Hypertext and Deixis 
Dan Loehr 
Georgetown University 
loehrd @gusun.georgetown.edu 
Abstract 
Hypertext linking is akin to deixis 
in that each points out of the dis- 
course to the context in which the 
discourse occurs. The context of a 
hypertext document such as a 
World Wide Web page is the 
World Wide Web itself. This study 
investigates some linguistic aspects 
of hypertext, using as data Web 
pages from college students and 
commercial enterprises. I first look 
at the syntactic forms of hypertext 
anchors, finding them mostly noun 
phrases. Next, noting the intuitive 
similarities between hypertext and 
place deictics, I attempt to find a 
quantitative relationship between 
the two. I show that certain place 
deictics do occur more frequently 
inside of hypertext anchors than in 
the ordinary text of Web pages, but 
this is due to the popularity of the 
deictic here, as in click here. Ob- 
serving that hypertext style guides 
prescribe against the usage of click 
here, I hypothesize that more expe- 
neaced hypertext authors will dis- 
prefer this usage. Using the as- 
sumption that college students are 
less experienced than commercial 
Web authors, I am unable to con- 
firm this hypothesis statistically. 
Finally, I examine the various types 
ofdeixis Web authors use in hyper- 
text anch ors, finding the majority to 
be person deictics. I close by not- 
ing that the variety and creativity of 
deixis within hypertext anchors 
suggests that the relatively new 
medium of hypertext is a fully pro- 
ductive genre of natural language. 
1. Introduction 
Deixis is, in Lyons' \[1979\] simple phrase, 
"identification by pointing". Specifically, deictic 
terms point out of the current discourse to facts con- 
cerning the context of the utterance. Deictics in- 
clude words such as I, here, and now. Sentences 
with these terms can not be understood fully unless 
one knows the people, place, and time involved in 
the utterance of the sentence. As an example, imag- 
ine finding a bottle in the ocean with the note, 
"Meet me here at noon tomorrow" (from Fillmore 
\[1975\]). Unless we know when, where, and by whom 
the note was written, we are at a loss to interpret the 
sentence. 
With the evolution of hypertext, we find another type 
of deixis. Hypertext, such as is found in World Wide 
Web pages, is electronic text which "points" to other 
locations in the same or different electronic doc u- 
ments. When hypertext is selected by means of a 
mouse or keyboard, the document "pointed to" is 
retrieved and displayed on the screen. An example 
of hypertext might be: I am a student at Georeetown 
University. (The hypertext is underlined, a conve n- 
tion I will follow in this paper). If a reader of that 
sentence clicked on the words Georgetown Univer- 
sity, an electronic document pointed to by the hyper- 
text would be displayed: presumably (but not nece s- 
sarily) relating to Georgetown University. 
This, then, is also deictic. Just as ordinary deicfics 
can not be fully evaluated without leaving the utter- 
ance and looking to the outside world, neither can 
hypertext be fully evaluated without leaving the text 
and looking at its context (the World Wide Web 
itself) to see what is being pointed at. We might 
call this hypertext deixis hyperdeixis (term credited 
to Dr. Cathy Ball), referring to the fact that hypertext 
"points", just as natural language expressions such 
as I, here, and now point. It is interesting to note 
that hypertext links are often referred to as 
"pointers", while the word deixis itself is from the 
Greek word for pointing. 
It is important to make a distinction here. The dei c- 
tic aspects of hypertext need not have anything to do 
with the words used as the anchor of the hypertext 
(i.e., the words which, when selected, activate the 
link). The deictic aspect has solely to do with the 
fact that the hypertext link points elsewhere. The 
words in the anchor do not have to relate to where 
the link takes you - in fact, they do not have to be 
words at all. The anchor can be a picture, or simply 
a blank spot on a page. No linguistic analysis of the 
anchor is needed to follow the hypertext link. 
Little is known about the linguistic properties of hy- 
pertext. Although many questions present them- 
30 D. Loehr 
selves, this paper will look at four: 
(1) Given that hypertext - text which 
"jumps somewhere" - is a relatively 
new phenomenon, just what forms 
do authors use for hypertext an- 
chors? 
(2) Given the intuitive "pointing" simi- 
larity between place deictics 
(expressions which point to a place) 
and hypertext, is there any relatio n- 
ship between the two? Specifi- 
cally, do authors of the new me- 
dium of hypertext tend to use place 
deictics more frequently inside of 
hypertext anchors, since hypertext 
"reminds" them of place deictics? 
Or, for the same reason, do authors 
tend to avoid using place deictics 
inside .of hypertext anchors? 
(3) If any relationship is found in (2) 
above, do writers' preferences for 
using place deictics in hypertext 
change as they become more expe- 
rienced with this new genre? 
(4) Again noting the similarity between 
place deictics and hypertext, are 
place deictics the most common 
form of deixis found within hyper- 
text anchors? Or are other types, 
such as person deictics or time 
deictics, more prevalent? 
This paper, then, represents an exploratory foray into 
the little-known world of hypertext and deixis. It will 
undoubtedly generate more questions than answers. 
2. Background 
The earliest major work on deictics (also called in- 
dexicals) is by Btihler. He begins with an elegant 
image: 
The arm and the finger gesture of 
man, to which the index finger owes 
its name, recurs when the signpost 
imitates the outstretched 
"arm"...and the question \[is\] posed 
as to whether spoken language con- 
tains signs that function as sign- 
posts. The answer is yes, deictic 
words such as here and there have 
a similar function. \[1934\] 
Biihler did not originate the idea of deixis. He notes 
that classical Greek grammarians were aware of the 
concept. In the modern era, he credits turn-of-the- 
century linguists such as Brugmann and Wegener, 
who, while reconstructing Proto-Indo-European, de- 
scribed four types of deixis: I, you, this, and that. 
Biihler notes that land you are referred to as person- 
alia (Latin persona), which is from the Greek for 
mask or role. The meaning is that I and you, like 
actors, switch roles. BiJhler then advances a two- 
field theory: a deictic field, in which de ictics point to 
the actual entity in the world, and a symbolic field, 
in which all other terms refer symbolically to entities 
in the world. 
Fillmore \[1975\] gives a very thorough taxonomy of 
deixis. He enumerates five types: person, place, 
time, discourse, and social. 
Person deixis refers to the speaker, addressee, and 
audience of the utterance. Examples are pronouns 
and possessives such as I, we, yours, my, etc. 
Place deixis, including the locative adverbs here and 
there and the demonstratives this and that, are di- 
vided into two types. Gestural place deixis requires 
the addressee to physically monitor the utterance for 
an accompanying gesture: "Put the flower pot here". 
Symbolic place deixis requires only knowledge of the 
conditions of the utterance, without physical monitor- 
ing. My friend from London may write, "It's raining 
here again." Though I did not see her write the let- 
ter, I know where she was when she wrote it, and 
thus I know where here is. As an interesting side- 
light, Fillmore notes that there is only one word in 
English which requires gestural deixis for its use: 
yea, as in "She is yea tall". 
Time deixis, including the adverbs now, last night, 
and tomorrow, involves three different times. The 
encoding time is the time at which the utterance is 
made. The decoding time is the time at which the 
utterance is received (which will be delayed in writ- 
ten text). Finally, the reference time is any time 
referred to in the topic of the sentence. A typical 
voice-mail message might illustrate all three types. 
"This is Dan calling at 9 AM (encoding time). Give 
me a call back when you get this message (decoding 
time). I'd like to set up a meeting for Tuesday the 
14th (reference time)." 
Discourse deixis, according to Fillmore, "has to do 
with the choice of lexical or grammatical elements 
which ... refer to some portion ... of the ongoing dis- 
course -- something like, for example, 'the former'." 
Discourse deixis (also called textual deixis by Lyons 
\[1979\]) can be both written ¢' This sentence is writ- 
ten in English") and spoken ("She spoke about this 
loud"). Fillmore notes that discourse deixis is de- 
Hypertext and Deixis 31 
scribed temporally, as "any point in discourse can be 
thought of as a point in time". For this reason, we 
can say "in the preceding paragraph", or "the follow- 
ing announcement is a paid advertisement". In a:kl i- 
tion, written discourse has unique terms due to pages 
being read top-to-bottom: thus, we say, "as men- 
tioned above", or "see below". 
Finally, social deixis encodes, in Fillmore's words, 
"the social relationships on the part of the partic i- 
pants in the conversation that determine, for exa m- 
ple, the choice of honorific or polite or intimate or 
insulting speech levels". In addition to honorifics, 
examples include nicknames, informal names, and 
titles of respect. 
Fillmore also notes that certain verbs, such as come 
and go, are also deictic, as they imply a here and a 
there. 
Considerable work in the field has also been by Ly- 
ons. He makes note of the deictic center- the "zero- 
point of deictic context", which is always egocentric 
for the speaker \[1968, 1977\]. This zero-point is in 
relation to person, time, and space, with the center 
representing I, here, and now. Deictic terms are ex- 
pressed in relation to this deictic center. Here and 
this are closer than there and that, and now is closer 
than tomorrow. It is illustrative that the first, sec- 
ond, and third person radiate outward from the dei c- 
tic center. Lyons reminds us that deictic terms may 
also have non-deictic uses \[1973\]. "This idea is not 
new" refers anaphorically to some previously ex- 
pressed idea, while "This book is dedicated to my 
husband" refers deictically to the book containing 
the dedication. Finally, Lyons discusses tensed 
verbs as time deictics \[1977\]. Note that we need to 
know the time of utterance to work out any truth 
conditions of a tensed sentence. 
I have thus far summarized work on deixis. As for 
deixis in relation to hypertext, no previous work ex- 
ists that I am aware of. I hope that this study (a 
deictic!) will shed some light on this new phenome- 
non. 
3. Forms of Hypertext Anchors 
Before addressing my first question the forms 
authors use for hypertext anchors - I will briefly dis- 
cuss my general methodology. Methodology specific 
to each of the other three questions will be covered 
in that question's section, in turn. 
The data used for all four questions consists of pages 
from the World Wide Web collected in April 1996. 
There are of course many types of Web pages 
(Roberts \[1997\] discusses the different genres found). 
For this study, I collected pages from two types: the 
"home pages" of college students, and the "home 
pages" of commercial enterprises. A home page is 
the initial, top-level page at a "web site", with a 
web site typically consisting of multiple related 
pages. For consistency, I used home pages only. 
For college student data, 126 different Web pages 
(containing roughly 1000 anchors) were collected 
randomly from three institutions: Cornell University, 
the University of Texas at Austin, and the University 
of Oregon. For commercial data, 31 Web pages 
(containing roughly 500 anchors) were collected. 
The commercial businesses included well-known 
firms in varying fields: examples are Apple Com- 
puter, Prudential Insurance, Pepsi-Cola, Sony, 
Whirlpool, Disney, and the American Stock Ex- 
change. 
After each page was downloaded, an automatic 
script stripped out the words used as anchors. 
(Fortunately, Web pages are written in HyperText 
Markup Language (HTML), which has specific tags 
surrounding the anchors). The anchors were saved to 
a separate file, as was the original Web page, for 
reference. Graphical anchors were ignored, unless 
they had "alternate" text associated with them, for 
the use of non-graphical Web browsers. 
To answer question (1), I then went through the list 
of anchors, and manually c~ed them per the follow- 
ing categories: 
Noun Phrase 
Verb Phrase 
Adjective Phrase 
Adverb Phrase 
Prepositional Phrase 
Relative Clause 
Full Sentence 
Imperative 
Question 
Declarative 
Other 
The results are shown in Table 1, which lists in 
roughly decreasing order the forms authors used for 
hypertext anchors, by number and percentage. The 
table is further broken down by source (college stu- 
dent and commercial), for comparison. Note that the 
"Full Sentence" category has been subdivided into 
"Imperative", "Question", and "Declarative". 
32 D. Loehr 
College 
Student 
Home Pages 
Commercial 
Home Pages 
Form 
NP 
Full Sentence 
Imperative 
Question 
Declarative 
VP 
AdvP 
AdjP 
PP 
Other 
Relative Clause 
# -~ # -~ 
893 88.9 364 72.4 
48 4.8 69 13.7 
32 3.2 41 8.1 
9 0.9 15 3.0 
7 0.7 13 2.6 
19 1.9 26 5.2 
18 1.8 4 0.8 
10 1.0 7 1.4 
8 0.8 12 2.4 
7 0.7 4 0.3 
2 0.2 17 3.4 
Totals 1005 =100 503! --100 
Examples 
Cornell University Home Page 
What's New? 
General Motors Announces I~ Electric Vehicle 
click here 
click here 
New at National 
About Me 
If Ooeratin~ Systems Were Beers,.. 
T Who We Are 
Table 1. Forms of hypertext anchors, number and approximate percentage, by source 
(Percentages do not exactly total 100.0, due to rounding) 
We can now answer our first question, as to the 
forms hypertext anchors take. The overwhelming 
majority are noun phrases, with full sentences taking 
a distant second. Note however, that commercial 
Web authors tend to use fewer noun phrases and 
more full sentences than college students (the diffe r- 
ence is significant, per a chi-squared test, with p <= 
0.01). Perhaps this is because commercial sites try 
to present a more polished image. 
Our first answer regarding hypertext immediately 
raises another questio n: Why are the overwhelming 
majority of anchors noun phrases? An intuitive reply 
is that authors are viewing hypertext anchors as re- 
ferring expressions. Since referring expressions in 
more familiar non-hypertext are almost always noun 
phrases, authors choose noun phrases for their an- 
chors. Continuing this theory, to the extent that 
authors are not using noun phrases as anchors, to just 
this extent are they "branching out" in the new phe- 
nomenon of hypertext, and discovering linguistic 
forms other than noun phrases for encoding hypertext 
links. Confirmation of these speculations, however, 
must await fu~er research. 
4. Relationship Between Place Deictics and 
Hypertext 
Let us now turn to our second question, reprinted 
here: 
(2) Given the intuitive "pointing" simi- 
larity between place deicfics and 
hypertext, is there any relationship 
between the two? Specifically, do 
authors of the new medium of hy- 
pertext tend to use place deictics 
more frequently inside of hypertext 
anchors, since hypertext "reminds" 
them of place deictics? Or, for the 
same reason, do such authors tend 
to avoid using place deictics inside 
of hypertext anchors? 
To determine this, we need to determine the fre- 
quency, or "density", of place deicfics inside of hy- 
pertext anchors versus that outside of hypertext an- 
chors (i.e. in the regular text of the Web page). 
A digression is needed to motivate my methodology 
for determining these frequencies. At first blush, the 
task seems relatively easy. One need simply go 
through the Web pages and code for each occurrence 
of a place deictic. Although time-consuming, this 
can be aided by heuristics such as searching for the 
most common place deictics: here, there, this, and 
that. Next, a simple programming script can split 
apart the Web pages into two files: one file contai n- 
ing all the words inside of hypertext anchors, and the 
other containing all the words outside of the hyper- 
text anchors. Finally, one can calculate the relative 
Hypertext and Deixis 33 
frequency of the place deicfics in each file, by divid- 
ing the number of place deicfics by the number of 
words in the file. 
There is a problem, however, in determining the fre- 
quency of something by dividing by the number of 
words in a text. As Ball explains: 
In a word-based frequency analysis, 
to say that a phenomenon occurs 
with equal relative frequency in two 
samples is to say that equal 
amounts of text, measured in words, 
will yield the same number of to- 
kens. But relative frequency should 
be a measure of the number of 
times something occurs within the 
number of opportunities for it to oc- 
cur (emphasis mine) \[1994\]. 
To give an example: If we are comparing the fre- 
quency of, say, interrogative sentences in the writing 
styles of Hemingway and Dostoevski, it would be 
meaningless to divide the number of interrogatives 
by the number of words in each author's text. The 
results of this would be skewed by the length of sen- 
tences which each writer produces. Hemingway may 
pen fully twice the number of sentences in the same 
number of words as his Russian counterpart. Thus, if 
both authors had the same absolute number of inter- 
rogatives, we would be wrong to conclude that they 
both favored this form equally. Rather, Hemingway, 
who produced twice as many sentences, used the 
interrogative only half as frequently as Dostoevski. 
The key, then, is to measure a phenomenon against 
the number of opportunities for the phenomenon to 
occur, not against raw word count. In this example, 
the opportunity for an interrogative to occur is a sen- 
tence - hence, we must count the number of sen- 
tences as well. 
We may now return to our problem of measuring the 
frequency of place deictics. We must decide what 
an appropriate opportunity for occurrence for a place 
deicfic would be. As Ball advises, selecting this 
category is "crucial and requires considerable 
thought \[1994\]", or else our results will be invalid. 
With this in mind, let us look first at the place dei c- 
tics here and there. Although technically these may 
be thought of as adverbs ("Come here"), I will argue 
that they pattern with prepositional phrases ("Come 
to this place"). As further evidence, consider "Put 
the book here". Replacing here with an adverb is 
ungrammatical (* "Put the book slowly"), while a 
prepositional phrase is acceptable ("Put the book on 
the table"). Thus, when an author chooses to use 
here or there, the author is choosing to use these 
words instead of using a prepositional phrase. 
Hence, I will consider an opportunity for occurrence 
of here and there to be a prepositional phrase. 
Although searching for prepositional phrases is time- 
consuming, it is certainly doable - and in fact, I did 
just that for this study. As a heuristic, I started by 
searching for a list of 50 common prepositions, which 
I estimate covered at least 90% of all raw occur- 
rences of prepositional phrases (the first six alone 
accounting for 70%, according to Mindt and Weber" 
\[1989\], reported in Crystal \[1995\]). I then visually 
scanned the text for any remaining prepositions. (A 
more efficient method would be to scan for the 81 
prepositions found in both the Brown and LOB cor- 
pora, said by Mindt and Weber to account for 99.9% 
of all raw occurrences). 
We must also consider the other types of place dei c- 
tics. While sometimes acting as determiners, the 
place deictics this and that can also be (and pattern 
with) both pronouns ("Look at that") and adjectives 
(" This morning"). Unfortunately, this means that an 
opportunity for an author to use this or that would be 
a noun phrase or an adjective phrase. While searc h- 
ing for prepositions is difficult, it is feasible because 
prepositions are a finite class. Yet searching for the 
open classes of nouns and adjectives I judged to be 
undoably time-consuming for this research. Equally 
difficult would be considering place deictics occu r- 
ring as motion verbs (i.e. come), as that requires 
searching for all verbs to find opportunities for occu r- 
rence. I have therefore chosen to restrict my study of 
place deicfics to only here and there. Hence, my 
research will have nothing to say about other place 
deictics. A strategy for resolving this dilemma would 
be to use an automatic parser, which could identify 
most nouns, adjective, and verbs. I leave that task to 
future work. 
Having covered the necessary background for com- 
paring the frequency of place deictics inside of and 
outside of hypertext anchors, I may now present the 
results in Table 2: 
33 D. Loehr 
Inside of 
Hypertext Anchors 
Outside of 
Hypertext Anchors 
Totals 
Number of Occurrences of Here 42 66 108 
and There 
577 2433 3010 
619 
Number of PPs (Opportunities for 
Occurrence of Here and There) 
2.7 
2499 
Frequency (in Percentage) 
Totals 
7.3 
3118 
Table 2. Frequency of the place deictics here and there 
inside of and outside of hypertext anchors 
In percentages, the place deictics here and there 
occur at a frequency of 7.3% inside of hypertext 
anchors, versus 2.7% outside hypertext anchors (the 
difference is significant, per chi-squared with p <= 
0.001). For simplicity, I have grouped the edue a- 
tional and commercial pages together in this table - 
broken apart, there is no statistically significant dif- 
ference between them in this regard. 
Thus, based on the place deicfics I've looked at, we 
can answer our second question: place deicfics do 
occur more frequently inside of hypertext anchors 
than in the regular Web page text. 
However, the conclusion that this is so because of 
the place-deictic quality of hypertext is not so easily 
reached. It turns out that of the 42 place deictics 
found in anchors, 41 of them were the word here. 
This abundance is due to the popularity of the hyper- 
text expression click here. This usage may be a 
simple Graphical User Interface (GUI) issue, in 
which authors are merely instructing their audience 
to click a certain button. Furthermore, if authors 
want their readers to go to some other place there, it 
is counterintuitive to place deixis to use the word 
here. Thus, the usage of here in hypertext is proba- 
bly not due to place deixis, but rather a GUI artifact. 
5. Changes in Hypertext Authors' Style as a 
Function of Experience 
In answering question (2), we could not prove a rela- 
tionship between hypertext anchors and place deic- 
tics. We did, however, see a relationship between 
the word here and hypertext anchors. Going with 
this, let us modify our third question to be: 
(3) Do writers' preferences for using the 
word here in hypertext change as a 
function of experience in this new 
genre? 
This question is not unfounded. Many hypertext style 
guides preach against what has been labeled the 
"click-here syndrome". The argument is that the 
word here has nothing to do with the target of the 
link, and should be avoided in favor of a more de- 
scriptive term. To quote one HTML manual: 
STYLE TIP: Try to create links 
that don't muck up the flow of the 
text .... Your text should stand on its 
own. You might want to avoid 
words like "click your mouse on 
this sentence to read about the his- 
tory of skyscrapers." Every Web 
user knows how to click .... Instead, 
say something like, "There's an ex- 
cellent article on the history of sky- 
scrapers by T.O. Tall" - and make 
the word "skyscraper" your link. 
(Fox and Downing \[1995\]) 
If this is what is prescribed, it is reasonable to hy- 
pothesize that professional writers will try to avoid 
the click-here syndrome. To test this, we can look at 
the frequency of the word here within the hypertext 
anchors of novice writers versus that of professional 
writers. To this end, I make the following assump- 
tion- that college students are less experienced at 
writing Web pages than those who author comme r- 
cial pages. This is of course a generalization, but 
given the fact that commercial outfits are concerned 
about the image they present, and that their Web 
page authors are most probably paid for their work 
(and hence professionals), it seems reasonable that 
they will hire experienced authors. No experience, 
however, is required for college students' authorship. 
Table 3 presents the frequency of here in the hyper- 
text anchors of college student home pages versus 
commercial home pages. 
Hypertext and Deixis 35 
College 
Home Pages 
Student Totals 
Number of Occur~nces of Here 34 7 41 
512 64 576 
6.6 
Number of PPs (Opportunities for 
Occurrence of Here) 
Frequency (in Percentage) 
Commercial 
Home Pages 
10.9 
546 71 617 Totals 
Table 3. Frequency of the term here inside of hype rtext anchors 
for college student home pages versus commercial home pages 
At first glance, the opposite of our hypothesis seems 
to be true - that the more experienced commercial 
authors use here more frequently. However, the dif- 
ference is not significant (per chi-squared at p <= 
0.05), so any discussion is moot. Hence, I cannot 
confirm or reject the hypothesis bas~ on this data~ 
Greater quantifies of data may be necessary to de- 
termine any real patterns. In addition, future work 
might also employ a more verifiable method of ide n- 
tifying novice and experienced Web authors than 
simply generalizing about college students and 
commercial webmasters. 
6. Types of Deictics Found in Hypertext 
Finally, let us turn to our fourth question: 
(4) Again noting the similarity between 
place deiedc s and hypertext, are 
place deicfics the most common 
form of deixis found within hyper- 
text anchors? Or are other types, 
such as person deictics or time 
deictics, more prevalent? 
To answer this, I looked at each anchor in the data to 
see if it contained any deicfics. These deictics were 
then categorized as one of: 
Person 
Speaker 
Audience 
Both 
Place 
Time 
Discourse 
Social 
More than one of the above 
Some discussion is necessary regarding my coding. 
First of all, the common Click here. as noted, was 
classified as a place deictic, as it refers to a physical 
location in the universe (a spot on a computer screen 
which may be clicked on). Yet as a sidelight, one 
could also consider it a discourse deictic, as it refers 
to the word here itself, which is part of the discourse. 
This interesting case would be the purest discourse 
deixis imaginable. Most discourse deixis refers to 
the text in which the deictic word occurs (this book, 
the paragraph below). We even have cases where 
sentences refer to themselves (This sentence is in 
English). But the deicfic here in click here would be 
a word which points to itself. Verbally, it is difficult 
to find an equivalent. One possibility would be, 
"The word I say now has three letters and rhymes 
with cow". Though certainly possible before hyper- 
text was developed, it is the advent of hypertext 
which has made this "ultimate discourse deixis" 
commonplace in language. 
(Note, however, that person deixis has always had 
this reflective deixis, at least verbally, in the first 
person. The speaker who says 'T' is referring to her- 
self. Thus, the word here (in a manner somehow 
reminiscent of Lewis Carroll) is personified, by cal l- 
ing attention to itself..) 
There were cases where here was not used in the 
click here sense. One example was an anchor la- 
beled The weather here, referring to the weather at 
Cornell University, the author's home. Another was 
the following: Here I am in Roslyn, WA, (where 
Northern Exposure was filmed). The hyperlink 
brought up a picture of the author in Roslyn, WA. 
This here is not referring to the word here. Nor is it 
referring to the location in which the author wrote the 
sentence (at Cornell University, in Ithaca, NY). 
Rather, we have a case in which the author has 
shifted his deictic center mentally, much as Thomas 
\[1996\] reports that speakers do when giving direc- 
tions. 
In terms of person deixis, I collapsed the addressee 
36 D. Loehr 
and audience types under the latter category. 
Though typically audience refers to others who hear 
the utterance (in addition to the intended addressee), 
Web pages are intended for everyone to read, not a 
specific addressee. Thus, there are no "bystanders" 
or, more accurately, the bystanders are the addres s- 
ees. Thus, for the Web, addressee and audience are 
the same thing. I have chosen the latter category as 
more accurately reflecting how the Web works. 
Lyons' notion of deictic and non-deictic uses of 
words was needed in several cases. For instance, 
this in Why is this oaee black? is clearly deictic. But 
in another anchor, Come to Think of It. Who Is Thi¢ 
Phil Guy?. the same word is not. Rather, it is ana- 
phoric, since Phil has introduced himself at the be- 
ginning of the Web page. 
Send an Electronic Postcard. Imperatives imply the 
second person, as in (You} Send an Electronic Post- 
card. The second form of implied deicfic was the 
implied first person possessive, as in Favorite Web 
Sites. It is probable that the author meant My Favor- 
ite Web Sites. However, though both of these types 
implied a person deictic, the actual deictic lexical 
item was missing. There were other cases where a 
deictic word (my) was just outside the anchor, as in 
My father. In all these instances, since there was no 
deictic word in the anchor, I did not classify it as 
containing an ordinary deictic. 
Now, to answer our fourth question. Table 4 lists the 
distribution of deictics (by number and percentage) 
found in hypertext anchors. 
Finally, when coding, I resisted the temptation to 
collect implied deictics. These occurred in two 
forms. The first was the imperative: an example was 
Type 
Person Only 
Speaker 
Audience 
Both 
Place Only 
Discourse Only 
Multiple Types 
Time Only 
Social Only 
Totals 
Number 
69 
60 
Approx. 
Percent 
-age 
54.8 
47.6 
Examples 
Mall Me 
4 3.1 Your feedback 
5 4.0 Tell Us Your Storv 
38 30.2 Return to Jane Austen's info DaMe 
10 8.0 Whv is this oaee black? 
7 5.6 alt=Make vour visit count, load this imaee 
2 1.6 I'm in the lab 
0 0.0 
126 =100 
Table 4. Types of deictics found in hypertext a nchors 
(Percentages do not exactly total 100.0, due to rounding) 
For simplicity, I have again grouped the educational 
and commercial pages together in this table - broken 
apart, there is no statistically significant difference 
between them on this question. 
We can now answer our fourth question. Of those 
deictics used in hypertext anchors, the majority are 
not place deictics, but person deictics. And of these, 
the majority refer to the speaker. Of course, the data 
for this effort was home pages, and the preponde r- 
ante of self-referenc es may reflect the fact that 
home page authors use their home page as a vehicle 
to tell the world about themselves. As research into 
the various genres of Web pages progresses, it will 
be interesting to see which is the most common type 
of home page. This data suggests that personal nar- 
ratives may be the most common. 
Though we have answered our fourth question quant i- 
tatively, there remain some interesting qualitative 
examples to discuss. 
First of all, following Fillmore's sense of come and 
go as deictic verbs, the common anchor return is 
also deictic, as in Return to Jane Austen's info oaee. 
To return implies a starting point - a here - to return 
from. Though the place we are "going" is specified 
(Jane Austen's info page), the place we are 
Hypertext and Deixis 37 
"leaving" (here) is not. We can only know where 
we are returning from by deixis - by knowing the 
place of utterance. Note that this is another c~ase 
where the deictic center is mentally shifted. The 
reader never actually moves anywhere - she is still in 
her chair before her computer. But mentally, she is 
traveling to wherever her Web documents are re- 
trieved from. Such is the magic of the World Wide 
Web. 
Another interesting area has to do with time deictics. 
There were only three of these found in anchors (two 
on their own, and one in combination with other 
deictics). The first rather novel example had a blink- 
ing anchor, in the sentence "Click n0w!" Perhaps 
the author meant for the reader to click at the mo- 
ment the blinking anchor was lit up. In the second 
example, the author writes, I may be in the computer 
lab, in fact, you can check if I'm in the lab right now. 
When this anchor is selected, a snapshot comes up 
which is taken from a live video camera in the 
author's laboratory (the author was not in when this 
data was collected). Even though the word now is 
not within this anchor, we can still find time deixis 
in the tensed verb am. The present tense implies just 
that -now- and the sentence cannot be evaluated 
unless one knows the time it is read. Note that both 
of these examples use decoding time (the time the 
sentence is read). 
The third, very clever, example is: 
I'd like to send a special hello to 
SteDhen Crawford. Why don't you, 
too? And watch here for a link to 
his future home page. Then you, 
too, can meet the greatest trombon- 
ist in the world 
Still watching? 
His home page isn't future anymore, 
it's here ! 
With his final here, this author has scored a deicfic 
triple play. First, he invokes here as a place deixis, 
meaning Stephen Crawford's home page is now avail- 
able from this place here. (Note the sentence above - 
Watch here for a link - this can be taken to mean 
Watch this place on the Web). Second, he uses the 
familiar click-here phrase - which, as mentioned, 
could be thought of as discourse deixis. Finally, the 
author uses a sense of here which means now, as in 
It's December 25, and Christmas is here. This invokes 
the third deixis (time). Note that this example uses 
encoding time (the time the sentence was written). 
Another interesting point regards gestural place 
deixis. One may think that gestures (actual pointing) 
are not possible in a textual medium. Yet Web 
pages are more than text - they support animation, 
and many modern pages contain "moving arrows" 
acting as anchors. Even non-graphical, text-based 
pages can cream arrows such as these: --> <--. 
However, I concentrated only on textual anchors 
(avoiding the difficult problem of classifying im- 
ages), and found no gestural deixis in my data. Yet 
temptingly close was the sentence Here I am (which 
brought up a picture of the author). This might be 
thought of as a presentative, much like the French 
word voila, which requires a gesture for its use. The 
words Here I am, by convention, almost seem to im- 
ply a gesture -the one of an actor stepping from be- 
hind a curtain, sweeping his hands before him and 
bowing his head in presentation to the audience. 
Note that no occurrences of social deixis were found. 
7. Conclusions 
Hypertext links behave much like ordinary deictics. 
Both require the context of the utterance for full in- 
terpretation. For hypertext, the "utterance" is the 
Web page, and the context of this Web page is the 
World Wide Web itself. Just as ordinary deictics 
like me make us look to the context to see who me 
really represents, so does hypertext make us look to 
the context of the Web to see what the hypertext link 
really represents. 
This study investigated several questions regarding 
hypertext and deixis, using as data Web pages col- 
lected from the Internet. In terms of form, hypertext 
anchors tend to be noun phrases, with a lesser con- 
centration of complete sentences. Although one 
might intuit a relationship between place deicfics 
and hypertext, none could be substantiated. Hyper- 
text anchors do contain a greater concentration of a 
certain place deicfic (here) than the rest of the Web 
page, but this is due to the popularity of the phrase 
click here. This popularity, in turn, may be due to 
the unique environment of hypertext, which makes 
the text itself a computer input "button". 
Because style guides prescribe the avoidance of the 
"click-here syndrome", it is reasonable to hypothe- 
size that more experienced hypertext authors will 
disprefer this in contrast with less experienced 
authors. Assuming that college students are less ex- 
perienced than professionals who author commercial 
Web pages, I was unable to confirm this statistically. 
As to the types of deixis found inside of hypertext 
anchors, the majority were person deictics referring 
to the speaker, probably due to the fact that the 
pages studied were home pages, in which authors 
38 D. Loehr 
present themselves to their audience. 
The only statistically valid difference between the 
two sets of authors is that professional authors tended 
to use more full sentences and less noun phrases than 
college students. This may have nothing to do with 
deixis. It may instead be due to the polished image 
which commercial sites hope to project to their cus- 
tomers. 
Fillmore, Charles. Santa Cruz Lectures on Deixis. 
Indiana University Linguistics Club, 1975. 
Fox, D. and Downing, T. HTML Web Publisher's 
Construction Kit. Waite Group Press, 1995. 
Lyons, John. Introduction to Theoretical Linguistics. 
Carnbndge University Press, 1968. 
Further work in this area might include the following. 
Our first question revealed that the majority of hyper- 
text anchors were noun phrases. This might be inves- 
tigated with an eye to the use of anchors as referring 
expressions, which are typically noun phrases. The 
second question dealt with determining the relative 
frequency of a variety of place deictics, noting the 
difficulty of calculating their opportunities for occu r- 
rence (which manifest as a variety of parts of 
speech). A robust parser might be used to identify 
these parts of speech. The third question - whether 
experienced authors disprefer the "click-here syn- 
drome"- remained unanswered. More data, and a 
more verifiable means of identifying experienced and 
novice authors, might prove fruitful. The final dis- 
covery -that the preferred deictics used in home 
page anchors were person deicfics - might shed light 
on the classification of home page genres. My find- 
ings suggest that home pages are of a narrative form, 
utilizing first person. 
Aside from any relationships between hypertext and 
deixis, we have seen that hypertext provides a 
unique environment for creative uses of deixis. Ges- 
tural deixis, time deixis, and even three-deixes-in- 
one-word were found. These examples, coupled with 
the fact that nearly all types of linguistic deixis ~xe 
found in hypertext, lend evidence that the relatively 
new medium of hypertext is a fully productive genre 
of natural language. 
Acknowledgements 
I gratefully acknowledge Hughes Electronics Corpo- 
ration, for providing a fellowship under which this 
research was carried out. I am also indebted to Dr. 
Cathenne Ball for the original idea of investigating 
hypertext and deixis. 

References 
Ball, Catherine. Automated Text Analysis: Caution- 
ary Tales. Literary and Linguistic Computing, Vol. 
9, No. 4, 1994. 
Btihler, Karl. Theory of Language (Sprachtheorie). 
Translation of 1934 original by D. Goodwin. John 
Benjamin Publishing, 1990. 
Lyons, John. Deixis as the Source of Reference. Lin- 
guistic Agency University Taler, 1973. 
Lyons, John. Semantics, Vol. 2. Cambridge Univer- 
sity Press, 1977. 
Lyons, John. Deixis and Anaphora. In T. Myers (ed.) 
The Development of Conversation and Discourse. 
Edinburgh University Press, 1979. 
Mindt, D and Weber, C. Prepositions in American 
and British English. World Englishes 8, 1989. 
Quoted in Crystal, D. The Cambridge Encyclopedia 
of the English Language. Cambridge University 
Press, 1995. 
Roberts, Greg. Manuscript. 1997 
Thomas, Katie. Manuscript. 1996 
