The Dictionary of Words in the Wild

The Dictionary of Words in the Wild [1] is an experiment
in social textuality and the perceptual dynamics of reading.
The Dictionary is a social image site where contributors can
upload pictures of words taken “in the wild” and tag them so
they are organized alphabetically as an online visual dictionary.
Currently the Dictionary has 2227 images of 3198 unique
words and 24 contributor accounts. The images uploaded
and tagged are of text, usually single words or phrases, that
appear in the everyday environment. Images uploaded include
pictures of signs, body tatoos, garbage, posters, graffi ti, labels,
church displays, gravestones, plastic bags, clothing, art, labels,
and other sights. The site is structured with an application
programming interface to encourage unanticipated uses.
In this paper we will,
• Give a tour through the online social site and its API,
• Discuss current scholarship on public textuality and the
perceptual dynamics of reading,
• Refl ect on the Dictionary in these contexts, and
• Conclude with speculations on its possible contributions
to our understanding of textuality and reading.
Outline of the Dictionary
The following is a narrative of the demonstration part of the
presentation.
The Dictionary was developed with Ruby on Rails by Andrew
MacDonald with support from the TAPoR project under the
direction of Geoffrey Rockwell. Users can get a free account
in order to start uploading images. (We discovered once the
project was visible for a few months that spambots were
automatically creating accounts so we have introduced a
CAPTCHAlike graphical challenge-response feature to weed
out false accounts.) When you upload a picture you are given the opportunity
to crop it and are prompted to provide a list of words that
appear in the text. You can also provide a brief description or
discussion of the word image.
Once uploaded the image is fi led in a database and the interface
allows you to access the images in different ways:
• You can click on a letter and browse images with words
starting with that letter. An image with “On” and “Off” will
be fi led twice, though at present the label is the fi rst word
tagged.
• You can search for a word and see the images that have
been tagged with that word.
• You can type in a phrase and get a sequence of images.
• You can see what comments have been left by others on
your images and respond to them.
The API allows the user to create other processes or text
toys that can query the dictionary and get back an XML fi le
with the URLs for word images like:
<phrase>
<word href=”href_for_image”>Word</word>
<word>Word_with_no_image</word>
</phrase>
One of the goals of the project is to support mashups that use
the dictionary for new projects. This is where the Dictionary
is different than other such projects, many of which use Flickr
to pull images of letters or words.
Theoretical Discussion
The Dictionary is an exploratory project designed to
encourage the gathering of images of words in the wild and
to provoke thinking about our encounters with these words.
It did not start with an articulated theory of text that it set
out to illustrate, nor in the contributors’ experience does
the actual collecting of words tend to be governed by this or
that theory. Rather, in its simplicity, the Dictionary encourages
participants to document public textuality as they encounter
and perceive it.
At least in the initial phase of the project, the designers have
imposed no rules or guidelines for the collecting, editing and
tagging of words. Although it is clear that certain requirements
for entering words into the Dictionary would make subsequent
research questions easier to pursue, the designers prefer not
to impose them so as to discover what in fact participants fi nd
to be interesting and practical to record. Because recording of
words is voluntary and would seem inevitably to be limited to
a few individuals, the time and effort required must be kept to
a minimum in order to have a collection suffi ciently large to
allow the research potential of the Dictionary to emerge.
The Dictionary is meant to provoke refl ection on the actual
verbal environment in its totality, on the moment-by-moment
encounter with individual words and phrases where one fi nds
them and on the experience of reading them as each reading
unfolds.
Collecting of verbal images for the Dictionary presents
collectors with a way of defamiliarizing familiar environments.
Conventional techniques for framing a photograph and digital
tools for cropping it give them limited but surprisingly powerful
means of recording defamiliarized sights. Additional means are
provided by a commentary feature, but the amount of time
required to compose this commentary tends to discourage
much of it. Theoretical refl ection on the encounter with words
in the wild would seem to require participation in the project
to be adequate to the data. For this reason collecting is not a
task to be done separately from theorizing. Whatever theory
is to emerge will come from participant observation.
In some cases, we have found, what appears interesting to
record is relatively static, requiring little compositional care or
subsequent cropping. In many cases, however, the experience
of reading is a dynamic event, as when part of a verbal message,
not necessarily fi rst in normal reading order, is perceived fi rst,
then submerged into the overall syntax of its verbal and/or
visual environment. In other cases, the experience may include
a specifi c detail of the environment in which the text of interest
is embedded but not be informed signifi cantly by other details.
Occasionally one has suffi cient time to frame a photograph to
capture the experience adequately, but most often photographs
must be taken quickly, as when unwelcome attention would be
drawn to the act of photography or the situation otherwise
advises stealth (an inscription on a t-shirt, for example). The
Dictionary, one might say, is a record of psycholinguistic events
as much or more than simply of environmental data.
In more theoretical terms, is the project aims to study how
language acts as a semiotic system materially placed in the real
word. In order to interpret this multidimensional, “semiotic”
role of language, our analysis focuses on how dictionary users
perceive different signs and attribute meanings to words by
referring to these signs. We will argue that through this kind
of visual dictionary contributors can interact and play with
language by using visual artifacts (photos, images, graffi ti etc)
to express and defi ne the meanings of words. We have strong
theoretical reasons for regarding text as co-created by the
reader in interaction with the verbal signs of the document
being read. The Dictionary gives the reader of words in the
wild a means of implementing the act of reading as co-creative,
but with a signifi cant difference from those acts that have
previously been the focus of theoretical work. The collector
of wild words, like the reader of literature, is obviously not so
much a viewer as a producer of meaning, but unlike the literary
reader, the collector is operating in a textual fi eld whose realworld
context is highly unlikely ever to be otherwise recorded.
It so obviously goes without saying that it also goes by and
vanishes without ever being studied. The title of the project suggests a distinction between text in
two different places: the kind at home and the kind in the wild.
But the collector’s gaze rapidly suggests that the distinction
is only in part one of place. Text at home can also be ‘wild’
if it can be defamiliarized, e.g. the title of a book on its spine
taken in poor light conditions inside a house. The wildness of
words, is then ‘in the eye of the beholder’, though the domestic
environment is usually so well regulated that opportunities
for perceiving wildness are far more limited than in relatively
undomesticated environments. Thus such opportunities
tend to occur far less frequently in well-kept or wealthy
neighbourhoods than in poorly kept ones, where rubbish is
frequently encountered and advertising of all kinds is evident.
This suggests ironically that poorer neighbourhoods are in
respect of the sheer amount of reading more rather than less
literate. But the correlation between wealth and verbosity is
not so straightforward. Airport lounges, for example, are rich
in examples of wild words. What would seem to matter in this
correlation is the acquisitive desire of the population: those
who have tend not to read for more.
Such theorizing, however, is clearly at a quite preliminary
stage. The project calls for a more systematic ethnography
of textuality and its everyday occurence. Insofar as it can
be conceptualized, the long-term goal of the project can be
put as a question: would it be possible to develop a panoptic
topology of the appearance of the legible in everyday life, if
even just for one person?
Similar Work
The Dictionary is one of a number of projects that use the
internet to share images of textuality. For example, Typography
Kicks Ass: Flickr Bold Italic [2] is Flash toy that displays messages
left by people using letters from Flickr. The London Evening
Standard Headline Generator [3] from thesurrealist.co.uk
generates headlines from a Flickr set of images of headlines.
IllegalSigns.ca [4] tracks illegal billboards in Toronto and has
a Clickable Illegal Signs Map [5] that uses Google Maps. On
Flickr one can fi nd sets like Its Only Words [6] of images of
texts.
What all these projects have in common is the photographic
gaze that captures words or phrases in a context whether
for aesthetic purposes or advocacy purposes. The Dictionary
is no different, it is meant to provoke refl ection on the wild
context of text as it is encountered on the street.
The Future
The success of the project lies in how the participants
push the simple assumptions encoded in the structure. The
project would have failed had no one contributed, but with
contributions come exceptions to every design choice. The
types of text contributors want to collect and formally tag has
led to the specifi cation of a series of improvements that are
being implemented with the support of TAPoR and SSHRC.
The paper will conclude with some images and the future
directions they have provoked:
• We need to parse phrases so that we remove punctuation.
For example, “faith,” won’t fi nd the image for “faith”.
• We need to allow implicit words to be entered with
parentheses where the word doesn’t appear, but is
implicit. An example would be http://tapor1-dev.mcmaster.
ca/~dictwordwild/show/694 which is fi led under “Average”
even though the word doesn’t appear.
• We need to allow short phrasal verbs and compounds to
be entered with quotation marks so they are fi led as one
item. An example would be “come up” or “happy days”.
• We need to allow images of longer passages to identifi ed
as “Sentences in the Sticks”, “Phrases in the Fields” or
“Paragraphs in the Pastures”. These would not be fi led
under individual words, but the full text could be searched.
• We need to allow people to control capitalization so that,
for example, “ER” (which stands for “Emergency Room”) is
not rendered as “Er”.
• We need to let people add tags that are not words so
images can be sorted according to categories like “Graffi ti”
or “Billboard”.
Links
1. Dictionary of Worlds in the Wild: http://tapor1-dev.
mcmaster.ca/~dictwordwild/
2. Typography Kicks Ass: http://www.typographykicksass.com/
3. The London Evening Standard Headline Generator: http://
thesurrealist.co.uk/standard.php
4. IllegalSigns.ca: http://illegalsigns.ca/
5. Clickable Illegal Signs Map: http://illegalsigns.ca/?page_id=9
6. Its Only Words: http://www.fl ickr.com/photos/red_devil/
sets/72157594359355250/ Bibliography
Davis, H. and Walton, P. (1983): Language, Image, Media,
Blackwell, Oxford.
Kress, G. and Van Leeuwen, T. (1996): Reading images: the
grammar of visual design, Routledge, London.
Kress, G. and Van Leeuwen, T. (2001): Multimodal Discourse:
the modes and media of contemporary communication, Arnold,
London.
McCarty, Willard. 2008 (forthcoming). “Beyond the word:
Modelling literary context”. Special issue of Text Technology,
ed. Lisa Charlong, Alan Burke and Brad Nickerson.
McGann, Jerome. 2004. “Marking Texts of Many Dimensions.”
In Susan Schreibman, Ray Siemens and John Unsworth, eds.
A Companion to Digital Humanities. Oxford: Blackwell. www.
digitalhumanities.org/companion/, 16. (12/10/07). 198-217.
Scollon, R. and Scollon, S. (2003): Discourses in Places: Language
in the material world, Routledge, London.
Van Leeuwen, T. (2005): Introducing social semiotics, Routledge,
London.