line with Middle-earth’s wider
disenchantment, per my keyness analysis (its transformation from a land of
mystery, adventure, and faerie; to one of darkness, hierarchy, and amnesia).
Conclusion
I began this article by describing Middle-earth as a world of journeys. In my use

of corpus linguistic techniques to pop the bonnet on Tolkien’s evolving
linguistic/stylistic choices across Third Age writings, I’ve found it was not only
Tolkien’s characters who undertook these journeys. But, also, the Ring itself.
The Ring was not a stable artefact in Tolkien’s unfolding imaginings. On the
contrary, it underwent a turbulent character arc of its own: suffering increasing
marginalisation from the wider narrative (reflected in declining keyness scores);
and undergoing significant transformations of quality with each Third Age
instalment (per shifting collocation relationships).
In both my keyness and collocation analyses, I’ve suggested these
ruptures in the Ring’s centrality and qualities tie into a disenchantment of
Middle-earth itself across the Third Age, as brought (artfully) into being by
Tolkien’s language use. My keyness analysis found the Ring was not the only
motif to suffer marginalisation from Tolkien’s narrative arc. Important motifs of
magic, adventure, memory, and mystery were also in decline as the Third Age
wore on – displaced by motifs of order, hierarchy, blood, and masculinity. This
displacement reflects Middle-earth’s generic Third Age disenchantment, which
is a master theme in Tolkien’s literature – encompassing the departure of elves,
the forgetting of past greatness/beauty, and the disappearance of ents/hobbits.
Whilst the Ring cannot be considered a positive phenomenon for Tolkien – the

way music and friendship unambiguously were – its declining relevance is also
part of this disenchantment. As I’ve shown in collocation analysis, not only did
the Ring slip down the pecking order of Middle-earth’s proper-noun beings
(replaced by highly-masculine items like the City or the King); it also underwent
a transformation in quality that carried it away from a profile of child-like
mystery, adventure, faerie – and towards one of darkness, power, and physical
burden.
The co-disenchantment between Middle-earth and its ubiquitous Ring is
telling. It reveals much regarding Tolkien’s changing attitude to Middle-earth,
as its stories unfolded. Based on my findings, it isn’t a stretch to conclude
Tolkien’s vision for Middle-earth fell from one of beauty, music, and magic to
one of shadow, blood, and hierarchy. For all that the forces of good won the War
of the Ring, we may ask: at what cost? Middle-earth after the Ring’s destruction
was a significantly more depressing place, per the mechanics of Tolkien’s word
choice, than it had been when Bilbo first stumbled across his golden ring.
There’s no better demonstration of this shift than the dissonance of Bilbo’s
wildly inappropriate request for the Ring, right at the end of the Third Age, in
Book 6, Chapter 6:
“Which reminds me: what’s become of my ring, Frodo, that you took
away?”
“I have lost it, Bilbo dear,” said Frodo. “I got rid of it, you know.”
“What a pity!” said Bilbo. “I should have liked to see it again…”
What makes this request so jarring is its abstraction from the Third Age journey
the Ring has been on – in its passage between Bilbo’s and Frodo’s hands, and
as traced in this article. In his senile state, Bilbo cannot recognise this journey.
He can only think of the Ring as the magical device he found in his youth – a
lucky and wondrous gift: rather than a dark and sinister burden, as in Frodo’s
experience. Bilbo’s senility (the absurdity of thinking it a pity that the Ring was
lost) throws the Ring’s journey into sharp relief. The exchange speaks to the
deliberate choices Tolkien made in evolving the Ring’s stylistic/lexical profile:
an evolvement he curated through evolving word choice as much as through
narrative techniques (keyness and collocates); and which he highlights to us, in
the end, through a comment on the follies of (Bilbo’s) old age.
The Bilbo story points to the artful way Tolkien curated his Third Age
narrative: taking his characters and universe on a journey through language –
and crystallising this journey’s parameters and breadth through a final, wry
exchange at its close Such intentionality, deliberation, thoughtfulness in
Tolkien’s techniques of authorial craft (at the level of word choice) returns us to
debates around his work’s status as “literature”, with which I opened this article.
Burton Raffel concluded his dismissal by arguing Tolkien’s works didn’t
deserve recognition as literature because language played second fiddle to
narrative structure within them – a submission of style to plot which, for Raffel,
epitomised Tolkien’s non-literary status. Per Raffel, language in LotR served
merely “a cog in some narrative machine… None of this [Tolkien’s written

word] has anything to do with what words as words can communicate; the
question of style is simply not an issue”.137 Yet, my analysis has shown that
Tolkien was indeed attentive to word and style. The words he used to lexicalise
the Ring operated not only as functions of wider plots’ unfolding. On the
contrary, Tolkien’s words were the unfolding. Whether in terms of volumes of
their use, or semantic configurations by which they appeared, the words by
which Tolkien told his story themselves played a calculated and impactful role:
shaping Middle-earth’s texture and experience, in ways that paralleled,
amplified, and advanced his mythopoesis.
In descending to the level of word, my analysis has come to significant
findings on Tolkien’s stylistic craft – findings which speak back, in turn, to
historical debates concerning his literary status. I want to make one final point
in this regard. Paying attention to the mechanics of word use, as I’ve done, can
illuminate similarly long-standing contests concerning Tolkien’s treatment of
modernity. In his foreword to LotR’s second edition, Tolkien famously claimed
a “cordial dislike” for “allegory in all its manifestations” – arguing the Third
Age had no “inner meaning or ‘message’… It is neither allegorical nor
topical”.138 This claim has long confounded critics seeking to make connections
between Tolkien’s real-world experience, and that of his legendarium.
On the other hand, in the same foreword, Tolkien also admitted no author
could “remain wholly unaffected by his experience”. And he went on to suggest
that, even though allegorical connections between the War of the Ring and the
wars Tolkien lived through were false, connections between Middle-earth’s fall
and that of his country’s “shabby” industrialisation/modernisation were more
warranted:
It has indeed some basis in experience, though… much further back

[than Tolkien’s immediate writing period]. The country in which I lived
in childhood was being shabbily destroyed before I was ten…
My corpus linguistic findings bring this dilemma between rejecting allegory and
making real world connections into alternative relief. They provide further
evidence of this tentative link between the Middle-earth experience and
Tolkien’s dissatisfaction with industry/modernity – noting meaningful patterns
in word use across Third Age texts, which also tend towards a vision of worldly
disenchantment in the face of historical (highly-masculine) development.
Regardless of assertions to the contrary, Tolkien’s own experiences in the world
he really inhabited did indeed bleed into his conception of Middle-earth – per
language choices he made, in crafting his legendarium’s diachronic evolution.
Noting these resonances between real and imagined world’s does not
lessen Middle-earth’s imaginative power, nor its timelessness (its non-limitation
137
Raffel, p. 227.
138
Tolkien, The Lord of the Rings: The Fellowship of the Ring.

to any one period or place). Equally, understanding Middle-earth’s connection

to Earth does not diminish Tolkien’s creativity and linguistic brilliance. Rather,
the relationship between worlds and experiences, which my corpus linguistic
findings illuminate, enhances these features. They help us read significant parts
of Tolkien’s legendarium. These include Middle-earth’s darker parts – per
excerpts from The Two Towers and The Return of the King:
For a while they sat without speaking under the shadow of a mound of
slag; but foul fumes leaked out of it, catching their throats and choking
them. Gollum was the first to get up. Spluttering and cursing he rose,
and without a word or a glance at the hobbits he crawled away on all
fours. Frodo and Sam crawled after him until they came to a wide almost
circular pit, high-banked upon the west. It was cold and dead, and a foul
sump of oily many-coloured ooze lay at its bottom. In this evil hole they
cowered, hoping in its shadow to escape the attention of the Eye…139
North amid their noisome pits lay the first of the great heaps and hills of
slag and broken rock and blasted earth, the vomit of the maggot-folk of
Mordor…140
We can understand these depictions of lands surrounding Mordor more fully by

situating them in connection to the country in which Tolkien grew up (many
elements of which remain active in our lives). But this resonance between real
and imagined worlds also helps us appreciate Middle-earth’s (and by
implication, our Earth’s) enchanting elements. These are elements we can see,
hear, sense in our own lives as much as on Tolkien’s page – and, therefore, which
we might learn to treat with more respect, as Tolkien sought his characters to
do. As Tolkien put it of his love for hobbits in The Lord of the Rings’ prologue:
they love peace and quiet and good tilled earth: a well-ordered and well-
farmed countryside was their favourite haunt. They do not and did not
understand or like machines more complicated than a forge-bellows, a
water-mill, or a hand-loom, though they were skilful with tools…
Peace, quiet, good tilled earth… These are things we know and can understand
– and, therefore, that we can learn to value more keenly thanks to the value to
which Tolkien ascribed them in his legendarium (and the darkness he associated
with their loss).
My use of corpus linguistics to assess Tolkien’s Third Age writings
clarifies these connections. Employing computational readings to shed
alternative, rhizomatic light on patterns of word use, I’ve uncovered significant
trends towards Tolkien’s world’s linguistic disenchantment: in the specific
139
Book 4, chapter 2: The Passage of the Marshes.
140
Book 5, chapter 10: The Black Gate Opens.
journey taken by the One Ring; and in Middle-earth’s global lexicalisation. My

study reveals artful mechanics of Tolkien’s language use – counting words to
elucidate how Tolkien brought his imagined world into alignment with personal
dissatisfactions on industry/modernity, through subtle authorial choices
engendering Middle-earth’s equivalent creeping disenchantment. This finding
represents a significant contribution to both Tolkien studies and wider literary
studies. It demonstrates the power of digital humanities approaches: helping
analysts read texts anew, by illuminating their mechanics from alternative, non-
human angles. In my application of this non-human reading, I’ve found patterns
clarifying long-held debates regarding Middle-earth’s allegorical relationship to
real-world experience. This outcome substantiates my claims in favour of
corpus linguistics for literary studies: revealing these techniques’ capacity to
reframe existing interpretive stalemates, and open new lines for knowing
familiar materials.
Appendix: formulae for calculating SMP and MI2
Formula for calculating Simple Maths Parameter keyness scores:141
𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑜𝑓 𝑤𝑜𝑟𝑑 𝑖𝑛 𝑓𝑜𝑐𝑢𝑠 𝑐𝑜𝑟𝑝𝑢𝑠 + 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡

= 𝑆𝑀𝑃 𝑠𝑐𝑜𝑟𝑒
𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑜𝑓 𝑤𝑜𝑟𝑑 𝑖𝑛 𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑐𝑜𝑟𝑝𝑢𝑠 + 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
Formula for calculating MI2 collocation scores:
𝑜𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑜𝑓 𝑐𝑜𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛!

𝑙𝑜𝑔! = 𝑀𝐼2 𝑠𝑐𝑜𝑟𝑒
𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 (𝑐ℎ𝑎𝑛𝑐𝑒) 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑜𝑓 𝑐𝑜𝑙𝑙𝑜𝑐𝑎𝑡𝑖𝑜𝑛
References
Almela, Ángela, and Irina Keshabyan, ‘A New Approach to Literature: Corpus

Linguistics’, International Journal of English Studies, 12.2 (2012), i–iv
Baker, Helen, Vaclav Brezina, and Tony McEnery, ‘Ireland in British
Parliamentary Debates 1803-2005’, in Exploring Future Paths for
Historical Sociolinguistics, ed. by Tanja Säily, Arja Nurmi, Minna
Palander-Collin, and Anita Auer (Amsterdam: John Benjamins, 2017),
pp. 83–107
Baker, Helen, Tony McEnery, and Andrew Hardie, ‘A Corpus-Based
Investigation into English Representations of Turks and Ottomans in the
Early Modern Period’, in Lexical Priming: Applications and Advances,
ed. by Michael Pace-Sigge and Katie Patterson (Amsterdam: John
Benjamins, 2017), pp. 42–66 <https://doi.org/10.1075/scl.79.02bak>
Baker, Paul, ‘The BE06 Corpus of British English and Recent Language
141
Default constant of 100 in LancsBox 6.0 (the so{ware used in this analysis).

Change’, International Journal of Corpus Linguistics, 14.3 (2009), 312–

37 <https://doi.org/ijcl.14.3.02bak>
———, ‘Times May Change, But We Will Always Have Money: Diachronic
Variation in Recent British English’, Journal of English Linguistics, 39.1
(2011), 65–88 <https://doi.org/10.1177/0075424210368368>
Baker, Paul, Costas Gabrielatos, Majid Khosravinik, Michał Krzyzanowski,
Tony McEnery, and Ruth Wodak, ‘A Useful Methodological Synergy?
Combining Critical Discourse Analysis and Corpus Linguistics to
Examine Discourses of Refugees and Asylum Seekers in the UK Press’,
Discourse & Society, 19.3 (2008), 273–306
<https://doi.org/10.1177/0957926508088962>
Balossi, Giuseppina, Corpus Linguistic Approach to Literary Language and
Characterization: Virginia Woolf’s The Waves (Amsterdam: John
Benjamins, 2014)
Bolding, Sharon, ‘Review: Tolkien as a Literary Artist by Thomas Kullmann
and Dirk Siepmann’, Mythlore, 41.1 (2022), 277–84
Brezina, Vaclav, Statistics in Corpus Linguistics: A Practical Guide
(Cambridge: Cambridge University Press, 2018)
Brezina, Vaclav, Tony McEnery, and Stephen Wattam, ‘Collocations in
Context: A New Perspective on Collocation Networks’, International
Journal of Corpus Linguistics, 20.2 (2015), 139–73
<https://doi.org/10.1075/ijcl.20.2.01bre>
Brezina, Vaclav, Pierre Weill-Tessier, and Tony McEnery, ‘#LancsBox
v.6.0.0’ (Lancaster: Lancaster University, 2021)
Colebatch, Hal, Return of the Heroes: The Lord of the Rings, Star Wars, Harry
Potter and Social Conflict (Perth: Australian Institute, 1990)
Curry, Patrick, Defending Middle-Earth: Tolkien, Myth, and Modernity (New
York: Mariner Books, 2004)
Demmen, Jane, Nick Hartshorne-Evans, Elena Semino, and Rajiv
Sankaranarayanan, ‘Language Matters: Representations of “Heart
Failure” in English Discourse - a Large-Scale Linguistic Study’, Open
Heart, 9.1 (2022), e001988 <https://doi.org/10.1136/openhrt-2022-
001988>
Desagulier, Guillaume, ‘Plotting Collocation Networks with R and Ggraph’,
Around the World, 2020
Drout, Michael D. C., ‘Tolkien’s Prose Style and Its Literary and Rhetorical
Effects’, Tolkien Studies, 1 (2004), 137–63
Drout, Michael D. C., Namiko Hitotsubashi, and Rachel Scavera, ‘Tolkien’s
Creation of the Impression of Depth’, Tolkien Studies, 11 (2014), 167–
211
Drout, Michael D. C., and Hilary Wynne, ‘Tom Shippey’s J.R.R. Tolkien:
Author of the Century and a Look Back at Tolkien Criticism since 1982’,
Envoi, 9.2 (2000), 101–67
Eve, Martin Paul, Close Reading with Computers: Textual Scholarship,
Computational Formalism, and David Mitchell’s Cloud Atlas (Redwood

City: Stanford University Press, 2019)
Fairclough, Norman, Language and Power, 3rd edn (Abingdon: Routledge,
2015)
Firth, John Rupert, Papers in Linguistics, 1934-1951 (Oxford: Oxford
University Press, 1957)
Gablasova, Dana, Vaclav Brezina, and Tony McEnery, ‘Collocations in
Corpus-Based Language Learning Research: Identifying, Comparing, and
Interpreting the Evidence’, Language Learning, 67.S1 (2017), 155–79
<https://doi.org/10.1111/lang.12225>
Ivar Agøy, Nils, ‘Vague or Vivid? Descriptions in The Lord of the Rings’,
Tolkien Studies, 10 (2013), 49–67
Kilgarriff, Adam, ‘Simple Maths for Keywords’, in Proceedings of Corpus
Linguistics Conference CL2009 (Liverpool: University of Liverpool,
2009), pp. 1–6
Kirk, Elizabeth, ‘“I Would Rather Have Written in Elvish”: Language, Fiction
and “The Lord of the Rings”’, Novel, 5.1 (1971), 5–18
Kučera, Henry, and W. Nelson Francis, Computational Analysis of Present-
Day American English (Providence: Brown University Press, 1967)
Kullmann, Thomas, and Dirk Siepmann, Tolkien as a Literary Artist:
Exploring Rhetoric, Language and Style in The Lord of the Rings
(London: Palgrave Macmillan, 2021)
Livesey, Michael, ‘Introducing the “Conceptual Archive”: A Genealogy of
Counterterrorism in 1970s Britain’, European Journal of International
Security, 8.4 (2023), 471–92 <https://doi.org/10.1017/eis.2023.10>
Mahlberg, Michaela, Corpus Stylistics and Dickens’s Fiction (Abingdon:
Routledge, 2013)
Mautner, Gerlinde, ‘Checks and Balances: How Corpus Linguistics Can
Contribute to CDA’, in Methods of Critical Discourse Studies, ed. by
Ruth Wodak and Michael Meyer, 3rd edn (London: SAGE, 2016), pp.
154–79
———, ‘Corpora and Critical Discourse Analysis’, in Contemporary Corpus
Linguistics, ed. by Paul Baker (London: Bloomsbury, 2012), pp. 32–46
McEnery, Tony, Helen Baker, and Vaclav Brezina, ‘Slavery and Britain in the
19th Century’, in Time in Languages, Languages in Time, ed. by Anna
Čermáková, Thomas Egan, Hilde Hasselgård, and Sylvi Rørvik
(Amsterdam: John Benjamins, 2021), pp. 9–38
Mejia-Ramos, Juan, Lara Alcock, Kristen Lew, Paolo Rago, Christopher
Sangwin, and Matthew Inglis, ‘Using Corpus Linguistics to Investigate
Mathematical Explanation’, in Methodological Advances in Experimental
Philosophy, ed. by Eugen Fischer and Mark Curtis (London:
Bloomsbury, 2019), pp. 239–64
Milom, Vanessa, ‘Corpus Linguistic Analysis of the Idiolects of Gollum and
Sméagol’, Journal of Linguistics and Literature, 5.1 (2022), 1–5
Moretti, Franco, Distant Reading (London: Verso, 2013)

———, Graphs, Maps, Trees (London: Verso, 2005)

O’Halloran, Kieran, ‘Performance Stylistics: Deleuze and Guattari, Poetry and
(Corpus) Linguistics’, International Journal of English Studies, 12.2
(2012), 171–99
Raffel, Burton, ‘The Lord of the Rings as Literature’, in Tolkien and the
Critics, ed. by Neil Isaacs and Rose Zimbardo (London: University of
Notre Dame Press, 1968), pp. 218–46
Rayson, Paul, ‘Computational Tools and Methods for Corpus Compilation and
Analysis’, in The Cambridge Handbook of English Corpus Linguistics,
ed. by Douglas Biber and Randi Reppen (Cambridge: Cambridge
University Press, 2015), pp. 32–49
<https://doi.org/10.1017/CBO9781139764377.003>
Reid, Robin Anne, ‘Mythology and History: A Stylistic Analysis of The Lord
of the Rings’, Style, 43.4 (2009), 517–38
Rheindorf, Markus, Revisiting the Toolbox of Discourse Studies: New
Trajectories in Methodology, Open Data, and Visualisation (London:
Palgrave Macmillan, 2019)
Roberts, Adam, ‘The One Ring’, in Reading The Lord of the Rings: New
Writings on Tolkien’s Classic, ed. by Robert Eaglestone (London:
Continuum, 2005), pp. 59–72
Rosebury, Brian, Tolkien: A Cultural Phenomenon (Basingstoke: Palgrave
Macmillan, 2003)
Salama, Amir, ‘Ideological Collocation and the Recontexualization of
Wahhabi-Saudi Islam Post-9/11: A Synergy of Corpus Linguistics and
Critical Discourse Analysis’, Discourse & Society, 22.3 (2011), 315–42
Shippey, Tom, J.R.R. Tolkien: Author of the Century (London: HarperCollins,
2000)
Siepmann, Dirk, ‘Tolkien as a Stylist: A Corpus-Based Investigation into “The
Lord of the Rings”’, in 9th International Corpus Linguistic Conference
(University of Birmingham, 2017)
Tauber, James, Chiara Palladino, Francesco Mambrini, Ugo Truffelli, Damian
Romero, Sarah Monnier, and others, ‘Digital Tolkien Project’
<https://digitaltolkien.com/> [accessed 15 December 2023]
Tolkien, J.R.R., The Hobbit, 2nd edn (London: HarperCollins, 2020)
———, The Lord of the Rings: The Fellowship of the Ring, 2nd edn (London:
HarperCollins, 2020)
———, The Lord of the Rings: The Return of the King, 2nd edn (London:
———, The Lord of the Rings: The Two Towers, 2nd edn (London:
Underwood, Ted, Distant Horizons: Digital Evidence and Literary Change
(Chicago: University of Chicago Press, 2019)
Zhao, Hai, Yan Song, and Chunyu Kit, ‘How Large a Corpus Do We Need:
Statistical Method Versus Rule-Based Method’, in Proceedings of the
Seventh International Conference on Language Resources and

Evaluation (Valletta: European Language Resources Association, 2010),
pp. 1672–77

The Ring Cycle

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

The Ring Cycle

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Ring Cycle

Uploaded by

Copyright:

Available Formats

Journal of Tolkien Research

Volume 18 Issue 1 Article 4

The Ring Cycle: Journeying Through the Language of Tolkien’s

Follow this and additional works at: https://scholar.valpo.edu/journaloftolkienresearch

The Ring Cycle: Journeying Through the Language of Tolkien’s Third

Journeying is a master theme in J.R.R. Tolkien’s published works.1 Indeed,

Published by ValpoScholar, 2023 1

Tolkien studies: in the beginning was the word

I begin by situating this article in relation to literatures on Tolkien’s

Published by ValpoScholar, 2023 3

antagonistic schools. On the one hand, Tolkien’s “detractors”10 amongst

Published by ValpoScholar, 2023 5

Corpus linguistics: reading rhizomatically

As a method of digital humanities scholarship, corpus linguistics can advance

Published by ValpoScholar, 2023 7

can be useful in checking out impressionistic conclusions about which

Corpus linguistics can bring systematicity and multiplicity to the analysis of

Published by ValpoScholar, 2023 9

studies,64 a handful of scholars have employed corpus linguistics to gain insight

Published by ValpoScholar, 2023 11

Methodology: keyness, collocation, weaknesses

As flagged in my introduction, my analysis of Third Age texts drew on two

These questions map sympathetically onto corpus linguistics’ two core

Keyness analysis: focus corpus, reference corpus, Simple Maths Parameter

In corpus linguistics, keyness analysis represents the assessment of a text

process to clarify language use’s diachronic evolution: by comparing a corpus

Published by ValpoScholar, 2023 13

Collocation analysis: MI2

Keyness metrics illuminate the quantity of word use in a corpus. By contrast,

by which English language conceptualises time and money.83) Collocation

Published by ValpoScholar, 2023 15

research. I merely used it to acquire a baseline for defining statistical

Weaknesses: decontextualisation, chronology, pagination

Finally, a caveat on weaknesses of my study (which also illuminate generic

Avoiding statistics’ misinterpretation, as in these examples, requires the

quantitative and qualitative readings should be considered complementary –

Published by ValpoScholar, 2023 17

Keyness analysis: the disenchantment of Tolkien’s world?

Having established my study’s argument, its relationship to scholarly literatures,

Published by ValpoScholar, 2023 19

The Hobbit The The Two The Return of

king 1.19 0.18 0.71 3.45

The Hobbit The The Two The Return of

Published by ValpoScholar, 2023 21

Published by ValpoScholar, 2023 23

Collocation analysis: the disenchantment of the Ring?

Within my search parameters,113 my collocation analysis returned a list

Published by ValpoScholar, 2023 25

The Hobbit: a magic ring

Table three: words collocating with ring in The Hobbit (MI2>9.00)

Collocate MI2 Frequency of

These collocation returns exemplify Tolkien’s operationalisation of ring in The

Hobbits are clever at quietness, especially in woods, as I have already

Published by ValpoScholar, 2023 27

Of a similar character is the Ring’s collocation with magic (also manifested in

In short, when lexicalising ring in The Hobbit, Tolkien did so according to an

The Fellowship of the Ring: a powerful ring

Published by ValpoScholar, 2023 29

Table four: words collocating with ring in The Fellowship (MI2>9.00)