An Anatomy of Anchors: Abstract, Introduction, and Conclusions

Deena Larsen
633 Oak St. #1
Lakewood CO 80215
Richard E. Higgason
Blue River Community College
20301 E. 78 HWY
Independence, MO. 64057


How are the properties of anchors used to influence reading, interpretive, and navigational strategies?

While much attention is paid to defining and examining interactions with links, little is paid to the front end: the anchor. We examine what an anchor is, describe six spectra of anchor properties (density, location, function, decoration, format, and uniformity), and present examples of these properties. This paper describes the state-of-practice—anchors not as they may be or should be but as they are in 67 sites and works.

Efferent (informational and promotional) and aesthetic (literary) sites and works tend to use anchors in different ways. Anchor property combinations affect the way readers interpret content and navigate.


The classic picture of a hypertext is node/ link, where nodes contain content and links connect nodes. However, the link always has something that trigggers it: an anchor. Anchors can range from hidden words and spaces that yield to animated features that grab the reader's attention. Anchors inform readers of possible meanings and semiotic connections and entice readers to follow the link to explore other places. Anchor properties contribute to a reader's overall impressions of a site or work, setting the tone for the work and determining how a reader will engage with the work.

We hope readers will use these resources to consider ways of displaying and using anchors for various purposes, propose studies in reading and information gathering effectiveness using current examples of anchors, and discover even more ways of breaking the boundaries of anchor/link/node expectations.


Anchor properties are deeply intertwingled, and it is not one but all of the properties of the anchors on a given site that can affect readers' actions and reactions. Readers approach the twin tasks of interpreting and determining whether to trigger the anchor differently for the various anchor property combinations. Moreover, because an anchor exists only as a part of a link, the anchor cannot always be discussed separately from the link. This is especially true when discussing anchor functions.

Efferent and aesthetic approaches

We found significant differences between efferent sites whose primary purpose was to provide information or sell a product and aesthetic works whose purpose was to entertain or to engage readers in conceptual exploration. (See methodology for classification process). The difference between these sites was not always a black and white boundary, but more of a tendency. In general, efferent sites tended to use standard anchor decoration, had dense navigation anchors rather than dense embedded content anchors, and used denotative rather than connotative anchors. Aesthetic sites, on the other hand, exhibited a broader range of anchor properties, showing more willingness to experiment and explore.

Much of the emphasis in anchor design has been on making the destination clearer. That is, in Landow's terms, making the site more efficient [94]. Information-providing sites are maligned (with the disparaging term mystery meat) for using highly ergodic techniques that do not guide the reader to useful information. Yet, Harpold reminds us that efficiency may not always be the goal. Instead, hypertext offers an opportunity for an exploratory journey and anchors can also serve as a way to delay the goal and entice readers to experience more of the text [90].

Anchor uniformity and density

The more consistent the approach, the more consistent reader interaction and expectation. Multiple schemas can be used to perform different functions (e.g., a menu for navigation and embedded text for content). Efferent sites tended to use mainly primary or multiple approaches while aesthetic works used a wider range of strategies. No efferent site used inconsistent approaches, but two aesthetic works did.

The density, visibility, and location of embedded links determine the amount of words or content a reader processes before encountering the anchor's incipient invitation to leave the space.

While the number of anchors needed to constitute a dense site is contested, we found that many sites have a high number of anchors. We found that most efferent sites had a menu-driven density. This reliance may be because efferent sites are serving information that can readily fit into a hierarchical menu system, and thus can offer readers the particular information they are seeking. Aesthetic sites, on the other hand were more evenly divided between menu, embedded, and sparse strategies.

Reader interpretation of anchors

Merely creating an anchor is a semiotic act which invites reader interaction and interpretation. Just as readers connect a signifier with the signified, an anchor becomes a signifier for a destination node to which it provides access (in terms of Pierce's trichotomy, the anchor serves as an index sign. As an index, the anchor signifies the relationship between the arrival and departure nodes.) On the origin node, the presence of the anchor provides a promise and therefore emphasizes and privileges the anchor content. This encourages readers to interpret both the anchor and the origin node in the context of the destination node. Thus, an anchor works interpretively to bridge the origin and destination.

Up to a point of diminishing returns, the more the anchor decoration varies from the moderate or standard (e.g., hunting for hidden or staring at striking anchors), the more significance is attached to that anchor and its interpretation and navigation.

Semiotic functions are independant of the other properties we have identified for anchors. For example, two sites can have the same type of anchor decoration, location, density, and uniformity, while one site uses anchors for denotative functions (e.g., Wikipedia [68]) the other site uses anchors for connotative functions (e.g., the Unknown [56]). However, clusters of anchors (i.e. maps and menus) tend to be more heirarchical and denotative as they are viewed as navigation for the site. Isolated anchors (i.e. embedded) are more likely to be contrasted with the non-anchored content and thus provide readers with clues to contextual relationships and connotations.