Lexicography 2.

0: Reimagining Dictionaries for

the Digital Age

Ben Zimmer

O n November 4, 2012, Macmillan Education announced

that it would cease publishing print dictionaries and
that Macmillan Dictionaries would henceforth live only online. The
announcement was hardly a sorrowful one: an upbeat press release
framed the move as “a cause for celebration,” and an accompanying
video sunnily declared, “The Macmillan Dictionary is going places.”
Editor-in-chief Michael Rundell contrasted Macmillan’s decision with
that of an equivalent move in the world of magazine journalism, the final
print issue of Newsweek (before it was revived more than a year later).
“Newsweek’s announcement was tinged with regret,” Rundell wrote,
quoting Newsweek’s press release: “‘Exiting print is a difficult moment for
all of us.’ … But at Macmillan, we take the opposite view: exiting print is
a moment of liberation, because at last our dictionaries have found their
ideal medium” (Rundell 2012a).
The reaction was mixed on lexicography mailing lists. “What a sad
day,” despaired Dan Pratt, a retired Defense Department mathematician
and linguistics PhD whose personal mission has been an overhaul of the
official dictionary used in Scrabble tournaments (Fatsis 2011). “When
looking up anything in a print dictionary, you generally stumble across
all sorts of delightful material you never would have known to look
for,” Pratt wrote. “With an electronic dictionary, generally speaking,
what you search is what you get, and nothing beyond.” Adam Kilgarriff,
whose company, Lexical Computing Ltd., develops corpus tools for
dictionary publishers, countered Pratt’s perspective, calling it “a day of
liberation from the straitjacket of print” (Kilgarriff 2012). In their talk

Dictionaries: Journal of the Dictionary Society of North America 35 (2014), 275–286

276 Ben Zimmer

of “liberation,” Kilgarriff and Rundell were echoing their like-minded

colleague Sue Atkins, who as early as 1996 said of electronic dictionaries,
“At last we are liberated from the straitjacket of the printed page and
alphabetical order” (Atkins 1996, 516).
Whether one views such developments as positive, negative, or a
mixture of both, there is a growing consensus among lexical reference
publishers that the move away from print is an inevitable one. The editors
of the Oxford English Dictionary have yet to make an announcement as
drastic as Macmillan’s, but even the suggestion that the OED might no
longer publish a print edition has caused consternation in both the
British and American press in recent years (Heffernan 2008, Mount
2010, Flanagan 2014). Though the OED’s official line is that “no decision
has yet been made” about the ultimate format of the edition currently
in revision, it is hard to imagine circumstances in which the dwindling
market for print dictionaries could support the publication of the third
edition (in 40 or so volumes) when the work is finally completed a
decade or two hence.
What is gained and what is lost in the shift from page to pixel?
Here I argue that the digital transformation of dictionaries and
thesauruses provides fresh opportunities for lexicographers to engage
with generations coming of age in the electronic era. I will survey some
of the latest advances in the field, with an eye to making electronically
based dictionaries and thesauruses maximally useful to a twenty-first-
century readership.
While there are many reasons to be hopeful in the new digital
landscape for lexical reference, such optimism needs to be tempered
by an understanding of how electronic formats for dictionaries and
thesauruses are still a work in progress, with many growing pains along
the way. Rundell, the Macmillan editor, provided a more even-handed
perspective in his contribution to the 2012 edited volume, Electronic
Lexicography. He observes that for advanced learner’s dictionaries (ALDs)
of the type that have been published by Macmillan, users tend to be in
the age range of 17–24 and as such are “digital natives” who “routinely
go to the Web for information of any kind, and generally expect to
get it for nothing” (Rundell 2012b). This chimes with an earlier study
by Muffy Siegel (2007) on the use of online dictionaries by American
undergraduates.
Writing before Macmillan had made its decision to abandon the
print market, Rundell allowed that “for the time being there remains
a residual market for ALDs in print form, mainly in places with poor
connectivity.” But over the long haul, he wrote, the digital format “looks
 Lexicography 2.0: Reimagining Dictionaries for the Digital Age 277

like the only serious contender.” Candidly, Rundell stated that “for
dictionary publishers, these developments are not altogether welcome: a
reliable revenue stream is drying up, and an alternative business model
has not yet emerged.”
Leaving aside for a moment the business concerns that are
afflicting the entire publishing industry, I would like to focus on exactly
how the move to digital can become a “liberating” moment for the
creators of dictionaries and thesauruses, and how users stand to benefit
from the electronic experience. In her introduction to the Electronic
Lexicography volume, Sylviane Granger lays out six areas of innovation that
lexicographers are currently exploring in online reference: hybridization,
more and better data, efficiency of access, customization, corpus integration, and
user input (Granger 2012). Let us consider each of these innovations in
turn, and how online resources are taking advantage of them.

Fusing Dictionaries with Other Resources

What Granger terms hybridization refers to the combination of different

types of reference resources, such as a dictionary and a thesaurus, that
is generally easier to accomplish online. Many dictionary publishers,
including Merriam-Webster, Oxford, Collins, Macmillan, Cambridge,
Wordnik, and Dictionary.com, have unified their dictionary and
thesaurus resources in online interfaces, so that a user arriving at the
dictionary entry for happy can easily click through to a list of synonyms
and other related words.
While such hybridization has generally been limited to the
synchronic offerings of modern dictionaries, the integration of the
Historical Thesaurus of the Oxford English Dictionary with the online
OED demonstrates that hybridization can work diachronically as well.
HTOED was conceived in 1965 at the University of Glasgow as a project
that would index all the words in the OED, organizing them by their
meanings and by their first known date of use. In 2009, at long last,
HTOED was published in two massive volumes, the first providing a
chronological listing of words in different conceptual classes and the
second providing an index to find particular meanings of words in the
book’s elaborate Roget-style hierarchy, from the abstract to the specific.
While it is easy to get lost in its pages, HTOED clearly needed an
online home to maximize its practicality for both casual and scholarly
readers. Thus it was incorporated into the online OED in 2010, and
there it truly thrives. Because the categories of words are presented
chronologically, one can quickly see, for instance, how 149 terms for a
278 Ben Zimmer

“contemptible person” extend from worm and wretch in Old English to

late-twentieth-century slang offerings like scuzzbag and sleazeball. For a
writer, searching for just the right word can turn into an adventure in
historical verisimilitude. A novelist or playwright seeking epithets for
dialog set in the early seventeenth century can zero in on such terms
as viliaco (1600), snotty-nose (1604), sprat (1605), wormling (1605), and
shag-rag (1611) (Zimmer 2012a).
In my own recent work at Thinkmap, Inc., I have been involved
in integrating the Vocabulary.com Dictionary with Thinkmap’s Visual
Thesaurus; from any word page in the dictionary, the user can view
an interactive word map encouraging the exploration of words and
meanings in a dynamic interface. But Vocabulary.com also illustrates a
different kind of hybridization by incorporating its dictionary into an
adaptive learning program. Thousands of words in the Vocabulary.com
Dictionary have been made learnable: by clicking on the “learn” button,
the word is immediately added to the queue of words that the user is
learning through Vocabulary.com’s gameplay, the “Challenge.” For each
learnable word, there are on average ten multiple-choice questions that
have been developed to cover the word’s different meanings, using a
variety of different question types. Thus, a word look-up need not be a
transient moment, but can instead open a door to discovery, introducing
the word to the user’s permanent vocabulary.
Conversely, dictionary resources are built into the Vocabulary.com
Challenge: for instance, if a user incorrectly answers a question, a helpful
explanation of the word is immediately presented, and the rest of the
word’s page in the Vocabulary.com Dictionary becomes available for a
deeper dive into semantics, usage, spelling, and pronunciation. This
two-way integration ensures that game-based learning and dictionary
resources work together in a mutually enriching fashion.

Presenting the Most Useful Lexical Data

Compared to their print counterparts, online dictionaries promise

to deliver more and better data. A reasonably sized print dictionary
or thesaurus may be able to squeeze in a few example sentences per
entry, but the limitations of page extent are no longer a concern in the
online space. Breaking the “straitjacket of print” requires a thorough
rethinking of what features might be most useful for a user, when an
entire web page (or even a series of pages) is available for explicating a
particular lexical item. Sites like Dictionary.com and Wordnik aggregate
 Lexicography 2.0: Reimagining Dictionaries for the Digital Age 279

dictionary data from multiple licensed sources, with Wordnik adding not
only corpus examples but real-time results from Twitter (for text) and
Flickr (for images). The danger then becomes overwhelming the user
with a surfeit of information about a lexical item, which can become
even more unwieldy if advertisements are also running on a web page.
Judicious use of web design requires knowing which kinds of
language-related data (text, audio, or visual) will be of most use and
how best to feature them in an online environment. This will depend
largely on the goals of the dictionary or thesaurus in question. Lexical
reference sites targeting English language learners, for instance, may
make use of simpler definitions and information about how words are
used in collocations.
By way of example, word pages in the Vocabulary.com Dictionary
feature more conversational explanations of words in addition to typical
dictionary definitions supplied by WordNet, a lexical database developed at
Princeton University. This more discursive approach to defining words by
means of “full-sentence definitions” has been explored in the past primarily
in dictionaries designed for learners of English as a foreign language
(Atkins and Rundell 2008, 441). But the value of explaining words in an
engaging style is just as important for native speakers of English who would
not be able to appreciate the full range of how a word is typically used from
a traditional dictionary definition. Researchers in how students engage
with dictionaries have long called for “definitions that characterize a word’s
use and are phrased in language that makes meaning easily accessible to
young learners” (McKeown 1991, 154). The Vocabulary.com Dictionary
takes exactly this approach to elucidate vocabulary words with the goal of
ensuring that the explanations will be memorable.
Consider the word quadruped, which a standard dictionary would
define simply as “an animal with four feet.” In the Vocabulary.com Dictionary,
the word page for quadruped leads with this light-hearted explanation:
A squirrel, a zebra, a deer, a wolf, and a grizzly bear meet in a
field. Yes, a disaster in the making, but also a bunch of quadrupeds
— animals that walk on four feet.

The second part of the explanation goes into more detail,

providing morphological and etymological background to help the
meaning stick in the learner’s mind, while still maintaining an easy-
going, discursive tone:
Cut quadruped in half and it makes sense: quadru means four, like
when a woman births four babies they are called quadruplets. And
280 Ben Zimmer

-ped is for the feet: think of centipedes and millipedes, insects that
have so many feet it’s disturbing. A human is a biped because
they walk on two feet. If you meet a human with four feet, you
could call him a quadruped. You could also call the circus and let
them know their quadruped is loose.

Along with these lexical explanations, Vocabulary.com word pages

also include audio pronunciations, “word families” of inflectionally and
derivationally related forms (indicating relative corpus frequency for
each form), and corpus example sentences arranged by genre and topic.
Additionally, if a word is part of the user’s Vocabulary.com learning
program, the dictionary page can graphically indicate progress toward
mastery of that word.
The flexible nature of web design also means that the selection
and arrangement of such features can continue to be tinkered with in
future iterations, a kind of adaptability that print reference works can
only take advantage of in editions typically scheduled several years apart
—and even then an institutional conservatism of function and design
rules the day. As Atkins (1996) observed, “change is not something that
people tend to associate with dictionaries.”

The Right Words at the Right Time

An online dictionary or thesaurus may be packed with useful features, but

it still requires efficiency of access for optimal user experience. Here web-
based reference works are still finding their footing. Cluttered web pages
full of banner ads and autoplay videos do users no favors, and often reflect
the tension between the content and marketing demands of an online
publisher. Ease of navigation and search are also primary concerns. Lew
(2012) rightly notes that “it is perfectly possible to produce an online
dictionary where access is more cumbersome than in a paper book.” But
when hyperlinked navigation and search functionality are deployed well,
it can in fact be much easier to home in on the desired lexical item
than it would be by paging through a printed work. Vocabulary.com, for
instance, uses autocomplete and autocorrect to provide “results as you
type” and corrects spelling errors by means of the Soundex phonetic
algorithm. Other online dictionaries also make algorithmically based
suggestions as the user types in the search box.
Large-scale lexicographical projects that have successfully
navigated the print/digital divide have developed new paths of
accessibility for users to explore the underlying lexical data. The online
 Lexicography 2.0: Reimagining Dictionaries for the Digital Age 281

OED, which provides access to revised entries as they are updated, is a

prime example of this. One can craft sophisticated search queries to
target entries according to language of origin, region of use, subject,
date of first citation, and so forth. But it is also possible to browse nearby
entries at any time, approximating the pleasure of thumbing through
the pages of the print edition, to a certain extent. The Dictionary of
American Regional English (DARE) has recently placed its voluminous data
online as well, and it too has made the transition to the digital realm
smoothly, thanks to an easy navigability among such features as regional
maps, topical indexes, and survey responses.
The geographical displays of Digital DARE remind us that the
Web excels as a visual medium, and that effective data visualization
can be achieved in any number of creative ways, using different spatial
metaphors. The Visual Thesaurus, for example, creates interactive
displays of the relationships between words and between senses of words
by means of elastic, spring-like graphs. The displayed relationships, based
on the WordNet hierarchy, include not just the usual thesaurus fare of
synonymy and antonymy, but also such semantic linkages as hyponymy
(the “type-of” relationship between a general and specific term) and
holonymy (the relationship between a whole and a part). The complex
networking between “synsets” (sets of synonyms that are considered
semantically equivalent) would be hard to visualize in any flat, textual
rendering. The Visual Thesaurus interface turns the whole semantic
network into a web of associations to explore, with navigation easily
accomplished with the click of a mouse—or, on a multi-touch screen,
with the swipe of a finger. Moving from node to node through this type
of semantic visualization, the jumps can be unexpected, allowing for the
emergence of a different kind of serendipity than the kind that a print
reference normally affords.

Personalizing the Dictionary Experience

The move from the static printed page to the dynamic web page
encourages a responsive design that is tailored to a user’s needs. This
becomes a crucial concern in creating resources for English language
learners. As Granger (2012) suggests, customization in an online dictionary
should ideally be both adaptive (automatically adjusting to user input)
and adaptable (open to the user’s own personalization).
The Vocabulary.com Dictionary is fully integrated in an adaptive
gameplay environment that quickly adjusts to the level of the user by
means of item response theory (IRT), so that the learner’s time is not
282 Ben Zimmer

wasted by vocabulary items too far above or below his or her proficiency.
IRT is the central model in computerized adaptive testing (currently
used in the US for such standardized tests as GRE and GMAT), and
it allows appropriate questions to be presented to test-takers, zeroing
in quickly to their skill level based on their successive responses. For
such an approach to work effectively, a large pool of questions must be
continually calibrated, scored according to difficulty and discrimination
(that is, how well a question discriminates between users of different
ability levels). Vocabulary.com has an ever-increasing pool of questions,
and the voluminous user data collected over millions of sessions provides
accurate models for the appropriate questions to present to a vocabulary
learner at any given time (Lohr 2013).
The Vocabulary.com Dictionary is not only integrated into an
adaptive learning system but is adaptable as well. Users can create their
own custom-made vocabulary lists to study from and then learn those
lists in the same gamified environment as the main Vocabulary.com
Challenge. This also allows teachers to focus on those vocabulary items
that are appropriate to their students’ particular needs. A learning
environment that is both adaptable and adaptive can lend itself to
vocabulary instruction in a way that transcends static flash cards and the
rote memorization of word lists.

Harnessing the Power of the Corpus

The advent of corpus integration in lexicography has transformed both

print and online reference works. The latest edition of the Oxford
American Writer’s Thesaurus demonstrates how corpus findings can be
incorporated on the printed page. “Word Toolkits” compile information
gleaned from the Oxford English Corpus in order to convey the results
of actual language data to the reader, making the information visually
engaging and transcending a dry list of synonyms. Word clouds illustrate
the most common companions of a given term, with the size of the words
indicating their relative frequency. Thus, at a glance, one can appreciate
that the adjective mighty typically modifies phenomena of the natural
world (river, wind, oak), political entities (empire, nation), and abstract
effects (power, force) (Zimmer 2012b).
But it is online where corpus-based language data can truly shine.
The creation and easy manipulation of corpora consisting of billions of
tokens (individual instances of lexical items) opens up possibilities for
“back-end” improvements in the building of dictionary and thesaurus
 Lexicography 2.0: Reimagining Dictionaries for the Digital Age 283

entries, and also for innovative “front-end” features that allow users to
plumb corpus data on their own. Rather than relying on the manual
accumulation of citation slips, lexicographers can now consult a
seemingly limitless supply of relevant text to illuminate the usage of any
given item. This embarrassment of riches presents new challenges, of
course. Corpus data is most useful when lexical items are identified by
part of speech; fortunately, automated part-of-speech taggers are making
great strides in accuracy. But this type of pre-processing must still be
accompanied by tasks that, at least thus far, only humans can adequately
perform, such as “word sense disambiguation” (WSD) to determine
which definition of a word in a dictionary or thesaurus entry corresponds
with a usage in a particular sentence culled from the corpus.
The payoff from this semi-automated work is sizable, however.
The online entries for Oxford Dictionaries, for instance, are enriched
by numerous corpus-derived example sentences broken down by
individual senses, reflecting the results of a major WSD project enlisting
freelance lexicographers in both the US and UK. A similar project has
been undertaken for Vocabulary.com, where the goal is not simply to
assign corpus sentences to appropriate word senses, but then to use such
sentences (derived largely from literary and journalistic sources) as the
basis for adaptive vocabulary instruction in a quiz format.

Incorporating the Wisdom of the Crowds

The final innovation outlined by Granger, user input, is perhaps the

most contentious one of all. While user-generated dictionary sites such
as Wiktionary and Urban Dictionary have sprung up in recent years,
the notion of “crowdsourcing” lexicographical work is not exactly new
(Zimmer 2007). In 1879, the OED’s first editor, James Murray, made an
appeal to “the English-speaking and English-reading public” to join in
the dictionary’s reading program, calling on them to scour books for
citations that could be used as historical evidence in the entries. The
call to arms was wildly successful, ultimately generating about a million
citation slips that served as the backbone of the OED. In more recent years,
editor-in-chief John Simpson renewed the call for public participation as
part of the OED’s revision project, launching a “Wordhunt” to drum up
early examples of high-profile words and phrases.
Other dictionaries have used their online presence to solicit
suggestions for new words and senses. Merriam-Webster’s Open
Dictionary has fielded nearly 20,000 contributions from users since it
284 Ben Zimmer

was begun online in 2005. The suggestions serve as inspiration for the
annual additions of new words, alerting editors to fresh formations and
shifts in meaning and usage. Collins has recently gone even further in
this direction. User suggestions are now tightly integrated into their
online dictionary, with contributors credited by name in the new entries.
And the Collins publishing schedule is greatly accelerated, with user
contributions folded into online updates about once a month. The
Collins new-words initiative, while clearly successful as a marketing
gambit, raises some questions about the shaky status of dictionaries in
the digital era. Is this kind of crowdsourcing a worthwhile endeavor for
dictionary makers, beyond providing valuable publicity for publishers
facing a tough consumer market? Or could the reliance on the wisdom
of the crowds end up diluting the authority that the leading print
dictionaries have traditionally held?
As reference publishers grapple with these questions, it is clear
that dictionaries and thesauruses need to adapt to their new online
homes and understand the opportunities that web-based interaction
presents. The lessons of search engine optimization, for instance, can be
applied to boost the online visibility of a dictionary or thesaurus, and web
traffic analytics can be studied to determine exactly how users approach
a site, down to which words are being searched for at any given moment
(Lannoy 2010). Such a fine-grained understanding of how online lexical
references build and maintain an audience will surely be a key to their
future survival. It is also vital that dictionary and thesaurus sites be
enriched with data-driven and visually stimulating features in order to
engage with “digital natives.” This means meeting the digital natives on
their own technologized terrain, fulfilling expectations of interactivity
and fluidity in sophisticated interfaces.
I believe that lexicographers, working with web and app
developers, are up to the challenge. Dictionaries and thesauruses
are being reinvented and in the process are becoming more valuable
resources than ever. This is a golden opportunity to build much better
lexical references, ones that honor the legacy of the print tradition while
taking full advantage of the benefits of online content and design. The
content can now more accurately reflect language as it is actually used,
with insights from corpus-driven methods illuminating semantic nuances
as never before. Improvements in design can present this linguistic data
more engagingly. And in so doing, lexicographers will introduce new
generations to the interwoven splendor of language.
 Lexicography 2.0: Reimagining Dictionaries for the Digital Age 285

References

Atkins, B[eryl] T. S[ue] 1996. Bilingual dictionaries: Past, present and future.
Euralex ’96 Proceedings, Part II, 515–546. Göteburg University, Department
of Swedish.
———, and Michael Rundell. 2008. The Oxford Guide to Practical Lexicography.
Oxford: Oxford University Press.
Dziemianko, Anna. 2012. On the usefulness of paper and electronic dictionaries.
In Electronic Lexicography, edited by Sylviane Granger and Magali Paquot,
319–342. Oxford: Oxford University Press.
Fatsis, Stefan. 2011. Word Freak: Heartbreak, Triumph, Genius, and Obsession in the
World of Competitive Scrabble Players. Tenth Anniversary Edition. Penguin.
Flanagan, Padraic. 2014. RIP for OED as world’s finest dictionary goes out of
print. The Telegraph, Apr. 20. <http://www.telegraph.co.uk/culture/
culturenews/10777079/RIP-for-OED-as-worlds-finest-dictionary-goes-
out-of-print.html>
Granger, Sylviane. 2012. Electronic lexicography: From challenge to opportunity.
In Electronic Lexicography, edited by Sylviane Granger and Magali Paquot,
1–11. Oxford: Oxford University Press.
Heffernan, Virginia. 2008. The medium: Lexicographical longing. New York
Times Magazine. May 11. <http://www.nytimes.com/2008/05/11/
magazine/11wwln-medium-t.html>
Kilgarriff, Adam. 2012. Re: End of print dictionaries at Macmillan. Euralex
mailing list, Nov. 6. <http://www.freelists.org/post/euralex/DSNA-RE-
End-of-print-dictionaries-at-Macmillan,2>
Lannoy, Vincent. 2010. Free online dictionaries: Why and how? In eLexicography
in the 21st Century: New Challenges, New Applications (Proceedings of eLex
2009), edited by Sylviane Granger and Magali Paquot, 173–182. Presses
Universitaires de Louvain.
Lew, Robert. 2012. How can we make electronic dictionaries more effective? In
Electronic Lexicography, edited by Sylviane Granger and Magali Paquot,
343–361. Oxford: Oxford University Press.
Lohr, Steve. 2013. A vocabulary site shows how to tailor online education.
The New York Times Bits blog, Apr. 10. <http://bits.blogs.nytimes.
com/2013/04/10/a-vocabulary-site-shows-how-to-tailor-online-
education/>
McKeown, Margaret G. 1991. Learning word meanings from definitions:
Problems and potential. In The Psychology of Word Meanings, edited by
Paula J. Schwanenflugel, 137–156. Hillsdale, NJ: Lawrence Erlbaum
Associates.
Mount, Harry. 2010. R.I.P. the OED: As the world’s greatest dictionary goes out
of print, why it spells disaster for everyone who loves words. Daily Mail.
Aug. 31. <http://www.dailymail.co.uk/news/article-1307580/R-I-P-
Oxford-English-Dictionary-spells-disaster-loves-words.html>
286 Ben Zimmer

Pratt, Dan. 2012. Re: End of print dictionaries at Macmillan. Euralex mailing list,
Nov. 5. <http://www.freelists.org/post/euralex/DSNA-RE-End-of-print-
dictionaries-at-Macmillan>
Rundell, Michael. 2012a. Stop the presses—the end of the printed
dictionary. Macmillan Dictionary Blog, Nov. 5. <http://www.
macmillandictionaryblog.com/bye-print-dictionary>
———. 2012b. The road to automated lexicography: An editor’s viewpoint. In
Electronic Lexicography, edited by Sylviane Granger and Magali Paquot,
15–30. Oxford: Oxford University Press.
Siegel, Muffy E. A. 2007. What do you do with a dictionary? A study of
undergraduate dictionary use. Dictionaries 28: 23–47.
Zimmer, Ben. 2007. Charting the digital future of dictionary research: Prospects
for online collaborative lexicography. Paper presented at the 2007
Dictionary Society of North America biennial conference, Chicago, IL.
———. 2012a. Reconsideration: The thesaurus. Lapham’s Quarterly 5, 2: 207–213.
<http://www.laphamsquarterly.org/reconsiderations/word-for-word.
php>
———. 2012b. Remodeling “the Warehouse of Roget”: How the thesaurus
is being reinvented as a writer’s tool. Introduction to Oxford American
Writer’s Thesaurus. Third edition. Compiled by Christine A. Lindberg, xi–
xvi. Oxford: Oxford University Press.