My introduction to instructional technology arose as a response to the collegial and disciplinary excitement emerging in Romantic studies over the first text archives being mounted on the World Wide Web. These archives, put up before text scanners were affordable and web page software available, involved the sheer labor of typing in texts and tags by hand using simple html commands. But the possibilities for scholarship and teaching that these web archives opened up were astounding; suddenly the new research into lesser known writers and texts of the Romantic period, as well as contextual materials, could be made easily accessible not only to faculty but to students. The web project I want to describe and meditate on in this article, La Belle Assemblée was my answer to this excitement, but that initial experience and my ongoing work with the project have led me to consider the intersection of technology and pedagogy as a complex problem, one that needs to be fleshed out and probed as we continue to experiment with differing levels of pedagogic invention and praxis. As scholars of literature, the Web poses us a highly significant problematic in which the textual and the visual compete with each other at the same time that the linear and the hyper model do. This perceptual matrix rests on cognitive modes that, if we are not attentive as textual providers, can resolve into a singular, less complex way of thinking through textual material for our students, and perhaps for us as well. How are we asking our students and ourselves to view—and to read—hypertexts? Are we aware of the ways in which technologically assisted models for presenting material can help us rethink the student's relation to that material, or do we deceive ourselves into disguising old approaches with new formats? And are we taking advantage of the new cognitive research to adjust how we use instructional technologies to new models of thinking? We have learned much about intertextuality from deconstruction, but are we asking the right questions about the student's intertextuality—her interaction with and interpellation by texts as ideological, as well as artistic, agents— in ways that usefully deconstruct the hypertextual moment?
I. The Project: Construction and Constructivism
La Belle Assembée ("The Bluestocking Archive" hereafter) is the website I have mounted that focuses on the role of the Bluestockings and of gender in the rise of Romanticism. The archive is essentially constructed as a hyper-book, by which I mean it is not quite or not yet a hypertext. Instead, it is conceptually simple: a front-ended site organized by a "working" table of contents, rather than constructed according to a dimensionality that we think of when we talk about hypertexts. It is thus neither linear nor nonlinear, displaying neither authorial intent nor readerly intervention (that is, neither hypertext nor cybertext), but is instead what we might call constructivist. "Constructivism" as a term purposely counters "deconstruction" and even "structuralism" to designate the kind of intertextual, anti-authorial, and anti-formal growth that increasingly populates meta-websites. The purpose of its organization is to make it open to additions by students in different courses over time. The archive is designed to work as an anthology, with the archive contents page as the access point for each text, and the links acting as page locators. In this sense, the archive mediates print and media culture to make the Web a useful tool for students who still need a bookish orientation in order to understand the print culture of Romanticism. My concern to bridge students' textual relations arose from my own experience with hypertext, where I am easily lost after following just a few links, unsure of how to map where I have been, how to situate the text, how to map it in my own mind. I worried that students would become immersed in the idea of hypertext without addressing the fact of textuality, a problem similar to that of immersing them in critical theory. My solution, which I still endorse, was to foreground the textual and minimize hypertextuality; the hypertextual became what students made of it, so that in construing and constructing links to and from the textual they themselves would produce hypertext, but as a by-product. This shift of attention away from the dazzle of the hyper and the cyber, and toward an enforced simplicity, did not meet with resistance from my students. I think it at once made the project more serious and something even the technophobes among them could clearly be part of. I believe this simplicity can be maintained as the project grows and the levels of connection expand because the connectivity will always be subsumed to the level of the text. It is this student-to-hypertext relation that I call " constructivist" in the sense that in an elementary way the student's contribution to the website structurally reproduces the mental activity of a reader confronting a text, as she construes it through the varying levels of textuality available from annotations, appendices, contextual materials, criticism, and so on.
My claim that the textual and the visual compete with each other at the same time that the linear and the hyper model do in hypertext needs to be addressed, I believe, at the level of textuality. Such competition can be exciting and energizing for certain endeavors, particularly gaming, but not (or not necessarily) for literary studies. At the same time, I do believe that the way we learn to read is changing as children's confrontation with text occurs on mixed fronts: in the linear book form, the internet hypertext form, the CD-ROM, the video game and televideo form, and so forth. Linearity and associative thinking are contending for cognitive patterning in young minds. Wherever current students happen to be in this transition from book culture to techno culture we are experiencing, they need some kind of integrative model for understanding the products of a prior book culture within the context of their own textual culture. This integration should certainly be premised at the level of the text, as I have already made clear, but it should also be adapted at the level of structure. Therefore, despite my claim for simplicity, the archive I am building uses a deeper structural analogy, that of the salon.  Eighteenth-century Bluestocking salons were carefully yet casually engineered so that participants gathered in select groups within small rooms and areas within the hostess's house, all of these conversational spaces leading effortlessly toward the central space where the hostess reclined with a few select companions. Eventually everyone had to pass through her room. Salon structure, then, orchestrates conversational interaction, directing its flow back to the centering space, in my case, the table of contents. No matter how deep into the connective links a reader wanders (especially as this project grows in size), the structure maps itself for her by necessitating a return to the center. This is only one aspect of "salon theory" as I will discuss it below, but it emphasizes the need to consider pedagogically how web projects incorporate and relate form, cognitive behavior, structure, and conceptual aides.
At the present time any tools that aid students in making the transition from how to think about printed texts to how to think in terms of hypertext are useful, whether they are software resources, projects designed to increase student confidence in using the internet and the Web, or discussion listservs and chatrooms organized around course subject matter. The tools we choose, and how we choose to employ them also need to be carefully considered in terms of how we want students to understand the relation between historical texts and their technologies and current technologies and the texts we make. My original response to this concern was to construct the Bluestocking Archive with an open-ended structure, but one that is designed at first to be so only linearly rather than in depth. (I use linearly here not in terms of the "linear/nonlinear" hypertextual distinction,  but to distinguish a book-oriented approach from true hypertext). As the linear documentation is more and more filled in, I then want to think about the intertextual implications more closely, and to design class projects that tackle the deep ends of the archive by constructing textual and interpretive connections through a dimensional or matrix effect that is more truly what is meant by hypertext. At this point the project will transition from an internet archive-a mere repository of texts-to a website, a hypertextual and interactive site of production. What students produce here through their relation to texts marks the website off from the archive of products. To prevent this transition from product to productivity from becoming something less than scholarly, I want to focus on the student as a site of intertextuality, something quite different from the intertextuality of the text or even the intertextualized archive. How the student is bombarded by and relates to an historical culture, current culture, ideologies of text and of technology must certainly take precedence in designs for web projects.
The difference between archive and website can be pertinently posed another way, and the following subtitles from Cyberspace: First Steps—"from Euclidean space to cyberspace" and "the mind is a leaking rainbow"—indicate the conceptual difference I want to highlight here.  The design of web projects has to seriously consider the difference between technology orchestrated as lecture (Euclidean space), as a resource guide (architectural space), and as an interactive agent (the leaking rainbow). While this juxtaposition points to the bleeding of imaginative color as the appropriate metaphor for what a web project should accomplish, an internet search for pedagogical approaches to the web turns up projects heavily concentrated on the first and second models. In an article in the Journal of Asynchronous Learning Networks, for instance, Skip Knox describes his internet project [Appendix 1] in terms that differentiate it very little from a standard lecture course. Several other sites provide extremely interesting resource collections [Appendix 2] for pedagogy and technology, but their site design cannot be considered course-worthy. For the same reason that these models should not be the goal of project designers, because they have no "bleeding" or permeable conceptual boundaries, the third, interactive, model should not be rigidly constrained by its original conception. The leaking rainbow asserts an interactivity that presupposes an ongoing intertextuality between project designer and student as much as between student and text, and between texts themselves. In my own case, what happened to the project was that my plans for delineating the connections between the Bluestocking intellectuals and the rise of Romanticism have transformed into a less rigidly or bookishly conceived range of writers. It now centers on those writers and texts the students are fascinated by rather than those I chose to represent. I believe this to be an important concession to student learning and to students' understanding of the dynamics of scholarship.
My choice of a simple approach that could then complicate itself as a part of its own process responds to this concern; it was also in part my own wish to see what students would make of the site, where their intellectual curiosity would take it, and in part a rebellion against websites and software programs that are so completely designed, so over-designed, that they can be a tool of a relatively unhelpful sort only, because there is no place to go with them, no way for the student to grapple with the material on their own as there is in a book. Such websites and projects do not consider the student's intertextual experience, wherein linguistic, visual, social, intrapsychic, and previous textual relations vie for dominance within the student's readerly activity. Certainly it is this experience that should not only direct the student's contribution to the building of a project such as the Bluestocking Archive, but should also formulate the accompanying classroom dynamic. The first time I taught Romantic period literature in relation to the Bluestocking Archive, I found I had to redesign my standard seminar course around the archive, calling it "Romanticism, Gender, and Technology". The course became a way to study archival materials (or at least materials not usually included in anthologies), since the archive is an ideal way to begin to design and use websites. Furthermore, by choosing the Bluestocking Circle as the centering topic and structuring metaphor for both the archive and class, I pushed the focus of both toward gender and women's intellectuality, re-orienting the course specifically in this direction. However, the choice was also scholarly, since I wanted to collect and "publish" those texts associated with the Bluestocking Circle, and I believe this initial scholarly drive was extremely important in creating a sense for students of real work being done in the archive. I would certainly want to add the scholarly dimension to the list of pedagogic imperatives that must be considered when using instructional technology, and when using hypertexts to study print texts.
Students in this course were required to use the archive as a stepping off point from which to rethink the bounds of the Romantic period, as well as its most marked characterizations, based on materials they found on the Web, in course materials, and in the library. They were asked to compare a recently issued anthology of Romantic literature to several standard anthologies that use traditional definitions and boundaries for Romanticism and then to use that comparison as a ground against which to fit their findings from the Web. Their oral presentations to the seminar involved taking a text, analyzing its place in relation to the canon and to recent reconfigurations of the canon and period, and arguing for its inclusion in the archive. Their final projects worked through the implications of these arguments for the text, and included their own preparation of a text for mounting on the archive. Finally, the class presented their Web texts along with interpretation and defense to the department at a colloquium. While not every student was entirely successful in meeting the goals of the course, the majority were so successful that I considered their work to be at the doctoral level, yet these were Masters students with little or no background in romanticism or Romantic studies. Because few students in this class were familiar with using the internet or even with sending email, and most had not encountered hypertexts and were not familiar with the concept, I cannot attribute the success of the course, which was far higher in terms of student learning than my usual Romantic studies syllabi, simply to the addition of technology to historical material. Far from offering immediate interest to the course, the use of technology proved a stumbling block to many students. Contrary to those who criticize the use of technology as destructive to literacy, bibliophilia, and print culture in general, some students are still far more comfortable with books than with computerized texts, with linear texts than with hypertexts, with considering technology as a transparent medium that cannot impinge on or challenge their confrontations with history. Rather than an easy answer of technology providing the impetus for the course's success, I attribute that success to the effort many of the students were forced to make to come to terms with the relation between hypertext and book, between a prior discourse and discourse protocol, and their own.
Recent offerings of related courses have allowed me to begin expanding the ways in which the archive will be used. As the number of texts the archive includes increases, I will add a secondary level of links to the table of contents for scholarship by students, rationales and supporting research for the inclusion of texts in the archive, footnotes and editorial notes, conference papers, and links to Web-published articles relating to a particular text. (For recent gestures in this direction see student papers on slavery and Wedgwood.) Once this process becomes the work of the archive, it will become more functionally three-dimensional, a more fully media-culture rather than print-culture tool. Nevertheless, even in its mid-way stage, the Bluestocking Archive offers a vehicle for rethinking the way that women created a cultural and textual space for a female version of what Romanticism represents. In asking students to work out how a text of their choice does or does not relate to the archive, I am asking them to consider gender at a very basic level in texts where the issue may not otherwise be raised. In situating these women's texts as the point of departure for the archive, the archive asks users to reconsider basic beliefs about the autonomy, rebellion, and masculinity of Romantic texts.
The act of considering such issues can be built into the project itself. This is also part of what I mean by a constructivist approach. The site and the student should interact over and through the text. This approach reappraises hypertext from something one might view as essentially text integrated with links to something George Landow describes as "always virtual, always a simulacrum for which no physical instantiation exists".  I am not suggesting that such web projects can provide the ideal reader response situation, but that the text the reader produces through the act of reading highly resembles the hypertext as simulacrum. It is this analogous experience of textual reading and hypertext modality which I believe the reader subliminally recognizes as she encounters the hypertextual moment. It is this recognition that should be instantiated by reproducing it as the student's contribution to the project through the linking of self-produced texts (that is, the mounting of self-selected texts and the mounting of research, annotations, and other student work).
How we think about web projects pedagogically will depend, I believe, on the metaphors and models we choose to explain such projects to ourselves. I have used several here already, such as reader response, in order to help me think through and describe what a constructivist approach can look like and how it can function. It seems to me that our prior, bookish metaphoric ways of thinking must simply fail when confronted with the scholarly and pedagogic possibilities offered by the internet. Howard Gardner's research shows that children's developing linguistic creativity leads them to a different conception and use of metaphor, and that once a child is old enough to comprehend adult systems of metaphoric use, their earlier usage is replaced by their apprenticeship into the older order.  Without appearing too Romantic by applying human development teleologically to technology, we might say that the shift from paper books to hyper or electronic (or even cyber) books, from print technology to electronic technology, also requires a shift in metaphoric invention but one that reverses Gardner's model. Here we must give up our "adult" metaphoric or schematic order to create new ones to fit the new possibilities awaiting our discovery. While we are discovering these new schematics through experimental fiction, interactive sites designed for multiple forms of exchange, and new uses for visuality, it is important to find bridging forms that can modify old metaphoric structures and generate new ones. Espen Aarseth notes that how we think about a text (the metaphoric schema we set up for textuality) affects our reading behavior: "the stability of paper-based documents is as much a product of our metaphysical belief in a transcendental text as an inherent quality of the physical object".  This stability is a much a contextual effect as the "instability" of e-texts and hypertexts, in other words, but it is a stability we are comfortable with because we have grown into its metaphoric structure. While we are in the process of learning to get comfortable with instability as an inherent textual quality, we need something intermediary. The electronic archive can function in just this way. The Bluestocking Archive is designed to work as an anthology, as I have said, with the centering archive contents page as the access point for each text, and the links acting as page locators. All other links build from these primary ones, to produce a multi-level site that is, finally, determinate in that all new links come back to the body of the site (versus an indeterminate site whose additive links continually open up new exploratory matter). In this sense, the archive site mediates print and electronic media culture to make the Web a useful tool for students who still need a bookish orientation in order to understand the print culture of Romanticism while modulating into hypertextual architecture. At the same time, it raises pertinent issues of gender and textuality, since the body of the site, like the body of the salon hostess, organizes the textual conversation and must therefore confront the student with the problematic of female texts. Again, our metaphors for presenting and analyzing such aspects of a web projects will mobilize how and how intently our students engage constructively (constructivist-ly) in textual relations.
This model raises several technical issues as they impact the pedagogy. The first technical issue I had to address in creating the Bluestocking Archive was the need create a connection between book technologies that are often unlinked and even isolated from each other and electronic technologies. Certain pedagogical issues reflect back on such considerations and raise others. In thinking about the website as a book, I am for a variety of reasons going against the standard view that the intrinsic quality of the Web will move us toward a more poststructuralist hypertextual format that encourages dimensionally structured thinking, so that associations are matrixed rather than linear. But in making the case for associative thinking, as many of those thinking through the implications of hypertext do, it is also hard to know how to plug this debate into technologically impacted pedagogies. That is, should we encourage students to think even more associatively than they already do because of their ready experience with the new technologies at home and in their previous schooling? I ask this because of the fact that human experience is foremost associative, and because the work of education thus far has been to train students out of the associative and intuitive mode and into the linear and logical mode. Was the emphasis on logic wrong? Or should we strike a balance between hyper-thought processes and the formal constraints of logical linearity? Lately, logic has been construed as male and association as female thought behaviors, with an increasing emphasis in some pedagogies on the positive valence of the associative as female. But technology takes the intuitive and associational thought process and, in combination with visual formatting, makes it quite often the only thought mode available—so that viewers follow links for hours at a time instead of studying one page through to the end. With the double valence on associative thought of being both female-oriented and technologically useful (to say nothing of its interesting resonance for Romantic scholars), the question becomes one of a critical nature: Should students be versed in both modes of thinking, not as "male" and "female" mentalities (to use this term loosely, but I think aptly) but as different styles of mental literacy? The largest issue we have to face in instructional technology, it seems to me, is educating students in visual literacy as well as textual literacy (and eventually aural literacy as well); shouldn't we be thinking through how to teach (either adjacently or integrally) visual and textual literacy behaviors, so that students understand the difference not as a reading behavior only, but as a mode of thought that can be consciously applied to materials that may seem to demand different modes?
Similarly, most uses of instructional technology that I have seen simply mount more and more sophisticated lectures that continue to use the same teacher-owned pedagogy that drives a lecture with visual-aid styles of teaching. But the uses of a technology that pushes the pedagogy into a different level of interaction can actually change the learning process. My experience with the seminar mentioned above was that students with no former knowledge of the Romantic period came away with so strong a sense of the period's thinking and writing, that this could have been a doctoral seminar. Part of this substantive difference in the learning that went on in the course was due to the pressure on the students to do their own research, but part of it was also do to the reflexive nature of the archive. Students could see the "hyper book" change, be in process, as they themselves added new features to their texts. Students decided to markup their oral presentations to link to the poems they mounted, or to add footnotes, or to add visual enhancements (such as scans and audio clips of the sections of Mozart's Requiem that seemed to be indicated by Felicia Hemans's poem on that score). Some students added more information after the course was over because it became clear that, at least in the case of the Hemans poem, the technological linkages between text, image, and sound were not enough; a critical apparatus had to intervene in order to reveal the logic of the whole. This student's work in particular demonstrates what I mean by a combined use of linear and associative cognitive modes both in the pedagogical sense and in terms of a usage tool for hypertextual production. What I see happening here is something between Barthes' readerly and writerly textualities.  Interactive hypertexts would be writerly in this sense, but these students were not re-encoding or elaborating texts, their writing did not affect plot or the literariness of the text, only structure (in the strictest sense) and interpretation. Yet they were not performing in a purely readerly way; their writing did change the text-but only in an editorial, critical, and contextual way, so to affect their own reading and compel readerly/writerly behavior from others.
There is another way in which the mix of cognitive modes that hypertexts push on else challenge pedagogical considerations. Initially, I was interested in the way a web archive would be conducive to team projects, and although the term projects turned into individual and not team research, I discovered that the seminar itself worked as a team with the same group dynamics and mutual encouragement that occurs in teams. This revision of the expected became one of the constants of the course, and is perhaps one of the pleasures of instructional technology. In another form this revision is apparent in my current concern with cognitive modes. Initially this concept was for me one concerning women's intellectuality and the Bluestockings' social and educative goals; it became one about the incongruity of using the most advanced pedagogic methods to teach such forgotten and undervalued texts. But in the space of that incongruity is an intersecting of teaching and revision that made me reconceive what it is beyond the textual/contextual that the Bluestockings have to offer us. I believe it is their emphasis on how determinedly we use our minds, in what way and to what end.
II. Employing Salon Theory
In thinking about the website I created to teach Romantic literature three things come together: salon conversation with its relation to both associative and linear thinking; hypertext as a resuscitation of associative thinking and associative thought as a Romantic theory of mind; the forgotten connections between literary texts as an analog of sorts for the forgotten connections between philosophies of language and thought. Putting these three topics in relation to each other makes more of my site than it warrants, and provides too much ground for me to cover here, but it does sketch out the size of what The Bluestocking Archive has meant to my thinking about Romanticism and its relation to present-day technologies and pedagogies.
I would like to remark first on what I mean by "salon theory. " The eighteenth-century salon, which was so crucial to the development and intellectual projects of the first generation Bluestockings, was more than anything a place where the mind could flourish through the exchange of ideas. If witticisms and literary flourishes despoiled some of this exchange, so much the better since this provided a socially acceptable cover for what could be talked about. To think that, like the earlier coterie gatherings, men and women could get together and talk about ideas. Talk, rather than writing and its exchange, fosters thought by association; ideas can spiral, vine about, bloom in odd ways when no one Socrates figure is directing the flow of things. Because salons encouraged people to cluster together in small groups, many conversations occurred simultaneously. Less than a room of one's own, more than a seat in a family parlor, the salon offered women in particular a place to be intellectual. Moreover, the role of conversation—the exchange of words and ideas, rather than a lecture; the equitable pitting of topics, the easy chance of digressions, the richness of associable ideas and inferences—is valued in Bluestocking salons. It should be valued today for its open-endedness, its plausible leads, its intellectual heritage. It is invaluable for women students who find the tension of seminar discussions a block to verbalizing their ideas, but it is equally helpful to men students by providing a platform for hashing out ideas before seminar meetings, while provoking them to examine their assumptions in light of their colleagues' reactions.
We are said to have lost the art of conversation in the present age, but that is a reference to parlor talk; we have indeed lost the tool of conversation, and that is something we could and should remedy, because students should not have to think in isolation, should not have wrestle with ideas on their own, should not come to erroneous versions of history or literature because they did not have to try out their version on someone else.
When I describe rooms, conversations, fostering, I am describing not only Bluestocking salons but how I hope graduate seminars may function given just such a place (albeit fragmented—the seminar room as well as the computer lab as well as the class listserv, all of which provide pockets for small conversations that in themselves happily frustrate my Socrates-type directives for organized class discussion). Salon theory means taking the literal salon and applying its organizing principle to pedagogical imperatives and simple technological tools. (Although the Salon website attempts to put together versions of a conversational format, this is essentially an electronic magazine and does not correspond to my conception of salon theory.) And although I believe the salon to be a feminist mode, with its frustration of the master narrative, its discouragement of linear thinking, its fostering of growth through small group participation, its placement of female intellectuality, that was not my original reason for experimenting with the form. Rather, I believe that in our postmodern age, end of the millenium angst has combined with speculative philosophy and technological achievements to produce a climate in which it has become almost necessary to think in relative terms. Bad word that it is, "relative" nevertheless describes the mindset we bring to nearly everything in the nineties, including our intellectual endeavors. The salon provides a place for relativity or for relative thinking by opening up a challenge to modernist views of texts, of knowledge, of literary inheritance, of genealogies. When students have to think through connections on their own, especially in relation to what salons started out as being (as well as by whom and for whom), then canons of texts make less sense than they did when modernism was the privileged approach to intellectual pursuit.
For Romantic studies, putting the salon form in relation to those texts produced by Bluestockings, those texts influenced by these women, and those texts directly opposed to them or rejecting them while building on their work, provides a challenge to old canons that students can explore and delineate for themselves no matter what goes on in the course syllabus. Some find they embrace the old standards for literary importance even more fully while others decide that the connections they find between texts new to them and familiar ones provide deep opportunities for study and learning. Both these directions are positive results of salon theory because both indicate choices made based on individual grappling with material and contexts. However students arrive at their choices, they not only have to defend them, but because of the salon format they also have to make decisions based on a (however conscious) use of associative thinking alongside linear thinking, and they have to adjust their defenses to these modes of thought as companion modes.
I have described briefly above how I believe hypertext encourages (or demands) associative thinking, but we tend to assess this thinking today as very postmodern, very hip. We forget it is what modernism had stigmatized as unacceptable for serious thought. But we need to remember two things: what associative thinking was believed to accomplish in the eighteenth century, and how the modernist theory that the mind works like a computer (that is, linearly) has been disproved. I will sketch both issues very briefly, but a deeper examination would provide interesting paths of study. (For instance, one such path, tracing the early development of neuroscience as it grew out of associationist theories of how the mind works, is already being pursued by Alan Richardson, whose website "Literature, Cognition and the Brain" on the subject organizes and links together web resources on neuroscience.)
Eighteenth century philosophy divided into two camps: empiricist and rationalist. Empiricists believed that sensory perception forms as well as informs mental activity. We hear the echoes of this today in the various claims for how different technologies, such as the pen or pencil, typewriter, and computer keyboard affect the rhythm and length of one's sentences, the sound and style of one's prose. Rationalists believed that the mind preexists external objects, and forms, interprets and organizes sensory perceptions in such a way that the external world cannot really be known. We see this today in linguistic theories that claim the brain is pre-wired for language; that a child does not form his thinking to a given language, but that he is already wired to learn fundamental grammatical rules common to all languages. Associationist theories work between these two camps to describe how sensations impress themselves on the mind, and are retained and complicated through associated experiences, associable either through the present context of the impression, through remembered impressions, or through juxtaposed memories. Today we might laugh at this description of how thought works, but we do pay attention to how thoughts "associate" through similarity, connection, metonymy, juxtaposition, bricolage. And essentially there is more similarity than difference in these two forms of association. Both lead to conceptions of thought that are web-like, where multiple connectives can be spun out and explored on a variety of levels and time-space formations, against and across linear thinking that closes down on possibility in order to formulate one assertion only. Clearly we conceptualize the World Wide Web in terms of this later form of associationism; clearly we view the world wide web as an analog of the human mind with its webbing thought processes; clearly we privilege associative thinking when we accept both the web image and the computer analog to describe our postmodern cognitive state. It seems clear to me, then, that we need to explore the pedagogical implications of these terms; the classroom cannot be rooted in a prior epistemological standpoint when the student is footed in the next point after that.
This brings us to the next point, that we do assume a fundamental analogy between computers and human brains. Yet this belief, which now because of the interlinking of computers has provided a web image for how computers work, initially instead provided a linear conception of cognition. Howard Gardner, who has been influential in pushing educators to consider cognitive models in pedagogic terms, notes that cognitive science is built on what he has called the "computational paradox": the attempt to make the computer and the human brain cognates has led not to its proof but to an understanding of how the mind does not function like a computer, or on the order of computer logic. "This is not to say that no cognitive processes are computerlike, " but that "the kind of systematic, logical, rational view of human cognition that pervaded the early literature of cognitive science does not adequately describe much of human thought and behavior".  Indeed, cognitive science itself had to become interdisciplinary to come to terms with the complexity of understanding the mind and its workings. That is, the science itself had to become associative, had to divest itself of its adherence to the study of linearity. If we understand this from a pedagogic perspective, it makes sense to provide students with associable approaches to the subject matter in order that their own web of impressions, inferences, and interpretive possibilities can accrete, providing them in the end with a complex view of how literature interacts with culture and history to produce something that we are endlessly coming to terms with, and about which no one holds the absolute interpretation, the absolute answer. Hypertext in particular, it seems to me (and despite all the hysterical disclaimers to the contrary), provides students with the sensory experience that can re-present salon discourse so that their own associative cognition can come into play. The intersection of salon theory and a constructivist frame of mind should, it seems to me, provoke or energize an interpellation of students so that their engagement with course/archive materials and thus their salon conversations become dialogized. I do not mean this in the fashionable and sometimes far too facile way dialogism is used in composition theory as voice on voice, but as a way to reorganize (and hopefully intensify) student-text relations.
Finally, by constructing with my students an archive of forgotten literary texts, and putting these texts in relation to canonical texts, or to texts by canonical authors (that is, texts we would not ordinarily read by authors we revere), we are constructing a genealogy. It is a web of literary inheritance that reminds us where mastery may be grounded, where knowledge comes from, how literary culture grows, succumbs, rebirths. It seems to me more and more that to study Romanticism is to come at it from this angle, in order to challenge the view the Romantics put forth themselves of their Athena-like origins, sprung full grown, if not from the head of Zeus then from the body of Nature; originary, original, founding a new worldview. I want students to understand that to study Romanticism is to see gender inflected in every word, every inherited image and issue, not merely in the quantitative difference between women and men authors, or in the essential representations of male and female characters. It seems to me that when students are responsible for constructing their own anthologies to study they are forced to come at the literary material from such a perspective.
In the same way that the Romantic male poets strove to forget their female inheritance, postmodern theories about cognition and language ignore their eighteenth-century predecessors, striving to position themselves as original. Long before Chomsky, Hobbes was claiming the mind was prewired for language. Of course this is not the same thing really, but it is the same attitude toward the mind that produced both thinkers' theories, and certainly Chomskian theory owes a great deal to the long line of thinkers beginning with Hobbes and carrying on through Locke and Berkeley and Hume. That sort of forgetting can, like the forgetting of the Bluestocking intellectuals, be remedied by careful use of technology in the classroom. By that I mean that when we use such things as internets, websites, and listservs in the classroom, we need to be careful about what these technologies mean to us. We need to ponder the pedagogic weights of these things, of course, but also their histories, or rather, what they imply for us. We can then remind our students of connections such as the genealogies of literary texts and the genealogies of the theories that allow us to study these texts in different ways; or of the ways these theories allow us to pose text against hypertext, and to come up with a differential. Billie Wahlstrom contends that computers in the composition classroom cannot provide a feminist pedagogy. The conception and construction of computers and the software we use on them are grounded in a patriarchal world, as Ruth Perry, Lisa Greber, and Ruth Hubbard have argued.  Wahlstrom builds on this observation to show that if teacher-student hierarchies reform into the less authoritative coach-student relation, nevertheless certain patriarchal behaviors are encoded in composition software that reinforce teacher authority and privilege. "On some networks, for instance, teachers can read students' files, but students do not have access to teachers' files. . . Additionally, networks are sometimes organized so that teachers can read students' unfinished documents". 
Wahlstrom urges that we study the "gendered processes" underlying software and hardware before employing them in the classroom so that we don't reinforce through our tools the very thing we believe we are challenging and rethinking through our class procedures, or in my case, through our subject matter. If many view the use of technology an always already male intrusion in the classroom, I would rather view the careful use of such technology a boon to feminist pedagogies. First, however, such a study as Wahlstrom promotes must be undertaken and applied in each course. We must use technology as a form of remembering, not as yet another version of Romantic forgetting. It seems to me that the salon can be used to repair the issue of technology's "anti-feminism" when salon theory is applied to hypertextual ways of reading to produce conversation, real conversation.
I have been trying to think about how a constructivist frame of mind would rethink and redesign this article so that becomes its own interactive website, its own invitation into salon theory. There is a way in which reconceptualizing monologic articles as conversations, if this is actually possible, would provide an alternative version of the imaginary architecture of cyberspace. It would go beyond the mere sprinkling of hyperlinks I have installed here to involve opportunities for feedback and discussion by readers, ways to read the difference between the sites I located by hyperlink into another space so that comparisons can be fleshed out by anyone as they read this piece. Not being able to go so far into such a dimensionalization of the article form, I have backtracked and built at least a basement to it through links to the Bluestocking Archive and other web projects, as well as the following page [Appendix 3], a description of what happened when I used the archive as research tool only for undergrads, resulting in missed opportunities for constructionism and a temporary failure of the project. Although I have not discussed this aspect of my project in this article, I do believe that the failures as much as the successes will aid how we define and refine new pedagogies of leaking borders.
Appendix 1: Pedagogy and Web Site Design
Description of the ALN: The objectives of the ALN Web are to provide (1) a focal point for information interchange among researchers and practitioners in the field of asynchronous learning networks and (2) a scholarly reviewed on-line journal which captures the archival knowledge of the field. The ALN Web contains: the Journal of Asynchronous Learning Networks (JALN), the ALN e-zine, columns, news posting areas, workshops and other features directly or indirectly related to ALN. The JALN provides viewable and downloadable journal articles in the format of a traditional journal."from the ALN Guidelines
The following is not a critique of "The Pedagogy of Web Site Design," but rather an attempt to illustrate how the Web can be used for traditional pedagogies as opposed to the kind of "constructivist" pedagogy I explore in "The Bluestocking Archive: Constructivism and Salon Theory Revisited."
Skip Knox describes the long distance learning course, Humanities 101, that he developed out of the "traditional" lecture course of the same title in his article ""The Pedagogy of Web Site Design." Knox has taught this course in its Web version six times as of summer 1997, giving him much more data for his comments than is normally the case for published meditations on teaching Web courses. He notes that his conclusions "apply best to a fully virtual asynchronous course in the humanities. A course that has a local, physical component will have a different dynamic, as will a course that uses synchronous tools like a MOO or chat."
The course is set up with recognizably traditional components ("syllabus, readings, registration information, study questions, and 18 or so complete lectures") on a Web site, with the site functioning as the focal point for all course activity. In addition, students have a textbook in hard copy and a listserv list for class discussion. The Web site has:
a cover page whose main function is to pre-load some graphics and to serve as a visually attractive first page. The true main page shows the main divisions of the course: the Visitor Center, the Registration area, and the Classroom. The latter is where all the content resides. In all, the site comprises 51Mb in 1,336 files. Of these, 691 files are HTML documents, about two-thirds of which are actual content pages and one-third is administrative material. The other 645 pages are graphics, sounds and text files.
Knox notes that in his first on-line teaching course, Humanities 309 (the Italian Renaissance), the reading was partially book dependent, with the Web site offering additional readings and the internet used for discussion and email. His initial approach was to free up instruction from the classroom lecture dynamic: Rather than "teaching" them, I would let my students find their own way through the information; I would serve as coach and resource, not as authority figure. . .So I provided objectives, identified the required reading, and even seeded the class with discussion questions, to help them get started, but otherwise I stood to the side, waiting to guide." The result was chaos and student distress. In a traditionally offered course, Knox says he would organize the course around topics: the snarled politics of 15th century Italy, philosophy, the crisis in the Catholic Church, art, literature. But Knox assumed that by providing course material on his Web site, students would be able to organize themselves according to their interests:
In creating a course when anything might be done at any time, I had created one where it appeared that everything needed to be done all the time. The Web site sat there, seeming to insist upon being read at once. If someone made a reference in discussion, then you felt obliged to go read that, or else not to participate. Students were either overwhelmed with work or were too intimidated to keep trying.
Instead of drawing the students together into an identified group, the Web site "anatomized" students. Knox's account of this first course points out the importance of student/reader conversation as an interactive aspect of the Web site: "we had no common basis for discussion. And without that, we literally had no class; we were instead separate individuals. . .who might by accident have something to say to one another." It seems to me that the difficulty was that conversation was not structured into the course site. He describes the course as being "unstructured, webbed," but I would argue that "webbed" is not synonymous with being "unstructured" and that the difficulty arose precisely because the site was not webbed.
Certainly the structure of an instructional Web site can be successful whether traditionally or "constructively" designed. When Knox describes how he has translated this experience in his successful course, Humanities 101, he focuses on how he has organized his lecture and source material on his Web site. He chose an organizing principle that clearly refers back to lecture style teaching. In explaining how the new site works he comes closest to discussing a "hyperbook" in the terms I analyze in my essay. However, his focus on linearity in particular outlines the difference between using the Web for a traditional pedagogy, versus using it to develop a new orientation toward teaching through the technology. The following points organize his argument:
Each lecture is presented as a series of pages, rather than as one file. There are a number of good reasons for this, one of which has to do with rhetoric. The common wisdom of Web design says no document should be more than three screens long because people begin to lose interest and focus. So an online lecture certainly needs to be broken up, but where to make the breaks?
Each Web page serves to present a thought, a concept, a scene in a narrative. The link between one page and the next is a caesura, and the end of a page is a dramatic moment, rather like a dramatic pause in public speaking. The reader has to click on the mouse button and wait a moment (but not too long!) for the next screen to appear. Just as the end of a chapter in a book should propel the reader forward to the next chapter, so the words at the end of one Web page should create a little tension and lead the reader forward. Not every page lends itself equally to this, but being aware of the technique can help in the presentation.
Even a casual visitor to the on-line lectures will notice that I have almost no external links within a lecture. This is an important component of Web rhetoric and of Web pedagogy. I have seen enough other teaching sites that argue for lots of hyperlinks that I want to explain my position and experience here.
The student comes to the Web site with roughly the same expectations she brings to a textbook. She expects to be able to understand quickly what the work is about, to be able to move through it readily, to have a clear idea of the boundaries of the work, and to be informed by the content. A site that is filled with hyperlinks violates almost all these expectations. While in theory we are offering the student the opportunity to explore, in practice the site consists of an unknown number of reading assignments.
Most of my lectures therefore have no links at all except to the next page, previous page, and Table of Contents. Students learn quickly that a lecture is a known quantity and can plan their time accordingly. Many of the lectures have sound file links, but these are simply to help the students pronounce names and do not take the student away from the page. I do not embed pictures in the lectures because the download time would disrupt the rhythm of the reading.
From these comments, it is clear that for Knox, pedagogy means layout rather than a cognitively-oriented theory:
The important lesson here is not whether linear is better than webbed. So many variables enter into the equation-the personality and tastes of the teacher, the constraints of the discipline and subject matter, the technology itself, the type of students, and so on-that one cannot reach general conclusions. The important lesson is that a form works best with a conscious pedagogy underpinning it. We are simply putting content on the Web, we are putting it there in a particular form. And, even more important, we ask the students to address that content in particular ways.
The notion that "We are simply putting content on the Web" is the primary principle I would contest in a constructivist pedagogy. Knox's comment that "a Web site is relatively static" further underscores the difference in the two approaches. Although Knox is describing a long distance learning course, which is extremely different from the classroom-oriented course I discuss, and which demands very different things from a Web site in order to supplement the missing context of the classroom, his useful analysis of his experiences helps point out how a pedagogy impacts Web design and orients, even directs, the student's use of and interaction with the site. For Knox, the Web site "plays an important part in creating and sustaining the community of a given class," besides "deliver[ing] content"; in a constructivist approach the Website should enable the creation of content, reenact the epistemology of student-text relations, and a component of these two processes sustain the community of a given class.
Appendix 2: Resources for Pedagogy and Technology
Pedagogy for Web-based Education
This site has interesting pages such as "Problem-based Learning," "Cognitive Apprenticeship," "Conversation Theory."
Live Text Institute for Learning Technologies
http://www.ilt.columbia.edu/k12/livetext-nf/webcurr.html [no longer online]
This site is based at Teachers College Columbia University. The site "is a guide to initiate experienced educators into designing constructivist, cooperative learning projects around the World Wide Web. Examples of projects inspired by this guide include The American History Archive Inquirer and the Dalton."
"Simple Start Pedagogical Resources"
This is a list of helpful urls on theory, practical tips and planning, and evaluation.
The Sloan Center for Asynchronous Learning Environments (SCALE)
"SCALE faculty are participating in a three-year project of restructuring undergraduate courses to integrate various techniques associated with asynchronous learning networks (ALN). The goals of this project are to create efficiencies in the educational process (cost, time, faculty productivity), to increase student retention, and to decrease time-to-degree."
Links to Teaching Resources on the Web: Education Pedagogy Learning
Connections: Scholarship on Hypertext and Computer-Assisted Pedagogy
From a site devoted to Gordon Pask (1928-1996), theorist, cybernetician and developer of Conversation Theory:
Conversation Theory is a constructivist, relativist theory based in the understanding that each student must accept responsibility for his own learning, that his understanding is uniquely his, and that such understandings are communicable and learnable (only) through conversation. He developed these ideas in numerous learning environments, and through their ramifications in, for instance, how matter to be learnt should be structured and differences in how we choose to learn (learning styles). He also saw clearly the limitations of his initial approach, and attempted to generalise Conversation Theory is two ways. Firstly, by considering the conversation as part of a gigantic and endless continuum of interaction (life)-the Interaction of Actors Theory-and secondly through the generalisation of structuring rules for the modification of bodies of topics to be learnt.by Ranulph Glanville. "This obituary by, originally written for the Bulletin of theInternational Federation for Systems Research, appears bypermission of the editor, and the author, and may be freely copied."
Godwin-Jones, Bob. "EMERGING TECHNOLOGIES: Dynamic Web Page Creation." Language Learning & Technology, Vol. 1, No. 2, January 1998, pp. 7-13"
Landow, George P. "Hypertext in Literary Education, Criticism, and Scholarship." Computers and the Humanities 23 (1989): 173-198.
Yoes, Catherine. "The Science Fiction Web Project: Adventures in Teaching with Storyspace." The Electronic Journal for Computer Writing, Rhetoric and Literature. (2) number 1.
Appendix 3: Coda
In an undergraduate course on Romantic period literature in spring 97, EN405, I decided to use the archive as research tool only, in addition to a lengthy bibliography of Web resources for learning, research, and interactive exploration on Romanticism, 19th-century history, maps, listservs, and so forth. I made this choice in part because of the large size of the class, another determinant factor in Web project design, but one which I now see should not have deterred me from designing the course around my Web site. I want to take this opportunity to think about why using the Web as a resouce guide resulted in missed opportunities and a temporary failure of the Bluestocking Archive project.
I decided to teach the course as a standard lecture and discussion course based on an anthology with supplementation from the Web with texts not included in the course book. However, because I did not provide class instruction on how to use the Web, those students not yet conversant with internet use resisted learning how to do so on their own time (mostly out of fear of the new, but also out of time deficit) and in the end never engaged this part of the course. These students prepared their oral presentations on an author or poet, as well as their research papers, using hard copy materials from the library stacks only. Those students already familiar with the Web were able to prepare their oral presentations using Web materials, but because they found this easier to do than using library materials, their presentations were extremely unbalanced and based on whatever they were able to find on the Web.
This combination of resistance (the difficulty technology imposes) and fudging (the facility it brings) contributed to a temporary death of the Bluestocking Archive. I was unable to think my way out of the tangle the course had brought to the project and to its goals. Although the students' term research papers were happier combinations of Web and library materials, they still fell short of what I had hoped students could accomplish with very little guidance from me. I am convinced now that this was because the resources I provided students with superceded the Bluestocking Archive for them; in fact, they visited the Web site very little throughout the semester. When I asked students to contribute one short text to the Web site, only a handful of students were interested in doing so (this was not a requirement of the course), and even fewer actually did so.
Furthermore, because I was not thinking of the Web site as central to the course, there were many missed opportunities for constructivism. I remember in particular an extremely interesting oral presentation on slavery in which the student had discovered some fascinating Web sites on the subject. But because of the setup of the course, she had to print the sites out and use them as hard copy. Not only did I fail to see the potential for this presentation as a page off the archive site at the time, but students reacted to what she had done as if she had simply read a book and was reporting someone else's research. I missed the chance to feed their excitement, but more importantly, I failed to translate her own excitement into a constructive interaction with the Web site. I believe that had I done so, many more of the presentations would have proved equally stimulating and educative.
See Sylvia Myers, The Bluestocking Circle (Oxford: Clarendon Press, 1990).
See Espen J. Aarseth, "Nonlinearity and Literary Theory," in Hyper/Text/Theory ed. George P. Landow (Baltimore: The Johns Hopkins University Press, 1994) pp. 51-86.
Michael Benedikt, ed., Cyberspace: First Steps (Cambridge, Mass.: The MIT Press, 1994).
George P. Landow, ed., Hyper/Text/Theory (Baltimore: The Johns Hopkins University Press, 1994) 6.
Howard Gardner, Art, Mind and Brain: A Cognitive Approach to Creativity (New York: Basic Books, 1982) 158-167.
Espen J. Aarseth, "Nonlinearity and Literary Theory," 55.
See Roland Barthes, S/Z (Paris: Seuil, 1970).
Howard Gardner, The Mind's New Science: A History of the Cognitive Revolution (New York: Basic Books, 1987) 44.
Ruth Perry and Lisa Greber, "Women and Computers: An Introduction." Special Issue Signs 16 (1990): 74-101.
Billie J. Wahlstrom, "Communication and Technology: Defining a Feminist Presence in Research and Practice." in Literacy and Computers 178.